Persuasion Theory and Research 3rd Edition [PDF]

Citation preview

Persuasion Third Edition



2



SAGE was founded in 1965 by Sara Miller McCune to support the dissemination of usable knowledge by publishing innovative and highquality research and teaching content. Today, we publish more than 750 journals, including those of more than 300 learned societies, more than 800 new books per year, and a growing range of library products including archives, data, case studies, reports, conference highlights, and video. SAGE remains majority-owned by our founder, and after Sara’s lifetime will become owned by a charitable trust that secures our continued independence. Los Angeles | London | Washington DC | New Delhi | Singapore | Boston



3



Persuasion Theory and Research Third Edition Daniel J. O’Keefe Northwestern University



Los Angeles London New Delhi Singapore Washington DC Boston



4



Copyright © 2016 by SAGE Publications, Inc. All rights reserved. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the publisher.



FOR INFORMATION: SAGE Publications, Inc. 2455 Teller Road Thousand Oaks, California 91320 E-mail: [email protected] SAGE Publications Ltd. 1 Oliver’s Yard 55 City Road London, EC1Y 1SP United Kingdom SAGE Publications India Pvt. Ltd. B 1/I 1 Mohan Cooperative Industrial Area Mathura Road, New Delhi 110 044 India SAGE Publications Asia-Pacific Pte. Ltd. 3 Church Street #10-04 Samsung Hub



5



Singapore 049483 Acquisitions Editor: Matthew Byrnie Digital Content Editor: Gabrielle Piccininni Editorial Assistant: Janae Masnovi Production Editor: Jane Haenel Copy Editor: Lynn Weber Typesetter: Hurix Systems Pvt. Ltd. Proofreader: Gretchen Treadwell Cover Designer: Gail Buschman Marketing Manager: Liz Thornton Printed in the United States of America Library of Congress Cataloging-in-Publication Data O’Keefe, Daniel J., 1950– Persuasion : theory and research / Daniel J. O’Keefe. — 3rd edition. pages cm Includes bibliographical references and index. ISBN 978-1-4522-7667-0 (pbk. : acid-free paper) 1. Persuasion (Psychology) I. Title. BF637.P4054 2016 153.8’52—dc23         2015000192 This book is printed on acid-free paper. 15 16 17 18 19 10 9 8 7 6 5 4 3 2 1



6



Brief Contents Preface 1. Persuasion, Attitudes, and Actions 2. Social Judgment Theory 3. Functional Approaches to Attitude 4. Belief-Based Models of Attitude 5. Cognitive Dissonance Theory 6. Reasoned Action Theory 7. Stage Models 8. Elaboration Likelihood Model 9. The Study of Persuasive Effects 10. Communicator Factors 11. Message Factors 12. Receiver Factors References Author Index Subject Index About the Author



7



Detailed Contents Preface 1 Persuasion, Attitudes, and Actions The Concept of Persuasion About Definitions: Fuzzy Edges and Paradigm Cases Five Common Features of Paradigm Cases of Persuasion A Definition After All? The Concept of Attitude Attitude Measurement Techniques Explicit Measures Semantic Differential Evaluative Scales Single-Item Attitude Measures Features of Explicit Measures Quasi-Explicit Measures Implicit Measures Summary Attitudes and Behaviors The General Relationship Moderating Factors Correspondence of Measures Direct Experience Summary Encouraging Attitude-Consistent Behavior Enhance Perceived Relevance Induce Feelings of Hypocrisy Encourage Anticipation of Feelings Summary Assessing Persuasive Effects Attitude Change Beyond Attitude Change Conclusion For Review Notes 2 Social Judgment Theory Judgments of Alternative Positions on an Issue The Ordered Alternatives Questionnaire The Concept of Ego-Involvement Ego-Involvement and the Latitudes 8



Measures of Ego-Involvement Size of the Ordered Alternatives Latitude of Rejection Own Categories Procedure Reactions to Communications Assimilation and Contrast Effects Attitude Change Effects Assimilation and Contrast Effects Reconsidered The Impact of Assimilation and Contrast Effects on Persuasion Ambiguity in Political Campaigns Adapting Persuasive Messages to Recipients Using Social Judgment Theory Critical Assessment The Confounding of Involvement With Other Variables The Concept of Ego-Involvement The Measures of Ego-Involvement Conclusion For Review Notes 3 Functional Approaches to Attitude A Classic Functional Analysis Subsequent Developments Identifying General Functions of Attitude Assessing the Function of a Given Attitude Influences on Attitude Function Individual Differences Attitude Object Situational Variations Multifunctional Attitude Objects Revisited Adapting Persuasive Messages to Recipients: Function Matching The Persuasive Effects of Matched and Mismatched Appeals Explaining the Effects of Function Matching Commentary Generality and Specificity in Attitude Function Typologies Functional Confusions Some Functional Distinctions Conflating the Functions Reconsidering the Assessment and Conceptualization of Attitude Function 9



Assessment of Attitude Function Reconsidered Utilitarian and Value-Expressive Functions Reconsidered Summary Persuasion and Function Matching Revisited Reviving the Idea of Attitude Functions Conclusion For Review Notes 4 Belief-Based Models of Attitude Summative Model of Attitude The Model Adapting Persuasive Messages to Recipients Based on the Summative Model Alternative Persuasive Strategies Identifying Foci for Appeals Research Evidence and Commentary General Correlational Evidence Attribute Importance Belief Content Role of Belief Strength Scoring Procedures Alternative Integration Schemes The Sufficiency of Belief-Based Analyses Persuasive Strategies Reconsidered Belief Strength as a Persuasion Target Belief Evaluation as a Persuasion Target Changing the Set of Salient Beliefs as a Persuasion Mechanism Conclusion For Review Notes 5 Cognitive Dissonance Theory General Theoretical Sketch Elements and Relations Dissonance Factors Influencing the Magnitude of Dissonance Means of Reducing Dissonance Some Research Applications Decision Making Conflict 10



Decision and Dissonance Factors Influencing the Degree of Dissonance Dissonance Reduction Regret Selective Exposure to Information The Dissonance Theory Analysis The Research Evidence Summary Induced Compliance Incentive and Dissonance in Induced-Compliance Situations Counterattitudinal-Advocacy–Based Interventions The “Low, Low Price” Offer Limiting Conditions Summary Hypocrisy Induction Hypocrisy as a Means of Influencing Behavior Hypocrisy Induction Mechanisms Backfire Effects Revisions of, and Alternatives to, Dissonance Theory Conclusion For Review Notes 6 Reasoned Action Theory The Reasoned Action Theory Model Intention The Determinants of Intention Attitude Toward the Behavior Injunctive Norm Descriptive Norm Perceived Behavioral Control Weighting the Determinants The Distinctiveness of Perceived Behavioral Control The Predictability of Intention Using the RAT Model Influencing Intentions Influencing Attitude Toward the Behavior The Determinants of AB Changing AB Influencing the Injunctive Norm The Determinants of IN Changing IN 11



Influencing the Descriptive Norm The Determinants of DN Changing DN Influencing Perceived Behavioral Control The Determinants of PBC Changing PBC Altering the Weights Intentions and Behaviors Factors Influencing the Intention-Behavior Relationship Correspondence of Measures Temporal Stability of Intentions Explicit Planning The Sufficiency of Intention Adapting Persuasive Messages to Recipients Based on Reasoned Action Theory Commentary Additional Possible Predictors Anticipated Affect Moral Norms The Assessment of Potential Additions Revision of the Attitudinal and Normative Components The Attitudinal Component The Normative Components The Nature of the Perceived Control Component PBC as a Moderator Refining the PBC Construct Conclusion For Review Notes 7 Stage Models The Transtheoretical Model Decisional Balance and Intervention Design Decisional Balance Decisional Balance Asymmetry Implications of Decisional Balance Asymmetry Self-Efficacy and Intervention Design Intervention Stage-Matching Self-Efficacy Interventions Broader Concerns About the Transtheoretical Model The Distinctive Claims of Stage Models Other Stage Models 12



Conclusion For Review Notes 8 Elaboration Likelihood Model Variations in the Degree of Elaboration: Central Versus Peripheral Routes to Persuasion The Nature of Elaboration Central and Peripheral Routes to Persuasion Consequences of Different Routes to Persuasion Factors Affecting the Degree of Elaboration Factors Affecting Elaboration Motivation Personal Relevance (Involvement) Need for Cognition Factors Affecting Elaboration Ability Distraction Prior Knowledge Summary Influences on Persuasive Effects Under Conditions of High Elaboration: Central Routes to Persuasion The Critical Role of Elaboration Valence Influences on Elaboration Valence Proattitudinal Versus Counterattitudinal Messages Argument Strength Other Influences on Elaboration Valence Summary: Central Routes to Persuasion Influences on Persuasive Effects Under Conditions of Low Elaboration: Peripheral Routes to Persuasion The Critical Role of Heuristic Principles Varieties of Heuristic Principles Credibility Heuristic Liking Heuristic Consensus Heuristic Other Heuristics Summary: Peripheral Routes to Persuasion Multiple Roles for Persuasion Variables Adapting Persuasive Messages to Recipients Based on the ELM Commentary The Nature of Involvement Argument Strength One Persuasion Process? The Unimodel of Persuasion 13



Explaining ELM Findings Comparing the Two Models Conclusion For Review Notes 9 The Study of Persuasive Effects Experimental Design and Causal Inference The Basic Design Variations on the Basic Design Persuasiveness and Relative Persuasiveness Two General Challenges in Studying Persuasive Effects Generalizing About Messages Ambiguous Causal Attribution Nonuniform Effects of Message Variables Designing Future Persuasion Research Interpreting Past Persuasion Research Beyond Message Variables Variable Definition Message Features Versus Observed Effects The Importance of the Distinction Conclusion For Review Notes 10 Communicator Factors Communicator Credibility The Dimensions of Credibility Factor-Analytic Research Expertise and Trustworthiness as Dimensions of Credibility Factors Influencing Credibility Judgments Education, Occupation, and Experience Nonfluencies in Delivery Citation of Evidence Sources Position Advocated Liking for the Communicator Humor Summary Effects of Credibility Two Initial Clarifications Influences on the Magnitude of Effect Influences on the Direction of Effect 14



Liking The General Rule Some Exceptions and Limiting Conditions Liking and Credibility Liking and Topic Relevance Greater Effectiveness of Disliked Communicators Other Communicator Factors Similarity Similarity and Liking Similarity and Credibility: Expertise Judgments Similarity and Credibility: Trustworthiness Judgments Summary: The Effects of Similarity Physical Attractiveness Physical Attractiveness and Liking Physical Attractiveness and Credibility Summary About Additional Communicator Characteristics Conclusion The Nature of Communication Sources Multiple Roles for Communicator Variables For Review Notes 11 Message Factors Message Structure and Format Conclusion Omission Recommendation Specificity Narratives Complexities in Studying Narrative and Persuasion The Persuasive Power of Narratives Factors Influencing Narrative Persuasiveness Entertainment-Education Summary Prompts Message Content Consequence Desirability One-Sided Versus Two-Sided Messages Gain-Loss Framing Overall Effects Disease Prevention Versus Disease Detection Other Possible Moderating Factors Summary 15



Threat Appeals Protection Motivation Theory Threat Appeals, Fear Arousal, and Persuasion The Extended Parallel Process Model Summary Beyond Fear Arousal Sequential Request Strategies Foot-in-the-Door The Strategy The Research Evidence Explaining FITD Effects Door-in-the-Face The Strategy The Research Evidence Explaining DITF Effects Conclusion For Review Notes 12 Receiver Factors Individual Differences Topic-Specific Differences General Influences on Persuasion Processes Summary Transient Receiver States Mood Reactance Other Transient States Influencing Susceptibility to Persuasion Reducing Susceptibility: Inoculation, Warning, Refusal Skills Training Inoculation Warning Refusal Skills Training Increasing Susceptibility: Self-Affirmation Conclusion For Review Notes References Author Index Subject Index About the Author 16



Preface This preface is intended to provide a general framing of this book and is particularly directed to those who already have some familiarity with the subject matter. Such readers will be able to tell at a glance that this book is in many ways quite conventional (in the general plan of the work, the topics taken up, and so forth) and will come to see the inevitable oversimplifications, bypassed subtleties, elided details, and suchlike. Because this book is pitched at roughly the level of a graduateundergraduate course, it is likely to be defective both by having sections that are too shallow or general for some and by having segments that are too detailed or technical for others; the hope is that complaints are not too badly maldistributed across these two categories. This book aims at a relatively generalized treatment of persuasion; in certain contexts in which persuasion is a central or recurring activity, correspondingly localized treatments of relevant research literatures are available elsewhere, such as for consumer advertising (e.g., Armstrong, 2010) and for certain legal settings (e.g., Devine, 2012). Readers acquainted with the second edition will notice the addition of chapters concerning social judgment theory and stage models, revision of the treatment of the theories of reasoned action and planned behavior, and new attention to subjects such as reactance and the use of narratives as vehicles for persuasion. This edition also gives special attention to questions of message adaptation. One broad theme that recurs in theoretical treatments of persuasion is the need to adapt persuasive messages to their audiences: different recipients may be persuaded by different sorts of messages. Thus one way of approaching any given theoretical framework for persuasion is to ask how it identifies ways in which messages might be adapted to audiences. For this reason, a number of the chapters concerning theoretical perspectives contain a section addressing this issue (and, as appropriate, this matter also arises in other chapters). Some readers will see the relationship of this theme to concepts such as “message tailoring” and “message targeting.” In the research literature, these labels have often been used to apply quite loosely to any sort of way in which messages are adapted to (customized for) recipients, although sometimes there have been efforts to use different labels to describe different degrees or kinds of message customization (e.g., sometimes 17



“targeting” is described as adaptation on the basis of group-level characteristics, whereas “tailoring” is based on individual-level properties). But no matter the label, there is a common underlying conceptual thread here, namely, that different kinds of messages are likely to be persuasive for different recipients—and hence to maximize persuasiveness, messages should be adapted to their audiences. As should be apparent, there are quite a few different bases for such adaptation: messages might be adapted to the audience’s literacy level, cultural background, values, sex, degree of extroversion, age, regulatory focus, level of self-monitoring, or race/ethnicity. A message may be customized to the audience’s current psychological state as described by, say, reasoned action theory (e.g., is perceived behavioral control low?), protection motivation theory (is perceived vulnerability sufficiently high?), or the transtheoretical model (which stage is the recipient in?). It may be superficially personalized (e.g., by mentioning the recipient’s name in a direct mail appeal), mention shared attitudes not relevant to the advocacy subject, and so on. For this reason, it is not fruitful to pursue questions such as “are tailored messages more persuasive than non-tailored messages?” because the answer is virtually certain to be “it depends”—if nothing else, the answer may vary depending on the basis of tailoring. For example, it might be that adapting messages through superficial personalization typically makes very little difference to persuasiveness, but adapting messages by matching the message’s appeals to the audience’s core values could characteristically substantially enhance persuasiveness. Still, the manifest importance of adapting messages to recipients recommends its prominence. Aristotle was right (in the Rhetoric): the art of persuasion consists of discerning, in any particular situation, the available means of persuasion. Those means will vary from case to case, and hence maximizing one’s chances for persuasion will require adapting one’s efforts to the circumstance at hand. Whether one calls this message adaptation, message tailoring, message targeting, message customization, or something else, the core idea is the same: different approaches are required in different persuasive circumstances. Adding material (whether about audience adaptation or other matters) is an easy decision; omitting material is not, because one fears encouraging the loss of good (if imperfect) ideas. Someone somewhere once pointed out 18



that in the social and behavioral sciences, findings and theories often seem to just fade away, not because of any decisive criticisms or counterarguments but rather because they seem to be “too old to be true.” This apt observation seems to me to identify one barrier to social-scientific research synthesis, namely, that useful results and concepts somehow do not endure but rather disappear—making it impossible for subsequent work to exploit them. As an example: If message assimilation and contrast effects are genuine and have consequences for persuasive effects, then—although there is little research attention being given to the theoretical framework within which such phenomena were first clearly conceptualized (social judgment theory) —we need somehow to ensure that our knowledge of these phenomena does not evaporate. Similarly, although it has been some time since substantial work was done on the question of the dimensions underlying credibility judgments, the results of those investigations (the dimensions identified in those studies) should not thereby fail to be mentioned in discussions of credibility research. To sharpen the point here: It has been many years since the islets of Langerhans (masses of endocrine cells in the pancreas) were first noticed, but medical textbooks do not ignore this biological structure. Indeed, it would be inconceivable to discuss (for example) mechanisms of insulin secretion without mentioning these structures. Now I do not mean to say that social-scientific phenomena such as assimilation and contrast effects are on all fours with the islets of Langerhans, but I do want to suggest that premature disappearance of social-scientific concepts and findings seems to happen all too easily. Without forgetting how grumpy old researchers can sometimes view genuinely new developments (“this new phenomenon is just another name for something that used to be called X”), one can nevertheless acknowledge the real possibility that “old” knowledge can somehow be lost, misplaced, insufficiently understood, unappreciated, or overlooked. It is certainly the case that the sheer amount of social-scientific research output makes it difficult to keep up with current research across a number of topics, let alone hold on to whatever accumulated information there might be. In the specific case of persuasion research—which has seen an explosion of interest in recent years—the problem is not made any easier by the relevant literature’s dispersal across a variety of academic locales. Yet somehow the insights available from this research and theorizing must 19



not be lost. Unfortunately, there are not appealing shortcuts. One cannot simply reproduce others’ citations or research descriptions with an easy mind (for illustrations of the attendant pitfalls, see Gould, 1991, pp. 155–167; Gould, 1993, pp. 103–105; Tufte, 1997, p. 71). One hopes that it would be unnecessary to say that as in the previous editions, I have read everything I cite. I might inadvertently misrepresent or misunderstand, but at least such flaws will be of my own hand. Moreover, customary ways of drawing general conclusions about persuasive effects can be seen to have some important shortcomings. One source of difficulty here is a reliance on research designs using few persuasive messages, a matter addressed in Chapter 9. Here I will point out only the curiosity that generalizations about persuasive message effects— generalizations intended to be general across both persons and messages— have commonly been offered on the basis of data from scores or hundreds of human respondents but from only one or two messages. One who is willing to entertain seriously the possibility that the same manipulation may have different effects in different messages should, with such data in hand, be rather cautious. Another source of difficulty has been the widespread misunderstandings embedded in common ways of interpreting and integrating research findings in the persuasion literature. To illuminate the relevant point, consider the following hypothetical puzzle: Suppose there have been two studies of the effect on persuasive outcomes of having a concluding metaphor (versus having an ordinary conclusion that does not contain a metaphor) in one’s message, but with inconsistent results. In Study A, conclusion type made a statistically significant difference (such that greater effectiveness is associated with the metaphorical conclusion), but Study B failed to replicate this result. In Study A, the participants were female high school students who read a written communication arguing that most persons need from 7 to 9 hours of sleep each night. The message was attributed to a professor at the Harvard Medical School; the communicator’s identification, including a photograph of the professor (an attractive, youthful-looking man), was provided on a cover sheet immediately 20



preceding the message. The effect of conclusion type on persuasive outcome was significant, t(60) = 2.35, p < .05: Messages with a concluding metaphor were significantly more effective than messages with an ordinary (nonmetaphorical) conclusion. In Study B, the participants were male college undergraduates who listened to an audio message that used a male voice. The message advocated substantial tuition increases (of roughly 50% to 60%) at the students’ university and presented five arguments to show the necessity of such increases. The communicator was described as a senior at the university, majoring in education. Although the means were ordered as in Study A, conclusion type did not significantly affect persuasive outcome, t(21) = 1.39, ns. Why the inconsistency (the failure to replicate)? A typical inclination has been to entertain possible explanatory stories based on such differences as the receivers’ sex (“Women are more influenced by the presence of a metaphorical conclusion than are men”), the medium (“Metaphorical conclusions make more difference in written messages than in oral messages”), the advocated position (“Metaphorical conclusions are helpful in proattitudinal messages but not in counterattitudinal ones”), and so on. But for this hypothetical example, those sorts of explanatory stories are misplaced. Not only is the direction of effect identical in Study A and Study B (each finds that the concludingmetaphor message is more effective) but also the size of the advantage enjoyed by the concluding-metaphor message is the same in the two studies (expressed as a correlation, the effect size is .29). The difference in the level of statistical significance achieved is a function of the difference in sample size, not any difference in effect size. Happily, recent years have seen some progress in the diffusion of more careful understandings of statistical significance, effect sizes, statistical power, confidence intervals, and related matters. (Some progress—but not enough. It remains distressingly common that even graduate students with statistical training can reason badly when faced with a problem such as that hypothetical.) With the hope of encouraging greater sensitivity concerning specifically the magnitude of effects likely to be found in persuasion research, I have tried to include mention of average effect sizes where appropriate and available. 21



But there is at present something of a disjuncture between the available methods for describing research findings (in terms of effect sizes and confidence intervals) and our theoretical equipment for generating predictions. Although research results can be described in specific quantitative terms (“the correlation was .37”), researchers are currently prepared to offer only directional predictions (“the correlation will be positive”). Developing more refined predictive capabilities is very much to be hoped for, but significant challenges lie ahead (for some discussion, see O’Keefe, 2011a). Even with increasing attention to effect sizes and their meta-analytic treatment, however, we are still not in a position to do full justice to the issues engaged by the extensive research literature in persuasion, given the challenges in doing relevant, careful, reflective research reviews. For example, research reviews all too often exclude unpublished studies, despite wide recognition of publication biases favoring statistically significant results (see, e.g., Dwan, Gamble, Williamson, Kirkham, & the Reporting Bias Group, 2013; Ferguson & Heene, 2012; Ioannidis, 2005, 2008). Similarly, meta-analytic reviews too often rely on fixed-effect analyses rather than the random-effects analyses appropriate where generalization is the goal (for discussion, see Card, 2012, pp. 233–234). All these considerations conspire to encourage a rather conservative approach to the persuasion literature (conservative in the sense of exemplifying prudence with respect to generalization), and that has been the aim in this treatment. Of course, one cannot hope to survey the range of work covered here without errors, oversights, and unclarities. These have been reduced by advice and assistance from a number of quarters. Students in my persuasion classes have helped make my lectures—and so this book— clearer than otherwise might have been the case. Many good insights and suggestions came from the reviewers arranged by Sage Publications: Jonathan H. Amsbary, William B. Collins, Julia Jahansoozi, Bonnie Kay, Andrew J. Kirk, Susan L. Kline, Sanja Novitsky, Charles Soukup, Kaja Tampere, and Beth M. Waggenspack. Jos Hornikx also provided especially useful commentary on drafts of this edition’s chapters. And I thank Barbara O’Keefe both for helpful conversation and for an unceasingly interesting life: “Age cannot wither her, nor custom stale / Her infinite variety.”



22



Chapter 1 Persuasion, Attitudes, and Actions The Concept of Persuasion About Definitions: Fuzzy Edges and Paradigm Cases Five Common Features of Paradigm Cases of Persuasion A Definition After All? The Concept of Attitude Attitude Measurement Techniques Explicit Measures Quasi-Explicit Measures Implicit Measures Summary Attitudes and Behaviors The General Relationship Moderating Factors Encouraging Attitude-Consistent Behavior Assessing Persuasive Effects Attitude Change Beyond Attitude Change Conclusion For Review Notes



This book surveys social-scientific theory and research concerning persuasive communication. The relevant work, as will become apparent, is scattered across the academic landscape—in communication, psychology, advertising, marketing, political science, law, and so on. Although the breadth and depth of this literature rule out a completely comprehensive and detailed treatment, the main lines of work are at least sketched here. This introductory chapter begins, naturally enough, with a discussion of the concept of persuasion. But because social-scientific treatments of persuasion have closely linked persuasion and attitude change, the concept of attitude is discussed as well, some common attitude assessment procedures are described, and the relationship of attitudes and behavior is considered; a concluding section discusses the assessment of persuasive effects. 23



The Concept of Persuasion About Definitions: Fuzzy Edges and Paradigm Cases A common way to clarify a concept is to provide a definition of the concept. But definitions can be troublesome things, precisely because they commonly are treated as providing sharp-edged distinctions between what is included in the category and what is not. What is troublesome about such sharp lines is that no matter where they are drawn, it is possible to sustain objections to their location; for some the definition will be too broad and for others too narrow. Definitions are almost inevitably open to such criticisms, no matter where the definitional lines are drawn, because most concepts have fuzzy edges, that is, gray areas in which application of the concept is arguable. For any concept, there are some cases that virtually everyone agrees are cases of the concept (few would deny that a chair is an instance of the category “furniture”), and there are some cases that virtually everyone agrees are not cases of the concept (a pencil is not an instance of furniture)—but there are also some cases that fall in a gray area and can give rise to disagreements (is a television set a piece of furniture? or perhaps is it an appliance?). No matter how the line is drawn, some objection is possible. So, for example, if one defines persuasion in such a way as to distinguish cases of persuasion from cases of manipulation by requiring that in genuine instances of persuasion, the persuader “acts in good faith” (as do Burnell & Reeve, 1984), then some will object that the definition is too narrow; after all, such a definition almost certainly excludes at least some instances of advertising. But including manipulation as instances of persuasion will meet objections from those who think it important to exclude instances of sheer manipulation from the definition of persuasion. Happily, it is possible to clarify a concept without having to be committed to a sharp-edged definition of the concept (and thus without having to settle such border disputes). Such clarification can be obtained by focusing on the shared features of paradigm cases of the concept. Paradigm cases of a concept are the sorts of instances that nearly everyone would agree were instances of the concept in question; they are straightforward, uncontroversial examples. By identifying the common features of paradigm cases, one can get a sense of the concept’s ordinary central 24



application, without having to draw sharp-edged definitional lines.



Five Common Features of Paradigm Cases of Persuasion Consider, then: What is ordinarily involved when we say that someone (a persuader) has persuaded someone else (a persuadee)? In such straightforward applications of the concept of persuasion, what sorts of shared features can be observed? (For an alternative to the following analysis, see Gass & Seiter, 2004.) First, when we say that one person persuaded another, we ordinarily identify a successful attempt to influence. That is, the notion of success is embedded in the concept of persuasion. For instance, it does not make sense to say, “I persuaded him but failed.” One can say, “I tried to persuade him but failed,” but to say simply “I persuaded him” is to imply a successful attempt to influence.1 Second, in paradigm cases of persuasion, the persuader intends to influence the persuadee. For example, if I say, “I persuaded Sally to vote for Jones,” you are likely to infer that I intended to obtain that effect. For just that reason, it is entirely understandable that someone might say, “I accidentally persuaded Mary to vote for Brown” precisely in the circumstance in which the speaker does not want a hearer to draw the usual inference of intent; absent such mention of accident, the ordinary inference will be that the persuasion was purposeful. A third feature shared by paradigm cases of persuasion is some measure of freedom (free will, free choice, voluntary action) on the persuadee’s part. Consider, for example, a circumstance in which a person is knocked unconscious by a robber, who then takes the victim’s money; one would not (except humorously) say that the victim had been “persuaded” to give the money. By contrast, being induced by a television ad to make a donation to a charitable cause is obviously an instance of persuasion. When the persuadee’s freedom is minimized or questionable, it becomes correspondingly questionable whether persuasion is genuinely involved; one no longer has a straightforward exemplary case of persuasion. Suppose a robber threatens to shoot the victim if the money is not forthcoming, and the victim complies: Is this an instance of persuasion? We need not settle this question here, as it requires a sharp line of 25



definition that we are avoiding.2 It is enough to notice that such cases are borderline instances of persuasion, precisely because the persuadee’s freedom is not so clear-cut as in paradigm instances. Fourth, paradigm cases of persuasion are ones in which the effects are achieved through communication (and perhaps especially through the medium of language). My physically lifting you and throwing you off the roof of a building is something quite different from my talking you into jumping off the same roof; the latter might possibly be a case of persuasion (depending on the circumstances, exactly what I have said to you, and so on), but the former is certainly not. What distinguishes these two instances is that communication is involved in the latter case but not in the former. Finally, paradigm cases of persuasion involve a change in the mental state of the persuadee (principally as a precursor to a change in behavior). Some ordinary instances of persuasion may be described as involving only a change in mental state (as in “I persuaded Jan that the United States should refuse to recognize the authority of the World Court”). But even when behavioral change is involved (as in “I persuaded Tom to take golf lessons”), there is ordinarily presumed to be some underlying change in mental state that gave rise to the behavioral change (e.g., Tom came to believe that his golf skills were poor, that his skills could be improved by taking lessons, etc.). Thus even when a persuader’s eventual aim is to influence what people do (how they vote or what products they buy), at least in paradigm cases of persuasion that aim is ordinarily accomplished by changing what people think (what they think of the political candidate or of the product). That is, persuasion is ordinarily conceived of as influencing others by influencing their mental states (rather than by somehow influencing their conduct directly). In persuasion theory and research, the relevant mental state has most commonly been characterized as an attitude (and thus the concept of attitude receives direct discussion later in this chapter).3 Even when a persuader’s ultimate goal is the modification of another’s behavior, that goal is often seen to be achieved through a process of attitude change—the presumption being that attitude change is a means of behavioral change.



A Definition After All? These shared features of exemplary cases of persuasion can be strung 26



together into something that looks like a definition of persuasion: a successful intentional effort at influencing another’s mental state through communication in a circumstance in which the persuadee has some measure of freedom. But it should be apparent that constructing such a definition would not eliminate the fuzzy edges of the concept of persuasion. Such a definition leaves open to dispute just how much success is required, just how intentional the effort must be, and so on. Hence, by recognizing these shared features of paradigm cases of persuasion, one can get a sense of the central core of the concept of persuasion, but one need not draw sharp definitional boundaries around that concept. Indeed, these paradigm case features permit one to see clearly just how definitional disputes can arise—for instance, disputes about the issue of just how much, and what sorts, of freedom the persuadee must have before an instance qualifies as an instance of persuasion. It is also easy to see that there can be no satisfactory definitive solution to these disputes, given the fuzzy edges that the concept of persuasion naturally has. Definitions of persuasion can serve useful functions, but a clear sense of the concept of persuasion can be had without resorting to a hard-edged definition.



The Concept of Attitude As mentioned above, the mental state that has been seen (in theory and research) to be most centrally implicated in persuasion is that of attitude. The concept of attitude has a long history (see D. Fleming, 1967). Early uses of the term “attitude” referred to posture or physical arrangement (as in someone’s being in “the attitude of prayer”), uses that can be seen today in descriptions of dance or airplane orientation. Gradually, however, attitudes came to be seen as “orientations of mind” rather than of body, as internal states that exerted influence on overt behavior. Perhaps it was inevitable, thus, that in the early part of the 20th century, the emerging field of social psychology should have seized on the concept of attitude as an important one. Attitude offered to social psychologists a distinctive psychological mechanism for understanding and explaining individual variation in social conduct (Allport, 1935). And although for a time there was considerable discussion of alternative definitions of attitude (e.g., Audi, 1972; Eagly & Chaiken, 1993, pp. 1–21; McGuire, 1985), a broad consensus emerged that an attitude is a person’s general evaluation of an object (where “object” is understood in a broad sense, as 27



encompassing persons, events, products, policies, institutions, and so on). Even when conceptual treatments of attitude differ in other ways, a common theme is that an attitude is an evaluative judgment of (reaction to) an object (Fishbein & Ajzen, 2010, pp. 75–79). Understood this way, it is perhaps obvious why attitude should so often be a mental state of interest to persuaders. What products people buy, which candidates they vote for, which policies they endorse, what hobbies they pursue, which businesses they patronize—influencing such things will often involve influencing people’s attitudes. Precisely because attitudes represent relatively stable evaluations that can influence behavior, they are a common persuasive target.



Attitude Measurement Techniques If persuasion is conceived of as fundamentally involving attitude change, then the systematic study of persuasion requires means of assessing persons’ attitudes. A great many attitude measurement techniques have been proposed, and a large literature addresses the use of attitude measures in specific circumstances such as public opinion polling and survey research. The intention here is to give a brief overview of some exemplary attitude measurement procedures (for more detailed information and reviews, see Banaji & Heiphetz, 2010, pp. 359–370; Krosnick, Judd, & Wittenbrink, 2005; Schwarz, 2008). Attitude assessment procedures can be usefully distinguished by the degree of explicitness (directness) with which they assess the respondent’s evaluation of the attitude object. Some techniques directly obtain an evaluative judgment; others do so in more roundabout ways.



Explicit Measures Explicit attitude measurement techniques directly ask the respondent for an evaluative judgment of the attitude object. Two commonly employed explicit assessment procedures are semantic differential evaluative scales and single-item attitude questions.



Semantic Differential Evaluative Scales One popular means of directly assessing attitude is to employ the 28



evaluative scales from the semantic differential scale of Osgood, Suci, and Tannenbaum (1957). In this procedure, respondents rate the attitude object on a number of (typically) 7-point bipolar scales that are end-anchored by evaluative adjective pairs (such as good-bad, desirable-undesirable, and so forth). An example appears in Figure 1.1. The instructions for this scale ask the respondent to place a check mark at the point on the scale that best represents the respondent’s judgment. The investigator can straightforwardly assign numerical values to the scale points (say, +3 for the extreme positive point, through 0 for the midpoint, to −3 for the extreme negative end) and then sum each person’s responses to obtain an indication of the person’s attitude toward (general evaluative judgment of) the object. Figure 1.1 Example of a semantic differential scale.



Single-Item Attitude Measures Another explicit means of assessing attitude is simply to have the respondent complete a single questionnaire item that asks for the relevant judgment (see Figure 1.2 for an example). There are various ways of wording the question and of anchoring the scale (e.g., “In general, how much do you like the United Nations?” with end anchors “very much” and “not at all”), and it is possible to vary the number of scale points, but the basic procedure is the same. A single-item attitude measure familiar to U.S. survey researchers is the “feeling thermometer,” which asks respondents to report their evaluation on a scale akin to a Fahrenheit thermometer; the endpoints of the scale are zero degrees (very “cold” or unfavorable feelings) and 100 degrees (very “warm” or favorable feelings; see, e.g., Alwin, 1997). Figure 1.2 Example of a single-item attitude measure.



A single-item attitude measure is an understandably attractive technique for circumstances such as public opinion polling. The attitude assessment 29



can be undertaken orally (as in telephone surveys or face-to-face interviewing); the question is typically straightforward and easily comprehended by the respondent; the question can be asked (and answered) in a short time. The central drawback of single-item assessments of attitude is potentially weak reliability. That is, a person’s response to a single attitude question may not be as dependable an indicator of attitude as the person’s response to three or four items all getting at roughly the same thing.



Features of Explicit Measures Explicit attitude measurement techniques obviously offer the advantage of being simple and straightforward, easy to administer, and so forth. Another advantage of these techniques is that they are relatively easy to construct. For instance, a public opinion survey of attitudes toward possible presidential candidates can easily accommodate some new possible candidate: The surveyor simply asks the standard question but inserts the name of the new candidate. General evaluative scales from the semantic differential can be used for rating all sorts of attitude objects (consumer products, political candidates, government policies, etc.); to assess attitudes toward Crest toothpaste rather than toward the United Nations, one simply makes the appropriate substitution above the rating scales. (This may be a false economy, however: for arguments emphasizing the importance of customizing semantic differential evaluative scales for each different attitude object, see Fishbein & Ajzen, 2010, pp. 79–82.) One salient disadvantage of these explicit techniques is that because they are so direct, they yield an estimate only of the respondent’s attitude. Of course, this is not a drawback if all the researcher wants to know is the respondent’s attitude. But investigators will often want other information as well (about, for example, beliefs that might lie behind the attitude), and in such circumstances, direct attitude assessment techniques will need to be supplemented or replaced by other procedures.



Quasi-Explicit Measures Quasi-explicit attitude measurement techniques assess attitude not by directly eliciting an evaluative judgment of the attitude object but by eliciting information that is obviously attitude-relevant and that offers a straightforward basis for attitude assessment. For example, paired30



comparison procedures and ranking techniques do not ask directly for an evaluation of any single attitude object but ask for comparative judgments of several objects. In a paired-comparison technique, the respondent is asked a series of questions about the relative evaluation of each of a number of pairs of objects (e.g., “Which candidate do you prefer, Archer or Barker? Archer or Cooper? Barker or Cooper?”); in a ranking procedure, the respondent ranks a set of attitude objects (e.g., “Rank these various leisure activities, from your most favorite to your least favorite”). The obtained responses obviously permit an investigator to draw some conclusions about the respondent’s evaluation of a given object. The two most common and well-known quasi-explicit attitude measurement procedures are those devised by Louis Thurstone and by Rensis Likert. In their procedures, the respondent’s attitude is inferred from agreement or disagreement with statements that are rather obviously attitude-relevant. The attitude assessment instrument, then, consists of statements to which the respondent reacts (by agreeing or disagreeing with each one), and the respondent’s attitude is inferred from the pattern of responses. Obviously, however, if a researcher is going to gauge respondents’ attitudes by examining respondents’ reactions to a set of statements, not just any statements will do; for example, one is not likely to learn much about attitudes toward the United Nations by assessing persons’ agreement with a statement such as “Baseball is a better game than football.” Thus the task faced in constructing a Thurstone or Likert attitude scale is the task of selecting items (statements) that appear to serve as suitable indicators of attitude. One may start with a large pool of statements that might possibly be included on a final attitude instrument, but the problem is to somehow winnow that pool down. This winnowing is accomplished by gathering and analyzing data about respondents’ reactions to a large number of possible items. Detailed descriptions of these procedures are available elsewhere (for some specifics, see Green, 1954; Likert, 1932; Thurstone, 1931), but the key is to identify those items (statements) that can dependably be taken as indicators of attitudes. For example, if the topic of investigation concerns attitudes toward the First Federal Bank, suitable statements might turn out to be ones such as “This bank is reliable,” “This bank is inefficient,” “This bank has unfriendly personnel,” and so on; by contrast, a statement such as “This bank has a branch at the corner of Main and Elm” would be unlikely 31



to be included (because knowing whether a respondent agreed or disagreed with such a statement would not provide information about the respondent’s attitude). Given a set of suitable statements, one elicits respondents’ agreement with each statement. This can be accomplished in various ways. For example, respondents can be given a list of statements and asked to check the ones with which they agree (this is Thurstone’s, 1931, procedure). Or the strength of agreement with each statement can be assessed through some appropriate scale (this is Likert’s, 1932, procedure; see Figure 1.3). An overall attitude score can then be obtained straightforwardly for each respondent, to serve as estimate of that person’s overall attitude. Figure 1.3 Example of an item on a Likert quasi-explicit attitude measure.



There is a good deal of variation in quasi-explicit attitude assessment techniques, but as a rule, these procedures provide more information than do explicit attitude measurement techniques. For example, when a Thurstone or Likert scale has been employed, a researcher can see what specific items were especially likely to be endorsed by respondents with particular attitudes; an investigator who finds, for instance, that those with unfavorable attitudes toward the bank very often agreed with the statement that “this bank has unfriendly personnel” may well have learned about a possible cause of those negative attitudes. Similarly, ranking techniques can give information about a large number of attitudes and so provide insight about comparative evaluations. Precisely because quasi-explicit procedures involve acquiring attitude-relevant information (rather than the attitude itself), these procedures offer information not available with explicit measurement techniques. But this additional information is obtained at a cost. Thurstone and Likert attitude scales have to be constructed anew for each attitude object; obviously, one cannot use the First Federal Bank attitude scale to assess attitudes toward other objects. (Indeed, the substantial effort needed to obtain a sound Thurstone or Likert scale is often a deterrent to the use of such techniques.) Procedures such as paired-comparison ratings or ranking tasks may take more time to administer than would direct attitude measures.



32



Implicit Measures Explicit and quasi-explicit measures are overwhelmingly the most common ways of measuring attitudes. But a variety of other techniques have been developed that assess attitude not by directly eliciting an evaluation of the attitude object or even by eliciting information obviously relevant to such an overall evaluation but instead by some more roundabout (implicit, indirect) means. Quite a few different implicit measures of attitude have appeared (for collections and general discussions, see De Houwer, Teige-Mocigemba, Spruyt, & Moors, 2009; Goodall, 2011; Petty, Fazio, & Briñol, 2009a; Wittenbrink & Schwarz, 2007). These include physiological indices, such as autonomic responses (e.g., heart rate) and measures of brain activity (for general reviews, see Cunningham, Packer, Kesek, & Van Bavel, 2009; Ito & Cacioppo, 2007); priming measures, in which attitudes are assessed by examining the speed (reaction time) with which people make evaluative judgments when those judgments are preceded (primed) by the attitude object (for a review, see Wittenbrink, 2007); the Implicit Association Test (IAT), in which attitudes are assessed by examining the strength of association (as measured by reaction time) between attitude objects and evaluative categories (for reviews and discussion, see Fiedler, Messner, & Bluemke, 2006; Greenwald, Poehlman, Uhlmann, & Banaji, 2009; Lane, Banaji, Nosek, & Greenwald, 2007; Oswald, Mitchell, Blanton, Jaccard, & Tetlock, 2013); and a variety of others (for examples and discussion, see Kidder & Campbell, 1970; Tykocinski & Bareket-Bojmel, 2009). What is common to all implicit measures is that it is generally not obvious to respondents that their attitudes are being assessed. For that reason, implicit measures are likely to be most attractive in circumstances in which one fears respondents may, for whatever reason, distort their true attitudes. In most research on persuasion, however, these circumstances are rather uncommon (respondents are ensured anonymity, message topics are generally not unusually sensitive ones, etc.); consequently, implicit attitude measures are rarely employed (for examples and discussion, see Briñol, Petty, & McCaslin, 2009; Hefner, Rothmund, Klimmt, & Gollwitzer, 2011; Maio, Haddock, Watt, & Hewstone, 2009).4



Summary



33



As this survey suggests, a variety of attitude measurement techniques are available. The overwhelmingly most frequently used attitude measurement procedures are explicit or quasi-explicit techniques; reliability and validity are more readily established for attitude measures based on these techniques than for measures derived from implicit procedures. Explicit procedures are often preferred over quasi-explicit techniques because of the effort required for constructing Thurstone or Likert scales. But which specific attitude assessment procedure an investigator employs in a given instance will depend on the particulars of the situation. Depending on what the researcher wants to find out, the time available to prepare the attitude assessment, the time available to question respondents, the sensitivity of the attitude topic, and so forth, different techniques will recommend themselves.



Attitudes and Behaviors The General Relationship Attitude has been taken to be a key mental state relevant to persuasion because of a presumed relationship between attitudes and actions. The assumption has been that attitudes are important determinants of behavior and, correspondingly, that one avenue to changing a person’s behavior will be to change that person’s attitudes.5 This assumption is generally wellfounded: A number of systematic reviews have found that attitudes and behaviors are commonly reasonably consistent (for some reviews, see Eckes & Six, 1994; Glasman & Albarracín, 2006; M.-S. Kim & Hunter, 1993a; Kraus, 1995).6



Moderating Factors The degree of attitude-behavior consistency has been found to vary depending on other “moderating” factors—factors that moderate or influence the relationship between attitudes and behaviors. A large number of possible moderating variables have been explored, including the degree to which the behavior is effortful or difficult (Kaiser & Schultz, 2009; Wallace, Paulson, Lord, & Bond, 2005); the perceived relevance of the attitude to the behavior (Snyder, 1982; Snyder & Kendzierski, 1982); attitude accessibility (Smith & Terry, 2003); attitudinal ambivalence (Conner et al., 2002; Jonas, Broemer, & Diehl, 2000); having a vested 34



interest in a position (Crano & Prislin, 1995); the extent of attituderelevant knowledge (Fabrigar, Petty, Smith, & Crites, 2006); and many others. In what follows, two well-studied factors are discussed as illustrative: the correspondence between the attitudinal and behavioral measures, and the degree of direct experience with the attitude object.



Correspondence of Measures One factor that influences the observed consistency between an attitudinal measure and a behavioral measure is the nature of the measures involved. Good evidence indicates that substantial attitude-behavior correlations will be obtained only when the attitudinal measure and the behavioral measure correspond in specificity (Ajzen & Fishbein, 1977). A general attitude will probably not be especially strongly correlated with any one particular specific behavior. A general attitude measure corresponds to a general behavioral measure, not to a specific one. For example, general attitudes toward religion might or might not be strongly correlated with performance of the particular act of (say) reading books about religious philosophy. But attitudes toward religion may well be strongly correlated with a general religious behavior index—an index based on multiple behaviors (whether the person reads books about religious philosophy, attends religious services, watches or listens to religious programs, owns religious music, donates money to religious institutions, consults clergy about personal problems, and so on). No one of these behaviors may be very strongly predicted by religious attitude, but the overall pattern of these behaviors might well be associated with religious attitude. That is, although the correlation of the general attitude with any one of these behaviors might be relatively small, the correlation of the general attitude with a multiple-act behavioral measure may be much greater.7 Several investigations have yielded information about the relative strength of the attitude-behavior association when single-act and multiple-act behavioral measures are predicted on the basis of general attitudes. In these studies, the average correlation between general attitude and any single-act index of behavior was roughly .30; by contrast, the average correlation between general attitude and a multiple-act behavioral measure was approximately .65 (Babrow & O’Keefe, 1984; Fishbein & Ajzen, 1974; O’Keefe & Shepherd, 1982; Sjoberg, 1982; Weigel & Newman, 1976; see also Bamberg, 2003). These findings plainly indicate that 35



attitudinal measures and behavioral measures are likely to be rather more strongly associated when there is substantial correspondence between the two measures and underscore the folly of supposing that a single specific behavior will necessarily or typically be strongly associated with a person’s general attitude (for some relevant reviews, see Ajzen & Cote, 2008; Eckes & Six, 1994; M.-S. Kim & Hunter, 1993a; Kraus, 1995). Correspondingly, these findings underscore the importance of carefully considering the focus of persuasive efforts. For instance, to encourage participation in a community recycling program, it might seem natural to construct persuasive messages aimed at inducing favorable attitudes toward protecting the environment. But this is not likely to be a particularly efficient persuasive strategy. Even if the messages succeed in producing positive environmental protection attitudes, those general attitudes may not be especially strongly associated with the specific behavior that is wanted (recycling program participation). A more effective focus for persuasive efforts might well be specific attitudes toward participation in the recycling program, rather than general environmental attitudes.8



Direct Experience A second factor influencing attitude-behavior consistency is the degree of direct experience with the attitude object. Attitudes based on direct behavioral experience with the attitude object have been found to be more predictive of later behavior toward the object than are attitudes based on indirect experience. (For some examples and discussion, see Doll & Ajzen, 1992; Doll & Mallu, 1990; Eagly & Chaiken, 1993, pp. 194–200; Glasman & Albarracín, 2006; Kraus, 1995; Steffen & Gruber, 1991. For some complexities, see Millar & Millar, 1998.) For example, during a housing shortage at Cornell University—wellpublicized on campus—some new students had to be placed in temporary quarters (thus giving them firsthand experience with the problem); other new students were given permanent dormitory rooms (and so knew of the problem less directly). The two groups had equally negative attitudes regarding the housing crisis, but the strength of the attitude-behavior relationship differed. Those whose attitudes were formed on the basis of direct experience exhibited greater consistency between their attitudes and behaviors aimed at alleviating the crisis than did students whose attitudes were based on indirect experience (Regan & Fazio, 1977). 36



A similar effect was observed in a study comparing attitude-behavior consistency for product attitudes that were based either on a trial experience with a sample of the product (direct experience) or on exposure to advertising messages about the product (indirect experience). Much greater attitude-behavior consistency was observed for those persons who had had the opportunity to try the product than for those who had merely read about it. For example, purchase of the product was more highly correlated with attitudes based on product trial (.57) than with attitudes based on product advertising (.18) (R. E. Smith & Swinyard, 1983). This finding does not mean that product trial influence strategies (e.g., providing free samples through the mail, offering grocery store shoppers a taste of a new food product, etc.) will necessarily be more effective in producing sales than will advertising strategies: Direct experience strengthens both positive and negative attitudes. The shopper who has a negative attitude toward a food product because of having read about it might still come to purchase the product; the shopper whose negative attitude is based on tasting the product, however, is much less likely to do so. In short, attitudes induced by direct experience will be more strongly correlated with behavior than attitudes induced by indirect experience. Two persons may have equally positive attitudes but may differ in whether they act consistently with those attitudes because of underlying differences in the ways in which the attitudes were formed.



Summary Research has examined a great many possible moderators of attitudebehavior consistency (for some general discussions, see Ajzen & Sexton, 1999; Eagly & Chaiken, 1993, pp. 193–215; Fazio & Roskos-Ewoldsen, 2005; Fazio & Towles-Schwen, 1999; Glasman & Albarracín, 2006; Wallace, Paulson, Lord, & Bond, 2005). The two mentioned here, although relatively prominent, are only illustrative.



Encouraging Attitude-Consistent Behavior Sometimes a persuader’s challenge is not so much to change a person’s attitude but to get that person to act on their attitude. For example, it is not enough to convince people to have favorable attitudes toward good health (indeed, they probably already have such attitudes); what’s needed is to 37



convince people to make attitude-consistent behavioral choices about exercise, diet, medical care, and the like. Similarly, persons who express positive attitudes toward energy conservation and environmental protection may nevertheless need to be induced to act consistently with those views—to engage in recycling, consider packaging considerations when buying products, choose appropriate thermostat settings, and so on. Thus the question arises of how persuaders might approach such tasks. At least three related strategies can be identified.



Enhance Perceived Relevance One strategy for enhancing attitude-behavior consistency is to encourage people to see their attitudes as relevant to their behavioral choices; people are likely to act more consistently with their attitudes when they do. For example, in a study by Snyder and Kendzierski (1982), participants were undergraduates known to have attitudes favorable to psychological research; they were asked to volunteer to participate in extra sessions of a psychology experiment. This was an especially demanding request (involving returning on different days, at inconvenient times, and so on). Indeed, in the control condition only 25% of the participants volunteered, despite their favorable attitudes. Before responding to the request, each participant overheard a conversation between two other students (confederates of the experimenters) who were discussing the request. The first student said, “I don’t know if I should volunteer or if I shouldn’t volunteer. What do you think?” In the control condition, the second student responded, “Beats me—it’s up to you.” In the experimental condition, the response was, “Well, I guess that whether you do or whether you don’t is really a question of how worthwhile you think experiments are”—a response designed to underscore the relevance of attitudes toward psychological research as guides for decision making in this situation. Although only 25% of the control condition participants agreed to volunteer, 60% of the experimental condition participants agreed. Obviously, then, one means of influencing behavior is the strategy of emphasizing the relevance of an existing attitude to a current behavioral choice. Little systematic research evidence concerns this strategy (see Borgida & Campbell, 1982; Prislin, 1987; Shepherd, 1985; Snyder, 1982), but as a testimony to the strategy’s potential effectiveness, consider the many products (DVDs, computer programs, tutoring sessions, and so on) purchased by parents who were prodded by sellers asking, “You want your 38



children to have a good education, don’t you? To have an edge in school? To get ahead in life?” Fundamentally, these questions reflect the seller’s understanding that enhancing the perceived relevance of an attitude to an action can be a means of increasing attitude-behavior consistency.



Induce Feelings of Hypocrisy A second strategy for encouraging attitude-behavior consistency can be appropriate in situations in which people have previously acted inconsistently with their attitudes: the strategy of hypocrisy induction. As discussed more thoroughly in Chapter 5 (concerning cognitive dissonance theory), a number of studies suggest that when persons have been hypocritical (in the sense of believing one thing but doing something different), one way of encouraging attitude-consistent behavior can be to draw persons’ attention to the hypocrisy (Stone, 2012). Specifically, when both the existing attitude and the previous inconsistency are made salient, persons are likely subsequently to act more consistently with their attitudes. For example, Stone, Aronson, Crain, Winslow, and Fried (1994) varied the salience of participants’ positive attitudes about safe sex practices (by having some participants write and deliver a speech about the importance of safe sex) and varied the salience of previous behavior that was inconsistent with such attitudes (by having some participants be reminded of their past failures to engage in safe sex practices, through having to list circumstances surrounding their past failures to use condoms). The combination of salient attitudes and salient inconsistency induced greater subsequent attitude-behavior consistency (reflected in greater likelihood of buying condoms, and buying more condoms, at the end of the experiment) than either one alone. Thus one means of inducing attitude-behavior consistency may be to lead people to recognize their hypocrisy.9



Encourage Anticipation of Feelings A third strategy for enhancing attitude-behavior consistency is to invite people to consider how they will feel if they fail to act consistently with their attitudes. As discussed more extensively in Chapter 6 (concerning reasoned action theory), feelings of anticipated emotions such as regret and guilt can shape people’s behavioral choices—and hence one way of influencing such choices is precisely by activating such anticipated feelings. A number of studies have influenced the salience of anticipated 39



emotions simply by asking about such feelings, with consequent effects on intention or behavior. For example, Richard, van der Pligt, and de Vries (1996b) asked people either to indicate how they would expect to feel after having unprotected sex (by rating the likelihood of experiencing various positive and negative emotions) or to indicate how they felt about having unprotected sex (using similar ratings). Those participants whose attention was drawn to their anticipated feelings were more likely to intend to use condoms (and subsequently were more consistent condom users) than the other participants. Such results plainly suggest that making salient the emotion-related consequences of contemplated attitude-inconsistent behavior may have the effect of enhancing attitude-behavior consistency.



Summary These three strategies all seek to tap some general desire for consistency as a way of influencing behavior in a circumstance in which persons will have an opportunity to act consistently with some existing attitude. But the strategies vary in the means of engaging that motivation. The perceivedrelevance strategy amounts to saying, “You might not have realized it, but this really is an opportunity to act consistently with your attitude.” The hypocrisy-induction strategy says, in effect, “You haven’t been acting consistently with your attitude, but here is an opportunity to do so.” The anticipated-feelings strategy implicitly says, “Here is an opportunity to act consistently with your attitude—and think how bad you’ll feel if you don’t.”10



Assessing Persuasive Effects Attitude Change Attitude measurement procedures obviously provide means of assessing persuasive effects. To see whether a given message changes attitudes, an investigator can assess attitudes before and after exposure to the message (perhaps also collecting parallel attitude assessments from persons not exposed to the message, as a way of reducing ambiguity about the potential causes of any observed changes). Indeed, such attitude assessment procedures are the most common ones used in studies of persuasive effects. The concrete realizations of attitude assessment may vary depending on the particulars of the research design; for example, in an experiment in which participants are randomly assigned to conditions, 40



one might dispense with the premessage attitude assessment and examine only postmessage differences, on the assumption that random assignment makes substantial initial differences unlikely. But effects on attitude are the effects most frequently considered in persuasion research.



Beyond Attitude Change Although attitude has historically been considered the key mental state relevant to persuasive effects, attitudes are not the only possible focus for persuasive efforts. Obviously, when other psychological states are of interest, other assessments will be useful or necessary. (For a general discussion, see Rhodes & Ewoldsen, 2013.) Sometimes the focus of a persuasive effort will be some determinant of attitude, such as a particular belief about the attitude object. For example, an advertising campaign might try to persuade people that a product is environmentally friendly (as a means of influencing persons’ attitudes toward the product and, eventually, product purchase). The appropriate assessment of the campaign’s persuasive effectiveness would involve changes in that specific belief about the product, not changes in the overall attitude toward the product. The belief that the product is environmentally friendly might well influence the overall attitude, but to see whether the target belief is changed by the persuasive effort, assessments of that belief will be needed.11 Sometimes persuaders want to influence some property of an attitude other than its valence (positive or negative) and extremity. That is, rather than influencing whether (or the degree to which) an attitude is positive or negative, a persuader might want to influence the salience (prominence, accessibility) of the attitude, the confidence with which it is held, the degree to which it is linked to other attitudes, and so forth (for discussions of some such attitudinal properties, see Conner & Armitage, 2008; Fabrigar, MacDonald, & Wegener, 2007; Petty, Briñol, Tormala, & Wegener, 2007, pp. 260–262; van Harreveld, Schneider, Nohlen, & van der Pligt, 2012; Visser, Bizer, & Krosnick, 2006). For example, when consumers already have positive attitudes toward one’s product, the persuasive task may be to ensure that those attitudes are salient (activated) at the right time, perhaps by somehow reminding people of their attitudes. Some such attitudinal properties have sometimes been grouped together under the general heading of “attitude strength” (for some discussions, see 41



Petty & Krosnick, 1995; Visser, Bizer, & Krosnick, 2006). Conceptualizations of attitude strength vary, but (as an example) Krosnick and Petty (1995) proposed that attitude strength is best understood as an amalgam of persistence (stronger attitudes are more persistent than are weaker ones), resistance (stronger attitudes are more resistant to change than are weaker ones), impact on information processing and judgments (stronger attitudes are more likely to affect such processes than are weaker attitudes), and impact on behavior (stronger attitudes will have more effect on behavior than will weaker ones). It should be apparent that persuaders might have an interest in influencing not merely attitude (valence and extremity) but attitude strength as well.12 Finally, persuasive efforts sometimes will be concerned not with any aspect of attitudes but rather with other mental states. For example, the key to changing some behaviors might involve not influencing persons’ attitudes but rather changing their perceived ability to perform the desired behavior. (For discussion of such persuasion targets, see Chapter 6 concerning reasoned action theory.) Consider, for instance, smokers who have a positive attitude toward quitting but have not yet really tried to do so because of a belief that quitting would be impossible; one can imagine such people finally making a serious attempt to quit if they are persuaded that they are indeed capable of quitting (e.g., by seeing examples of similar people who have managed to quit). In short, it should be apparent that persuasive efforts might seek changes in mental states other than attitude, and hence researchers will want correspondingly different outcome assessments. Attitude change will often, but not always, be a persuader’s goal.13



Conclusion This introductory chapter has elucidated the concepts of persuasion and attitude, described some common attitude assessment procedures, sketched the relationship of attitudes and behavior, and discussed the assessment of persuasive effects. In the following chapters, extant social-scientific theory and research about persuasion are reviewed. Several theoretical perspectives that have been prominent in the explanation of persuasive effects are discussed in Chapters 2 through 8. Research on various factors influencing persuasive effects is explored in Chapters 9 through 12.



42



For Review 1. What is a paradigm (exemplary) case? Give examples. Describe how the shared features of paradigm cases of a concept can provide clarification of the concept. Explain how the “sharp edges” of a definition can lead to disputes over borderline cases. 2. What are the shared features of exemplary cases of persuasion? Explain how a successful attempt to influence is such a feature. Explain how the persuader’s intending to influence is such a feature. Explain how some measure of freedom on the persuadee’s part is such a feature. Explain how having the effects be achieved through communication is such a feature. Explain how a change in the persuadee’s mental state is such a feature. Explain how features present in full-fledged ways in paradigm cases can, when present in only some diminished fashion, make for borderline cases of a concept. 3. Identify one important mental state often changed in persuasion. What is an attitude? Explain why attitudes are a common target for persuasive messages. 4. What are explicit attitude measurement techniques? What are semantic differential evaluative scales? Explain how they work. What are single-item attitude measures? What is the feeling thermometer? Identify a circumstance in which single-item attitude measures are especially useful. Identify and explain a weakness of such measures. 5. What are quasi-explicit attitude measurement techniques? Explain how a respondent’s agreement or disagreement with belief statements can serve as a measure of attitude. Describe the process of identifying suitable belief statements for such attitude measurement procedures. Identify an advantage (and accompanying disadvantage) of using such attitude measures. 6. What are implicit attitude measurement techniques? Give examples. How are implicit attitude measures different from explicit and quasiexplicit measures? 7. Are attitudes and behaviors generally consistent? What factors influence the degree of attitude-behavior consistency? How does the correspondence between the attitudinal measure and the behavioral measure influence attitude-behavior consistency? How does the degree of direct experience with the attitude object influence attitudebehavior consistency? Describe three general ways of encouraging attitude-consistent behavior. Explain how increasing the perceived relevance of an attitude to a behavior might enhance attitude-behavior 43



consistency. Explain how inducing feelings of hypocrisy might enhance attitude-behavior consistency. Explain how encouraging anticipation of feelings might enhance attitude-behavior consistency. 8. How can persuasion be assessed using attitude measurement techniques? Explain why other assessments (that is, other than attitude) may be useful or necessary.



Notes 1. This point can also be expressed by saying that to persuade is a perlocutionary act, whereas (for example) to urge is an illocutionary act. In this regard there is a difference between “A persuaded B” and “A attempted to persuade B” (Gass & Seiter, 2004, p. 27n3). 2. For discussion of some challenges in sharply distinguishing persuasion and coercion, see Powers (2007). 3. Descriptions of the relationship between persuasion and attitude change vary. For instance, sometimes attitude change is treated as a necessary aspect of persuasion (e.g., in the claim that “persuasion inherently has attitude change as its goal”; Beisecker & Parson, 1972, p. 5); sometimes persuasion is defined as one species of attitude change (e.g., as “a modification in one’s attitude that is the consequence of exposure to a communication”; Levy, Collins, & Nail, 1998, p. 732); and sometimes persuasion is simply treated as identical with attitude change generally, no matter how such change arises (e.g., Chaiken, Wood, & Eagly, 1996, p. 702). Whatever the particular characterization of the relationship, however, persuasion and attitude change have long been seen as closely linked. 4. Recent work on implicit measures has invited important new lines of investigation concerning the nature of attitude, but many open questions remain (for some discussion, see Bodenhausen & Gawronski, 2013; Bohner & Dickel, 2011; De Houwer, 2009; Gawronski & Bodenhausen, 2007; Petty, Fazio, & Briñol, 2009b). 5. For a period of time, it appeared as if the assumption of a close relationship between attitudes and behaviors was mistaken. Some classic studies (e.g., LaPiere, 1934) and some reviews (e.g., Wicker, 1969) suggested that people’s actions were commonly inconsistent with their attitudes. But these pessimistic conclusions about the attitude-behavior 44



relationship were overdrawn. 6. These reviews have used different procedures and analyzed different numbers of studies, but their estimates of the mean attitude-behavior correlation range from roughly .40 to .50. Larger mean correlations are reported when various methodological artifacts are corrected or with optimal levels of moderator variables. 7. Ajzen and Fishbein’s (1977) analysis specifies four ways in which attitudes and behaviors might correspond (action, target, context, and time). So, for example, the behavior of attending church services on campus this Sunday corresponds most directly to the attitude toward attending church services on campus this Sunday; this attitude and behavior correspond in the action specified (attending), the target toward which the action is directed (church services), the context of the action (on campus), and the time of the action (this Sunday). A more general behavior (e.g., one without a specified context or time, such as attending church services) corresponds most directly to a more general attitude (obviously, the attitude toward attending church services). Thus for an attitude toward an object (e.g., a consumer product), the corresponding behavioral measure would include assessments involving various actions, contexts, and times—which is the point of the multiple-act behavioral measure. 8. Notably, reasoned action theory (discussed in Chapter 6) includes attitudes toward specific behaviors as a key determinant of behavioral intentions. 9. As discussed in Chapter 5 (on cognitive dissonance theory), however, hypocrisy induction efforts can also backfire as a behavioral influence mechanism; instead of changing their future behaviors to be more consistent with their attitudes, people might change their attitudes to be consistent with their previous behavior (Fried, 1998). 10. Actually, some sense of hypocrisy may be a deeper connecting thread among these strategies. The perceived relevance strategy and the anticipated feelings strategy might be described as alerting people to hypocrisy (or to potential hypocrisy or hypocrisy-related feelings). So, for example, the reason that heightening the perceived relevance of an attitude to an action enhances attitude-behavior consistency may be precisely that such enhanced perceived relevance leads to an increased recognition of past inconsistency (and thus to feelings of hypocrisy, guilt, and so on— 45



which then motivate attitude-consistent future behavior) and/or to an increased expectation that negative feelings (guilt, regret, and so forth) will arise if attitude-inconsistent behavior is undertaken (with attitudeconsistent behavior then motivated by a desire to avoid such negative feelings). 11. As discussed in Chapter 8, the elaboration likelihood model (ELM) has pointed to several “metacognitive” states that might influence attitude, such as thought confidence (Briñol & Petty, 2009a, 2009b; Petty & Briñol, 2010, pp. 230–231). 12. Research on attitude strength is somewhat unsettled conceptually. For example, Krosnick and Petty’s (1995, p. 4) approach treats strength’s effects (persistence, resistance, and impact on information processing, judgments, and behavior) as the “defining features” of strength. But if strength is defined as (say) resistance, then it is necessarily true that “strong” attitudes are resistant. That is, this leaves unanswered the question of what makes attitudes resistant (saying “these attitudes are resistant because they are strong” would be akin to saying “these men are single because they are unmarried”). An alternative approach might define attitude strength not by its effects but by the conjunction of various effectindependent properties of attitude (e.g., an attitude’s interconnectedness with other attitudes, its importance, and the certainty with which it is held) or even dispense with any overarching concept of attitude strength in favor of studying the particular individual effect-independent features (see Visser, Bizer, & Krosnick, 2006). It will obviously be a substantial undertaking to distinguish these various features conceptually and to investigate their empirical interrelationships and effects (for work along these lines, see Petty, Briñol, Tormala, & Wegener, 2007). 13. In experimental persuasion research, the most common outcome assessments have been of attitudes, intentions, and behaviors. The persuasiveness of a given message might vary across these outcomes (e.g., a message might produce greater attitude change than behavioral change). However, where the research question concerns the relative persuasiveness of two message kinds, those three outcomes yield substantively identical conclusions. Carefully expressed: The mean effect sizes (describing the difference in persuasiveness between two message types) for attitudinal outcomes, for intention outcomes, and for behavior outcomes are statistically indistinguishable and hence functionally interchangeable (O’Keefe, 2013b). Hence meta-analyses aimed at drawing conclusions 46



about relative persuasiveness need not (should not) distinguish studies on the basis of the outcome measure used; similarly, in formative message design research (e.g., campaign planning) that tests the relative persuasiveness of alternative possible messages, assessments of appropriate intentions will provide a perfectly suitable guide to the relative behavioral persuasiveness of the messages.



47



Chapter 2 Social Judgment Theory Judgments of Alternative Positions on an Issue The Ordered Alternatives Questionnaire The Concept of Ego-Involvement Ego-Involvement and the Latitudes Measures of Ego-Involvement Reactions to Communications Assimilation and Contrast Effects Attitude Change Effects Assimilation and Contrast Effects Reconsidered Adapting Persuasive Messages to Recipients Using Social Judgment Theory Critical Assessment The Confounding of Involvement With Other Variables The Concept of Ego-Involvement The Measures of Ego-Involvement Conclusion For Review Notes



Social judgment theory is a theoretical perspective most closely associated with Muzafer Sherif, Carolyn Sherif, and their associates (C. W. Sherif, Sherif, & Nebergall, 1965; M. Sherif & Hovland, 1961; for a classic review, see Kiesler, Collins, & Miller, 1969, pp. 238–301; for additional discussions, see Eagly & Chaiken, 1993, pp. 363–382; Granberg, 1982; C. W. Sherif, 1980). The central tenet of social judgment theory is that messages produce attitude change through judgmental processes and effects. More specifically, the claim is that the effect of a persuasive communication depends upon the way in which the receiver evaluates the position it advocates. Hence attitude change is seen as a two-step process: First, the receiver makes an assessment of what position is being advocated by the message. Then attitude change occurs after this judgment—with the amount and direction of change dependent on that judgment. The plausibility of this general approach should be apparent: Our reaction 48



to a particular persuasive communication will depend (at least in part) on what we think of—how favorable we are toward—the point of view that it advocates. But this suggests that, in order to understand a recipient’s reaction to a given message, it is important to understand how the receiver assesses the various positions on that issue (that is, the various different stands that a message might advocate). Hence the next section discusses the nature of people’s judgments of the alternative positions on an issue. Subsequent sections discuss receivers’ reactions to persuasive messages, how social judgment theory suggests adapting messages to recipients, and some weaknesses of social judgment theory.



Judgments of Alternative Positions on an Issue On any given persuasive issue, a number of different positions or points of view are likely to be available. Consider, for example, some different possible stands on an issue such as gun control: One might think that there should be very few restrictions on ordinary citizens’ possession of firearms or (the other extreme) that almost no ordinary citizen should be permitted to possess a firearm; or one might hold any number of intermediate positions varying in the degree of restriction. A person is likely to have different assessments of these various positions, finding some of them acceptable, others objectionable, perhaps some neither particularly acceptable or unacceptable. If, as social judgment theory suggests, a person’s reaction to a persuasive message depends on the person’s judgment of the position being advocated, then it is important to be able to assess persons’ judgments of the various possible positions.



The Ordered Alternatives Questionnaire For obtaining person’s judgments of the various different positions, social judgment theory researchers developed the Ordered Alternatives questionnaire. An Ordered Alternatives questionnaire provides the respondent with a set of statements, each representing a different point of view on the issue being studied. The statements are chosen so as to represent the range of positions on the issue (from the extreme view on one side to the extreme view on the other) and are arranged in order from one extreme to the other—hence the name “Ordered Alternatives.” For example, the following Ordered Alternatives questionnaire was developed for research on a presidential election campaign (M. Sherif & Hovland, 49



1961, pp. 136–137; for other examples, see Hovland, Harvey, & Sherif, 1957; C. W. Sherif, 1980). ______ (A) The election of the Republican presidential and vicepresidential candidates in November is absolutely essential from all angles in the country’s interests. ______ (B) On the whole the interests of the country will be served best by the election of the Republican candidates for president and vice-president in the coming election. ______ (C) It seems that the country’s interests would be better served if the presidential and vice-presidential candidates of the Republican party are elected this November. ______ (D) Although it is hard to decide, it is probable that the country’s interests may be better served if the Republican presidential and vice-presidential candidates are elected in November. ______ (E) From the point of view of the country’s interests, it is hard to decide whether it is preferable to vote for the presidential and vice-presidential candidates of the Republican party or the Democratic party in November. ______ (F) Although it is hard to decide, it is probable that the country’s interests may be better served if the Democratic presidential and vice-presidential candidates are elected in November. ______ (G) It seems that the country’s interests would be better served if the presidential and vice-presidential candidates of the Democratic party are elected this November. ______ (H) On the whole the interests of the country will be served best by the election of the Democratic candidates for president and vice-president in the coming election. ______ (I) The election of the Democratic presidential and vicepresidential candidates in November is absolutely essential from all angles in the country’s interests. The respondent is asked first to indicate the one statement that he or she finds most acceptable (for example, by putting a + + in the corresponding blank). The respondent is then asked to indicate the other statements that are acceptable to the respondent (+), the one statement that is most objectionable (XX), and the other statements that are unacceptable (X). The respondent need not mark every statement as acceptable or unacceptable; that is, some of the positions can be neither accepted nor rejected by the respondent (and so be left blank, or marked with a zero). (For procedural details, see Granberg & Steele, 1974.) 50



These responses are said to form the person’s judgmental latitudes on that issue. The range of positions that the respondent finds acceptable form the respondent’s latitude of acceptance, the positions that the respondent finds unacceptable constitute the latitude of rejection, and the positions that the respondent neither accepts or rejects form the latitude of noncommitment. The structure of these judgmental latitudes can vary from person to person. In fact, two people might have the same “most preferred” position on an issue, but differ in their assessment of the other positions on the issue and hence have very different latitudes of acceptance, rejection, and noncommitment. For example, suppose that on the presidential election issue, Carol and Mary both find statement B most acceptable: their own most-preferred position is that, on the whole, the interests of the country will be best served by the election of the Republicans. Mary finds statements A, C, D, and E also acceptable, is noncommittal toward F, G, and H, and rejects only the extreme Democratic statement I; Carol, on the other hand, thinks that A is the only other acceptable statement, is noncommittal regarding C and D, and rejects E, F, G, H, and I. Mary thus has a larger latitude of acceptance than Carol (Mary finds five positions acceptable, Carol only two), a larger latitude of noncommitment (three positions as opposed to two), and a smaller latitude of rejection (only one position is objectionable to Mary, whereas five are to Carol). Notice (to jump ahead for a moment) that even though Carol and Mary have the same most preferred position, they would presumably react very differently to a message advocating position E: Mary finds that to be an acceptable position on the issue, but Carol finds it objectionable. As this example suggests, from the point of view of social judgment theory, a person’s stand on an issue involves not merely a most preferred position, but also assessment of all the other possible positions on the issue—as reflected in the set of judgmental latitudes (the latitudes of acceptance, rejection, and noncommitment). Social judgment theory proposes that the structure of the judgmental latitudes systematically varies depending on one’s level of egoinvolvement with the issue. Before discussing this relationship, some attention to the concept of ego-involvement is required.



The Concept of Ego-Involvement The concept of ego-involvement has been variously described in social judgment theory, and there is room for some uncertainty about just what it 51



comes to (for discussion, see Wilmot, 1971a). However, very broadly speaking, what is meant by “ego-involvement” is roughly the same as would be meant in colloquially referring to someone’s being “involved with an issue.” Thus a person might be said to be ego-involved when the issue has personal significance to the individual, when the person’s stand on the issue is central to his or her sense of self (hence ego-involvement), when the issue is important to the person, when the person takes a strong stand on the issue, when the person is strongly committed to the position, and so forth. Ego-involvement is thus in a sense an omnibus concept, meant to refer to this constellation of properties. Notice, however, that ego-involvement is distinct from the extremity of the most-preferred position (C. W. Sherif, 1980, p. 36; M. Sherif & Hovland, 1961, p. 171). That is, to be ego-involved on an issue is not the same thing as holding an extreme position on the issue. For example, one might take an extreme stand on an issue without being highly ego-involved (e.g., a person might hold an extreme position on the issue of controlling the federal deficit without being especially ego-involved in that stand). And one can be highly ego-involved in a middle-of-the-road position (“I’m strongly committed to this moderate position, my sense of self is connected with my being a moderate on this issue,” and so on). Thus egoinvolvement and position extremity are conceptually different. Social judgment theory does suggest that ego-involvement and position extremity will be empirically related, however, such that those with more extreme positions on an issue will tend to be more ego-involved in that issue (M. Sherif & Hovland, 1961, pp. 138–140). But this empirical relationship should not obscure the conceptual distinction between ego-involvement and position extremity.



Ego-Involvement and the Latitudes Social judgment theory suggests that one’s level of ego-involvement on an issue will influence the structure of one’s judgmental latitudes on that issue. Specifically, the claim is that as one’s level of ego-involvement increases, the size of the latitude of rejection will also increase (and the sizes of the latitudes of acceptance and noncommitment will decrease). Hence highly involved persons will have a relatively large latitude of rejection and relatively small latitudes of acceptance and noncommitment. That is, the more involved person will find relatively few stands on the issue to be acceptable (small latitude of acceptance), won’t be neutral or noncommittal toward very many positions (small latitude of 52



noncommitment), and will find many positions objectionable (large latitude of rejection). To gather evidence bearing on this claim, one needs a way to assess the relative sizes of the judgmental latitudes (which the Ordered Alternatives questionnaire provides) and a procedure for assessing ego-involvement. Two such ego-involvement measurement procedures are described in the next section.



Measures of Ego-Involvement Several different techniques have been devised for assessing egoinvolvement. Two particular measures can serve as useful examples.



Size of the Ordered Alternatives Latitude of Rejection In early studies of the relationship of ego-involvement to the structure of the judgmental latitudes, the participants were often persons whose involvement levels could be presumed on the basis of their group memberships.1 For example, to locate people who could be presumed to be relatively highly involved in an election campaign, researchers might go to local Democratic and Republican party headquarters. For comparison, other participants could be obtained from unselected samples (e.g., undergraduate students) that presumably would be comparatively lower in ego-involvement. In studies such as these, persons in the presumably higher-involvement groups had larger latitudes of rejection than did presumably less involved participants (for a general review of such work, see C. W. Sherif et al., 1965). On the basis of such results, the size of the latitude of rejection on the Ordered Alternatives questionnaire has been recommended as a measure of ego-involvement (e.g., Granberg, 1982, p. 313; C. W. Sherif et al., 1965, p. 234): The larger one’s latitude of rejection, the greater one’s degree of ego-involvement. Of course, as the latitude of rejection increases, the combined size of the latitudes of acceptance and noncommitment must necessarily decrease. It appears that the latitude of noncommitment tends to shrink more than does the latitude of acceptance. That is, as the latitude of rejection increases, the latitude of noncommitment decreases but there is sometimes little change in the latitude of acceptance (for a review, see C. W. Sherif et al., 1965). 53



This regularity has sometimes led to the suggestion that the size of the latitude of noncommitment might serve as a measure of ego-involvement (e.g., C. W. Sherif et al., 1965, p. 234), but the size of the latitude of rejection is the far more frequently studied index.



Own Categories Procedure A second measure of ego-involvement was derived from what is called the Own Categories procedure. Participants are presented with a large number of statements (60 or more) on the topic of interest and are asked to sort these statements into however many categories they think necessary to represent the range of positions on the issue. They are told to sort the items such that those in a given category seem to reflect the same basic viewpoint on the topic (for procedural details, see C. W. Sherif et al., 1965, pp. 92–126). What is of central interest is the number of categories a respondent creates. As in the studies of the Ordered Alternatives questionnaire, results were compared from selected and unselected respondents whose involvement levels could be presumed on independent grounds. Systematic differences were observed in the number of categories created. Those participants who were presumably highly involved created fewer categories than did low-involvement participants.2 Such results suggested the use of the Own Categories procedure as an index of ego-involvement: The fewer categories created, the greater the degree of ego-involvement (e.g., C. W. Sherif et al., 1965, p. 126). This result can seem to be counterintuitive, but it makes good sense from the perspective of social judgment theory (particularly against the backdrop of assimilation and contrast effects, to be discussed shortly). With increasing ego-involvement, increased perceptual distortion is likely. When involvement is exceptionally high, the individual’s thinking takes on an absolutist, black-or-white quality; in such a case, only two categories might be thought necessary (“Here are the few statements representing the right point of view—the one I hold—and here are all the wrongheaded ones”).2



Reactions to Communications Social judgment theory holds that a receiver’s reaction to a given 54



persuasive communication will depend centrally on how he or she evaluates the point of view it is advocating. That implies that, in reacting to a persuasive message, the receiver must initially come to decide just what position the message is forwarding. Social judgment theory suggests that, in making this judgment, the receiver may be subject to perceptual distortions called assimilation and contrast effects.



Assimilation and Contrast Effects Assimilation and contrast effects are perceptual effects concerning the judgment of what position is being advocated by a message. An assimilation effect is said to occur when a receiver perceives the message to be advocating a position closer to his or her own position than it actually does; that is, an assimilation effect involves the receiver minimizing the difference between the message’s position and the receiver’s position. A contrast effect is said to occur when a receiver perceives the message to be advocating a position farther away from his or her position than it actually does; thus a contrast effect involves the receiver’s exaggerating the difference between the message’s position and the receiver’s position.3 Social judgment theory offers a rule of thumb concerning the occurrence of assimilation and contrast effects (C. W. Sherif et al., 1965, p. 129). Broadly speaking, a communication advocating a position in the latitude of acceptance is likely to be assimilated (perceived as even closer to the receiver’s own view), and a communication advocating a position in the latitude of rejection is likely to be contrasted (perceived as even more discrepant from the receiver’s view). In the latitude of noncommitment, either assimilation or contrast effects might be found; the location of the boundary in the latitude of noncommitment (the point at which assimilation effects stop and contrast effects begin) is not clear, but it seems likely to occur somewhere closer to the latitude of rejection than the latitude of acceptance (Kiesler et al., 1969, p. 247). Notice, thus, that the perceived position of a persuasive communication may be different for persons with differing stands on the issue. An illustration of this phenomenon is provided by a study in which participants saw a message concerning a presidential election. The communication briefly listed the claims of the two major parties on various campaign issues, but did not take sides or draw clear conclusions. When pro-Republican respondents were asked what position the message advocated, they characterized it as being slightly pro-Democratic; pro55



Democratic respondents, on the other hand, saw the message as being slightly pro-Republican. Both groups of respondents thus exhibited a contrast effect, exaggerating the difference between the message and their own position (M. Sherif & Hovland, 1961, p. 151). (For other research illustrating assimilation and contrast effects, see Atkins, Deaux, & Bieri, 1967; Hurwitz, 1986; Manis, 1960; Merrill, Grofman, & Adams, 2001; C. W. Sherif et al., 1965, pp. 149–163.) Assimilation and contrast effects appear to be magnified by egoinvolvement. That is, there is a greater degree of perceptual distortion (regarding what position a message is advocating) as the receiver’s degree of involvement increases (C. W. Sherif, Kelly, Rodgers, Sarup, & Tittler, 1973; C. W. Sherif et al., 1965, p. 159). This relationship can be seen to underlie the previously described involvement-related differences revealed in the Own Categories procedure: Because higher ego-involvement means a greater propensity toward perceptual distortion, the higher-involvement perceiver finds it difficult to discern fine differences between advocated positions—and thus needs fewer categories to represent (what appear to be) the range of different positions on the issue. However, assimilation and contrast effects are minimized by messages that make clear what position is being advocated. That is, only relative ambiguous messages are subject to assimilation and contrast effects (see Granberg & Campbell, 1977; C. W. Sherif et al., 1965, p. 153; M. Sherif & Hovland, 1961, p. 153). When a persuader makes clear just what view is being forwarded, assimilation and contrast effects are minimized.4



Attitude Change Effects Whether receivers will change their attitudes following reception of a persuasive communication is said by social judgment theory to depend on what position the message is perceived to be advocating—that is, the perceived location of the communication with respect to the latitudes of acceptance, rejection, and noncommitment. The general principle offered by social judgment theory is this: A communication that is perceived to advocate a position that falls in the latitude of acceptance or the latitude of noncommitment will produce attitude change in the advocated direction (that is, in the direction sought by the message), but a communication that is perceived to advocate a position that falls in the latitude of rejection will produce no attitude change and may even provoke “boomerang” attitude change (i.e., change in the direction opposite that advocated by the 56



message). A number of studies have reported results consistent with this general principle (Atkins et al., 1967; Eagly & Telaak, 1972; B. T. Johnson, Lin, Symons, Campbell, & Ekstein, 1995; Sarup, Suchner, & Gaylord, 1991; C. W. Sherif et al., 1973; Siero & Doosje, 1993). This principle has important implications for the question of the effects of discrepancy (the difference between the message’s position and the receiver’s position) on attitude change. A persuader might advocate a position very discrepant from (very different from) the receiver’s own view, thus asking for a great deal of attitude change; or a persuader might advocate a position only slightly discrepant from the receiver’s, so seeking only a small amount of change. The question is: What amount of discrepancy (between the message’s position and the receiver’s position) will produce the greatest amount of attitude change in the advocated direction? Social judgment theory suggests that with increasing discrepancy, more favorable attitude change will occur—up to a point, namely, the latitude of rejection. But beyond that point, increasing discrepancy will produce less favorable reactions (indeed, may produce boomerang attitude change). Thus the general relationship between discrepancy and attitude change is suggested to be something like an inverted-U-shaped curve, and indeed the available research evidence is largely consistent with that suggestion. (For a general discussion of this view, see Whittaker, 1967. For findings of such a relationship—at least under some conditions—see, e.g., E. Aronson, Turner, & Carlsmith, 1963; Freedman, 1964; Sakaki, 1980; M. J. Smith, 1978; Whittaker, 1963, 1965. For complexities and additional discussion, see Chung, Fink, & Kaplowitz, 2008; Clark & Wegener, 2013; Fink & Cai, 2013; Fishbein & Lange, 1990; Kaplowitz & Fink, 1997.) However, for social judgment theory, any effects of discrepancy on attitude change are simply indirect reflections of the role played by the judgmental latitudes. Correspondingly, the inverted-U curve (relating discrepancy to attitude change) is only a crude and general guide to what persuasive effects may be expected in a given circumstance.5 To illustrate this complexity, consider the interplay of discrepancy and ego-involvement. As receivers become increasingly ego-involved in an issue, their latitudes of rejection presumably grow larger. Thus for lowinvolvement receivers, a persuader might be able to advocate a very discrepant viewpoint without entering the (small) latitude of rejection; by contrast, for high-involvement receivers, a very discrepant message will 57



almost certainly fall into the (large) latitude of rejection. Thus with any one influence attempt, a persuader facing a highly involved receiver may be able to advocate safely only a small change; obtaining substantial change from the highly involved receiver may require a series of small steps over time. By contrast, considerable attitude change might be obtained from the low-involvement receiver rather rapidly, through advocating a relatively discrepant (but not too discrepant) position (as suggested by Harvey & Rutherford, 1958). The larger point for persuaders is this: Effective persuasion requires knowing more than the receiver’s most preferred position; one needs to also know the structure of the judgmental latitudes. As noted earlier, two people with the same most preferred position might nevertheless have very different evaluations of other positions on the issue—and so a given persuasive message might fall in one recipient’s latitude of acceptance but the other’s latitude of rejection, even though the two people have the same most preferred position on the issue.



Assimilation and Contrast Effects Reconsidered The attitude-change principles discussed in the preceding section refer to what position the message is perceived to advocate. It thus becomes important to reconsider the role of assimilation and contrast effects in persuasion, since these influence the perceived position of a message. The key point to be noticed is this: Assimilation and contrast effects reduce the effectiveness of persuasive messages.



The Impact of Assimilation and Contrast Effects on Persuasion Consider first the case of a contrast effect. If a message that advocates a position in the receiver’s latitude of rejection—and so already is unlikely to produce much favorable attitude change—is perceived as advocating an even more discrepant position, then the chances for favorable attitude change diminish even more (and the chances for boomerang attitude change increase). Obviously, then, contrast effects impair persuasive effectiveness. But assimilation effects also reduce persuasive effectiveness. When an assimilation effect occurs, the perceived discrepancy between the 58



message’s stand and the receiver’s position is reduced—and hence the communicator is seen as asking for less change than he or she actually seeks.6 Consider the case of a message that advocates a position in the latitude of acceptance or the latitude of noncommitment; with increasing perceived discrepancy, the chances of favorable attitude change presumably increase. But an assimilation effect will reduce the perceived discrepancy between the message’s view and the receiver’s position, and so it will reduce the amount of attitude change obtained. Indeed, in the extreme case of complete assimilation (when the receivers think that the message is simply saying what they already believe), no attitude change will occur, because the audience has misperceived the communicator’s position. That is, when the recipient mistakenly believes (because of the perceptual distortion of assimilation) that the message advocates the recipient’s current position, then the recipient’s attitude will not change.7 However, persuaders can minimize assimilation and contrast effects by being clear about their position on the persuasive issue at hand. As mentioned previously, only relatively ambiguous communications (that is, messages that aren’t clear about their stand on the persuasive issue) are subject to assimilation and contrast effects. Thus social judgment theory emphasizes for persuaders the importance of making one’s advocated position clear.



Ambiguity in Political Campaigns One might think that the prevalence (and apparent success) of ambiguity in political campaigns suggests that something is amiss (with social judgment theory, if not the political campaign process). After all, if ambiguity reduces persuasive effectiveness, why is it that successful political campaigners so frequently seem to be ambiguous about their stands on the issues? To understand this phenomenon, it is important to keep in mind the persuasive aims of election campaigns. Ordinarily the candidate is not trying to persuade audiences to favor this or that approach to the matter of budget policy or gun control or any other “campaign issue.” Rather, the persuasive aim of the campaign is to get people to vote for the candidate— and candidates are never ambiguous about their stand on that question. Thus on the topic on which candidates seek persuasion (namely, who to vote for), candidates obviously take clear positions.



59



Candidates do sometimes adopt ambiguous positions on “campaign issues” (economic policy, social issues, and so on). If a candidate were trying to persuade voters that “the right approach to the issue of gun control is thusand-so,” then being ambiguous about the candidate’s position on gun control would reduce the chances of successful persuasion on that topic. Such ambiguity would encourage assimilation and contrast effects, thereby impairing the candidate’s chances of changing anyone’s mind about that issue. But, ordinarily, candidates don’t seek to persuade voters about the wisdom of some particular policy on some campaign issue. Usually, the candidate hopes to encourage voters to believe that the candidate’s view on a given issue is the same as the voter’s view. That is, candidates hope that with respect to campaign issues, voters will assimilate the candidate’s views (overestimate the degree of similarity between the candidate’s views and their own). Social judgment theory straightforwardly suggests how such an effect might be obtained. Suppose—as seems plausible—that for many voters, the positions around the middle of the scale on a given campaign issue commonly fall in the latitude of noncommitment or the latitude of acceptance; for a small number of voters (e.g., those with extreme views and high ego-involvement on that topic), such positions might fall in the latitude of rejection, but most of the electorate feels noncommittal toward, if not accepting of, such views. In a such a circumstance, if the message suggests some sort of vaguely moderate position on the issue, without being very clear about exactly what position is being defended, then the conditions are ripe for assimilation effects regarding the candidate’s stand on that topic. Voters who themselves have widely varying views on the issue may nevertheless all perceive the candidate’s issue position to be similar to their own. (For research concerning assimilation and contrast effects in political contexts, see Drummond, 2011; Granberg, 1982; Granberg, Kasmer, & Nanneman, 1988; Judd, Kenny, & Krosnick, 1983; M. King, 1978.)



Adapting Persuasive Messages to Recipients Using Social Judgment Theory One recurring theme in theoretical analyses of persuasion is the idea that to maximize effectiveness, persuasive messages should be adapted (tailored, 60



adjusted) to fit the audience. From the perspective of social judgment theory, this especially means adapting messages to the recipient’s judgmental latitudes. As mentioned earlier, social judgment theory emphasizes that a persuader needs to know more than simply the receiver’s most preferred position; the structure of the judgmental latitudes —the sizes and locations of the latitudes of acceptance, rejection, and noncommitment—is also important. Even if two receivers have the same most preferred position, a given persuasive message might fall in the latitude of acceptance for one person but in the latitude of rejection for another, leading to quite different reactions to a given message. Persuaders are often not in a position to vary their advocated view for different audiences. (For example, politicians who attempt to do so can find themselves accused of “flip-flopping” or “talking out of both sides of the mouth.”) But in some circumstances, persuaders can be free to vary what they ask of audiences. For example, a charity might vary how large a donation is requested depending on the recipient’s financial circumstance; people who are financially better-off may be asked for larger sums. In such a circumstance, having some sense of the recipient’s judgmental latitudes —what requested amounts might seem outrageously large to them (latitude of rejection) and which might seem at least worthy of considering (latitude of noncommitment)—can be crucial. More broadly, one might think of the judgmental latitudes as identifying what sorts of claims might be found plausible by the intended audience. For example, S. Smith, Atkin, Martell, Allen, and Hembroff (2006) were planning a campus campaign to reduce alcohol abuse. Because people can overestimate the frequency of campus drinking, the campaign wanted to convey accurate information about alcohol consumption (accurate information about the “descriptive norm,” the actual frequency of a given behavior). However, if the audience were to perceive the (accurate) information to be unbelievable, then presumably the campaign would not be successful. So in preliminary campaign planning research, Smith and colleagues adapted the Ordered Alternatives questionnaire to assess how believable respondents found various percentages of students who drink five or fewer drinks when they party. The accurate description of the frequency of such drinking fell within the audience’s latitude of noncommitment—that is, the audience was apparently not predisposed to reject that percentage as unrealistic. The researchers were thus able to confidently design campaign communications using accurate descriptivenorm information, knowing that such claims would not be thought 61



unreasonable. Social judgment theory also plainly suggests that messages may need to be adapted to the audience’s level of ego-involvement. Where message recipients are not very involved in the issue, a persuader might be able to advocate a relatively discrepant position without encountering the latitude of rejection; where the audience is highly involved, on the other hand, the large latitude of rejection is likely to necessitate a smaller discrepancy if the message is to be effective.



Critical Assessment Social judgment theory obviously offers a number of concepts and principles useful for illuminating persuasive effects. But several weaknesses in social judgment theory and research have become apparent.



The Confounding of Involvement With Other Variables One weakness in much social judgment theory research stems from the use of participants from preexisting groups thought to differ in involvement (e.g., in research on a presidential election, using committed members of a political party to represent persons high in ego-involvement). This research procedure has created ambiguities in interpreting results, because the procedure has confounded involvement with a number of other variables. Two variables are said to be confounded in a research design when they are associated in such a way as to make it impossible to disentangle their separate effects. In the case of much social judgment theory research, the persons selected to serve as high-involvement participants differed from the low-involvement participants not just in involvement but in other ways as well. For example, the high-involvement participants had more extreme attitudes than the low-involvement participants (e.g., M. Sherif & Hovland, 1961, pp. 134–135). In such a circumstance, when the highinvolvement group displays a larger latitude of rejection than the lowinvolvement group, one cannot unambiguously attribute the difference to involvement (as social judgment theory might propose). The difference in latitude size could instead be due to differences in position extremity. The problem is that, according to social judgment theory, ego-involvement 62



and position extremity are distinct concepts. Involvement and extremity are often correlated (such that higher involvement is characteristically associated with more extreme views) but nevertheless conceptually distinct. Hence it is important to be able to distinguish the effects of egoinvolvement from the effects of position extremity. Social judgment theory claims that larger latitudes of rejection are the result of heightened egoinvolvement, not the result of extreme positions per se (e.g., C. W. Sherif et al., 1965, p. 233); but because the research evidence in hand confounds ego-involvement and position extremity, the evidence is insufficient to support such a claim. In fact, the groups used in much social judgment research differed not only in involvement and position extremity but in age, educational achievement, and other variables. As a result, one cannot confidently explain observed differences (e.g., in the size of the latitude of rejection, or in the number of categories used in the own-categories procedure) as being the result simply of involvement differences; one of the other factors, or some combination of other factors, might have been responsible for the observed effects. (A more general discussion of this problem with social judgment research has been provided by Kiesler et al., 1969, pp. 254–257.)



The Concept of Ego-Involvement The concept of ego-involvement is a global or omnibus concept, one that involves a constellation of various properties—the person’s stand on the issue being central to the person’s sense of self, the issue’s importance to the person, the issue’s personal relevance to the person, the degree of commitment the person has to the position, and the degree of intensity with which the position is held, and so on (for a useful discussion, see Wilmot, 1971a). But these are distinguishable properties. For instance, I can think an issue is important without my stand on that issue being central to my selfconcept (e.g., I think the issue of controlling the federal deficit is important, but my sense of identity isn’t connected to my stand on this matter). I can hold a given belief intensely, even though the issue isn’t very important to me (e.g., my belief that the Earth is round). An issue may not be personally relevant to me (e.g., abortion), but I could nonetheless be strongly committed to a position on that issue, and my stand on that issue could be important to my sense of self. I can hold a belief strongly (e.g., about the superiority of a given basketball team), even 63



though that belief isn’t central to my self-concept. The general point is that the notion of ego-involvement runs together a number of distinct concepts in an unsatisfactory manner. It is possible to distinguish (conceptually, if not empirically) commitment to a position, importance of the issue, personal relevance of the issue, and so forth, and hence a clear understanding of the roles these play in persuasion will require separate treatment of each. (For examples of efforts at clarifying one or another aspect of involvement, see B. T. Johnson & Eagly, 1989; Levin, Nichols, & Johnson, 2000; Slater, 2002; Thomsen, Borgida, & Lavine, 1995.)8



The Measures of Ego-Involvement Research concerning the common measures of ego-involvement—the size of the latitude of rejection in the Ordered Alternative questionnaire and the number of categories created in the Own Categories procedure—has revealed some worrisome findings, of two sorts.9 First, the measures are not very strongly correlated with each other. Two instruments that measure the same property ought to be strongly correlated. For example, in the case of the two common measures of egoinvolvement, the two measures should be strongly negatively correlated: as the size of the latitude of rejection increases, the number of categories created should decrease. But research that has examined the correlations among various involvement measures (including, but not limited to, the size of the latitude of rejection on the Ordered Alternatives questionnaire and the number of categories in the Own Categories procedure) have commonly yielded correlations that are roughly zero (e.g., Wilmot, 1971b). The implication is that the different measures of involvement cannot all be measuring the same thing. Maybe one of them is measuring involvement and the others are not, or maybe none of them is measuring involvement. But plainly these assessments are not all measuring the same thing. Second, the measures do not display the expected patterns of association with other variables. For example, ego-involvement measures are not strongly correlated with such variables as the perceived importance of the issue to the respondent, the perceived importance of the issue to society, the respondent’s perceived commitment to their most acceptable position, or the respondent’s self-reported certainty, intensity of feeling, or interest 64



in the topic (R. A. Clark & Stewart, 1971; Krosnick, Boninger, Chuang, Berent, & Carnot, 1993; Wilmot, 1971b). In short, there are good empirical grounds for concern about the adequacy and meaning of the common measures of ego-involvement. This is perhaps to be expected, however, given the lack of clarity of the concept of egoinvolvement; one cannot hope to have a very satisfactory assessment procedure for a vague and indistinct concept. In any case, the empirical evidence suggests that the various indices of ego-involvement ought not be employed unreflectively.



Conclusion In some ways social judgment theory is too simplified to serve as a complete account of persuasive effects. From a social judgment theory point of view, the only features of the message that are relevant to its impact are (a) the position it advocates and (b) the clarity with which it identifies its position. It doesn’t matter whether the message contains sound arguments and good evidence or specious reasoning and poor evidence; it doesn’t matter what sorts of values the message appeals to, how the message is organized, or who the communicator is. Everything turns simply on what position the message is seen to defend. And surely this is an incomplete account of what underlies persuasive message effects. But a theory can be useful even when incomplete. Social judgment theory does draw one’s attention to various important facets of the process of persuasion. For example, realizing the possibility of assimilation and contrast effects can be crucially important to persuasive success. A persuader who is not sufficiently clear about his or her advocated position may think persuasion has been achieved because the message recipient professes complete agreement with the message, but if the recipient misperceived what view was being advocated (an assimilation effect), such expressions of agreement will be misleading indicators of persuasive success. Similarly, recognizing that people commonly have not only a mostpreferred position but also have assessments of other positions on the issue —the judgmental latitudes—can help persuaders understand their goals and challenges more clearly. For example, the persuader’s objective does not always have to be to induce the recipients to change their mostpreferred positions so as to match the persuader’s most-preferred position. 65



Sometimes it may be enough to get recipients to see that the persuader’s position falls in their latitude of noncommitment. For example, where public policy issues are the subject of advocacy, “the battle is not to convince citizens that one’s policy is right, but simply that it is not unreasonable” (Diamond & Cobb, 1996, p. 242). And even if social judgment theory’s notion of ego-involvement is too much of an omnibus concept, it nevertheless points to the importance of variations in how individuals relate to a given persuasive issue. (Indeed, social judgment theory here provides a useful example of how difficult it can be to take a broad, common-sense concept like involvement and articulate it in a careful, empirically well-grounded way.) So although social judgment theory must be judged something of an historical relic at present, in the sense that it is not the object of much current research attention, it nevertheless is a framework that offers some concepts and principles of continuing utility—and it offers some instructive object lessons.



For Review 1. What is the central tenet of social judgment theory? Upon what is the effect of a persuasive communication said to centrally depend? According to social judgment theory, what are the two steps involved in attitude change? 2. Explain the idea that people have judgments of the alternative positions available on an issue. How can one obtain such judgments? Describe the Ordered Alternatives questionnaire. What instructions are respondents given for completing the Ordered Alternatives questionnaire? 3. What are the judgmental latitudes? What is the latitude of acceptance? The latitude of rejection? The latitude of noncommitment? Explain how, for social judgment theory, a person’s stand on an issue is represented by more than the person’s mostacceptable position. 4. What is ego-involvement? What is the conceptual relationship of egoinvolvement and position extremity? Is being ego-involved in a issue the same thing as holding an extreme position on the issue? According to social judgment theory, what is the empirical relationship of ego-involvement and position extremity? 5. How is ego-involvement predicted to influence the structure of the 66



6.



7.



8.



9.



10.



11.



12.



judgmental latitudes? What latitude structure is said to be characteristic of a person high in ego-involvement? Of a person low in ego-involvement? Explain how, in social judgment theory research, group membership was used to validate the use of the size of the latitude of rejection (on the Ordered Alternatives questionnaire) as a measure of egoinvolvement. What is the Own Categories procedure? Explain how ego-involvement is thought to influence the number of categories used in the Own Categories procedure. What are assimilation and contrast effects (broadly speaking)? What is a contrast effect? What is an assimilation effect? What is the rule of thumb concerning when each effect will occur? Explain how, because of assimilation and contrast effects, the perceived position of a persuasive message may be different for people with different positions on the issue. What is the relationship between egoinvolvement and assimilation and contrast effects? What kinds of messages are subject to assimilation and contrast effects? How can a persuader minimize assimilation and contrast effects? Describe social judgment theory’s rule of thumb concerning attitude change effects following persuasive communications. What is “discrepancy”? What is the relationship between discrepancy and attitude change, according to social judgment theory? Describe how this analysis suggests different approaches to persuading high- and low-involvement receivers. Explain how contrast effects reduce the effectiveness of persuasive messages. Explain how assimilation effects reduce the effectiveness of persuasive messages. How can political campaigns exploit assimilation effects concerning positions on policy issues? Explain how persuasive messages might be adapted to the recipient’s judgmental latitudes. Explain how messages might be adapted to the recipient’s level of ego-involvement. What does it mean to say that two factors (variables) are confounded? Describe how extremity and involvement have been confounded in social judgment research. Explain the implications of this confounding for interpreting social judgment research. Explain how the concept of ego-involvement conflates a number of different concepts. Identify and describe two worrisome findings concerning the measures of ego-involvement. What sort of correlation is expected between two instruments that measure the same property? If two measures of involvement do measure involvement, what correlations 67



would be expected between them (e.g., between the number of categories in the Own Categories procedure and the size of the latitude of rejection on the Ordered Alternatives questionnaire)? What correlations have been observed? Are the measures of egoinvolvement strongly correlated with each other? Do the measures of ego-involvement display the expected patterns of association with other variables? How are measures of ego-involvement related to assessments of perceived topic importance or commitment to one’s position?



Notes 1. The anchoring of attitudes in reference groups is emphasized in some social judgment theory conceptualizations of involvement (e.g., M. Sherif & Sherif, 1967, pp. 135–136), and hence this was an attractive research procedure. 2. Ego-involvement is thought to influence not only the number of categories but also the distribution of statements across categories. Where ego-involvement is lower, statements are likely to be distributed roughly evenly across whatever categories are created; the higher-involved person is likely to use categories disproportionately (C. W. Sherif et al., 1965, p. 239). 3. Assimilation and contrast effects, more broadly defined, are familiar psychophysical phenomena. If you’ve been lifting 20-pound boxes all day, a 40-pound box will feel even heavier than 40 pounds (contrast effect), but a 21-pound box will feel very similar (assimilation effect). The psychophysical principle involved is that when a stimulus (the 40-pound box) is distant from one’s judgmental anchor (the 20-pound boxes), a contrast effect is likely; when the stimulus is close to the anchor, an assimilation effect is likely. Indeed, social judgment theory was explicitly represented as an attempt to generalize psychophysical judgmental principles and findings to the realm of social judgment, with the person’s own most-preferred position serving as the judgmental anchor (see M. Sherif & Hovland, 1961). 4. For social judgment theory, the perceived position of a message thus is influenced by message properties (the advocated position and the clarity with which it is expressed) in conjunction with the recipient’s own position (which serves as a perceptual anchor). For discussion of other 68



factors influencing the perceived position of messages, see, for example, Kaplowitz and Fink (1997) and R. Smith and Boster (2009). 5. From a social judgment theory perspective, the relation of discrepancy and attitude change is presumably not a completely symmetrical invertedU-shaped curve but something rather more like half of such a curve: gradually increasing favorable attitude change up to the latitude of rejection, but with a sharp drop-off (not a gentle decline) at that point. 6. Concerning social judgment theory’s predictions about the impact of assimilation effects on persuasion, the current description—that assimilation effects reduce persuasion—parallels that of most commentators (e.g., Kiesler, Collins, & Miller, 1969, p. 260). Eagly and Chaiken (1993, p. 387n13), however, have argued that “social judgment theory posits a positive relation between assimilation and persuasion.” Social judgment theorists have certainly described assimilation effects as making it more likely that recipients will have certain sorts of evaluatively positive reactions to a message (thinking it fair, unbiased, and so forth; e.g., M. Sherif & Sherif, 1967, p. 130), but such positive assessments are not necessarily incompatible with a lack of persuasion (attitude change). Consider, for example, that C. W. Sherif et al. (1965) indicated that highinvolvement persons “are particularly prone to displace the position of a communication in such a way that their stand is unaffected” (p. 176). This displacement can involve either assimilation or contrast, but “in either case … their placement of communications is such that little effect on their attitudes could be expected” (p. 177). This passage certainly suggests that social judgment theory depicts assimilation effects as reducing attitude change. 7. Because involvement magnifies assimilation and contrast effects, highly involved people are understandably often difficult to persuade: It’s not just that they have small latitudes of acceptance and noncommitment (and so are predisposed to react negatively to many advocated positions) but also that they’re especially prone to misperceiving what view is being advocated. 8. The elaboration likelihood model, discussed in Chapter 8, invokes a rather narrower concept of involvement: direct personal relevance of the issue. 9. A third troubling finding, not discussed here, is that there is more crossissue consistency in an individual’s apparent level of ego-involvement (as 69



assessed by common measures of ego-involvement) than should be expected given that ego-involvement is an issue-specific property (that is, one that varies from issue to issue for a given individual), not a personality-trait-like disposition; for discussion and references, see O’Keefe (1990, pp. 42–43).



70



Chapter 3 Functional Approaches to Attitude A Classic Functional Analysis Subsequent Developments Identifying General Functions of Attitude Assessing the Function of a Given Attitude Influences on Attitude Function Adapting Persuasive Messages to Recipients: Function Matching Commentary Generality and Specificity in Attitude Function Typologies Functional Confusions Reconsidering the Assessment and Conceptualization of Attitude Function Persuasion and Function Matching Revisited Reviving the Idea of Attitude Functions Conclusion For Review Notes



One general approach to the analysis of attitudes focuses on the functions that attitudes can serve. The basic idea is that attitudes may serve various functions for persons, that is, may do different jobs, meet different needs or purposes. The relevance of this idea for understanding persuasion is that the most effective technique for changing an attitude may vary depending on the attitude’s function. Functional analyses of attitude have a long history; the treatment here first discusses one classic example of such an analysis and then turns to more recent developments. (For useful collections and discussions concerning functional approaches to attitude, see Carpenter, Boster, & Andrews, 2013; Eagly & Chaiken, 1993, pp. 479–490; Maio & Olson, 2000b; Shavitt & Nelson, 2002; Watt, Maio, Haddock, & Johnson, 2008.)



A Classic Functional Analysis In a well-known analysis, Katz (1960) proposed four attitude functions: 71



utilitarian, ego-defensive, value-expressive, and knowledge. The utilitarian function is represented by attitudes that help people maximize rewards and minimize punishments. For example, students who experience success with essay exams are likely to develop favorable attitudes toward such exams. Attitudes serving a utilitarian function, Katz suggested, will be susceptible to change when the attitude (and related activities) no longer effectively maximizes rewards and minimizes punishments. Thus utilitarian attitudes are probably most effectively changed by either creating new rewards and punishments (as when, for instance, a company creates a new incentive program to encourage suggestions by employees) or by changing what is associated with existing rewards and punishments (as when a company changes the basis on which salespeople’s bonuses are based). Attitudes serving an ego-defensive function do the job of defending one’s self-image. Ego-defensive attitudes are exemplified most clearly by prejudicial attitudes toward minorities; such attitudes presumably bolster the holder’s self-image (ego) by denigrating others (see, e.g., Fein & Spencer, 1997). The most promising avenues to changing attitudes serving such a function, Katz suggested, might involve removing the threat to the ego (thus removing the need for self-defense) or giving persons insight into their motivational dynamics (getting people to see that their attitudes are not substantively well-grounded but simply stem from ego-defensive needs). With attitudes serving a value-expressive function, persons get satisfaction from holding and expressing attitudes that reflect their central values and self-images. For example, a person whose self-image is that of a conservative Republican might get satisfaction from supporting a balanced budget amendment because such a viewpoint reflects the person’s selfimage. Attitudes serving a value-expressive function are thought to be likely to change either when the underlying beliefs and self-images change (because then there would be no need to express the old values) or when an alternative, superior means of expressing the values is presented (as when a political candidate says, “If you’re looking for a real conservative [or liberal or whatever], then vote for me, because I represent those values better than the other candidates do”). The knowledge function of attitudes reflects the role of attitudes in organizing and understanding information and events. For example, one way of making sense of complex sociopolitical situations (such as in the 72



Middle East) can be to, in effect, identify the “good guys” and the “bad guys.” That is, attitudes (evaluations) can serve as at least a superficial mechanism for organizing one’s understandings of such situations. Attitudes serving a knowledge function, Katz suggested, are especially susceptible to change through the introduction of ambiguity (as when the good guys do something bad or the bad guys do something good); such ambiguity indicates that the attitudes are not functioning well to organize information, thus making the attitudes more likely to change. Katz’s description of these four attitude functions provides a useful concrete example of a functional analysis of attitudes. It is appropriately nuanced; for example, it acknowledges that a given attitude might serve more than one function. It makes plain the connection between functional attitude analysis and the understanding of alternative persuasion mechanisms by suggesting means of influence especially well tailored to each functional attitude type. (The analysis does not claim that the recommended means of changing each type of attitude will be guaranteed to be successful, but only that a given attitude type is more likely to be changed when approached with the appropriate means of influence.) Katz’s analysis did not initially attract much research attention, in good part because of perceived difficulties in assessing attitude function (for some discussion, see Kiesler, Collins, & Miller, 1969, pp. 302–330; Shavitt, 1989). But functional analyses of attitude have subsequently flowered.



Subsequent Developments Identifying General Functions of Attitude Katz’s list of attitude functions is only one of many proposed function typologies. In other analyses, different functions have been proposed, the relationships among various functions reconsidered, and alternative organizational schemes suggested. For example, M. B. Smith, Bruner, and White (1956) suggested a socialadjustive function, in which attitudes help people adjust to social situations and groups. As described by Snyder and DeBono (1989), persons hold attitudes serving a social-adjustive function because such attitudes “allow them to fit into important social situations and allow them to interact 73



smoothly with their peers” (p. 341); expression of the attitude may elicit social approval, make it easier to adapt to social situations, and the like.1 Shavitt’s (1990) taxonomy distinguished a utilitarian function, a socialidentity function (understood as including both social-adjustive and valueexpressive functions), and a self-esteem maintenance function (including ego-defensive purposes). Gastil (1992) proposed six attitude functions: personal utility, social utility, value expressive, social adjustment (easing social interaction), social-identity (forging one’s identity), and self-esteem maintenance. But there is not yet a consensus on any one functional typology. This surely reflects the lack of any simple, easily assessed source of evidence for or against a given function list. An attitude function taxonomy presumably shows its worth by being broadly useful, across a number of applications, in illuminating the underlying motivational bases of attitude. Expressed generally, this illumination consists of showing that the scheme in question permits one to detect or predict relevant events or relationships, but this evidence can be quite diverse. A given typology’s value might be displayed by showing that knowledge of an attitude’s function (as captured by the typology in question) permits one to predict or detect (for example) the product features that persons will find most appealing, the relative effectiveness of various persuasive messages, the connection between personality traits and attitude functions, and so on. But because for any given typology there commonly is relatively little research evidence distinctively bearing on that scheme, there is at present little basis for supposing that any given specific typology is unquestionably superior to all others. (There is even less evidence comparing the usefulness of alternative taxonomies; for an example, see Gastil, 1992.) This lack of consensus makes for a rather chaotic and unsettled situation, one in which a genuine accumulation of results (and corresponding confident generalization) is difficult. If there was one widely agreed-on set of specific functions, then research could straightforwardly be accumulated; more could be learned about (say) what personality traits or situational features incline persons to favor this or that function, what sorts of messages are best adapted for changing attitudes serving the various functions, and so forth. Instead, most of the research evidence concerning functional attitude analyses is of a piecemeal sort: One study compares personality correlates of social-adjustive and value-expressive functions, another examines different means of influencing attitudes serving egodefensive functions, and so on. 74



In such a circumstance, one promising approach might be to paint in broader strokes, deferring matters of detailed functional typologies in favor of identifying some general functional differences. One broad functional distinction has been found widely useful and seems contained (implicitly or explicitly) in a great many attitude function analyses: a distinction between symbolic and instrumental attitude functions (see Abelson & Prentice, 1989; Ennis & Zanna, 2000, pp. 396–397). Briefly expressed, symbolic functions focus on the symbolic associations of the object; attitudes serving a symbolic function do the jobs of expressing fundamental moral beliefs, symbolizing significant values, projecting selfimages, and the like (e.g., Katz’s ego-defensive function). Instrumental functions focus on the intrinsic properties of the object; attitudes serving instrumental functions do the jobs of summarizing the desirable and undesirable aspects of the object, appraising the object through specific intrinsic consequences or attributes, and so forth (e.g., Katz’s utilitarian function). For example, concerning stricter gun control laws in the United States, a supporter’s positive attitudes might have a predominantly symbolic basis (beliefs such as “It represents progress toward a more civilized world”) or an instrumental basis (“It will reduce crime because criminals won’t be able to get guns so easily”); similarly, an opponent’s negative attitudes might be motivated by largely symbolic considerations (“It represents impingement on constitutional rights”) or by largely instrumental considerations (“It will increase crime because criminals will still have guns, but law-abiding citizens won’t”). Of course, it is possible for a person’s attitude on a given topic to have a mixture of symbolic and instrumental underpinnings. And an attitude’s function might change through time. For instance, an attitude might initially serve a symbolic function but subsequently come to predominantly serve instrumental ends (see Mangleburg et al., 1998). But the general distinction between symbolic and instrumental attitude functions appears to be a broadly useful one (see, e.g., Crandall, Glor, & Britt, 1997; Herek & Capitanio, 1998; A. Kim, Stark, & Borgida, 2011; Prentice & Carlsmith, 2000).



Assessing the Function of a Given Attitude Given a typology of attitude functions, the question that naturally arises is how one can tell what function an individual’s attitude is serving. Indeed, one recurring challenge facing functional attitude theories has been the assessment of attitude functions (see Shavitt, 1989). 75



One straightforward procedure for assessing the function of a given attitude involves coding (classifying) relevant free-response data (data derived from open-ended questions). For example, Shavitt (1990) asked participants to write down “what your feelings are about the attitude object, and why you feel the way you do. … Write down all of your thoughts and feelings that are relevant to your attitude, and try to describe the reasons for your feelings” (p. 130). Responses were then classified on the basis of the apparent attitude function. For example, responses concerning what the attitude communicates to others were coded as indicating a social-identity function, whereas responses focused on attributes of the attitude object were classified as reflecting a utilitarian function. Such free-response data can be elicited in various ways (participants might write essays or simply list their beliefs), and the classification system will vary depending on the functional typology being used (e.g., one might simply contrast symbolic and instrumental bases of attitudes; see Ennis & Zanna, 1993). But the general principle behind these procedural variants is that different attitude functions will have different characteristic clusters of affiliated beliefs, spawned by the different motivations behind (different functions of) the attitude, and hence examination of such freely elicited beliefs will illuminate attitude functions. (For other examples of such procedures, see Herek, 1987; Maio & Olson, 1994.) A second avenue to the assessment of attitude functions is the use of a questionnaire with standardized scale response items. The leading example is Herek’s (1987) Attitude Functions Inventory, which presents respondents with statements about different possible bases for their views (statements of the form “My views about X mainly are based on …”); respondents are asked to indicate the degree to which each statement is true of them (giving answers on a scale anchored by the phrases “very true of me” and “not at all true of me”). So, for instance, in an item assessing the value-expressive function, persons are asked about the degree to which it is true that their views are based on their “moral beliefs about how things should be”; for the ego-defensive function, one item asks whether the respondent’s views are based on “personal feelings of disgust or revulsion.” (Each attitude function is assessed using several items.) As another example of such procedures, Clary, Snyder, Ridge, Miene, and Haugen (1994) had participants rate the importance of 30 possible reasons for volunteering (six reasons for each of five attitude functions). For example, “I can gain prestige at school or work” was one utilitarian reason, 76



whereas “members of a social group to which I belong expect people to volunteer” was a social-adjustive reason. (For other examples of the use of these or similar instruments, see Clary et al., 1998; Ennis & Zanna, 1993; Gastil, 1992; Herek, 2000; Shavitt, 1990.) In much attitude function research, however, a third approach has been adopted, that of using proxy indices such as personality characteristics to stand in for more direct assessments of function (on the basis of associations between such characteristics and attitude functions). Among these, the most frequently employed has been the individual-difference variable of self-monitoring. Self-monitoring refers to the control or regulation (monitoring) of one’s self-presentation, and specifically to the tendency to tailor one’s behavior to fit situational considerations. Broadly speaking, high self-monitors are concerned about the image they project to others and tailor their conduct to fit the particular circumstances they are in. Low self-monitors are less concerned about their projected image and mold their behavior to fit inner states (their attitudes and values) rather than external circumstances (social norms of appropriateness). In a wellestablished questionnaire used to assess self-monitoring, high selfmonitoring is reflected by agreement with statements such as “I guess I put on a show to impress or entertain others” and “I would probably make a good actor”; low self-monitoring is reflected by agreement with statements such as “I have trouble changing my behavior to suit different people and different situations” and “I can only argue for ideas which I already believe” (see Gangestad & Snyder, 2000). Self-monitoring is taken to be broadly reflective of differences in likely attitude function. For example, as described by DeBono (1987), the expectation is that high self-monitors will emphasize social-adjustive functions (letting the high self-monitor behave in ways appropriate to the social situation), whereas low self-monitors will favor value-expressive functions (in the sense that the low self-monitor’s attitudes will be chosen on the basis of the degree to which the attitude is consistent with the person’s underlying values).2 So, for example, high self-monitors are likely to especially stress the image-related aspects of products (because of the social-adjustive function), whereas low self-monitors are more likely to focus on whether the product’s intrinsic characteristics and qualities match the person’s criteria for such products. For any such proxy measure, of course, the key question will be the degree to which the proxy is actually related to differences in attitude function, a 77



question probably best addressed by examining the relationship between proxy measures and more direct assessments. In the specific case of personality characteristics such as self-monitoring, presumably such characteristics merely incline persons (in appropriate circumstances) to be more likely to favor one or another function. For instance, it is surely not the case that all the attitudes of high self-monitors (whether toward aspirin or automobiles or affirmative action) serve social-adjustive functions. (See Herek, 2000, pp. 332–335, for commentary on the use of such proxy measures.)



Influences on Attitude Function A variety of factors might influence the function that a given attitude serves. Three such classes of factors merit mention: individual differences (from person to person), the nature of the attitude object, and features of the situation.



Individual Differences Different persons can favor different attitude functions, as straightforwardly illustrated by self-monitoring. As just discussed, high self-monitors appear to favor social-adjustive functions, whereas low selfmonitors seem more likely to adopt value-expressive functions. Other personality correlates of attitude function differences have not received so much recent research attention, although plainly it is possible that other individual-difference variables might be related to differences in attitude function (see, e.g., Katz, McClintock, & Sarnoff, 1957; Zuckerman, Gioioso, & Tellini, 1988). But apart from any underlying personality differences, people’s motivations can vary. For example, different people can have different reasons for volunteering, although those differences might not be systematically related to any general personality disposition.



Attitude Object The function of an attitude toward an object may also be shaped by the nature of the object because objects can differentially lend themselves to attitude functions. For example, air conditioners commonly evoke predominantly utilitarian thoughts (“keeps the air cool,” “expensive to run”), whereas wedding rings are more likely to elicit social-identity thoughts (“represents a sacred vow”; Shavitt, 1990). Similarly, attitudes 78



toward shampoo are determined more by instrumental attributes (such as conditioning hair) than by symbolic ones (such as being a high-fashion brand), whereas for perfume, symbolic attributes are more influential than instrumental ones (Mittal, Ratchford, & Prabhakar, 1990). Each of these objects (air conditioners, shampoo, wedding rings, and perfume) appears to predominantly encourage one particular attitude function (and so might be described as unifunctional). But other objects are multifunctional, in the sense of easily being able to accommodate different attitude functions. For instance, automobiles can readily permit both symbolic and instrumental functions; a person’s attitude toward an automobile might have a largely instrumental basis (“provides reliable transportation”), a largely symbolic one (“looks sexy”), or some mixture of these.



Situational Variations Different situations can elicit different attitude functions (for a general discussion, see Shavitt, 1989, pp. 326–332). For example, if the situation makes salient the intrinsic attributes and outcomes associated with an object, presumably instrumental (utilitarian) functions will be more likely to be activated; by contrast, social-identity functions might be engaged by “situations that involve using or affiliating with an attitude object, or expressing one’s attitude toward the object, in public or in the presence of reference group members” (p. 328). Thus attitude functions may vary depending on features of the immediate situation.



Multifunctional Attitude Objects Revisited As noted above, attitude objects differ in the degree to which they accommodate multiple attitude functions, and this influences the role that individual-difference variations and situational variations can play in determining attitude function. For unifunctional attitude objects (those eliciting predominantly one function), individual-difference variations and situational variations may not have much impact. For example, aspirin is likely to be generally (that is, across individuals and across situations) perceived largely in instrumental terms. But (as emphasized by Shavitt, 1989) multifunctional attitude objects (such as automobiles) represent objects for which individual-difference variations and situational variations are likely to have greater impact on attitude function. It is possible for the attitudes of high and low self-monitors toward automobiles 79



to serve different functions because the attitude object can accommodate different functions. Similarly, situational factors can influence the salience of various functions only if the attitude object permits different attitude functions. The larger point is that the attitude object, individual differences, and situational factors all intertwine to influence attitude function.3



Adapting Persuasive Messages to Recipients: Function Matching One recurring theme in theoretical analyses of persuasion is the idea that to maximize effectiveness, persuasive messages should be adapted (tailored, adjusted) to fit the audience. In functional approaches to attitude, this general idea is concretized as the matching of the persuasive appeal to the functional basis of the recipient’s attitude. Two recipients with identical overall evaluations may have different underlying functional bases for their attitudes, which will require correspondingly different persuasive approaches.



The Persuasive Effects of Matched and Mismatched Appeals Consistent with this analysis, a variety of investigations have found that messages with appeals matching the receiver’s attitude function are indeed more persuasive than messages containing mismatched appeals (Carpenter, 2012b). A good number of these studies have used self-monitoring as an indirect index of variation in attitude function and have focused on how self-monitoring differences are related to differential susceptibility to various types of appeals in consumer product advertising. As discussed above, high self-monitors are expected to favor social-adjustive functions and low self-monitors to favor value-expressive functions; the appeal variation consists of using arguments emphasizing correspondingly different aspects of the attitude object. In the specific domain of consumer products, the contrast can be characterized as a difference between appeals emphasizing the image of the product or its users (a social-adjustive appeal) and appeals emphasizing the intrinsic quality of the product (a value-expressive appeal; see, e.g., Snyder & DeBono, 1987). This contrast is exemplified by a pair of experimental magazine advertisements for Canadian Club whiskey (Snyder & DeBono, 1985, Study 1). Both ads showed a bottle of the whiskey resting on a set of blueprints for a house; 80



for the image-oriented advertisement, the slogan read, “You’re not just moving in, you’re moving up,” whereas the product quality-oriented advertisement claimed, “When it comes to great taste, everyone draws the same conclusion.” Across a number of studies, high self-monitors have been found to react more favorably to image-oriented advertisements than to product qualityoriented ads, whereas the opposite effect is found for low self-monitors (e.g., DeBono & Packer, 1991; Lennon, Davis, & Fairhurst, 1988; Snyder & DeBono, 1985; Zuckerman et al., 1988; cf. Bearden, Shuptrine, & Teel, 1989; Browne & Kaldenberg, 1997; for related work, see DeBono & Snyder, 1989; DeBono & Telesca, 1990).4 Outside the realm of consumer product advertising, parallel differences (between high and low selfmonitors) have been found with related appeal variations. For example, concerning the topic of institutionalization of the mentally ill, DeBono (1987) found that low self-monitors were more persuaded by valueexpressive messages (indicating what values were associated with positive attitudes toward institutionalization) and high self-monitors by socialadjustive messages (indicating that a substantial majority of the receiver’s peers favored institutionalization). Similar effects have been reported concerning dating attitudes (Bazzini & Shaffer, 1995), voting (Lavine & Snyder, 1996), HIV prevention (Dutta-Bergman, 2003), and breastfeeding (Sailors, 2011). Parallel findings have been reported in investigations that assessed individual attitude function differences in ways other than self-monitoring differences. For example, Clary et al. (1994) initially assessed attitude function through participants’ ratings of the importance of various possible reasons for volunteering. Participants were then presented with provolunteering messages that were matched or mismatched to their attitude functions; matched messages were more persuasive than mismatched messages in inducing intentions to volunteer (see also Clary et al., 1998). In Celuch and Slama’s (1995) research, variations in the degree to which persons’ self-presentation motives emphasized “getting ahead” or “getting along” were related to the persuasiveness of messages emphasizing either the self-advancement aspects of a product or the conformity-relevant aspects of a product.5 Similar results have been obtained concerning the effectiveness of different appeals for different types of products. Shavitt (1990, Study 2) compared the effects of a utilitarian or social-identity appeal for either a 81



utilitarian product (such as air conditioners) or a social-identity product (such as greeting cards); brands advertised with function-relevant appeals were preferred over brands advertised with function-irrelevant appeals (so that, for example, ads using utilitarian appeals were preferred over ads using social-identity appeals when air conditioners were advertised, but this preference was reversed when greeting cards were advertised). Finally, these same effects have been observed when the situational salience of attitude functions was varied experimentally. Julka and Marsh (2005) varied the degree to which a knowledge or value-expressive function was activated, and then they exposed participants to either a matched or mismatched persuasive appeal for organ donation; matched appeals produced more favorable attitude change and led participants to be more likely to take an organ-donor registration card. In short, substantial evidence suggests that persuasive appeals that are matched to the receiver’s attitude function will be more persuasive than mismatched appeals (for a review, see Carpenter, 2012b). And it is worth noticing that studies in this area often report relatively large effects, suggesting the substantive importance of functionally matched messages.6



Explaining the Effects of Function Matching Exactly why are function-matched appeals typically more persuasive than unmatched appeals? The answer to this question is not yet entirely clear, but there are two broad possibilities. One is that functionally matched appeals simply speak to a receiver’s psychological needs in ways that unmatched appeals do not. This explanation is unsurprising, really—after all, such correspondence is what makes the appeals matched. A receiver who values a vehicle’s gas mileage more than the image projected by the vehicle will naturally be more persuaded by appeals based on the former than by those based on the latter. Correspondingly, people are likely to perceive functionally matched messages as containing better arguments than mismatched messages (Hullett, 2002, 2004, 2006; Lavine & Snyder, 1996, 2000; see, relatedly, Shavitt & Lowrey, 1992). A second possibility is that function-matched messages are processed more carefully than mismatched messages. For example, in Petty and Wegener’s (1998b) study, high and low self-monitors read consumer product messages that varied in the functional match of the appeals (image-based versus product quality-based appeals) and in the quality of 82



the supporting arguments (strong versus weak arguments). Attitudes were more strongly influenced by argument quality when the message contained matched appeals than when it contained mismatched appeals. For instance, high self-monitors were more influenced by the strength of the arguments when the appeals were image-based than when the appeals were product quality-based. This effect suggests that receivers more carefully scrutinized messages with appeals matching their functional attitude bases than they did messages with mismatched appeals (see also DeBono, 1987; Lavine & Snyder, 1996, 2000; Petty, Wheeler, & Bizer, 2000). Several studies have reported related findings suggesting that high self-monitors more carefully process messages from attractive than unattractive (or expert) communicators (DeBono & Harnish, 1988; DeBono & Telesca, 1990), findings that might reflect the propensity for high self-monitors to favor social-adjustive functions. And this explanation is at least not inconsistent with evidence suggesting that function-matched messages may be better remembered than mismatched messages (DeBono & Packer, 1991, Study 3). If this second explanation is sound, then at least sometimes functionmatched messages should be less persuasive than mismatched messages, namely, when the messages contain poor-quality arguments (as indeed was observed by Petty & Wegener, 1998b). The weaknesses of such poorquality messages will be recognized by receivers who scrutinize the message closely but can go unnoticed by receivers who do not process so carefully. So this explanation supposes that the widely observed greater persuasiveness of function-matched messages is actually a consequence of the generally high argumentative quality of the appeals; with poorerquality arguments, function-matched messages might not have displayed such a persuasive advantage. Both explanations might turn out to have some merit. For example, it may be that when message scrutiny is already likely to be high, the different intrinsic appeal of matched and mismatched arguments will play a key role, whereas in other circumstances, the functional match or mismatch of arguments will influence the degree of scrutiny given (Petty et al., 2000; Ziegler, Dobre, & Diehl, 2007).7 Further research on these questions will be welcomed.



Commentary 83



Generality and Specificity in Attitude Function Typologies The general enterprise of functional attitude analysis is driven by the search for a small set of universal, exhaustive, and mutually exclusive attitude functions that can be used to dependably and perceptively distinguish any and all attitudes. But as discussed earlier, there is not yet a consensus on any such set of functions, and perhaps there never will be such consensus. The general idea of functional analysis, however, can still be of use in illuminating attitudes and persuasion processes, even without some universal scheme of attitude functions. It has proved possible, for at least some attitude objects or issues, to distinguish different attitudinal functions in a way that is dependable (reliable) and that provides insight into attitudes on that subject. For example, functional analyses have provided insight concerning attitudes on such matters as volunteering (Stukas, Snyder, & Clary, 2008), meat (M. W. Allen & Ng, 2003), and democracy (Gastil, 1992). In particular, a number of studies have illuminated attitudes on various HIV/AIDS-related issues by considering the relative contributions of symbolic and instrumental attitudinal underpinnings (e.g., Crandall et al., 1997; Herek & Capitanio, 1998; Hosseinzadeh & Hossain, 2011). That is, the functions served by attitudes on a given subject can be analyzed even in the absence of a general typology of attitude functions; insights into the motivational dynamics of particular attitudes are possible even without having some universal set of functions. Of course, there is variation in the specific functional typologies used to analyze these different particular attitudes. On one subject, it may be helpful to distinguish various subtypes of a function, but distinguishing those subtypes may not be necessary when considering attitudes on a different topic. For instance, a generalized utilitarian function might suffice when analyzing the functions of attitudes toward amusement parks, whereas when analyzing attitudes toward democracy, it is useful to distinguish personal utility functions (my beliefs about how democracy benefits me personally) and social utility functions (my beliefs about how democracy benefits society as a whole; see Gastil, 1992). The larger point is that although one may hope that continuing research will eventuate in a well-evidenced small set of general and well-articulated attitude functions, one need not wait for such an outcome to appreciate the value of functional attitude analyses. Indeed, even if there comes to be some 84



consensus on a general attitude function typology, analyses of specific attitudes will still almost certainly require the typology to be modified (elaborated, refined, adapted) to provide maximum illumination of the particular attitude under study.



Functional Confusions Some Functional Distinctions There is some underlying murkiness in the conceptualization of attitude functions. This can be displayed by considering that there are distinctions —often unappreciated—among the functions of an attitude, the functions of expressing an attitude, and the functions of the attitude object. Consider first that there is a distinction to be drawn between the functions of an attitude (that is, the functions of having that attitude) and the functions of expressing an attitude. For example, imagine that John has unfavorable views about minorities. His having that attitude might serve an ego-defensive or self-esteem maintenance function; having that attitude makes him feel better about himself—even if he never reveals his views to anyone else. On the other hand, his expressing that attitude around his bigoted friends might serve a social-adjustive function, one of letting him fit in with people who are expressing similar attitudes. Thus the job done by the attitude (the having of the attitude) and the job done by the expression of the attitude can be different.8 Similarly, there is a plain distinction between the functions of an attitude and the functions of an attitude object. After all, no one would give the same answer to the question, “What are attitudes toward amusement parks good for?” and the question, “What are amusement parks good for?” The functions of an attitude (the jobs that the evaluation does) and the functions of an object (the jobs that the object does) are plainly different. Finally, there is a distinction between the functions of an attitude object and the functions of expressing an attitude. The purposes served by abolition of capital punishment (the attitude object) are different from the purposes served by a person’s saying, “I support the abolition of capital punishment” (the expression of the attitude). Consider, for example, that a person might believe that abolishing capital punishment would serve an instrumental function (“Doing away with capital punishment would 85



prevent execution of the innocent”), but this does not mean that the person’s expressing opposition to capital punishment serves an instrumental function; depending on the circumstances, expression of that attitude might serve some thoroughly symbolic end (symbolizing one’s values, for instance). Thus there is a difference between the jobs done by the attitude object (the product, the policy) and the jobs done by expressing attitudes concerning that object.9



Conflating the Functions These three elements—the functions of an attitude, the functions of expressions of an attitude, and the functions of an attitude object—are commonly conflated in theory and research on attitude functions. For example, the functions of having an attitude and the functions of expressing an attitude are often not carefully distinguished. Such conflation can be detected in (among other places) Katz’s (1960, p. 187) discussion of voting behavior as reflecting a value-expressive function, in Snyder and DeBono’s (1989, p. 341) description of social-adjustive attitudes as allowing people to fit into social situations smoothly, and in Shavitt and Nelson’s (2000, p. 55) treatment of the meanings communicated by a person’s consumer product choices as an aspect of a social-identity function of attitudes. All these would seem to be more accurately characterized as describing functions of expressing attitudes rather than functions of having attitudes. Similarly, the functions of attitudes and the functions of attitude objects are often treated as if these were indistinguishable. For instance, Clary et al. (1994) quietly shift from discussing the idea that “attitudes could serve a variety of distinct social and psychological functions” (and that “the same attitude could serve very different motivational functions for different people”) to discussing “the relevant motivations underlying volunteerism” (note: motivations underlying volunteerism, not motivations underlying attitudes toward volunteerism) and “the specific functions served by volunteerism” (pp. 1130–1131; again, not the specific functions served by positive or negative attitudes toward volunteerism but the functions served by volunteerism itself). Similarly, Ennis and Zanna’s (1993) opening paragraphs slide from discussion of “the psychological functions of attitudes” to discussion of “the psychological functions of a product” (p. 662; see also Ennis & Zanna, 2000). In discussions of attitude function, such elision of the functions of attitudes, the functions of attitude objects, and the functions of attitude expressions is common (see, e.g., 86



Lavine & Snyder, 2000; Pratkanis & Greenwald, 1989; Shavitt, 1990). This state of affairs suggests that it may be useful to reconsider the assessment and conceptualization of attitude function with the relevant distinctions in mind.



Reconsidering the Assessment and Conceptualization of Attitude Function Assessment of Attitude Function Reconsidered The common conflation of attitude function and attitude object function naturally raises some questions about procedures for assessing attitude function. The procedures that assess attitude function through coding freeresponse data or through standardized scale item data rest on the idea that the function an attitude serves is reflected in the beliefs held about the attitude object, the reasons given for the attitude, the importance of alternative reasons for the attitude, and so forth. But all these may most directly reflect not the function served by the respondent’s attitude but rather the respondent’s perceptions of the functions served by the object (the attitude object’s useful properties, the purposes served by the attitude object, and so forth)—and hence the respondent’s perception of what is valuable about the object. To concretize this, imagine asking people questions of the form, “What’s important about X?” or “What’s right or wrong [or good or bad] about X?”—that is, questions of the sort one might use in an open-ended questionnaire designed to assess attitude function. Depending on the topic, a variety of answers might be received. For example, when asked, “What’s right or wrong about gun control?” Al says, “It encourages crime by disarming citizens,” but Bob says, “It infringes rights.” When asked, “What’s good or bad about this automobile?” Christine says, “It gets good gas mileage,” whereas Donna says, “It makes me look cool.” Asked, “What’s important about volunteering?” Ed says, “It helps me develop job skills,” but Frank says, “It contributes to the community.” One straightforward way of understanding the variation in these answers is to see it simply as reflecting differences in what people value (not only differences in general abstract values but also differences in what goals they seek, what attributes they value in particular types of objects, and so on). That is, these sorts of differences—which commonly have been taken 87



to indicate differences in attitude functions—might more lucidly be characterized as simply differences in what people value in (that is, what they want from) attitude objects. Consider, for example, the procedure in which attitude function assessment is based on respondents’ perceived importance of various reasons for volunteering (Clary et al., 1998; Clary et al., 1994). On the face of things, it seems that this procedure classifies persons on the basis of their perceived importance of various functions of the action of volunteering (jobs that volunteering performs, outcomes of volunteering), not on the basis of any functions of their attitudes. So, for instance, someone who says that improving one’s resume is an important reason for volunteering appears to be identifying a perceived important function of (job done by, purpose served by) the action—which is not the same as identifying a function of one’s positive or negative attitude toward that action. Proxy measures of attitude function such as self-monitoring are, if anything, even more susceptible to being understood in this way. For example, it is plain that high and low self-monitors can (in appropriate circumstances) have systematically different beliefs about attitude objects. But these different beliefs appear to correspond to differences in how selfmonitors value certain functions of the object. (For example, high selfmonitors value identity projection functions of automobiles more than do low self-monitors.) Just because high self-monitors value certain functions of objects more than do low self-monitors does not show that the attitudes of high self-monitors serve different functions than do the attitudes of low self-monitors. To put it another way: High and low self-monitors want different things from their consumer products. But this does not mean that high and low self-monitors want different things from their attitudes. One might plausibly say that high and low self-monitors want their attitudes to do the same job—the job of identifying good and bad products for them.10 Thus, instead of the supposition that high and low self-monitors have attitudes that serve different functions, what seems invited is the conclusion that although high and low self-monitors may sometimes differ in the criteria they use to appraise objects, the underlying function of the evaluation (the function of the attitude) is identical. Consider, as a concrete example, Wang’s (2009) study of attitudes toward 88



regular physical activity. Respondents rated the desirability and likelihood of various possible consequences of regular exercise. These consequences were then grouped into three attitude functions: a “utilitarian” function (e.g., “would help me reduce stress,” “would help me feel more energetic”), a “social-identity” function (e.g., “would provide me with more opportunities to socialize,” “would help me improve my social relationships”), and a “self-esteem maintenance” function (e.g., “would help me lose weight,” “would help me stay in shape”). The importance of these different functions of exercise (note: not functions of attitudes toward exercise, but functions of exercise) varied across people. In particular, the importance of social-identity outcomes (as influences on intention) varied depending on self-monitoring: high self-monitors placed more emphasis on social-identity outcomes than did low self-monitors. So although high and low self-monitors appear to want different things from exercise (they differently value various outcomes), they would seem to want the same thing from their attitudes (namely, serving the function of indicating whether regular exercise would be a good thing for them to do). In sum, the procedures commonly used for assessing attitude functions can instead be understood as assessing variations in the perceived value or importance of attributes (or functions) of the attitude object.11



Utilitarian and Value-Expressive Functions Reconsidered Against this backdrop, it may be useful to reconsider how utilitarian and value-expressive attitude functions have been conceptually differentiated. In Katz’s (1960) treatment of these two functions, utilitarian attitudes are exemplified by attitudes based on economic gain or other concrete rewards (p. 171), whereas value-expressive attitudes are concerned with abstract “central values” and self-images (p. 173). Indeed, Katz specifically described value-expressive attitudes as different from attitudes aimed at “gaining social recognition or monetary rewards” (p. 173). A similar way of distinguishing utilitarian and value-expressive functions appears in Maio and Olson’s (1994, 1995) research examining the hypothesis that persons with value-expressive attitudes will exhibit closer connections between attitudes and values than will persons with utilitarian attitudes. In this work, the values implicated in value-expressive attitudes are conceived of as abstract ends such as equality, honesty, and freedom (Maio & Olson, 1994, p. 302), as opposed to the narrower self-interested ends represented by utilitarian attitudes; values are “evaluations of abstract 89



ideas (e.g., equality, honesty) in terms of their importance as guiding principles in one’s life” (Maio & Olson, 1995, p. 268). From this point of view, a person considering whether to make a charitable donation who thinks about “the importance of helping others” has a value-expressive attitude, whereas people who think about “whether they can afford to donate” have utilitarian attitudes (Maio & Olson, 2000a, p. 251). That is, value-expressive attitudes have commonly been distinguished from utilitarian attitudes on the basis of the nature of the outcomes sought: Abstract, prosocial ends indicate value-expressive attitudes, whereas concrete, self-enhancing ends indicate utilitarian attitudes. But plainly this way of distinguishing attitudes seems less a matter of attitude function than a matter of the abstractness or nobility of the ends served by the object. On the conventional view, “protecting the environment” might be a value, but “protecting my savings account” would not be. Yet obviously each represents a potential outcome that can be valued, and hence each represents a basis of assessment of objects (assessment of objects for the degree to which the objects realize the outcome). Thus value-expressive attitudes and utilitarian attitudes arguably do not actually serve different attitude functions; the underlying attitude function is identical in the two cases (evaluative object appraisal in the service of obtaining satisfactory outcomes), although the criteria for assessing objects (that is, the outcomes of interest) may vary. This sort of reasoning led Maio and Olson (2000a, pp. 258–260) to introduce the idea of “goal-expressive” attitudes, precisely meant to “encompass what Katz referred to as value-expressive and utilitarian functions” (p. 259). By collapsing value-expressive and utilitarian attitudes into one functional category, this approach abandons the idea that valueexpressive and utilitarian attitudes serve different purposes; it recognizes their similarity in abstract attitude function (appraisal) while not losing sight of the variation possible in substantive motivational content (for a related view, see Eagly & Chaiken, 1998, p. 304; for a similar treatment of value-expressive and social-adjustive attitudes, see Hullett & Boster, 2001).



Summary Taken together, these considerations invite a simpler, more straightforward account of much research on attitude function variation. Specifically, this work might more perspicaciously be described as work identifying 90



variation in what people value (their wants, goals, evaluations of various properties of objects, and so on). As indicated above, both the procedures commonly used to differentiate attitude functions and the conceptual treatment of value-expressive and utilitarian functions can be seen to distinguish cases on the basis of persons’ values, not on the basis of attitude function.



Persuasion and Function Matching Revisited Existing research on persuasion and function matching is entirely congenial with the idea that apparent attitude function variation reflects variation in people’s values. Indeed, approached from such a perspective, it is hardly surprising that function-matched messages are so often more persuasive than unmatched messages—because the matched messages speak to what people want. Consider the case of self-monitoring: high and low self-monitors characteristically differ in their evaluation of various outcomes and object attributes. For instance, high self-monitors characteristically place a higher value on aspects of self-image presentation. Given this difference, it is perhaps unsurprising that high self-monitors find image-oriented appeals and certain normatively oriented appeals (concerning what their peers think) to be especially congenial (e.g., DeBono, 1987; Snyder & DeBono, 1985); such appeals fit their values (not their attitude functions). As another example, consider the previously mentioned finding that variations in the degree to which persons’ self-presentation motives emphasize “getting ahead” or “getting along” are related to the persuasiveness of messages emphasizing either the self-advancement aspects of a product or the conformity-relevant aspects of a product (Celuch & Slama, 1995). These results can obviously be straightforwardly described as a matter of matching appeals to the motivations (wants, desires, values, goals) of message receivers (specifically, motivations for self-presentation). The same holds true for appeals matched not to individual-difference variations but to variations in the nature of the object. Different objects are valued for different types of reasons. People generally want certain sorts of things from air conditioners and different sorts of things from greeting cards—and hence appeals matched to what people want from these objects (not to what they want from their attitudes toward those objects, but to 91



what they want from those objects) will naturally be likely to enjoy some persuasive advantage (as observed by Shavitt, 1990). Similarly for situational variations: When certain values (attributes, outcomes, etc.) are made more salient, persuasive appeals engaging those wants are likely to be more successful than appeals engaging nonsalient desires (e.g., Maio & Olson, 2000a, Study 4). This redescription is also congenial with the proposed account of functionmatching persuasion effects that suggests that functionally matched messages engender greater message scrutiny. It would not be surprising that a receiver’s attention be especially engaged by messages that appear to be discussing something important to the receiver. Finally, research directly exploring the role of function-relevant values has found that the strength with which recipients held such values influenced the degree to which recipients were persuaded by corresponding appeals: the stronger a recipient held the values engaged by a message’s appeals, the greater the message’s persuasiveness. The implication is that functionmatched appeals are more persuasive than mismatched appeals because the matched appeals engage values that the recipient holds more strongly (Hullett, 2002, 2004, 2006; Hullett & Boster, 2001; see, relatedly, Bailis, Fleming, & Segall, 2005). In short, existing research concerning attitude functions and persuasive appeals appears to be well captured by two core ideas: First, what is valued varies. Different persons can have different values (with systematic relationships here, such as connected with self-monitoring differences); different types of objects are characteristically valued for different reasons; and as situations vary so can the salience of different values. Second, persuasive messages are more effective when they engage what people value than when they do not. These two ideas are currently clothed in talk about variation in attitude function, but such talk is at least misleading and arguably dispensable in favor of talk about variation in values.12 In the long run, however, clear treatment of variation in values will require some typology of values, that is, some systematic analysis of the ways in which values (goals, desired properties of objects, etc.) can vary (for some classic examples, see Rokeach, 1973; Schwartz, 1992). The empirical success of research using attitude function categories suggests that these categories might provide some leads in this regard (although now the consequences of the lack of 92



agreement about a functional taxonomy may be more acutely felt). For example, a carefully formulated version of the symbolic-instrumental contrast might serve as one way of distinguishing variation in values (see Allen, Ng, & Wilson, 2002; Eagly & Chaiken, 1998, p. 304). It may be profitable, however, to consider other sources as well and, in particular, to consider independent work on typologies of general, abstract values as a possible source of further insight (as recommended by Maio & Olson, 2000a).



Reviving the Idea of Attitude Functions The analysis offered in the preceding section might appear to recommend jettisoning the idea of attitude functions and replacing it with an analysis of systematic differences in what people value. Such an approach does seem to capture much of the work conducted under the aegis of functional approaches to attitude. That reframing, however, arguably fails to appreciate the potential contribution afforded by considering genuine differences in attitude function. The value-based reframing of attitude functions implicitly focuses on only one attitude function, that of object appraisal (evaluative appraisal in the service of satisfaction of wants). But this overlooks another apparent general function of attitudes, a self-maintenance function, as exemplified by Katz’s (1960) ego-defensive function.13 The ego-defensive function is genuinely a function of an attitude, not a function of an attitude object. For example, the ego-defensive function of prejudicial attitudes toward a minority group is different from the function of the minority group itself: The minority group does not serve the function of bolstering the person’s self, but the negative attitude toward the minority group can.14 This suggests that there are at least two distinguishable broad functions of attitude.15 But most of the work on persuasion and attitude functions has implicitly addressed attitudes serving object appraisal functions and so has focused on adapting messages to different bases of object appraisal. Scant work is concerned with (for example) how persuasion might be effected when attitudes serve ego-defensive ends or with how to influence attitudes adopted because of the reference group identification purposes served by holding the attitude.16 Thus there is good reason to want to retain some version of the idea of 93



different attitude functions, as illustrated by the apparent usefulness of a contrast between object-appraisal functions and self-maintenance functions.17 But if the idea of attitude function is to be revived, a consistent and clear focus on the functions of attitudes (as opposed to the functions of objects or the functions of attitude expression) will be needed, accompanied by attention to the continuing challenge of attitude function assessment.



Conclusion Despite some conceptual unclarities, work on the functional approach to attitudes has pointed to some fundamentally important aspects of attitude and persuasion. In cases in which attitudes are primarily driven by an interest in object appraisal, persuaders will want to attend closely to the receiver’s basis for assessing the attitude object. What people value can vary, and hence the persuasiveness of a message can depend in good measure on whether the message’s appeals match the receiver’s values.



For Review 1. Explain the general idea behind functional approaches to attitude. 2. In Katz’s classic analysis of attitude function, what four attitude functions are identified? Explain the utilitarian function. What techniques are best adapted to changing attitudes serving a utilitarian function? Explain the ego-defensive function. What techniques are best adapted to changing attitudes serving an ego-defensive function? Explain the value-expressive function. Under what conditions are attitudes serving a value-expressive function likely to be susceptible to change? Explain the knowledge function. What is the primary mechanism of change for attitudes that serve a knowledge function? 3. Is there a consensus about a particular typology of attitude functions? Is there a broad distinction (among functions) that is common to alternative functional typologies? Explain symbolic functions of attitude. Explain instrumental functions of attitude. 4. Describe three ways of assessing the function of a given attitude. What is free-response data? Explain how free-response data can be analyzed to reveal attitude functions. Explain how standardized questionnaires can be used to assess attitude functions. Explain how proxy indices can be used to assess attitude functions; give an example. 94



5. Identify three kinds of factors that can influence attitude function. Describe how individual differences can influence attitude function; give an example. Explain how the nature of the attitude object can influence attitude function. Give examples of objects for which attitudes likely serve a generally instrumental function; give examples of objects for which attitudes likely serve a generally symbolic function. What are multifunctional attitude objects? Describe how situational variations can affect attitude function. For what kinds of attitude objects are individual differences and situational variations likely to have the greatest effect on attitude function? 6. Explain how functional approaches provide a basis for adapting persuasive messages to recipients. What is function matching? Are function-matched appeals generally more persuasive than mismatched appeals? Describe image-oriented advertising appeals. Describe product quality-oriented advertising appeals. Are high self-monitors generally more persuaded by image-oriented or by product qualityoriented appeals? Are low self-monitors generally more persuaded by image-oriented or by product quality-oriented appeals? Describe two possible explanations of the persuasive advantage of functionmatched appeals over mismatched appeals. 7. Explain how the general idea of attitude functions can be useful even in the absence of an agreed-upon universal typology of attitude functions. How might different functional typologies be useful for different specific attitudes? 8. Explain the distinction between the functions of an attitude and the functions of expressing an attitude. Explain the distinction between the functions of an attitude and the functions of an attitude object. Explain the distinction between the functions of an attitude object and the functions of expressing an attitude. Describe how these different functions have been confused in theory and research about attitude functions. 9. Explain how differences in attitude function as assessed through open-ended questions reflect differences in what respondents value in attitude objects. Explain how differences in attitude function as assessed through self-monitoring reflect differences in what respondents value in attitude objects. Explain how the distinction between utilitarian and value-expressive functions reflects differences in what respondents value in attitude objects. 10. How can the idea of function-matched appeals be redescribed in terms of matching the audience’s values? Describe the difference between object appraisal and self-maintenance as two broad attitude 95



functions. Which has been the focus of most research attention?



Notes 1. Snyder and DeBono’s (1989) description of the social-adjustive function implicitly focused not on the function of the attitude but on the function of the attitude expression. By contrast, M. B. Smith et al.’s (1956) discussion of this function emphasized that “one must take care to distinguish the functions served by holding an opinion and by expressing it” (p. 41). The potential social-adjustive function of attitude expression is straightforward enough (e.g., one can fit into social situations by expressing this or that opinion). The social-adjustive function of simply holding an attitude, on the other hand, is “at once more subtle and more complex” (p. 42). At base, it involves the creation of feelings of identification or similarity through attitudes; the mere holding of certain attitudes can be “an act of affiliation with reference groups” (M. B. Smith et al., 1956, p. 42), independent of any overt expression of the attitude. Unhappily, as discussed later in this chapter, the distinction between attitude functions and attitude expression functions has not commonly been closely observed. 2. The use of the term value-expressive is potentially confusing here. The kind of attitude function putatively favored by low self-monitors might better be described as value-matching (because the low self-monitor’s attitudes are influenced by the extent to which the object’s properties match the person’s values). This is different from the value-expressive function described by Katz (1960), in which satisfaction is had by holding attitudes that represent (express) fundamental values. Katz’s valueexpressive function seems rather more symbolic than instrumental (and hence more similar to the ego-defensive function than to the utilitarian or knowledge functions); the attitude’s basis is less the particular pros and cons of the attitude object and more the symbolic connection with core values. By contrast, DeBono’s (1987) value-expressive function seems more instrumental than symbolic; the attitude’s job is not to represent or display to others one’s central values but rather to summarize the object’s pros and cons as assessed against one’s desiderata for such objects. 3. As an illustration of the complexity that’s possible here, consider that attitudes toward organ donation appear to be unifunctional (specifically, serving the same attitude function for high and low self-monitors alike) but attitudes toward discussing organ donation appear multifunctional (serving 96



different functions depending on self-monitoring; Wang, 2012). 4. With respect to consumer product advertising, the differences between high and low self-monitors extend beyond the differential appeal of imagebased and product quality-based ads. There are also related differences in the ability to remember whether an ad has been seen before (e.g., high selfmonitors more accurately remember exposure to image ads than to quality ads; DeBono & Packer, 1991, Study 3); in how self-relevant ads are perceived to be (e.g., high self-monitors see image ads as more selfrelevant than quality ads; DeBono & Packer, 1991, Study 2); in the types of advertisements they create for multiple-function attitude objects such as watches (e.g., low self-monitors prefer to use utilitarian appeals, whereas high self-monitors prefer social-identity arguments; Shavitt & Lowrey, 1992); and in the impact of the appearance (DeBono & Snyder, 1989), name (Smidt & DeBono, 2011), or packaging (DeBono, Leavitt, & Backus, 2003) of the product (e.g., the better-looking the car, the higher the quality ratings given by high self-monitors). For a general discussion, see DeBono (2006). 5. Actually, there are a number of studies that (a) are not commonly treated as representing research on attitude function-matching and (b) may not even cite attitude function-matching research but that nevertheless (c) examine the relative effectiveness of persuasive appeals that have been designed to match variations in receivers’ psychological needs as extrapolated from some individual-difference variable (thus paralleling the research format of much function-matching research). Studies by Cesario, Grant, and Higgins (2004, Study 2), Orbell and Hagger (2006), and Aaker and Schmitt (2001)—examining, respectively, regulatory focus (prevention-focused versus promotion-focused), consideration of future consequences (temporally distant versus temporally proximate consequences), and individualism-collectivism (as reflected in cultural variations)—provide just three examples. For a general discussion of this point, see O’Keefe (2013a). 6. In a meta-analysis of the effect of function matching on persuasion, Carpenter (2012b) reported a mean effect size (expressed as a correlation) of .37 across 38 cases, and a mean effect size of .31 for the 29 cases in which the basis of functional matching was self-monitoring (as opposed to, e.g., the nature of the attitude object). But for two reasons these means should be interpreted cautiously. First, these mean effects were based on effect sizes that had been adjusted for measurement unreliability; the 97



application of such adjustments inflates effect sizes relative to those in other persuasion meta-analyses based on unadjusted effect sizes. Second, at least some studies in this research area have used designs in which a given participant saw both a matched and a mismatched appeal, often in close proximity; as Shavitt (1990, pp. 141–142) pointed out, such designs might be expected to yield larger effect sizes than would the more usual between-subjects designs (in which a participant sees only one kind of appeal). 7. Some readers will recognize fragments of elaboration likelihood model reasoning here (see Chapter 8), and specifically the idea that the variable of functional matching versus mismatching might (like many variables) play multiple roles in persuasion, depending on the circumstance; for some amplification, see Petty et al. (2000, p. 145). 8. To further cement that distinction, notice that people might express attitudes they do not hold simply because the (deceptive) expression of the attitude serves some purpose, some function. A lifelong committed Democrat, newly introduced to a group of Republicans, is asked by them about a preference among presidential candidates. The Democrat strongly prefers the Democratic candidate but—not wanting to initiate a potentially unpleasant discussion—says, “I prefer the Republican.” The function served by the negative attitude toward the Republican candidate is obviously different from the function served by expressing a positive attitude toward that candidate. The larger point is that the functions of attitudes should not be confused with the functions of expressing attitudes. (A complexity: The possessing of an attitude can serve the purpose of having the attitude available for ready expression. But this does not underwrite confusing the functions of attitudes with the functions of expressing attitudes.) 9. There is a complexity here, however. The functions of an object and the functions of expressing an attitude toward that object can sometimes overlap (or coincide), at least in the realm of attitude objects that can be possessed or used, such as consumer products. For such objects, possession or use of the attitude object presumably counts as expression of the corresponding favorable attitude (one’s ownership of the object presumably expresses one’s liking for the object), and hence similar jobs can be potentially done by the attitude object (that is, one’s having or using the attitude object) and by other means of expressing the attitude (e.g., saying one likes the object). 98



10. In a sense, of course, the consumer product attitudes of high and low self-monitors do different jobs, because the attitudes of high self-monitors focus on one type of product attribute and the attitudes of low selfmonitors focus on another type: High self-monitors want their attitudes to do the job of identifying objects that satisfy high self-monitor values, and low self-monitors want their attitudes to do the job of identifying objects that satisfy low self-monitor values. However, such a way of differentiating attitude functions could be taken to absurd lengths, in that whenever two persons differentially valued some attribute of an object, their attitudes could be said to serve different functions; if Alice values, but Betty does not, an automobile’s having a built-in navigation system, then their attitudes toward automobiles serve different functions (in that only Alice’s attitude would do the job of identifying cars that satisfy Alice’s valuing of navigation systems). The real question is how to group different possible attitude jobs (when to lump them together, when to distinguish them), and the suggestion here is that it will be useful to recognize that although high and low self-monitors may vary in what they value, there is a sense in which the fundamental job done by their attitudes —evaluative appraisal in the service of value satisfaction—is the same. (In particular, as will be suggested shortly, attitudes driven by this sort of interest look rather different from attitudes driven by an interest in ego protection.) 11. As Eagly and Chaiken (1993, p. 490; 1998, p. 308) have stressed, early functional approaches emphasized latent motivational aspects of attitudes, aspects not necessarily apparent in manifest belief content or conscious thought—and hence not necessarily well captured by coding the manifest content of answers to open-ended questions or by examining responses to standardized self-report instruments. 12. For a related attempt at reinterpreting function-matching appeal research in ways that do not involve reference to attitude functions, see Brannon and Brock (1994), who propose that “schema-relevance,” not attitude function relevance, actually underlies the findings of attitude function research. 13. A self-maintenance function might include not only ego-defensive functions but also those social-adjustive functions of holding (as opposed to expressing) attitudes, as described by M. B. Smith et al. (1956, p. 42) and mentioned above in note 1; having a given attitude can create feelings of identification or similarity, thus serving the function of creating and 99



maintaining one’s view of oneself. 14. The ego-defensive function of the attitude may also be shared, however, by the attitude expression. That is, expressing prejudicial attitudes may serve an ego-defensive function. But the focus here is on functions of attitudes (not functions of attitude expression), and the point is that ego defense is indeed one job that can be done by the holding of an attitude. 15. In fact, Pratkanis and Greenwald’s (1989) analysis proposed just these two functions: “First, an attitude is used to make sense of the world and to help the organism operate on its environment. … Second, an attitude is … used to define and maintain self-worth” (p. 249). This latter function unfortunately elides attitude function and attitude expression function: “We attach different labels to this self-related function of attitude, depending on the audience (public, private, or collective) that is observing the attitude and its expression” (p. 249). 16. Some work exists concerning attitudes about objects that serve selfrelated purposes (such as attitudes about class rings), but this is different from work concerning attitudes serving self-related purposes. 17. To be sure, even this contrast is contestable, in the following sort of way: “Self-maintenance is itself a value, something wanted. Thus even ego-defensive attitudes actually reflect an object appraisal function (evaluative appraisal in the service of value satisfaction); it’s just that ‘self-maintenance’ is the value that’s being served instead of some other value.” And it is certainly true that understood in a sufficiently abstract way, all attitudes presumably (indeed, perhaps by definition) serve some broad appraisal function. Still, there looks to be a difference between appraisal that is in some sense object-driven (I know what I’m looking for in aspirin or automobiles or whatever, and I engage in object appraisal to see how this candidate stacks up) and appraisal that seems somehow selfdriven (I want to ensure a certain sort of self-evaluation, and I engage in object appraisal to produce that outcome). Expressed differently: The attitude functions of high self-monitors (“I like the car because it’s sexy”) and low self-monitors (“I like the car because it’s reliable”) look rather similar when contrasted with those of the bigot’s ego-defensive prejudices. And that contrast is particularly sharp from a persuader’s point of view: The same general sort of approach might be taken to persuade high and low self-monitors (emphasizing different object attributes, to be sure, but 100



otherwise a similar approach), whereas persuading bigots likely requires something rather different. For a similar general conclusion, see Carpenter, Boster, and Andrews (2013).



101



Chapter 4 Belief-Based Models of Attitude Summative Model of Attitude The Model Adapting Persuasive Messages to Recipients Based on the Summative Model Research Evidence and Commentary General Correlational Evidence Attribute Importance Belief Content Role of Belief Strength Scoring Procedures Alternative Integration Schemes The Sufficiency of Belief-Based Analyses Persuasive Strategies Reconsidered Conclusion For Review Notes



This chapter discusses belief-based approaches to the analysis of attitude and attitude change. The central theme of these approaches is that one’s attitude toward an object is a function of the beliefs that one has about the object. There are a number of variants of this general approach, with the variations deriving from differences in what features of beliefs are seen to contribute to attitude and from differences in how beliefs are seen to combine to yield an attitude. One particular belief-based approach, the summative model of attitude, has enjoyed special prominence among students of persuasion and social influence and hence is the focus of the chapter’s attention. (The summative model also figures in reasoned action theory; see Chapter 6.)



Summative Model of Attitude The Model The summative model of attitude (Fishbein, 1967a, 1967b) is based on the claim that one’s attitude toward an object is a function of one’s salient 102



beliefs about the object. For any given attitude object, a person may have a large number of beliefs about the object. But at any given time, only some of these are likely to be salient (prominent)—and it is those that are claimed to determine one’s attitude. In, say, a public opinion or marketing questionnaire, one might elicit the respondent’s salient beliefs (e.g., about a product or a political candidate) by asking the respondent to list the characteristics, qualities, and attributes of the object. Across a number of respondents, the most frequently mentioned attributes represent the modally salient beliefs, which can be used as the basis for a standardized questionnaire. (For discussion of procedures for identifying salient beliefs, see Ajzen & Fishbein, 1980, pp. 68–71; Ajzen, Nichols, & Driver, 1995; Breivik & Supphellen, 2003; Fishbein & Ajzen, 2010, pp. 100–103; Middlestadt, 2012; van der Pligt & de Vries, 1998b; for some complexities, see Roskos-Ewoldsen & Fazio, 1997.) In particular, the model holds that one’s attitude toward an object is a function of belief strength (that is, the strength with which one holds one’s salient beliefs about the object) and belief evaluation (the evaluation one has of these beliefs).1 Specifically, the relation of belief strength and belief evaluation to attitude is said to be described by the following formula: Ao = Σbiei where AO is the attitude toward the object, bi is the strength of a given belief, and ei is the evaluation of a given belief. The sigma (Σ) indicates that one sums across the products of the belief strength and belief evaluation ratings for each belief. That is, one multiplies each belief evaluation by the strength with which that belief is held and then sums those products to arrive at an estimate of the overall attitude toward the object. If there are five salient beliefs about the object, then the attitude estimate is given by b1e1 + b2e2 + b3e3 + b4e4 + b5e5. The procedures for assessing the elements of this model are well established. One’s attitude toward the object (AO) can be obtained by familiar attitude measurement techniques. The strength with which a belief is held (Σbi) can be assessed through scales such as likely–unlikely, probable–improbable, and true–false. The evaluation of a belief (ei) is assessed through semantic-differential evaluative scales such as good–bad, desirable–undesirable, favorable–unfavorable, and the like. As an example: Suppose that a preliminary survey had indicated that the 103



most salient beliefs held about Senator Smith by the senator’s constituents were that the senator supports defense cuts, is helpful to constituents, is respected in the Senate, and is unethical. One might assess the strength with which the first of these beliefs was held by respondents through items such as those listed in Figure 4.1. The evaluation of that belief can be assessed with items such as those in Figure 4.2. Figure 4.1 Examples of questionnaire items assessing belief strength (bi).



Figure 4.2 Examples of questionnaire items assessing belief evaluation (ei).



Suppose (to simplify matters) that for each belief, belief strength and belief evaluation were assessed by a single scale (perhaps “likely–unlikely” for belief strength, “good–bad” for belief evaluation) scored from + 3 (likely or good) to −3 (unlikely or bad). A particular respondent might have the belief strength and belief evaluation ratings for the four salient beliefs about Senator Smith shown in Figure 4.3. The respondent in Figure 4.3 believes that it is quite likely that the senator supports defense cuts (belief strength of +3), and supporting defense cuts is seen as a moderately negative characteristic (evaluation of -2); the respondent thinks it very unlikely that the senator is helpful to constituents (helpfulness to constituents being thought to be a very good quality); the respondent thinks it moderately likely that Smith is respected in the Senate, and that is a slightly positive characteristic; and the respondent thinks it rather unlikely that Smith possesses the highly negative characteristic of being unethical. Figure 4.3 Estimating attitude from belief strength (bi) and belief evaluation (ei).



104



Because (in this example) each belief strength score (bi) can range from −3 to +3 and each belief evaluation score (ei) can range from −3 to +3, each product (biei) can range from -9 to +9, and hence the total (across the four beliefs in this example) can range from −36 to +36. A person who thought that the qualities of supporting defense cuts, being helpful to constituents, and being respected in the Senate were all very positive characteristics (belief evaluations of +3 in each case) and who thought it very likely that the senator possessed each of these qualities (belief strength of +3 for each), and who also thought it quite unlikely (−3 belief strength) that the senator possessed the strongly negative (−3 belief evaluation) characteristic of being unethical would have a total (Σbiei) of +36, indicating an extremely positive attitude toward the senator—as befits such a set of beliefs. By comparison, the hypothetical respondent with a total of-7 might be said to have a slightly negative attitude toward Senator Smith. Perhaps it is apparent how this general approach could be used for other attitude objects (with different salient beliefs, of course). In consumer marketing, for example, the attitude object of interest is a product or brand, and the salient beliefs typically concern the attributes of the product or brand. Thus, for instance, the underlying bases of consumers’ attitudes toward a given brand of toothpaste might be investigated by examining the belief strength and belief evaluation associated with consumers’ salient beliefs about that brand’s attributes: whitening power, taste, ability to prevent cavities, cost, ability to freshen breath, and so forth. As another example of application, people’s attitudes toward public policy proposals can be studied; here the salient beliefs might well include beliefs about the consequences of adoption of the policy. Consider, for instance, some possible cognitive bases of attitudes toward capital punishment. Does capital punishment deter crime (belief strength), and how good an outcome is that (belief evaluation)? Is capital punishment inhumane, and 105



how negatively valued is that? Is capital punishment applied inequitably, and how disadvantageous is that? And so forth. Two persons with opposed attitudes on this issue might equally value crime deterrence—that is, have the same evaluation of that attribute—but disagree about whether capital punishment has that attribute. Or two people with opposed attitudes might agree that capital punishment has the characteristic of satisfying the desire for vengeance but differ in the evaluation of that characteristic.



Adapting Persuasive Messages to Recipients Based on the Summative Model A recurring theme in theoretical analyses of persuasion is the idea that to maximize effectiveness, persuasive messages should be adapted (tailored, adjusted) to fit the audience. And one of the most common ways of persuading others is by making arguments that change the recipient’s beliefs in some way: an advertisement leads people to believe that a cleaning product has exceptional stain-removal properties, a speaker at a city council meeting convinces council members that additional dedicated bicycle paths would enhance road safety, and so forth. The summative model of attitude points to a number of alternative strategies for influencing attitude in this way—and provides a systematic way of adapting persuasive messages to audiences by identifying plausible foci for persuasive appeals.



Alternative Persuasive Strategies Because, on this view, one’s attitude is taken to be a function of the strength and evaluation of one’s salient beliefs about the object, attitude change will involve changing these putative bases of attitude. The model thus suggests three broad ways in which attitude might be changed. First, the evaluation of an existing salient belief might be changed. For example, to encourage a more positive attitude, a persuader might try to make some existing positive belief even more positively evaluated (“Senator Smith is, as you know, respected in the Senate, but you may not realize just how desirable that attribute is—it means Senator Smith can be more effective in passing legislation to help our state”) or to make some existing negative belief less negatively evaluated (“Sure, Senator Smith was only an average student—but then again, being an average student isn’t so bad”). Second, the strength (likelihood) of an existing salient belief might be 106



changed. For example, to encourage a more positive attitude, a persuader might try to weaken the strength of an existing negative belief (“It’s not likely that Senator Smith accepted bribes, because Senator Smith is already very wealthy”) or to enhance the strength of an existing positive belief (“You already know it’s true that Senator Smith has worked hard for the people of this state—but you don’t know just how true that is”). Third, the set of salient beliefs might be changed. This can be accomplished in two ways. One is to add a new belief of the appropriate valence (“You might not realize it, but Senator Smith has been quietly working to fix the government’s budget problems”). The other is to change the relative salience of existing beliefs (“Have you forgotten that five years ago Senator Smith helped keep XYZ Industries from moving out of state?”).



Identifying Foci for Appeals The summative attitude model can also be useful in identifying likely foci for persuasive appeals, that is, identifying what kinds of appeals will likely be most appropriate for influencing a given message recipient. This facet of the model is particularly apparent when considering mass persuasion contexts. In planning persuasive messages, a persuader might survey those favoring and opposing the advocated view, to get a clear sense of how the strength and evaluation of their beliefs differ. Such information can be helpful in identifying exactly which beliefs to address in one’s persuasive appeals—and more specifically whether to target belief strength or belief evaluation (or both). Indeed, such data can produce unexpected insights. Consider, for example, the challenge of persuading people to eat a low-fat diet. One apparent advantage of such diets is that they reduce the risk of cardiovascular disease—so a persuader might naturally think of constructing persuasive appeals around that consequence. But this argument might not necessarily be especially effective. A study of UK undergraduates distinguished participants on the basis of whether they intended to eat a low-fat diet over the next month (Armitage & Conner, 1999). Using modally salient beliefs (identified in a preliminary study), the mean belief strength (likelihood) and belief evaluation ratings were examined separately for “intenders” and “non-intenders.” Intenders and non-intenders had equally positive evaluations of reducing the risk of heart disease, which is perhaps not surprising. But they also had equally strong beliefs about the likelihood that eating a low-fat diet would lead to that 107



consequence. That is, non-intenders didn’t need to be convinced about the outcome of heart disease risk reduction; a persuader would only be wasting effort to construct persuasive appeals based on that benefit. But the data did suggest other, more likely targets for persuasive messages. For example, intenders and non-intenders had equally negative assessments of “eating boring food,” but intenders thought that to be much less likely a consequence than did non-intenders; a persuader thus might try to convince non-intenders that in fact eating a low-fat diet doesn’t mean having to eat boring food. Similarly, intenders and non-intenders had equally positive evaluations of “feeling healthier,” but non-intenders were not as convinced (as were intenders) that eating a low-fat diet would make them feel healthier. Notably, for beliefs about whether a low-fat diet “helps to maintain lower weight,” intenders and non-intenders differed with respect to both belief strength (intenders thought that outcome more likely than did non-intenders) and belief evaluation (intenders valued that outcome more than did non-intenders). A persuader who wanted to emphasize this advantage would face the task of convincing non-intenders both that this was a desirable outcome and that eating a low-fat diet would produce the outcome. The general point is this: The summative model of attitude offers a framework within which persuaders can think systematically about which persuasive appeals to make. Instead of selecting arguments haphazardly and hoping to stumble into an effective appeal, persuaders can methodically identify the most likely avenues to persuasive success. (For examples of such analyses, see Brown, Ham, & Hughes, 2010; Cappella, Yzer, & Fishbein, 2003; Fishbein & Yzer, 2003; Jung & Heald, 2009; Middlestadt, 2012; Parvanta et al., 2013; Rhodes, Blanchard, Courneya, & Plotnikoff, 2009; relatedly, see Niederdeppe, Porticella, & Shapiro, 2012.)



Research Evidence and Commentary The commentary that follows initially takes up some questions addressed specifically to the summative model (the general evidence concerning the model; the roles of attribute importance, belief content, and belief strength in the model; and the procedures for scoring the model’s scales), then turns to another belief-based model (offering an alternative image of how beliefs are related to attitudes) and to the general question of the sufficiency of belief-based analyses of attitude; a concluding section reconsiders the persuasive strategies suggested by the summative model. 108



General Correlational Evidence A number of investigations have examined the correlation between a direct measure of the respondent’s attitude toward the object (AO) and the predicted attitude based on the summative formula (Σbiei) using modally salient beliefs. Reasonably strong positive correlations have commonly been found, ranging roughly from .55 to .80 with a variety of attitude objects including public policy proposals (e.g., Peay, 1980; Petkova, Ajzen, & Driver, 1995), political candidates (e.g., M. H. Davis & Runge, 1981; Holbrook & Hulbert, 1975), and consumer products (e.g., Holbrook, 1977; Nakanishi & Bettman, 1974).2 That is, attitude appears to often be reasonably well predicted by this model. This correlational evidence, however, does not offer compelling support for the claim that attitude is determined by salient beliefs. Attitude can sometimes be equally well-predicted (using Σbiei) from nonsalient beliefs as from salient ones (e.g., Ajzen et al., 1995; A. J. Smith & Clark, 1973). Such findings might reflect respondents’ use of their current attitudes as guides to responding to items concerning nonsalient beliefs (e.g., if I have a negative attitude toward the object, and the standardized belief list asks for my reactions to statements associating the object with some attribute that I had not considered, I might well give relatively unfavorable responses precisely because I already have a negative attitude toward the object). Moreover, when standardized belief lists (that is, lists containing modally salient beliefs) and individualized belief lists (in which each respondent gets an individually constructed questionnaire, listing only his or her particular salient beliefs) have been compared as the basis for attitude prediction, often there is no dependable difference (e.g., Agnew, 1998; Bodur, Brinberg, & Coupey, 2000; O’Sullivan, McGee, & Keegan, 2008; see, relatedly, Steadman & Rutter, 2004). So the correlational evidence in hand certainly shows that belief assessments can indicate a person’s attitude (if only because persons give attitude-consistent responses to questionnaire items about beliefs) but falls short of showing that attitudes are determined by salient beliefs. (For a careful discussion of these matters, see Eagly & Chaiken, 1993, pp. 232–234.)



Attribute Importance Several investigations have explored the potential role of attribute importance or relevance in predicting attitude. The summative model, it 109



will be noticed, uses only belief strength and belief evaluation to predict attitude; some researchers have thought that the predictability of attitude might be improved by adding the importance or relevance of the attribute as a third variable. That is, in addition to assessing belief strength and belief evaluation, one would also obtain measures of the relevance or importance of each belief to the respondent; then some three-component formula such as ΣbieiIi (where Ii refers to the importance of the attribute) could be used to predict attitude.3 But the evidence in hand suggests that adding relevance or importance to the summative formula does not improve the predictability of attitude (e.g., L. R. Anderson, 1970; Hackman & Anderson, 1968; Holbrook & Hulbert, 1975; Kenski & Fishbein, 2005). In understanding this result, it may be helpful to consider the possibility that the attributes judged more important or relevant may also have more extreme evaluations; the assessment of belief evaluation (ei) may already involve indirect assessment of relevance and importance (e.g., Holbrook & Hulbert, 1975; cf. van der Pligt & de Vries, 1998a). Moreover, if an investigator is careful to select only salient attributes as the basis for attitude prediction, then presumably all the attributes assessed are comparatively relevant and important ones. So there appears to be little reason to suppose that the predictability of attitude from the original summative formula (Σbiei) can be improved by adding a belief importance or belief relevance component.4 Enhancing the predictability of attitude, however, is arguably not the main relevant research goal. The larger purpose is that of illuminating how beliefs contribute to attitude; given that attitude can be predicted even from nonsalient beliefs (as discussed above), the use of predictability as the relevant criterion is a bit misleading. Indeed, with that larger goal in mind, belief importance ratings can be seen to be valuable in another way (i.e., beyond whether they add to the predictability of attitude from Σbiei). Suppose, for example, that an investigator has not been careful to ensure (by pretesting) that the listed beliefs are modally salient for respondents. Although many of the listed beliefs are not actually ones that determine the respondents’ attitudes, it is still possible that the belief list will produce a reasonably strong correlation between Σbiei and attitude (because, as discussed earlier, respondents can use their current attitudes as guides to responding to these nonsalient belief items). In such a circumstance, importance ratings might indicate which of the listed beliefs are actually 110



salient for the respondents. Indeed, even if the belief list has been pretested to ensure that it contains modally salient beliefs, belief importance ratings may still give some insight into what underlies attitudes, especially if there is some reason to think that different respondents (or subgroups of respondents) may have importantly different sets of salient beliefs. In short, although belief importance might not add to the predictability of attitude from Σbiei, belief importance ratings may be crucial in permitting the identification of those beliefs (on a standardized list) that actually determine the respondent’s attitude—and hence the beliefs that warrant a persuader’s attention. (For illustrations of such a role for importance ratings, see, e.g., Elliott, Jobber, & Sharp, 1995; van der Pligt & de Vries, 1998a, 1998b. For a general discussion, see van der Pligt, de Vries, Manstead, & van Harreveld, 2000.)



Belief Content The summative model offers what might be called a content-free analysis of the underpinnings of attitude. That is, for the summative model’s analysis, the content of a belief is irrelevant; what matters is simply how the belief is evaluated and how strongly it is held. Ignoring content may indeed be appropriate given an interest simply in attitude prediction. But for other purposes, systematic attention to belief content may be important. Functional approaches to attitude (discussed in Chapter 3) provide a useful contrast here. Functional approaches identify different syndromes (coherent sets) of beliefs based on belief content; different attitude functions correspond to substantively different (sets of) salient beliefs. So, for example, two automobile owners may have equivalently positive attitudes about a car but very different kinds of underlying salient beliefs: One has beliefs about gas mileage and frequency-of-repair records, whereas the other has beliefs about what identity is conveyed by the automobile. Such differences in belief content can figure importantly in persuasion. Consider, for example, the persuasive strategy of creating a more positive attitude by inducing belief in some new positive attribute. If content is ignored, one positive attribute might seem as good as another for this purpose. But a functional perspective recommends considering the substantive content of the belief to be instilled; after all, if the receiver has predominantly image-oriented beliefs about the object, then trying to add some product quality–oriented belief (“gets good gas mileage”) may not be 111



successful. The point here is not that such considerations cannot be represented within a summative model framework (see, e.g., Belch & Belch, 1987). For example, one might say that the “good gas mileage” attribute is more likely to be salient for (or valued by) one person than another or that that attribute is more likely to be perceived as associated with the attitude object by one person than by another. Rather, the point is that the summative model provides no systematic ways of thinking about belief content, although such content is manifestly important. In a sense, then, one might think of these approaches as complementary: Functional approaches emphasize the (manifest or latent) content of beliefs, whereas belief-based attitude models (such as the summative model) are aimed at illuminating how underlying beliefs contribute to an overall attitude.



Role of Belief Strength There is good reason to think that the apparent contribution of belief strength scores to the prediction of attitude does not reflect a genuine role for belief strength in determining attitude but instead is a methodological artifact (something artificially arising from the research methods employed rather than from any genuine phenomenon). The relevant evidence comes from research comparing Σei (that is, the simple sum of the belief evaluations) with Σbiei (the summative formula) as predictors of attitude. The relative success of these two formulas varies, depending on the way in which the list of salient beliefs is prepared. The most common way of preparing the list of salient beliefs is by eliciting beliefs from a test sample, identifying the most frequently mentioned beliefs, and using these on the questionnaire. In this procedure, a standardized belief list is composed (i.e., every respondent receives the same set of modally salient beliefs). An alternative procedure is to elicit salient beliefs from each respondent individually and so have each respondent provide belief strength and belief evaluation ratings for his or her unique set of salient beliefs. That is, an individualized belief list can be constructed for each respondent. The research evidence indicates that when individualized belief lists are used, Σei and Σbiei are equally good predictors of attitude; adding belief strength scores to the formula does not improve the predictability of 112



attitude. With standardized belief lists, however, Σbiei is a better predictor than is Σei. That is, belief strength scores significantly improve the predictability of attitude only when standardized (as opposed to individualized) belief lists are used (Cronen & Conville, 1975; Delia, Crockett, Press, & O’Keefe, 1975; Eagly, Mladinic, & Otto, 1994).5 On reflection, of course, this result makes good sense. With individualized belief lists, the respondent has just indicated that he or she thinks the object possesses the attribute; only beliefs that the respondent already holds are rated for belief strength. By contrast, with standardized belief lists, belief strength scores distinguish those beliefs the respondent holds from those the respondent does not hold. The use of standardized lists thus creates a predictive role for belief strength scores (namely, the role of differentiating those beliefs the respondent holds from those the respondent does not hold), but the predictive contribution of belief strength scores is a methodological artifact, not an indication of any genuine place for belief strength in the cognitive states underlying attitude. To put the point somewhat differently: These results suggest that—insofar as the underlying bases of attitude are concerned—we may more usefully think of persons’ beliefs about an object as rather more categorical (“I think the object has the attribute,” “I don’t think the object has the attribute,” or “I’m not sure”) than continuous (“I think that the probability that the object possesses the attribute is thus-and-so”). The belief strength scales give the appearance of some continuous gradation of belief probability, but these scales make a contribution to attitude prediction only because standardized belief lists are used. When individualized belief lists are used, belief strength scores are unhelpful in predicting attitude because in each case the individual thinks that the object has the attribute—and it is that simple categorical judgment (not variations in the reported degree of probabilistic association) that is important in determining the individual’s attitude. (For some evidence suggesting a more categorical than purely continuous image of belief strength, see Weinstein, 2000, esp. pp. 72–73.)6



Scoring Procedures There has been a fair amount of discussion in the literature concerning how the belief strength and belief evaluation scales should be scored (e.g., Ajzen & Fishbein, 2008; Bagozzi, 1984; Fishbein & Ajzen, 2010, pp. 105– 110; Lauver & Knapp, 1993; J. L. Smith, 1996). By way of illustration, two common ways of scoring a 7-point scale are from −3 to +3 (bipolar 113



scoring) and from 1 to 7 (unipolar scoring). (There are possibilities in addition to −3 to +3 and 1 to 7, but these two provide a useful basis for discussion.) With belief strength and belief evaluation scales, one might score both scales −3 to +3, score both scales 1 to 7, or score one scale −3 to +3 and the other 1 to 7. But (because the scales are multiplied) these different scoring procedures can yield different correlations of Σbiei with attitude, and hence a question has arisen concerning which scoring procedures are preferable. Sometimes conceptual considerations have been adduced as a basis for choosing a scoring method. These arguments commonly take one of two forms. One is an appeal to the nature of the relevant psychological states. For example, it is sometimes suggested that evaluation is naturally better understood as bipolar rather than as unipolar or that belief strength scales should not be scored in a bipolar way because it is not psychologically meaningful for attitude objects to be negatively associated with attributes (for an example of such arguments, see Bagozzi, 1984). The other considers the plausibility of the consequences of employing various combinations of scoring procedures. For example, to take the simplifying case of a person with just one salient belief, a respondent who strongly believes that the object possesses a very undesirable characteristic should presumably have the least favorable attitude possible—but if both scales are scored from 1 to 7, that respondent will not have the lowest possible Strength × Evaluation product (which thus suggests the implausibility of such scoring). In particular, the combination of bipolar scoring for belief evaluation and unipolar scoring for belief strength has often been argued to be the theoretically most appropriate scoring combination (e.g., Steinfatt, 1977). But the main criterion for assessing scoring procedures has been the predictability of attitude thereby afforded. That is, the criterion has been the observed correlation between Σbiei and attitude.7 Several studies have compared the predictability of attitude using different scoring methods. Although results vary, the most common finding seems to have been that scoring both scales in a bipolar fashion yields larger correlations (of Σbiei, with attitude) than do alternative combinations and, in particular, is superior to the intuitively appealing bipolar evaluation and unipolar strength combination (e.g., Ajzen, 1991; Gagné & Godin, 2000; Sparks, Hedderley, & Shepherd, 1991; for discussion, see Fishbein & Ajzen, 2010, pp. 108–109).8 114



But now the task becomes explaining why bipolar scoring for both scales appears to maximize the correlation between Σbiei and attitude. Bipolar scoring seems to make intuitive psychological sense in the case of belief evaluation scales, but the general empirical success of bipolar scoring for belief strength scales may appear puzzling. One possibility is simply this: When standardized lists of modal salient beliefs are used, bipolar scoring of belief strength scales may permit participants to remove all effects of beliefs that they do not have (or beliefs that are not salient for them). When such beliefs appear on the standardized belief list, a mark at the midpoint of belief strength scales is a sensible response (the respondent does not know, or is not sure, whether the object has the attribute, so marks the midpoint rather than favoring either “likely” or “unlikely”). With bipolar scoring, such a response is scored as zero—which, when multiplied by the corresponding belief evaluation, will yield a product of zero (no matter what the evaluation is); this has the entirely appropriate effect of removing that belief from having any impact on the respondent’s predicted attitude.9 In short, the common superiority of bipolar (over unipolar) scoring of belief strength scales might be a consequence of the use of standardized lists of beliefs and so may be a methodological artifact rather than a source of substantive information about how belief strength perceptions operate.



Alternative Integration Schemes The summative model depicts beliefs as combining in a particular way to yield an overall evaluation, namely, in an additive way (summing across the biei products). Hence, for instance, everything else being equal, adding a new positive belief will make an attitude more favorable. But different integration schemes—that is, different images of how beliefs combine to yield attitudes—have been proposed. The most prominent of these is an averaging model, as embedded in Anderson’s information integration theory (N. H. Anderson, 1971, 1981b, 1991).10 Crudely expressed, an averaging model suggests that attitude is determined by the average, not the sum, of the relevant belief properties. An averaging model can produce some counterintuitive predictions. For example, it suggests that adding a new positive belief about an object will not necessarily make the attitude more positive. Suppose that a person’s current attitude toward the object is based on four beliefs evaluated +3, +3, +3, and +2. Imagine that the person acquires a new belief that is evaluated 115



+2 (and, to simplify matters, assume equal belief strength weights for each belief). A summative picture of belief combination expects the additional belief to make the attitude more positive (because the sum of the evaluations would be 13 rather than 11), but an averaging model predicts that the overall attitude would be less positive: The average of the initial four beliefs is 2.75, but the average of the set of five beliefs is 2.60 (that is, adding the new attribute lowers the average evaluation). For a time, a fair amount of research attention was devoted to comparing summative and averaging (and other) models of belief integration. But for various reasons, no general conclusion issues from this research. For one thing, in many circumstances, the models make equivalent predictions (and so cannot be empirically distinguished); moreover, in circumstances in which the models do make divergent predictions, each can point to some evidence suggesting its superiority over the other (e.g., N. H. Anderson, 1965; Chung, Fink, Waks, Meffert, & Xie, 2012; Fishbein & Hunter, 1964). Thus neither seems to provide an entirely satisfactory general account of how beliefs are related to attitudes; indeed, there may be no single simple rule by which persons combine beliefs into an attitude. (For some evidence and discussion, see Betsch, Kaufmann, Lindow, Plessner, & Hoffmann, 2006; Eagly & Chaiken, 1993, pp. 241–255; Harris & Hahn, 2009; R. Wyer, 1974, pp. 263–306.)11 An inability to display any decisive general superiority of one model over the other is in some ways unfortunate, as summative and averaging models can yield different recommendations to persuaders. Suppose, for example, that voters have a generally favorable attitude toward some policy issue (e.g., gun control) that appears as a referendum ballot item. The organizers of the campaign favoring that policy discover some new advantage to the proposed policy. Naturally enough, they undertake an advertising campaign to publicize this new positive attribute of the policy, hoping to make voters’ attitudes even more favorable toward their position. The initiation of this new campaign rests implicitly on a summative image of how this new information will be integrated: Adding a new positive belief about an object should make attitudes toward that object more favorable. But an averaging model will predict that, at least under some circumstances, the addition of this new positive attribute could make attitudes toward the policy less favorable than they had been—and hence might conclude that this new advertising campaign is ill-advised. In the absence of good evidence about just what sort of belief combination 116



principle might best describe what will occur in a circumstance such as this, one can hardly give persuaders firm recommendations.



The Sufficiency of Belief-Based Analyses Belief-based attitude models depict beliefs about the object as the sole determinants of attitudes. But the question has arisen whether such beliefs are a sufficient basis for understanding attitudes; the issue is whether some non–belief-based (noncognitive) elements might independently contribute to attitude (independent, that is, of representations of belief structure such as Σbiei). The central research evidence here takes the form of studies investigating whether a given noncognitive element makes a contribution to the prediction of attitude beyond that afforded by measures of belief structure (Σbiei). A convenient illustration is provided by research concerning the effects of consumer advertising. Advertising presumably attempts to influence the consumer’s beliefs about the product’s attributes or characteristics, thereby influencing the consumer’s attitude toward the product. But evidence suggests that at least under some circumstances, the influence of advertising on receivers’ attitudes toward a given brand or product may come about not only through receivers’ beliefs about the product’s characteristics but also through the receivers’ evaluation of the advertisement itself (the receivers’ “attitude toward the ad”). As receivers have more favorable evaluations of the advertising, they come to have more favorable attitudes toward the product being advertised. And several studies have reported that this effect occurs over and above the advertising’s effects on product beliefs—that is, attitude toward the ad and Σbiei jointly have been found to be more successful in predicting attitude than is Σbiei alone (e.g., Mitchell, 1986; for related findings, see MacKenzie, Lutz, & Belch, 1986; for a review, see S. P. Brown & Stayman, 1992). Such evidence appears to point to some influence on attitudes beyond beliefs about the object and hence suggests the insufficiency of a purely belief-based analysis of the determinants of attitude. The research evidence bearing on these matters is not uncontroversial. Fishbein and Middlestadt (1995) argued that most of the research purporting to show an independent effect for noncognitive elements (including attitude toward the ad) is methodologically flawed (for 117



discussion, see, e.g., Fishbein & Middlestadt, 1997; Herr, 1995; Miniard & Barone, 1997; Priester & Fleming, 1997). One illustration of such flaws is that if an investigator is not careful to ensure that salient beliefs are being assessed, then the apparent ability of some noncognitive factor to add to the predictability of attitude beyond Σbiei might reflect not some genuine influence of the noncognitive factor but rather a shortcoming in the assessment of beliefs; the suggestion is that with better belief assessment, the apparent noncognitive contribution might disappear.12 There does however seem to be good evidence pointing to an independent role for some noncognitive elements, namely, feelings (emotions, affect). The suggestion is that attitude might be influenced either by cognitive (belief-related) considerations or by affective (feeling-related) considerations. So, for example, a person’s evaluation of a politician might reflect cognitions concerning the politician’s personal attributes and issue positions or might be based on the feelings that the politician evokes in the person (hope, anger, disgust, pride, etc.). Consistent with this suggestion, several studies have reported that attitudes are often better predicted from a combination of affective and cognitive assessments than from either one alone (e.g., Abelson, Kinder, Peters, & Fiske, 1982; Agarwal & Malhotra, 2005; C. T. Allen, Machleit, Kleine, & Notani, 2005; Bodur et al., 2000; Crites, Fabrigar, & Petty, 1994; Eagly et al., 1994; Haddock & Zanna, 1998). Such evidence invites a picture of attitudes as potentially having both affective and cognitive determinants (as offered by, e.g., Eagly & Chaiken, 1993, pp. 14–16; Zanna & Rempel, 1988) and suggests the incompleteness of a purely belief-based analysis of attitude.13 Of course, if belief is understood sufficiently broadly, none of this is necessarily inconsistent with a belief-based model of attitude. The distinction between affect and cognition might sensibly be said to be one of emphasis: No cognition is free from affect (every belief has some evaluative aspect, even if neutral), and even self-reported feelings amount to reports about what people believe (what people believe their feelings are). Indeed, these kinds of considerations have led some commentators to suggest that instead of drawing a contrast between pure affect and pure cognition, it might be more useful to distinguish affective beliefs and cognitive beliefs (see, e.g., Trafimow & Sheeran, 1998) or instrumental beliefs and experiential beliefs (Fishbein & Ajzen, 2010, pp. 82–85). Even approached in such a fashion, however, the evidence in hand suggests the importance of being alert to the different types of beliefs that 118



might underlie attitudes.14 Some attitudes might be primarily based on affective or experiential considerations, others predominantly on cognitive or instrumental considerations, and still others on a mixture of these elements.15 And, of course, understanding the current basis of a person’s attitude is commonly a first step toward understanding how the attitude might be changed.



Persuasive Strategies Reconsidered The various research findings discussed above invite some reconsideration of the persuasive strategies suggested by the summative model. Those strategies involve changing the strength of a current belief, changing the evaluation of a current belief, or changing the set of salient beliefs (by adding new beliefs or by altering the relative salience of current ones). Scant evidence directly compares the relative effectiveness of these different strategies, but other research provides some insight into the likely utility of the various alternatives.16



Belief Strength as a Persuasion Target The apparently artifactual role of belief strength scores suggests the implausibility of certain persuasive strategies that the summative model might recommend. Consider a persuader who is trying to induce a favorable attitude toward Boffo Beer. Suppose that a particular respondent has the salient belief that Boffo Beer tastes good and on 7-point scales (scored −3 to +3) indicates that this attribute is highly desirable (+ 3 for belief evaluation) and that it is moderately likely (+ 2 for belief strength). The summative attitude model suggests that this respondent’s attitude could be made more positive by influencing the belief strength rating for this attribute—specifically, by getting the respondent to believe that it is very likely that Boffo Beer tastes good (+3 for belief strength). But if belief strength does not actually influence attitude, then such a strategy is misguided; if a person already has the relevant categorical judgment in place, trying to influence the degree of association between the object and the attribute will not influence attitude. Thus if our hypothetical respondent already believes that Boffo Beer tastes good, there appears to be little point in seeking changes in the exact degree of the respondent’s subjective probability judgment that Boffo Beer tastes good.



119



Of course, if our respondent thinks Boffo Beer does not taste good (or has no opinion about Boffo’s taste), then in seeking to induce a positive attitude toward the beer, a persuader may well want to induce the belief that Boffo does taste good. But this will be a matter of changing the relevant categorical judgment (e.g., from “Boffo Beer doesn’t taste good” to “Boffo Beer does taste good”) and need not be approached as though there is some psychologically real probabilistic degree of perceived association between object and attribute. That is, the key distinction will be between whether the person does or does not have the belief, not between finer gradations of belief strength. Indeed, a surprisingly large number of studies have found that messages varying in the depicted likelihood of consequences did not differentially influence persuasive outcomes (whereas messages varying in the desirability of depicted outcomes did correspondingly differ in persuasiveness). For example, in a series of studies, Smith-McLallen (2005) manipulated both likelihood information and desirability information, finding that attitudes were more influenced by variations in the desirability of the claimed consequences than by variations in the likelihood of those consequences’ occurrence (see also Johnson, SmithMcLallen, Killeya, & Levin, 2004; Levin, Nichols, & Johnson, 2000). Similar results have been reported in investigations of colorectal cancer screening (Lipkus, Green, & Marcus, 2003), consumer product safety perceptions (Wogalter & Barlow, 1990, Experiment 1), and energy consumption reduction (Hass, Bagley, & Rogers, 1975): Messages varying in the desirability of the depicted consequences varied correspondingly in persuasiveness, but messages varying in the depicted likelihood of the consequences did not (see, relatedly, Jerit, 2009; Mevissen, Meertens, Ruiter, & Schaalma, 2010).17 These findings thus are consistent with the idea that variation in belief strength may not be consequential—and thus may not be a useful target for persuasive efforts. (For a review, see O’Keefe, 2013a.)



Belief Evaluation as a Persuasion Target Little research directly addresses the persuasive strategy of changing the evaluation of a currently held belief. Lutz (1975) found that messages designed specifically to change laundry detergent attribute evaluations had little effect on those evaluations (and, correspondingly, little effect on attitudes). It may be that, as Eagly and Chaiken (1993, p. 237) suggest, some attribute evaluations are relatively stable by virtue of their basis in 120



prior experience. For example, the evaluation of the laundry detergent attribute “gets your clothes clean” might well be expected to be relatively stable for most people. Still, sometimes attribute evaluations appear to be a useful focus for persuaders. For example, a consumer may need to be convinced that a faster Internet connection would be desirable. Death penalty opponents might seek to convince the public that vengeance is a base and unworthy motive—and hence that capital punishment’s attribute of providing vengeance is less desirable than one might previously have thought. Although it may not be easy to influence the evaluations associated with existing beliefs, changing such evaluations may nevertheless sometimes be a key aspect of a persuader’s campaign.18



Changing the Set of Salient Beliefs as a Persuasion Mechanism Changing the set of salient beliefs might be accomplished in two (not mutually exclusive) ways, by adding new salient beliefs or by altering the relative salience of existing beliefs. Adding a new appropriately valenced belief (e.g., adding a belief that associates some new positive attribute with the object when the goal is to make the attitude more favorable) would appear to be relatively attractive as an avenue to attitude change.19 But the previously discussed research evidence suggests that two considerations should be kept in mind here. First, the content (not just the evaluation) of the advocated new belief may need to be considered closely, as some candidate new beliefs may be more compatible than others with the current set of beliefs. For instance, a receiver with predominantly concrete, instrumental beliefs about a given automobile (“It gets bad gas mileage” or “It gets good gas mileage”) may not be receptive to advertising invoking image-oriented appeals (“You’ll feel so sexy driving it”); for such a person, adding new instrumentally oriented beliefs (“It’s a very safe vehicle”) might be a more plausible approach.20 Second, nonsummative models predict that adding a new belief may fail to move the attitude in the desired direction. If beliefs combine in a way that involves averaging the evaluations of the individual beliefs (rather than summing them), then it is possible that (for example) adding a new 121



positive belief may not make the attitude more positive.21 The other broad way of changing the set of salient beliefs is to alter the relative salience of currently held beliefs. For example, a persuader might seek to make the audience’s beliefs about positive attributes more salient, thereby enhancing the attitude. There is little direct evidence about the effectiveness of implementing this strategy in persuasive messages (see Batra & Homer, 2004; Delia et al., 1975; Shavitt & Fazio, 1990). Nevertheless, it is easy to see that (for example) one purpose of point-of purchase displays (e.g., in grocery stores) can be to influence which of the product’s attributes are salient. In employing such a strategy, it is important to identify just which beliefs are already actually salient, and (as intimated earlier) belief importance ratings can be especially valuable for this purpose when standardized lists (of modally salient beliefs) are used. For example, as van der Pligt et al. (2000) have pointed out, smokers do not necessarily evaluate the undesirable health consequences of smoking any less negatively than do nonsmokers (nor do they necessarily give different belief strength ratings), but the health consequences are less important (less salient)—and other consequences more important—for smokers than for nonsmokers. Thus attempting to shift the relative salience of these beliefs, making the negative consequences more prominent and the positive consequences less salient, may be a more productive avenue for persuasion than attempting to influence belief evaluation or belief strength. An indication of the potential effects of belief salience manipulations can be seen in studies of “issue framing,” in which messages (such as news reports) concerning public policy issues vary in the interpretive framework through which the issue is reported. In one classic study, participants read one of two editorials concerning whether a hate group should be allowed to stage a public rally. Those who read an editorial emphasizing free speech were more inclined to permit the rally than those who read an editorial about the public safety risks of such a rally (Nelson, Clawson, & Oxley, 1997). One plausible mechanism for such effects is that the varying frames of the communication produce corresponding variation in the salience of frame-related beliefs. (For other examples and reviews, see Bolsen, Druckman, & Cook, 2014; Chong & Druckman, 2007; Druckman, 2011; Nelson, Oxley, & Clawson, 1997; Tewksbury & Scheufele, 2009.)



122



Conclusion The general idea that the beliefs one has about an object influences one’s attitude toward that object is enormously plausible, and, correspondingly, it seems obvious that one natural avenue to attitude change involves influencing beliefs. Hence it is not surprising that belief-based models of attitude have received such attention from students of persuasion. Indeed, the summative model of attitude obviously offers some straightforward recommendations to persuaders.22 Still, many particulars of the relationship of beliefs and attitudes remain elusive, with corresponding uncertainties for the understanding of persuasion.



For Review 1. Explain the general idea of belief-based approaches to attitude. What is a salient belief? How can one identify a person’s salient beliefs about a given object? Explain how, in a survey context, one might identify the modal (average) salient beliefs about an object. 2. According to the summative model of attitude, what are the two determinants of attitude? What is belief strength? Describe questionnaire items that might be used to assess belief strength. What is belief evaluation? Describe questionnaire items that might be used to assess belief evaluation. Explain the summative model’s description of how belief strength and belief evaluation combine to produce attitude; that is, describe and explain the summative model’s formula. Give an example that illustrates the model’s application. 3. Sketch three alternative strategies for attitude change suggested by the summative model. Explain (and give examples of) the strategy of changing the evaluation of an existing salient belief, the strategy of changing the strength of an existing salient belief, and the strategy of changing the set of salient beliefs. Describe two ways of changing the set of salient beliefs. Explain how the summative model can be useful in identifying possible foci for persuasive appeals. 4. What is the general pattern of correlations between the summative model’s predictions and direct measures of attitude? Explain why such correlational evidence does not necessarily show that attitude is determined by salient beliefs. What is attribute importance? Does adding attribute importance to the summative model’s formula improve the predictability of attitude? Explain how belief importance ratings can be useful even if they do not improve the predictability of 123



5.



6.



7.



8.



9.



10.



attitude. Is the summative model concerned with the content (as opposed to the evaluation) of beliefs? Explain the complementary relationship of functional approaches and belief-based models of attitude. Describe the difference between standardized and individualized belief lists. Do belief strength scores improve the predictability of attitude when individualized belief lists are used? Do belief strength scores improve the predictability of attitude when standardized belief lists are used? Explain. Explain why it matters whether belief strength and belief evaluation scales are scored in different ways. What is unipolar scoring? What is bipolar scoring? Which kind of scoring maximizes the correlation between Σbiei and attitude? Describe an averaging model of how beliefs combine to yield attitude. What does the research evidence indicate about whether an averaging model or a summative (adding) model is superior? Explain how adding models and summative models can have different implications for persuasive strategy. Explain the idea that non–belief-based (noncognitive) elements might independently contribute to attitude. What sort of evidence bears on such claims? Identify a noncognitive element that improves the predictability of attitude. Explain how such elements might be redescribed in belief-based terms. Explain how the artifactual role of belief strength scores (in predicting attitude) has implications for the persuasive strategy of changing belief strength. Do messages that vary in the depicted likelihood of consequences also vary correspondingly in persuasiveness? Describe the challenges of trying to influence attitude by changing belief evaluations. Explain why the strategy of adding new beliefs (as a way of changing attitudes) might require attending to belief content (not just evaluation). Why might adding a new positive belief not make attitudes more positive? Describe how belief importance ratings might be useful when trying to change attitudes by influencing the relative salience of beliefs.



Notes 1. The summative model of attitude is sometimes referred to as an expectancy-value (EV) model of attitude. An EV model of attitude represents attitude as a function of the products of the value of a given 124



attribute (e.g., the attribute’s desirability) and the expectation that the object has the attribute (e.g., belief strength). The summative model is only one version of an EV model, however; this basic EV idea has been formulated in various ways (e.g., Rosenberg, 1956). But the summative model is the best studied, appears to have been the most successful empirically, and indeed is the standard against which alternative EV models have commonly been tested. (For some general discussions of EV models of attitude, see Bagozzi, 1984, 1985; Eagly & Chaiken, 1993; Kruglanski & Stroebe, 2005.) 2. This attitude model is embedded in reasoned action theory (RAT; see Chapter 6). Research concerning RAT has produced evidence concerning the use of the summative formula to predict specifically attitudes toward behaviors (AB), with similar results (for some review discussions, see Albarracín, Johnson, Fishbein, & Muellerleile, 2001; Armitage & Conner, 2001; Conner & Sparks, 1996). 3. Although not discussed here, the potential complexities in assessing belief importance should not be underestimated; see Jaccard, Radecki, Wilson, and Dittus (1995), van der Pligt et al. (2000, pp. 145–155), and van Ittersum, Pennings, Wansink, and van Trijp (2007). 4. Relatedly, equally good correlations of Σbiei with attitude have been observed using a list of modally salient beliefs and using a smaller set of beliefs identified by the respondent as most important (Budd, 1986; Steadman & Rutter, 2004; van der Pligt & de Vries, 1998a; van Harreveld, van der Pligt, & de Vries, 1999; cf. Elliott et al., 1995). As with the previously mentioned findings concerning the predictability of attitude from nonsalient beliefs, these results might reflect the use of one’s current attitude as a guide to responding to whatever belief items are presented. 5. Esses, Haddock, and Zanna (1993) reported a related but slightly different finding. With individualized belief lists, attitudes toward various social groups were equally well-predicted by the average of the attribute evaluations (that is, Σei/n, where n is the number of beliefs) and by the average of a multiplicative composite in which each attribute evaluation was multiplied by the individual’s judgment of the percentage of the group to which each attribute applies (that is, ΣbiPi /n where P is the relevant percentage). 6. Convincing evidence of a continuous-probability role for belief strength 125



in contributing to attitude might be had by research that used individualized belief lists but replaced the conventional belief strength end-anchors (“likely–unlikely”) with ones more appropriate to the desired judgment, such as “slightly likely” and “very likely.” If, with individualized belief lists and such end-anchors, Σbiei were found to generally be superior to Σei as a predictor of attitude, the case for conceiving of belief strength in continuous probability terms would be strengthened. 7. This is a curious criterion, because the goal of maximizing the predictability of attitude from Σbiei (that is, maximizing the correlation between attitude and Σbiei) is at best an interim research goal. The goal is understandable, because predictive accuracy provides evidence bearing on the adequacy of the summative model. But (as intimated earlier) such predictability does not necessarily show that the model’s implicit depiction of the underlying psychological processes is correct (recall, for example, that attitude can be predicted even from lists of nonsalient beliefs). Some measure of predictability is surely a necessary condition for the model’s adequacy, but the more important question concerns the substantive adequacy of the model, that is, whether the model provides an accurate account of the relationship of beliefs and attitudes. 8. This conclusion is also recommended by several studies using optimal scaling procedures, which assign scale values in such a fashion as to maximize the resulting correlation. With these procedures, a constant is added to each belief strength score and another is added to each belief evaluation score; the constants are computed precisely to produce the largest possible correlation between Σbiei and attitude. Several studies have reported that the computed scaling constants suggest that both scales should be bipolar (Ajzen, 1991; Holbrook, 1977; cf. Doll, Ajzen, & Madden, 1991). The Σbiei-attitude correlations that result from optimal scaling are not themselves good evidence for the summative model (after all, the data have been manipulated to maximize the correlations) and are not meant to be used that way. Instead, the result of interest is the particular scaling constants recommended—not necessarily the specific numerical values (as these might bounce around from study to study) but rather what general sort of scale the constants recommend. For example, a researcher might start with a belief strength scale scored in a unipolar fashion, but the optimal scaling constants might be such as to transform it into a bipolar scale. That is, optimal scaling results can give evidence 126



about whether unipolar or bipolar scoring will maximize the correlation. 9. Notice the contrast: With bipolar belief strength scoring, it does not matter what the respondent’s evaluation is of an attribute for which the respondent has marked the midpoint of the belief strength scales (because the Strength × Evaluation product for that attribute will be zero). But with unipolar belief strength scoring, the Strength × Evaluation product will vary depending on the respondent’s evaluation of that attribute. Hence even if a respondent is completely uncertain about whether the object has the attribute (and so marks the midpoint of the belief strength scales), the respondent would nevertheless be predicted to have a relatively more favorable attitude if the attribute were evaluated positively than if the attribute were evaluated negatively. Scott Moore helped me see this point clearly. 10. Anderson’s information integration theory is much broader than a simple averaging model (see, e.g., N. H. Anderson, 1981a, 1991). The general notion is that there are many information integration principles that persons employ, one of which is a weighted-averaging principle (for useful general discussions, see Eagly & Chaiken, 1984, pp. 321–331; 1993, pp. 241–253). My purpose here is simply to introduce the idea that nonsummative images of belief combination are possible; an uncomplicated averaging model provides a convenient example. 11. As an indication of the hidden complexities here, consider that the summative model suggests that if a new belief is added to the set of salient beliefs, then (assuming constant belief strength) the increase in overall attitude resulting from the addition of a positively evaluated belief will be the same size as the decrease in overall attitude resulting from the addition of a negatively evaluated belief of equivalent extremity. But there is good evidence that new pieces of positive and negative information do not always have equal-sized effects on attitudes. In fact, negative information often has a disproportionate impact on evaluations or decisions compared to otherwise equivalent positive information (e.g., Hamilton & Zanna, 1972; Lutz, 1975; for reviews, see Cacioppo, Gardner, & Berntson, 1997; Rozin & Royzman, 2001; Skowronski & Carlston, 1989). 12. There are actually some rather difficult methodological challenges here. For instance, the very assessment of beliefs may create apparent consistency between attitudes and belief-based elements (as when existing attitudes guide one’s responses to nonsalient belief items). Such 127



consistency may suggest the operation of a belief-based attitude process even where none exists. For example, an attitude might be formed in a wholly non–belief-based way, then used to guide responses to belief items in such a way that the attitude appears to be largely determined by those belief elements (for discussion of such problems, see Fishbein & Middlestadt, 1997, pp. 112–113; Herr, 1995). 13. The idea that attitudes might have multiple underlying components— including both affective and cognitive ones—has a long history in the study of attitudes (e.g., Rosenberg & Hovland, 1960). But (as pointed out by Eagly et al., 1994) it took quite some time for research to explicitly take up the question of whether the predictability of attitude can be enhanced by including non–belief-based considerations. And although multicomponent views of attitude commonly treat affect and cognition as representing just two of three attitudinal bases (the third being conation or behavioral elements, as when one’s past behavior influences one’s attitudes through self-perception processes; see, e.g., Bem, 1972), research has come to focus on only the affective and cognitive elements (see, e.g., Haddock & Zanna, 1998, p. 328n4). 14. There is good reason to think that the wording of belief elicitation questionnaires may influence the types of beliefs that people report. For example, some common procedures may generally elicit predominantly instrumental-utilitarian beliefs rather than symbolic beliefs (see Ennis & Zanna, 1993, 2000; Sutton et al., 2003). This suggests the importance of careful questionnaire design that minimizes the chances of missing some important class of underlying beliefs. For example, it is possible to ask different questions to elicit affective considerations and cognitive ones (e.g., French et al., 2005; Haddock & Zanna, 1998). 15. A contrast between experiential/affective beliefs and instrumental/cognitive beliefs can be thought of as another way of analyzing belief content. That is, this distinction points to a substantive variation in the beliefs underlying attitudes, and in that sense it can be seen to be similar to elements of functional analyses of attitude (discussed in Chapter 3). However, functional analyses offer the idea that beliefs characteristically coalesce in substantively different motivationally coherent packages or syndromes; by contrast, a general distinction between affective and cognitive beliefs need not imply that an individual’s beliefs commonly cluster together on the basis of being affective or cognitive (but see Huskinson & Haddock, 2004; Trafimow & Sheeran, 128



1998). 16. Some general evidence indicates covariation between attitude change and change in the underlying bases of attitude. That is, changes in belief strength and evaluation (or Σbiei) have been found to be accompanied by corresponding changes in attitude (e.g., DiVesta & Merwin, 1960; Lutz, 1975; Peay, 1980). Such evidence is consistent with the model’s suggestions that attitude change can be influenced by changes in belief strength and belief evaluation; however, other uncertainties (e.g., the apparent artifactual contribution of belief strength scores) make such evidence less helpful than it might be. 17. Notably, in Hass et al.’s (1975) study, across message conditions, an energy crisis was perceived to be moderately likely; the perceived likelihood of an energy crisis (on a 10-point scale) was 5.9 in the lowlikelihood message condition and 6.9 in the high-likelihood message condition. This provides an illustration of the idea that when the relevant categorical judgment is in place, smaller belief strength variations may not matter much. 18. One way of enhancing the desirability of an attribute may be to emphasize its scarcity, because “opportunities seem more valuable to us when they are less available” (Cialdini, 2009, p. 200). For examples of scarcity-based phenomena, see Aggarwal, Jun, and Huh (2011), Brannon and Brock (2001), and Eisend (2008). For reviews and general discussion, see Brock (1968) and Lynn (1991). For some complexities, see Cialdini, Griskevicius, Sundie, and Kenrick (2007), Gierl and Huettl (2010), and Jung and Kellaris (2004). 19. Adding a new belief might be thought of as influencing belief strength but only in the sense that it involves changing some categorical judgment (e.g., from “I don’t know whether the object has attribute X” to “I think the object has attribute X”). That is, it need not be understood as necessarily involving some gradation of subjective probability. 20. Given the previously discussed distinction between affectively oriented and cognitively oriented beliefs, one might naturally hypothesize that attitudes would most effectively be changed by appeals that invoke the same sorts of considerations as underlie the attitude—that affectively oriented appeals would be more persuasive than cognitively oriented appeals for affectively based attitudes, for instance. A systematic review of 129



the relevant research is not in hand, but the evidence at least suggests a rather more complicated picture (see, e.g., Clarkson, Tormala, & Rucker, 2011; Conner, Rhodes, Morris, McEachan, & Lawton, 2011; Edwards, 1990; Fabrigar & Petty, 1999; Haddock, Mio, Arnold, & Huskinson, 2008; Ruiz & Sicilia, 2004). 21. As a complexity: Adding a new salient belief may cause some existing belief to become less salient. The number of beliefs that can be salient is surely limited (given that human information-processing capacity is not unbounded). If the current set of beliefs has exhausted that capacity, then the addition of some new salient belief will necessarily mean that some old belief has to drop from the set of salient beliefs. Presumably in such a circumstance, a comparison of the evaluations of the two beliefs in question (the new salient one and the previously salient one) will, ceteris paribus, indicate the consequences for attitude change. 22. The model, however, implicitly emphasizes message content as central to persuasive effects and does not directly speak to the roles played by such factors as communicator credibility, message organization, receiver personality traits, and so forth. From the model’s point of view, all such factors only indirectly influence message-induced attitude change— indirectly in the sense that their influence is felt only through whatever effects they might have on belief strength, evaluation, and salience.



130



Chapter 5 Cognitive Dissonance Theory General Theoretical Sketch Elements and Relations Dissonance Factors Influencing the Magnitude of Dissonance Means of Reducing Dissonance Some Research Applications Decision Making Selective Exposure to Information Induced Compliance Hypocrisy Induction Revisions of, and Alternatives to, Dissonance Theory Conclusion For Review Notes



A number of attitude theories have been based on the idea of cognitive consistency—the idea that persons seek to maximize the internal psychological consistency of their cognitions (beliefs, attitudes, etc.). Cognitive inconsistency is taken to be an uncomfortable state, and hence persons are seen as striving to avoid it (or, failing that, seeking to get rid of it). Heider’s (1946, 1958) balance theory was perhaps the earliest effort at developing such a consistency theory (for discussion and reviews, see Crockett, 1982; Eagly & Chaiken, 1993, pp. 133–144). Osgood and Tannenbaum’s (1955) congruity theory represented another variety of consistency theory (for discussion and reviews, see Eagly & Chaiken, 1993, pp. 460–462; R. Wyer, 1974, pp. 151–185).1 But of all the efforts at articulating the general notion of cognitive consistency, the most influential and productive has been Leon Festinger’s (1957) cognitive dissonance theory. This chapter offers first a sketch of the general outlines of dissonance theory and then a discussion of several areas of research application.



General Theoretical Sketch 131



Elements and Relations Cognitive dissonance theory is concerned with the relations among cognitive elements (also called cognitions). An element is any belief, opinion, attitude, or piece of knowledge about anything—about other persons, objects, issues, oneself, and so on. Three possible relations might hold between any two cognitive elements. They might be irrelevant to each other, have nothing to do with each other. My belief that university tuition will increase next year and my favorable opinion of Swiss chocolate are presumably irrelevant to each other. Two cognitive elements might be consonant (consistent) with each other; they might hang together, form a package. My belief that the Greater Chicago Food Depository is a worthy charity and my knowing I donate money to that organization are presumably consonant cognitions. Finally, two cognitive elements might be dissonant (inconsistent) with each other. The careful specification of a dissonant relation is this: Two elements are said to be in a dissonant relation if the opposite of one element follows from the other. Thus (to use Festinger’s classic example of a smoker) the cognition that “I smoke” and the cognition that “smoking causes cancer” are dissonant with each other; from knowing that smoking causes cancer, it follows that I should not smoke—but I do.2



Dissonance When two cognitions are in a dissonant relation, the person with those two cognitions is said to have dissonance, to experience dissonance, or to be in a state of dissonance. Dissonance is taken to be an aversive motivational state; persons will want to avoid experiencing dissonance, and if they do encounter dissonance, they will attempt to reduce it. Dissonance may vary in magnitude: one might have a lot of dissonance, a little, or a moderate amount. As the magnitude of dissonance varies, so will the pressure to reduce it; with increasing dissonance, there will be increasing pressure to reduce it. With small amounts of dissonance, there may be little or no motivational pressure.



Factors Influencing the Magnitude of Dissonance 132



Expressed most broadly, the magnitude of dissonance experienced will be a function of two factors. One is the relative proportions of consonant and dissonant elements. Thus far, dissonance has been discussed as a simple two-element affair, but usually two clusters of elements are involved. A smoker may believe, on the one hand, that smoking reduces anxiety, makes one appear sophisticated, and tastes good and, on the other hand, also believe that smoking causes cancer and is expensive. There are here two clusters of cognitions, one of elements consonant with smoking and one of dissonant elements. Just how much dissonance this smoker experiences will depend on the relative size of these two clusters. As the proportion of consonant elements (to the total number of elements) increases, less and less dissonance will be experienced, but as the cluster of dissonant elements grows (compared with the size of the consonant cluster), the amount of dissonance will increase. The second factor that influences the degree of dissonance is the importance of the elements or issue. The greater importance this smoker assigns to the expense and cancer-causing aspects of smoking, the greater the dissonance experienced; correspondingly, the greater importance assigned to anxiety reduction and the maintenance of a sophisticated appearance, the less dissonance felt. If the entire question of smoking is devalued in importance, less dissonance will be felt.



Means of Reducing Dissonance There are two broad means of reducing dissonance, corresponding to the two factors influencing the magnitude of dissonance. The first way to reduce dissonance is by changing the relative proportions of consonant and dissonant elements. This can be accomplished in several ways. One can add new consonant cognitions; the smoker, for instance, might come to believe that smoking prevents colds—a new consonant cognition added to the consonant cluster. One can change or delete existing dissonant cognitions; the smoker might persuade him- or herself that, say, smoking does not really cause cancer. The other way to reduce dissonance is by altering the importance of the issue or the elements involved. The smoker could reduce dissonance by deciding that the expense of smoking is not that important (devaluing the importance of that dissonant cognition), might come to think that reducing anxiety is an important outcome (increasing the importance of a consonant cognition), or might decide that the whole question of smoking just is not 133



that important. (For an illustration of this means of dissonance reduction, see Denizeau, Golsing, & Oberle, 2009.)



Some Research Applications Cognitive dissonance theory has produced a great deal of empirical work (for general reviews, see Brehm, 2007; Cooper, 2007; Harmon-Jones, 2002; Harmon-Jones & Mills, 1999; Stone & Fernandez, 2008a). In the study of persuasive communication, at least four research areas are of interest: decision making, selective exposure to information, induced compliance, and hypocrisy induction.3



Decision Making One application of dissonance theory concerns decision making (or choice making). Dissonance is said to be a postdecisional phenomenon; dissonance arises after a decision or choice has been made. When facing a decision (in the simplest case, a choice between two alternatives), one is said to experience conflict. But after making the choice, one will almost inevitably experience at least some dissonance, and thus one will be faced with the task of dissonance reduction. So the general sequence is (a) conflict, (b) decision, (c) dissonance, and (d) dissonance reduction.



Conflict Virtually every decision a person makes is likely to involve at least some conflict. Rarely does one face a choice between one perfectly positive option and one absolutely negative alternative. Usually, one chooses between two (or more) alternatives that are neither perfectly good nor perfectly bad—and hence there is at least some conflict, because the choice is not without some trade-offs. Just how much conflict is experienced by a person facing a decision will depend (at least in part) on the initial evaluation of the alternatives. When (to take the simplest twooption case) the two alternatives are initially evaluated similarly, the decision maker will experience considerable conflict; two nearly equally attractive options make for a difficult choice. This conflict stage is the juncture at which persuasive efforts are most obviously relevant. Ordinarily, persuasive efforts are aimed at regulating (either increasing or decreasing) the amount of conflict experienced by 134



decision makers. If one’s friend is inclined toward seeing the new actionadventure film, rather than the new romantic comedy, one can attempt to undermine that preference and so increase the friend’s conflict (by saying things aimed at getting the friend to have a less positive evaluation of the action film and a more positive evaluation of the comedy), or one can attempt to persuade the friend to follow that inclination and so reduce the friend’s conflict (by saying things aimed at enhancing the evaluation of the already preferred action film and at reducing further the evaluation of the comedy). Of course, a persuader might attempt to regulate a decision maker’s conflict by trying to alter the evaluation of only one (not both) of the alternatives; I might try to get you to have a more positive attitude toward my preferred position on the persuasive issue, although I do not attack the opposing point of view. But—perhaps not surprisingly—the research evidence suggests that persuasive communications that only make arguments supporting the persuader’s position (one-sided messages) are generally not as effective as messages that also refute arguments favoring the opposing side (refutational two-sided messages). This research (also discussed in Chapter 11) suggests that as a rule persuaders are most likely to successfully regulate the conflict experienced by the persuadee if they attempt to influence the evaluation not only of their preferred alternative but of other options as well (for a review, see O’Keefe, 1999a). In any case, by regulating the degree of conflict experienced, the persuader can presumably make it more likely that the persuadee will choose the option desired by the persuader. But after the persuadee has made a choice (whether or not the one wanted by the persuader), the persuadee will almost inevitably face at least some dissonance—and, as will be seen, the processes attendant to the occurrence of dissonance have important implications for persuasion.4



Decision and Dissonance Some dissonance is probably inevitable after a decision because in virtually every decision, at least some aspects of the situation are dissonant with one’s choice. Specifically, there are likely to be some undesirable aspects to the chosen alternative and some desirable aspects to the unchosen alternative; each of these is dissonant with the choice made. Consider, for example, a person choosing where to eat lunch. Al’s Fresco 135



Restaurant offers good food and a pleasant atmosphere but is some distance away and usually has slow service. The Bistro Cafe has so-so food and the atmosphere isn’t much, but it’s nearby and has quick service. No matter which restaurant is chosen, there will be some things dissonant with the person’s choice. In choosing the Bistro, for instance, the diner will face certain undesirable aspects of the chosen alternative (e.g., the poor atmosphere) and certain desirable aspects of the unchosen alternative (e.g., the good food the diner could have had at Al’s).



Factors Influencing the Degree of Dissonance The amount of dissonance that one faces following a choice depends most centrally on two factors. One is the similarity of the initial evaluations: The closer the initial evaluations of the alternatives, the greater the dissonance. Thus a choice between two nearly equally attractive sweaters is likely to evoke more dissonance than a choice between one fairly attractive and one fairly unattractive sweater. The other factor is the relative importance of the decision, with more important decisions predicted to yield more dissonance. A choice about what to eat for dinner this evening is likely to provoke less dissonance than a choice of what career to pursue. These two factors represent particularized versions of the general factors influencing the degree of dissonance experienced: the relative proportions of consonant and dissonant elements (because when the two alternatives are evaluated similarly, the proportions of consonant and dissonant elements will presumably approach 50–50) and the importance of the issue or elements (here represented as the importance of the decision).



Dissonance Reduction One convenient way in which a decision maker can reduce the dissonance felt following a choice is by reevaluating the alternatives. By evaluating the chosen alternative more positively than one did before and by evaluating the unchosen alternative less positively than before, the amount of dissonance felt can be reduced. Because this process of re-rating the alternatives will result in the alternatives being less similarly evaluated than they were prior to the decision, this effect is sometimes described as the “postdecisional spreading” of alternatives (in the sense that the alternatives are spread further apart along the evaluative dimension than they had been). If (as dissonance theory predicts) people experience 136



dissonance following decisions, then one should find dissonance reduction in the form of this postdecisional spreading of the alternatives, and one should find greater spreading (i.e., greater dissonance reduction) in circumstances in which dissonance is presumably greater. In simplified form, the typical experimental arrangement in dissonancebased studies of choice making is one in which respondents initially give evaluations of several objects or alternatives and are then faced with making a choice between two of these. After making the choice, respondents are then asked to reevaluate the alternatives, with these rankings inspected for evidence of dissonance reduction through postdecisional spreading of alternatives. The research appears to indicate that one does often find the predicted changes in evaluations following decisions (e.g., Brehm, 1956; G. L. White & Gerard, 1981). However, the evidence is not so strong that the magnitude of dissonance reduction is greater when the conditions for heightened dissonance are present (as when the two alternatives are initially rated closely, or the decision is important), because conflicting findings have been reported, especially for the effects of decisional importance (for discussion, see Converse & Cooper, 1979). However, much of the research evidence concerning postdecisional reevaluation suffers from methodological problems. In particular, M. K. Chen and Risen (2010; Risen & Chen, 2010) have argued that the research procedures common in this research area can produce the appearance of the postdecisional spreading of alternatives (attitude change) even if no change has actually occurred. The gist of the argument is that the initial ranking is an imperfect index (as can be seen by the fact that when faced with a choice between two differently ranked alternatives, people do not always choose the better-ranked one)—and this imperfection will generate the appearance of postdecisional spreading even if there is no genuine change in evaluations.5 (For some discussion, see M. K. Chen & Risen, 2009; Sagarin & Skowronski, 2009a,b.6) But alternative experimental procedures (aimed at addressing the apparent defects) appear to yield results similar to those previously obtained (e.g., Alós-Ferrer, Graníc, Shi, & Wagner, 2012; Sharot, Fleming, Yu, Koster, & Dolan, 2012), indicating that postdecisional preference shifts are in fact real. However, when researchers control for the artifacts identified by M. K. Chen and Risen (2010), the effects appear to be diminished or 137



weakened relative to those observed in earlier research (Izuma & Murayama, 2013). In sum, the sorts of postdecisional re-evaluations expected by dissonance theory appear to be genuine, but the effect is not as large as one might have supposed on the basis of the initial research findings. The general finding of postdecisional spreading in the evaluations of the alternatives suggests that decision maker satisfaction will “take care of itself” (Wicklund & Brehm, 1976, p. 289). Because people are likely to more positively value that which they have freely chosen, then if one can induce them to choose a given alternative, they will be likely to more positively value that alternative just because they have chosen it. For example, if people are induced to buy a product, they will likely have a more positive attitude toward the product just as a consequence of having chosen to buy it. Of course, this does not mean that every purchaser is guaranteed to end up a satisfied customer; it still may happen that (say) a new car buyer decides that the car is a lemon and returns it to the dealer. Nevertheless, there are forces at work that incline people to be happier with whatever they have chosen, just because they have chosen it. Indeed, in the service of postdecisional dissonance reduction, people may engage in selective information seeking or processing that confirms their choice. For example, purchasers of a given model of automobile may be especially drawn to advertisements for the vehicle they just purchased (the classic study is Ehrlich, Guttman, Schönbach, & Mills, 1957; for examples of related work, see Fischer, Lea, Kastenmüller, Greitemeyer, Fischer, & Frey, 2011; Fischer, Schulz-Hardt, & Frey, 2008; Keng & Liao, 2009; Shani & Zeelenberg, 2007). Information supportive of one’s decision will naturally be seen as a source of dissonance-reducing material (for more general discussions of studies of post-decisional information preferences, see D’Alessio & Allen, 2007; Fischer, 2011; Fischer & Greitemeyer, 2010).7



Regret Given these postdecisional dissonance-reduction processes, persuaders might naturally infer that once a persuadee has been induced to decide the way the persuader wants, then the persuader’s job is done; after all, having made the choice, the persuadee is likely to become more satisfied with it through the ordinary processes of dissonance reduction. However, this inference is unsound; persuaders who reason in this fashion may find their 138



persuasive efforts failing in the end, in part because of the occurrence of regret (see Festinger, 1964). When regret occurs, it arises after the decision has been made but before dissonance has been reduced (through postdecisional spreading of alternatives). When regret is happening, the alternatives are temporarily evaluated more similarly than they were initially. Then, following this regret phase (during which dissonance presumably increases), the person moves on to dissonance reduction, with the evaluations of the alternatives spreading farther apart (see Festinger & Walster, 1964). Regret is not inevitable, and research is only beginning to explore the factors influencing the arousal and resolution of regret (e.g., Keaveney, Huber, & Hemnann, 2007; Mannetti, Pierro, & Kruglanski, 2007; Rosenzweig & Gilovich, 2012; Zeelenberg & Pieters, 2007), but regret occurs sufficiently commonly to be quite familiar. (Indeed, some readers will have recognized “buyer’s remorse” in the preceding description.) One plausible account of this regret phenomenon is that having made the choice, the decision maker now faces the task of dissonance reduction. Naturally, the decision maker’s attention focuses on those cognitions that are dissonant with his or her choice—on undesirable aspects of the chosen option and on desirable aspects of the unchosen option—perhaps in the hope of eventually being able to minimize each. As the decision maker focuses on undesirable aspects of the chosen alternative, that alternative may seem (at least temporarily) less attractive than it had before; focusing on desirable aspects of the unchosen option may make that option seem (at least temporarily) more attractive than it had before. With the chosen alternative becoming rated less favorably, and the unchosen alternative becoming rated more favorably, the two alternatives naturally become evaluated more similarly than they had been. During this regret phase, it is even possible that the initial evaluations become reversed, so that the initially unchosen alternative becomes rated more favorably than the chosen option. In such a circumstance, the decision maker may back out of the original choice. This outcome becomes more likely when the two alternatives are initially evaluated rather similarly because in such a circumstance, comparatively small swings in absolute evaluations can make for reversals in the relative evaluations of the alternatives. There is a moral here for persuaders concerning the importance of follow139



up persuasive efforts. It can be too easy for a persuader to assume that the job is done when the persuadee has been induced to choose in the way the persuader wants, but the possibility of regret, and particularly the possibility that the decision maker’s mind may change, should make the persuader realize that simply inducing the initial decision may not be enough. A fitting example is provided by an investigation of automobile buying. In purchases of automobiles from a dealer, ordinarily some time elapses between the buyer’s agreeing to buy the car and the delivery of the car to the buyer. It sometimes happens that during this interval, the would-be purchaser changes his or her mind and backs out of the decision to buy the car. (There are likely any number of reasons why this happens, but it should be easy enough to imagine that at least some of the time regret is at work.) In this investigation, during the interval between decision and delivery, some automobile purchasers received two follow-up telephone calls from the seller; the calls emphasized the desirable aspects of the automobile that had been chosen, reassured the purchaser of the wisdom of the decision, and (one might say) encouraged the purchaser to move past the regret phase and on to the stage of dissonance reduction. Other purchasers received no such calls. Significantly fewer of the purchasers receiving the follow-up calls backed out of their decisions than did those not receiving the call (the back-out rate was cut in half), underscoring the potential importance of follow-up persuasive efforts (Donnelly & Ivancevich, 1970). (For another illustration of post-purchase follow-up messages, with complexities, see Hunt, 1970.)



Selective Exposure to Information A second area of dissonance theory research that is relevant to persuasion concerns persons’ propensities to expose themselves selectively to information. Below, the dissonance theory analysis of information exposure is presented, followed by a discussion of the relevant research.



The Dissonance Theory Analysis If dissonance is an aversive motivational state, then naturally persons will want to do what they can to avoid dissonance-arousing situations and will prefer instead to be in circumstances that do not arouse dissonance (or that even increase the consonance of their cognitions). This general idea finds specific expression in the form of dissonance theory’s selective exposure 140



hypothesis. Broadly put, this hypothesis has it that persons will prefer to be exposed to information that is supportive of (consonant with) their current beliefs rather than to nonsupportive information (which presumably could arouse dissonance).8 This hypothesis applies to information exposure in any circumstance, but it perhaps becomes especially pointed in the context of exposure to mass media political information. If persons generally seek out only media sources that confirm or reinforce their prior political beliefs (and, correlatively, avoid exposure to nonsupportive or inconsistent information), a polarized electorate might be the result. More generally, of course, the selective exposure hypothesis suggests that persuaders (through the mass media or otherwise) may need to be concerned about getting receivers to attend to their messages. If, as dissonance theory suggests, there is a predisposition to avoid nonsupportive information, then persuaders may face the task of somehow overcoming that obstacle so that their communications can have a chance to persuade.



The Research Evidence In the typical experimental research paradigm for the investigation of selective exposure, respondents’ attitudes on a given issue are assessed. Then respondents are given the choice of seeing (reading, hearing) one of several communications on the issue. These communications are described in a way that makes clear what position on the issue is advocated by each, and both supportive and nonsupportive messages are included. The respondent is then asked to select one of the messages. Support for the selective exposure hypothesis consists of respondents’ preferring to see supportive rather than nonsupportive communications. (For a discussion of methodological issues in such experiments, see Feldman, Stroud, Bimber, & Wojcieszak, 2013.) There is now considerable accumulated evidence indicating a general preference for supportive over nonsupportive information (for some reviews, see Hart et al., 2009; Smith, Fabrigar, & Norris, 2008).9 That is, people do commonly prefer to be exposed to information that is congenial with their existing beliefs and attitudes than to uncongenial information. Such a preference is seen across a variety of concrete circumstances, including preferences for news media outlets (e.g., Iyengar & Hahn, 2009; 141



Stroud, 2008). However, the strength of this preference can vary depending on a number of factors. A great many different such factors have been explored, the research evidence is sometime sparse, and there is not yet an entirely clear picture of how all these might fit into some larger structure (for some reviews, see Cotton, 1985; Frey, 1986; Hart et al., 2009; Smith, Fabrigar, & Norris, 2008).10 But as an example: The relevance of an issue to a person’s core values influences the strength of the preference for supportive information; the preference is stronger when the issue concerns core values (e.g., issues such as abortion or euthanasia) than when such values are not relevant (for a review, see Hart et al., 2009). Even in the latter circumstances, however, a (relatively weak) preference for supportive information is apparent. But research has also identified other, potentially competing, influences on information exposure—influences that can be strong enough to overcome the usual preference for supportive information. One such influence is the perceived utility of the information, with persons preferring information with greater perceived usefulness even if it is nonsupportive. Consider, for example, an investigation in which undergraduates initially chose to take either a multiple-choice exam or an essay exam. The students were then asked for their preferences among reading several articles, some supporting the decision and some obviously nonsupportive. For instance, for a student who chose the multiple-choice exam, the nonsupportive articles were described as arguing that students who prefer multiple-choice tests would actually be likely to do better on essay exams. Contrary to the selective exposure hypothesis—but not surprisingly—most of the students preferred articles advocating a change from the type of exam they had chosen (Rosen, 1961). Obviously, in this study, the nonsupportive communication offered information that might be of substantial usefulness to the students, and the perceived utility of the information could well have outweighed any preference for supportive information (for a review indicating the importance of information utility as a moderator of selective exposure, see Hart et al., 2009).11 Fairness norms may also play a role in information exposure. In certain social settings, there is an emphasis on obtaining the greatest amount of information possible, being fair to all sides, and being open-minded until all the evidence is in. One such setting is the trial. For example, in one study, participants received brief synopses of a murder case and then 142



rendered a judgment about the guilt of the defendant. They were subsequently offered a chance of seeing either confirming or disconfirming information. Participants showed a general preference for nonsupportive information, perhaps because the trial setting was one that made salient the norms of fairness and openness to evidence (Sears, 1965).



Summary All told, there is a general preference for supportive information. The strength of this preference may vary, and the preference can be overridden by other considerations (e.g., information utility). But dissonance theory’s expectations about general information preferences have certainly been confirmed. For that reason, persuaders who hope to encourage attention to their messages will want to be attentive to the factors influencing information exposure, as these may suggest avenues by which such attention can be sought (see, e.g., Flay, McFall, Burton, Cook, & Warnecke, 1993). However, the idea of selective exposure comprises two distinguishable processes: avoidance of dissonant information (“selective avoidance”) and attraction to consonant information (“selective approach”). Either (or both) might be responsible for the appearance of selective exposure effects. For example, in an experimental situation in which people choose between consonant and dissonant information, the choice of consonant information might be driven by selective avoidance (motivation to avoid the dissonant information), selective approach (desire to see the consonant information), or some combination of these. There is not yet much research evidence on the matter, but there is some reason to suspect that selective avoidance effects may be weaker than selective approach effects (e.g., Garrett, 2009; Garrett & Stroud, 2014; see also Cotton, 1985, p. 26; Frey, 1986, pp. 69–70). That is, people may actively look for confirming information, but not necessarily avoid disconfirming information. In any case, one ought not assume that selective approach and selective avoidance are equally powerful processes. So, for example, even though an information environment such as afforded by the Internet may enable selective approach, and although persons do generally have a preference for supportive information, people may nevertheless not actively avoid discrepant information—and under the right circumstances (e.g., high perceived information utility, a setting that prioritizes fairness) might even seek out nonsupportive information (for 143



some relevant work, see Valentino, Banks, Hutchings, & Davis, 2009; Wojcieszak & Mutz, 2009).



Induced Compliance Perhaps the greatest amount of dissonance research concerns what is commonly called induced compliance. Induced compliance is said to occur when an individual is induced to act in a way discrepant from his or her beliefs and attitudes. One special case of induced compliance is counterattitudinal advocacy, which is said to occur when persons are led to advocate some viewpoint opposed to their position. Most of the research on induced compliance concerns counterattitudinal advocacy because that circumstance has proved a convenient focus for study. (For a detailed discussion of induced compliance research, see Eagly & Chaiken, 1993, pp. 505–521.)



Incentive and Dissonance in Induced-Compliance Situations Obviously, induced compliance situations have the potential to arouse dissonance; after all, a person is acting in a way discrepant from his or her beliefs. Dissonance theory suggests that the amount of dissonance experienced in an induced compliance situation will depend centrally on the amount of incentive offered to the person to engage in the discrepant action. Any incentive offered for performing the counterattitudinal action (e.g., some promised reward or threatened punishment) is consistent (consonant) with engaging in the action. Thus someone who performs a counterattitudinal action with large incentives for doing so will experience relatively little dissonance. To use Festinger’s (1957) example: Suppose you are offered a million dollars to publicly state that you like reading comic books (assume, for the purpose of the example, that you find this offer believable and that you do not like reading comic books). Presumably, you would accept the money and engage in the counterattitudinal advocacy. You might experience some small amount of dissonance (from saying one thing and believing another) —but the million dollars is an important element that is consonant with your having performed the action, and hence overall there is little dissonance experienced. 144



But if the incentive had been smaller (less money offered), then the amount of dissonance experienced would have been greater. The greatest possible dissonance would occur if the incentive were only just enough to induce compliance. Suppose that you would not have agreed to engage in the counterattitudinal advocacy for anything less than $100. In that case, an offer of exactly $100—the minimum needed to induce compliance— would have produced the maximum possible dissonance. Any incentive larger than that minimum would only have reduced the amount of dissonance experienced. When substantial dissonance is created through induced compliance, pressure is created to reduce that dissonance. One easy route to dissonance reduction is to bring one’s private beliefs into line with one’s behavior. For example, if you declared that you liked reading comic books when offered only $100 (your minimum price) for doing so, you would experience considerable dissonance and could easily reduce it by deciding that you think reading comic books isn’t quite as bad as you thought. What happens if the incentive offered is insufficient to induce compliance? That is, what are the consequences if a person is offered some incentive for engaging in a counterattitudinal action, and the person does not comply? To continue the example, suppose that you had been offered only $10 to say that you like to read comic books. You would decline the offer, thereby losing the possibility of getting the $10—and hence you would experience some dissonance over that (“I could have had that $10”). But you would not experience much dissonance, and certainly not as much as if you had turned down an offer of $90. Faced with the dissonance of having turned down $90, one natural avenue to dissonance reduction would be to strengthen one’s initial negative attitude (“I was right to turn down that $90, because reading comic books really is pretty bad”). So the relationship between the amount of incentive offered and the amount of dissonance experienced is depicted by dissonance theory as something like an inverted V. With increasing incentive, there is increasing dissonance—up to the point at which the incentive is sufficiently large to induce compliance. But beyond that point, increasing incentive produces decreasing dissonance, such that with very large incentives, there is little or no dissonance experienced from engaging in the counterattitudinal action. Thus, so long as the amount of incentive is sufficient to induce compliance, additional incentive will make it less likely that the person will come to have more favorable attitudes toward 145



the position being advocated. In an archetypal experiment, Festinger and Carlsmith (1959) obtained striking evidence for this analysis. In this study, participants performed an exceedingly dull and tedious task. At the conclusion of the task, they were asked to tell a student who was waiting to participate in the experiment (the student was a confederate of the experimenter) that the task was enjoyable and interesting. As incentive for performing this counterattitudinal behavior, participants were offered money; half were offered $1 (low incentive), and half were offered $20 (high incentive). After engaging in the counterattitudinal advocacy, participants’ attitudes toward the task were assessed. Festinger and Carlsmith found, consistent with dissonance theory’s predictions, that those receiving $1 came to think that the task was significantly more enjoyable than did those who complied for $20. Those who complied under the influence of a large incentive ($20) presumably experienced less dissonance from engaging in the counterattitudinal act (because they had the $20 that was consonant with performing the act)— and so had little need for attitude change. By contrast, participants receiving the small incentive ($1) presumably experienced more dissonance and hence had more motivation to change their attitudes to reduce dissonance; they reduced their dissonance by coming to have a more favorable attitude toward the dull task. Subsequent investigations provided additional confirming evidence (for a classic collection of studies on induced compliance, see Elms, 1969). For example, E. Aronson and Carlsmith (1963) found that children prohibited from playing with an attractive toy by a mild threat (of punishment for disobedience) subsequently found the toy less attractive than did children prohibited by a severe threat. That is, those who engaged in the counterattitudinal action of avoiding the toy when given only mild incentive to do so apparently experienced greater dissonance (than did those who avoided the toy when given strong incentives to do so) and hence displayed greater underlying attitude change. There have been relatively fewer studies of circumstances in which the incentives offered are insufficient to induce compliance, but this evidence is also generally consistent with dissonance theory predictions. For instance, Darley and Cooper (1972) found that persons who were offered insufficient incentives (to engage in counterattitudinal advocacy) were inclined to strengthen their initial attitudes, and—as expected from dissonance theory—greater strengthening 146



occurred with larger incentives.



Counterattitudinal-Advocacy–Based Interventions The potential utility of induced-compliance processes as a basis for attitude change is nicely illustrated by counterattitudinal-advocacy interventions. In these interventions, participants are led to engage in counterattitudinal advocacy (under conditions of minimal incentive) as a means of producing attitude change. In particular, such interventions have formed the basis of an effective eating disorder prevention program. As background: Among young women, one source of eating disorders such as bulimia is internalization of an excessively thin ideal body image. Interventions have been designed in which at-risk women (ones with excessively elevated body image concerns) engage, voluntarily, in what amounts to counterattitudinal advocacy, in which they argue against the thin ideal (with this concretized in various ways, such as role-playing exercises in which they attempt to dissuade a friend from pursuing the thin ideal). A number of studies have found that this intervention significantly reduces various risk factors for eating disorders (e.g., degree of thin-ideal internalization, body dissatisfaction, bulimic symptoms, etc.); for details and discussion, see Becker, Smith, and Ciao (2006), Perez, Becker, and Ramirez (2010), Roehrig, Thompson, Brannick, and van den Berg (2006), Stice, Chase, Stormer, and Appel (2001), Stice, Marti, Spoor, Presnell, and Shaw (2008), and Stice, Shaw, Becker, and Rohde (2008). Similar counterattitudinal advocacy interventions have been explored in other contexts such as prejudice reduction (Eisenstadt, Leippe, Rivers, & Stambush, 2003; Heitland & Bohner, 2010) and attitudes toward online gaming (Wan & Chiou, 2010). The general idea is the same: Persons who, with minimal incentive, voluntarily engage in counterattitudinal advocacy will emerge with attitudes more favorable to the views they have just advocated.



The “Low, Low Price” Offer Another example of an induced compliance process is provided by the familiar marketing ploy of the “low, low price” offer. This offer is sometimes cast as a straightforward lower price (“fifty cents off”), sometimes as “two for the price of one” (“buy one get one free,” “three for 147



the price of two,” etc.). The central idea is that a lower price is offered to the consumer, making purchase more likely. Now imagine a situation in which a particular consumer is faced with competing brands of soap. This consumer does not have an especially positive impression of Brand A—it’s not the consumer’s usual brand—but Brand A is running a really good low-price special (“three bars for the price of one”). From a dissonance theory perspective, this lower price represents an increased incentive for the consumer to purchase Brand A. As the deal gets better and better—that is, as the price gets lower and lower—there is more and more incentive to comply (incentive to purchase). For example, there is more incentive to comply when the deal is “three for the price of one” than when the deal is “two for the price of one.” The key insight offered by dissonance theory here is this: The greater the incentive to comply, the less dissonance created by the purchase—and hence less chance for favorable attitude change toward the brand. This consumer might buy Brand A this time (because the price is so low), but the consumer’s underlying unfavorable attitude toward Brand A is not likely to change—precisely because the incentive to comply was so great. So while the “low, low price” offer might boost sales for a while, it can also undermine the development of more positive attitudes toward the brand. An illustration of these processes was offered in a set of five field experiments concerning the effects of low introductory selling prices. House brands of common household products (e.g., aluminum foil, toothpaste, light bulbs) were introduced in various stores in a chain of discount houses. In some of the stores, the brands were introduced at the regular price, whereas at other stores, the brands were introduced with a low introductory price offer for a short period (before the price increased to the regular price). As one might expect, when the low-price offer was in effect, sales were higher at the stores offering the lower prices. But when prices returned to normal, the subsequent sales were greater at the stores that had the initial higher prices (Doob, Carlsmith, Freedman, Landauer, & Tom, 1969). Introducing these products at low introductory prices proved to be harmful to long-run sales, presumably because there was relatively little brand loyalty established by the low introductory selling price. Thus the greater incentive created by the lower price apparently prevented the development of sufficiently positive attitudes toward the brand. 148



One should not conclude from this that the low-price offer is a foolish marketing stratagem that should never be used. The point is that this marketing technique can set in motion forces opposed to the development of positive attitudes toward the brand and that these forces are greater as the incentive becomes greater (as the deal gets better). But some low-price offers are better than others (from the view of creating favorable attitude change): A low-price offer that is only just barely good enough to induce purchase—an offer that provides just enough incentive to induce compliance—will create the maximum possible dissonance (and so, a marketer might hope, maximum favorable attitude change toward the product). Low-price offers may also be useful as strategies for introducing new brands; the marketer’s plan is that the low price would induce initial purchase and that this exposure to the brand’s intrinsic positive characteristics will create a positive attitude toward the brand. Of course, if the brand does not have sufficiently great intrinsic appeal (as was likely with the house brands studied by Doob et al., 1969), then using low introductory prices to induce trial will not successfully create underlying positive attitudes toward the brand. (For indications of the complexity of the effects of price promotion on attitudes, see DelVecchio, Henard, & Freling, 2006; Raghubir & Corfman, 1999; Yi & Yoo, 2011.)



Limiting Conditions Researchers have not always obtained the induced compliance effects predicted by dissonance theory. Two important limiting conditions have been identified. First, the predicted dissonance effects seem to occur only when the participants feel that they had a choice about whether to comply (e.g., about whether to perform the advocacy). That is, freedom of choice seems to be a necessary condition for the appearance of dissonance effects (the classic work on this subject is Linder, Cooper, & Jones, 1967; for a relevant review, see Preiss & Allen, 1998). Thus one can expect that inducing counterattitudinal action with minimal incentive will produce substantial dissonance (and corresponding favorable attitude change) only when the person freely chooses to engage in the counterattitudinal behavior.12 Second, the predicted dissonance effects are obtained only when there is no obvious alternative cause to which the feelings of dissonance can be attributed. Attributional processes are the (often nonconscious) methods by which people arrive at explanations for their feelings. If people can attribute their dissonance feelings to some cause other than their 149



counterattitudinal behavior, the usual dissonance effects will not be observed. For example, if people take a pill (actually a placebo) before engaging in counterattitudinal advocacy and are told the pill will probably make them feel tense or anxious, then counterattitudinal advocacy does not produce the usual changes in attitude (because people attribute their discomfort to the pill, not to the counterattitudinal action; Zanna & Cooper, 1974; for related work, see J. Cooper, 1998, Study 1; Fried & Aronson, 1995; Joule & Martinie, 2008).



Summary Dissonance theory’s expectations about the effects of incentive for counterattitudinal action on attitude change have been confirmed in broad outline—although not without the discovery of unanticipated limiting conditions. When a person freely chooses to engage in counterattitudinal action (without an apparent alternative account of the resulting feelings), increasing incentive for such action leads to lessened pressure for making one’s beliefs and attitudes consistent with the counterattitudinal act. Hence a persuader seeking long-term behavioral change (by means of underlying attitude change) ought not to create intense pressure to engage in the counterattitudinal behavior; rather, the persuader should seek to offer only just enough incentive to induce compliance and let dissonance reduction processes encourage subsequent attitude change.13 Consider, as an example, marketing contests in which consumers are invited to submit a slogan or ad for a product or to write an essay explaining why they like the product, with prizes (cash or goods) to be received by selected entries. When the behavior is counterattitudinal (“I don’t really like the product, I just want to try to win the prize”), larger prizes will likely minimize the development of more favorable attitudes toward the advocated product (compared with smaller prizes). Or consider some social influence tasks commonly faced by parents. In hoping to encourage the young child not to play with the expensive electronic equipment, parents ought to provide only just enough punishment to induce compliance; excessive punishment might produce short-term obedience but not underlying change (e.g., when the parents are present, the child will not play with the equipment—but the child will still want to, and when the back is turned …). Or in trying to encourage children to do their homework, parents ought to think carefully about offering extremely large rewards for compliance; such rewards can 150



undermine the development of positive attitudes toward homework (whereas a minimal reward can induce immediate compliance while also promoting the development of positive attitudes). All these examples illustrate the potential application of the general principle that smaller incentives for freely chosen counterattitudinal behavior are more likely than larger incentives to produce underlying favorable attitudes toward that behavior.



Hypocrisy Induction Hypocrisy as a Means of Influencing Behavior Sometimes a persuader’s task is not so much to encourage people to have the desired attitudes as it is to encourage people to act on existing attitudes. For example, people commonly express positive attitudes toward recycling, natural resource conservation, condom use, and so forth, yet often fail to act accordingly. Such inconsistencies might be exploited by persuaders, however, as suggested by dissonance research on hypocrisy induction. The basic idea is that calling attention to the inconsistency of a person’s attitudes and actions—that is, the person’s hypocrisy—can arouse dissonance, which then is reduced through behavioral change (altering the behavior to make it consistent with the existing attitude).14 For example, in a study of safer-sex practices, Stone et al. (1994) varied whether participants engaged in public proattitudinal advocacy about safe sex and varied whether they were made mindful of their past unsafe practices (by listing circumstances surrounding their past failures to use condoms). The combination of advocacy and mindfulness (the hypocrisy condition) was expected to induce greater dissonance—and so greater subsequent behavioral consistency—than either treatment alone (or neither treatment). Consistent with this expectation, hypocrisy-condition participants (compared with those in other conditions) were more likely to buy condoms (and bought more condoms on average) at the end of the experiment. That is, faced with the reality of their inconsistent actions, these persons reduced their dissonance by bringing their behaviors in line with their safer-sex attitudes. (For other examples and relevant discussion, see Dickerson, Thibodeau, Aronson, & Miller, 1992; Fointiat, 2004; Freijy & Kothe, 2013; Fried & Aronson, 1995; Hing, Li, & Zanna, 2002; Stone & 151



Fernandez, 2011; Stone, Wiegand, Cooper, & Aronson, 1997.)



Hypocrisy Induction Mechanisms The specific interventions or treatments that might induce hypocrisy have not yet been carefully distinguished or explored. Most hypocrisy induction studies to date have employed structured proattitudinal advocacy exercises of the sort used by Stone et al. (1994), but presumably the underlying mechanism involves the salience of attitude-behavior inconsistency. The general idea is that inconsistencies between beliefs and behavior can, if made sufficiently salient, lead individuals to seek consistency, such as by changing their actions to accord with their attitudes. Having persons engage in proattitudinal advocacy is one possible way, but surely not the only possible means, of enhancing the salience of an existing inconsistency. For instance, in the right circumstances, a simple reminder of one’s attitudinal commitments might be sufficient. Consider, for example, Aitken, McMahon, Wearing, and Finlayson’s (1994) research concerning water conservation. Households given feedback about their water consumption, combined with a reminder of their previously expressed belief in their responsibility to conserve water, significantly reduced their consumption. Feedback alone was useful in reducing consumption but not as effective as the combination of feedback and the (presumably hypocrisy-inducing) reminder. (For a similar intervention, with similar results, concerning electricity conservation, see Kantola, Syme, & Campbell, 1984.) To date, it appears that successful hypocrisy induction treatments involve a combination of two key elements: (a) ensuring the salience of the relevant attitude (e.g., through proattitudinal advocacy or through being explicitly reminded of one’s commitment to the attitude) and (b) ensuring the salience of past failures to act in ways consistent with that attitude (e.g., by having the person recall such failures or by giving feedback indicating such failures). When only one of these elements is present, hypocrisy effects are weaker or nonexistent (Aitken et al., 1994; E. Aronson, Fried, & Stone, 1991; Dickerson et al., 1992; Kantola et al., 1984; Stone et al., 1994; Stone, Wiegand, Cooper, & Aronson, 1997, Experiment 1). It will plainly be useful to have some clarification of alternative means of implementing these two elements, identification of circumstances in which one or another form of implementation is more powerful, and so forth. (For some discussion focused specifically on proattitudinal advocacy 152



mechanisms, see Stone, 2012; Stone & Fernandez, 2008b; Stone & Focella, 2011.)



Backfire Effects It might appear straightforward enough to use hypocrisy as a means of inducing behavioral change, but it is important to consider that, faced with evidence of inconsistency between attitudes and actions, people might change their attitudes rather than their behaviors. Fried (1998) had participants engage in public advocacy about the importance of recycling, under one of three conditions varying the salience of past inconsistent behavior. Some participants listed their past recycling failures anonymously (as in previous hypocrisy induction manipulations), some listed their past failures in ways that permitted them to be personally identified, and some did not list past failures (the no-salience condition). Persons in the anonymous-salience condition exhibited the usual behavioral effects of hypocrisy (e.g., they pledged larger amounts of money to a recycling fund than did persons in the no-salience condition), but persons in the identifiable-salience condition did not. These persons, instead of changing behaviors to become consistent with their prorecycling attitudes, changed their attitudes to become consistent with their recycling failures—specifically, they displayed a reduced belief in the importance of recycling. It is not yet clear exactly how to explain such reversal of effects, how general such outcomes are, the conditions under which they are likely to occur (perhaps, say, with relatively unimportant attitudes), and so forth. But persuaders will certainly want to take note of the potential dangers of hypocrisy induction as an influence mechanism. As a means of changing a person’s behavior, pointing out that the person’s conduct is inconsistent with the person’s professed beliefs might lead to the desired behavioral change—or might lead to belief revision (and so backfire on the persuader). One factor that might plausibly encourage such backfire effects is selfefficacy, people’s perceived ability to perform the behavior (perceived behavioral control, in the terminology of reasoned action theory as discussed in Chapter 6). If people think that behavior change is unavailable as a method to reduce dissonance, they may turn to attitude change instead. Thus one likely limiting condition on the effectiveness of hypocrisy induction is that the level of self-efficacy (perceived behavioral control) be 153



sufficiently high. In fact, if this limiting condition is not met, hypocrisy induction might well produce boomerang attitude change, that is, attitude change in a direction opposite that wanted by the persuader.15 This exemplifies a general point about arousing dissonance as a means of influence. The key to the successful use of dissonance arousal as an influence strategy is to arouse dissonance and then to shut off all the possible modes of dissonance reduction except the desired one. (For some classic discussion of this idea, see Abelson, 1968; E. Aronson, 1968, pp. 14–17. For discussion of this idea in the context of hypocrisy induction specifically, see Fried, 1998.) When hypocrisy induction arouses dissonance, but the only avenue to dissonance reduction is attitude change (not behavioral change), then naturally that’s the route people will choose.



Revisions of, and Alternatives to, Dissonance Theory A number of revisions to dissonance theory have been suggested, and several competing explanations have also been proposed. These various alternative possibilities are too numerous and individually complex (e.g., some are focused specifically on explaining induced compliance effects, whereas others offer broader reinterpretations of dissonance work) to be easily compared here. But by way of illustration, it may be useful to discuss one facet that is common to a number of the alternatives, namely, an emphasis on the centrality of the concept of self (self-concept, selfidentity) to dissonance phenomena. For example, E. Aronson (1992, 1999) has suggested revising dissonance theory to specify that dissonance arises most plainly from inconsistencies that distinctly involve the self. That is, “dissonance is greatest and clearest when it involves not just any two cognitions but, rather, a cognition about the self and a piece of our behavior that violates that self-concept” (E. Aronson, 1992, p. 305). Steele’s self-affirmation theory, offered more as a competitor to dissonance theory, suggests that the key motivating force behind dissonance phenomena is a desire for self-integrity (Steele, 1988; for reviews, see J. Aronson, Cohen, & Nail, 1999; Sherman & Cohen, 2006). The various attitudinal and behavioral changes attendant to dissonance are argued to reflect the desire to maintain an image of the self as “adaptively 154



and morally adequate, that is, as competent, good, coherent, unitary, stable, capable of free choice, capable of controlling important outcomes, and so on” (Steele, 1988, p. 262). Given the apparent centrality of the self to dissonance phenomena, perhaps it is unsurprising that several commentators have pointed to the close conceptual connections between dissonance and guilt (e.g., Baumeister, Stillwell, & Heatherton, 1995; Kenworthy, Miller, Collins, Read, & Earleywine, 2011; Klass, 1978; O’Keefe, 2000, pp. 85–88; Stice, 1992). Guilt paradigmatically arises from conduct that is inconsistent with selfstandards, and it commonly motivates actions aimed at restoring a sense of integrity and worth. Some dissonance phenomena (e.g., induced compliance) may be more amenable to guilt-based analyses than others (e.g., selective exposure; cf. Kenworthy et al., 2011), but continuing attention to the relationship of dissonance and guilt seems appropriate.16 It remains to be seen how successful these and other alternatives will prove to be. (For examples and discussion of various approaches, see Beauvois & Joule, 1999; J. Cooper, 2007; Eagly & Chaiken, 1993, pp. 505–552; Harmon-Jones, Amodio, & Harmon-Jones, 2010; Nail, Misak, & Davis, 2004; Stone & Cooper, 2001; Stone & Fernandez, 2008a; Van Overwalle & Jordens, 2002.) The general question is the degree to which a given framework can successfully encompass the variety of findings currently housed within dissonance theory, while also pointing to new phenomena recommending distinctive explanation. But no matter the particulars of the resolution of such issues, it is plain that dissonance-related phenomena continue to provide rich sources of theoretical and empirical development. For students of persuasion, these various alternatives bear watching because of the possibility that these new frameworks will shed additional light on processes of social influence.



Conclusion Dissonance theory does not offer a systematic theory of persuasion (and was not intended to). But dissonance theory has served as a fruitful source of ideas bearing on social influence processes and has stimulated substantial relevant research. To be sure, unanticipated complexities have emerged (as in the discovery of limiting conditions on induced compliance effects or the phenomenon of postdecisional regret). But cognitive dissonance theory has yielded a number of useful and interesting findings 155



bearing on processes of persuasion.



For Review 1. Explain the general idea of cognitive consistency. What is a cognitive element (cognition)? What are the possible relationships between two cognitions? Explain how two cognitions can be irrelevant to each other, consistent with each other, or inconsistent with each other. When are two cognitions said to be in a dissonant relationship? 2. What are the properties of dissonance? What sort of state is it? Can dissonance vary in magnitude? What factors influence the degree of dissonance experienced? Explain how the relative proportion of consonant and dissonant elements influences dissonance. Explain how the importance of the elements and the issue influence dissonance. Describe and explain two basic ways of reducing dissonance. 3. Explain how choice (decision making) inevitably arouses dissonance. Is dissonance a predecisional or postdecisional state? What state is a decision maker said to be in before having made the decision? What state is a decision maker said to be in after having made the decision? Identify two factors that influence the amount of postdecisional dissonance. How can dissonance be reduced following a decision? What is postdecisional spreading of alternatives? Has research commonly detected postdecisional spreading of alternatives? Describe how selective information seeking or processing can reduce postdecisional dissonance. How is regret manifest following a decision? Does regret precede or follow dissonance reduction? Explain how regret can lead to a reversal of a decision. Describe the function of follow-up persuasive efforts in the context of postdecisional processes. 4. What is the selective exposure hypothesis? Explain how the hypothesis reflects the main tenets of dissonance theory. Describe the usual research design for studying selective exposure. In such designs, what sort of result represents evidence of selective exposure? Is there evidence of a general preference for supportive information? Is this a strong preference? Explain how the strength of the preference for supportive information is related to the relevance of the issue to one’s core values. What others factors influence information exposure? Explain how perceived information utility and fairness norms can influence information exposure. What is the distinction 156



between selective avoidance effects and selective approach effects? Which appears to be stronger? 5. What is induced compliance? What is counterattitudinal advocacy? Explain the dissonance theory view of induced compliance situations. What is the key influence on the amount of dissonance experienced in such situations? Describe the relationship between incentive and dissonance in such situations. Explain how counterattitudinal advocacy interventions can be a way of changing attitudes. Explain, from a dissonance perspective, the effects of low-price offers for consumer goods. From the marketer’s point of view, what is the ideal amount of incentive to offer? Explain, from a dissonance perspective, the operation of promotions that invite consumers to send in essays explaining why they like the product (or to send in advertisements, etc.), in return for being entered in a prize drawing. Identify two limiting conditions on the occurrence of the predicted dissonance effects in induced compliance situations. How is freedom of choice such a condition? How is the lack of an apparent alternative cause (for the feelings of dissonance) such a condition? 6. What is hypocrisy induction? Explain how hypocrisy induction can lead to behavioral change. Identify a common persuasive situation in which hypocrisy induction might be useful to a persuader. What are the two key elements of successful hypocrisy inductions? Describe how and why hypocrisy induction efforts might backfire. Identify a limiting condition on the success of using hypocrisy induction to change behavior. 7. Explain the central role of the self in dissonance processes. What does self-affirmation theory identify as the motivation behind dissonance phenomena? Explain how dissonance and guilt might be related.



Notes 1. Despite their age and relatively narrowed focus, both balance theory and congruity theory continue to find useful application (see, e.g., Basil & Herr, 2006; E. Walther & Weil, 2012; J. B. Walther, Liang, Ganster, Wohn, & Emington, 2012; Woodside, 2004; Woodside & Chebat, 2001). More generally, cognitive consistency remains an enduring subject of research attention (e.g., Gawronski & Strack, 2012). 2. This “follows from” is, obviously, a matter of psychological implication, not logical implication. What matters is whether I think that 157



one belief follows from another—not whether it logically does so follow. 3. Among the lines of research not discussed here, work on “selfprophesy” effects (also called the “question-behavior” effect and the “mere measurement” effect) is worth mention. The effect of interest is that when people are asked to predict what they will do, such self-predictions can make performance of the predicted behavior more likely (for illustrations and discussion, see Conner, Godin, Norman, & Sheeran, 2011; Godin et al., 2010; J. K. Smith, Gerber, & Orlich, 2003; Spangenberg & Greenwald, 1999; Spangenberg, Sprott, Grohmann, & Smith, 2003; Sprott et al., 2006). One might imagine any number of possible explanations for such effects (e.g., Gollwitzer & Oettingen, 2008), among them a dissonancebased account (Spangenberg, Sprott, Grohmann, & Smith, 2003). 4. The biased processing that is characteristic of postdecisional dissonance reduction mechanisms (to be discussed shortly) can also be apparent at predecisional phases (for a review, see Brownstein, 2003). 5. So (the argument runs) although it may appear that people became more positive about an option after choosing it, they might in reality have had that more positive evaluation even before choosing that option—a more positive evaluation that went undetected because the initial evaluation assessment was imperfect. Thus the postdecisional evaluation may appear to have changed (appear to have become more positive) even though there has been no underlying change in the actual evaluations. M. K. Chen and Risen’s (2010) argument is more complex and nuanced than this (and in particular emphasizes the importance of choosing an appropriate comparison condition), but this will serve to convey the flavor of the argument. 6. M. K. Chen and Risen (2010) offer a mathematical proof showing how spreading can occur absent genuine change. That proof may not be entirely secure (Alós-Ferrer & Shi, 2012), but it is not clear that the proof is essential to their arguments. 7. Note that persons engaged in postdecisional dissonance reduction might exhibit a preference for supportive information, whereas those engaged in postdecisional regret (discussed shortly) might have different preferences (e.g., R. L. Miller, 1977). 8. The discussion in this section concerns information preferences generally. As mentioned earlier, selective information exposure might be a 158



way of pursuing specifically postdecisional dissonance reduction. 9. In Hart et al.’s (2009) review, the mean effect in a random-effects metaanalysis, across 300 cases, was a standardized mean difference (d) of .38, which corresponds to a correlation (r) of .19. 10. Some caution is appropriate in interpreting Hart et al.’s (2009) proffered conclusions, as these are marred by two procedural choices. One choice was to place more weight on results from fixed-effect analyses than on results from random-effects analyses; the random-effects results should be preferred, because only those results provide a basis for generalizing beyond the cases in hand (see, e.g., Borenstein, Hedges, Higgins, & Rothstein, 2010). The second was to use one-tailed tests for selected hypotheses when two-tailed tests should be preferred; as several commentators have suggested, no researcher actually endorses the beliefs logically implied by one-tailed tests (e.g., a belief that finding an extremely large effect in the direction opposite from that predicted is the same as finding no effect) and the use of one-tailed tests encourages “significance chasing” (for some discussion, see Abelson, 1995, pp. 57– 59; Burke, 1953; Dwan, Gamble, Williamson, Kirkham, & the Reporting Bias Group, 2013; Kaiser, 1960). As an illustration of the consequences: Hart et al. said that “past reviews concluded that attitudinal confidence and congeniality [that is, selective exposure to attitude-congenial information] are unrelated … but our results suggested that congeniality is weaker at high (vs. low or moderate) levels of confidence” (p. 581). But the results of the random-effects analyses showed no significant differences in selective exposure as a function of variations in attitudinal confidence (see Tables 3 and 4). 11. Indeed, in Hart et al.’s (2009) review, utility emerged as an especially powerful moderator of selective exposure effects—to all appearances, the only one capable of perhaps producing dependable preferences for nonsupportive information (this, even in the random-effects analyses). 12. This description adopts the conventional language that characterizes choice as a necessary condition for the induced compliance effects predicted by dissonance theory. But any claims about necessary conditions for dissonance-predicted effects have to be offered rather tentatively, if only because the research evidence (concerning not only choice but other putatively necessary conditions as well) has commonly not been analyzed in ways entirely conducive to supporting such claims. There is a difference 159



between saying (for example), “Choice is necessary for the appearance of dissonance-predicted effects” and “Dissonance-predicted effects are larger under conditions of choice than under conditions without choice.” The former depicts choice as a necessary condition, the latter as a moderating factor; the former thus predicts null (zero) effects in no-choice conditions, the latter only that the size of the effect will be smaller in no-choice conditions. Any hypothesis of zero effect, however, is almost certainly literally false; a more appropriate hypothesis would presumably be that the effects would be trivially small (or, perhaps, opposite in direction). There has not been discussion of what “trivially small” might be in this context, however. The evidence that is usually advanced to support necessary condition claims about choice commonly takes the form of (a) a finding that a dissonance-predicted effect is statistically significantly different from zero under choice conditions but not under no-choice conditions or (b) a finding that a dissonance-predicted effect is significantly larger under choice than under no-choice conditions. But neither of these is good evidence for the necessary condition claim (and only the latter is good evidence for a moderating factor claim). (For a first pass at a better approach, see Preiss & Allen, 1998.) The evidentiary situation is more complicated, but no more satisfactory, in the case of the claim that a particular combination of conditions is necessary. In short, although it has become customary to characterize research in this area as having identified various necessary conditions for the appearance of dissonance-predicted effects, these characterizations should be seen as deserving further attention (attention especially focused on matters of effect size and statistical power). 13. Conversely, a persuader seeking long-term commitment to an existing behavior—and so who wants to discourage behavioral change—might offer just-barely-insufficient incentive for change, thereby arousing dissonance that is resolved in favor of the existing behavior. Consider Amazon’s “Pay to Quit” program: Once a year, an employee who works in an Amazon fulfillment center is offered money to quit. In the first year the offer is for $2000, increasing $1000 each year to a maximum of $5000. The headline on the offer is “Please Don’t Take This Offer”—because Amazon hopes the employees will stay. If the financial incentive is not quite sufficient to induce quitting, then employees will experience dissonance (because of forgoing the money)—and to resolve that dissonance they will strengthen their existing positive attitudes (about being Amazon employees). But those stronger positive attitudes created each year mean that in subsequent years, the offer has to keep increasing 160



(in order to keep inducing significant dissonance). (This program is described in Amazon’s 2013 report to shareholders, widely available online including here: http://www.sec.gov/Archives/edgar/data/1018724/000119312514137753/d702518dex9 Thanks to Steve Booth-Butterfield for spotting this ploy.) 14. Notice the contrast with induced compliance situations, in which counterattitudinal behavior is induced by some incentive and (under appropriate conditions) leads to dissonance and subsequent attitudinal realignment. In hypocrisy induction circumstances, counterattitudinal behavior need not be induced because it already has occurred. But in the colloquial sense of hypocrisy, both counterattitudinal advocacy situations and hypocrisy induction situations represent cases of hypocrisy (in which a person says one thing but thinks or does another). 15. As a concrete illustration, consider recycling: Several studies have found that a key common barrier to recycling is a perceived lack of information about how to perform the behavior, a lack of appropriate facilities, the perception that the behavior is difficult to do, and so forth (see, e.g., M. F. Chen & Tung, 2010; De Young, 1989, 1990; Ojala, 2008). (Correlatively, several studies have found perceived behavioral control to be a strong predictor of recycling intentions—often the predictor with the largest zero-order correlation with intention. See, e.g., Mannetti, Pierro, & Livi, 2004; Nigbur, Lyons, & Uzzell, 2010; K. M. White, Smith, Terry, Greenslade, & McKimmie, 2009.) People holding such beliefs, if made to feel hypocritical about not recycling, might well resolve their dissonance by concluding that recycling is not all that valuable. 16. Concerning counterattitudinal advocacy specifically, consider that (a) one common source of guilt feelings is having told a lie (e.g., Keltner & Buswell, 1996, Study 1); (b) another common source of guilt is having inflicted harm on others (e.g., Baumeister, Reis, & Delespaul, 1995, Study 2); and (c) presumably doing both these things would (ceteris paribus) arouse greater guilt than doing either alone. From this perspective, it should not be surprising that (a) counterattitudinal advocacy, even without aversive consequences, has the capacity to arouse dissonance (e.g., Harmon-Jones, Brehm, Greenberg, Simon, & Nelson, 1996; for a review, see Harmon-Jones, 1999); (b) aversive consequences, even from proattitudinal advocacy, have the capacity to arouse dissonance (e.g., Scher & Cooper, 1989); and (c) the combination of counterattitudinal advocacy and aversive consequences arouses greater dissonance than does 161



either element alone (e.g., R. W. Johnson, Kelly, & LeBlanc, 1995).



162



Chapter 6 Reasoned Action Theory The Reasoned Action Theory Model Intention The Determinants of Intention The Distinctiveness of Perceived Behavioral Control The Predictability of Intention Using the RAT Model Influencing Intentions Influencing Attitude Toward the Behavior Influencing the Injunctive Norm Influencing the Descriptive Norm Influencing Perceived Behavioral Control Altering the Weights Intentions and Behaviors Factors Influencing the Intention-Behavior Relationship The Sufficiency of Intention Adapting Persuasive Messages to Recipients Based on Reasoned Action Theory Commentary Additional Possible Predictors Revision of the Attitudinal and Normative Components The Nature of the Perceived Control Component Conclusion For Review Notes



The behaviors of central interest to persuaders are voluntary actions, ones under the actor’s volitional control. The most immediate determinant of such an action is presumably the actor’s behavioral intention—what the person intends to do. Influencing behavior, then, is to be accomplished through influencing persons’ intentions. For example, getting voters to vote for a given political candidate will involve (at a minimum) getting the voters to intend to vote for the candidate. The question that naturally arises is, “What determines intentions?” This chapter discusses reasoned action theory (RAT), a framework that provides a broad, general account of the determinants of intention—thereby identifying underlying targets for persuasive messages.1



163



The Reasoned Action Theory Model Reasoned action theory (RAT) is a general model of the determinants of volitional behavior developed by Martin Fishbein and Icek Ajzen (Fishbein & Ajzen, 2010). In what follows, the RAT model is described and the current state of research on the theory is reviewed. Subsequent sections describe the theory’s implications for influencing intentions, discuss the relationship of intentions and behaviors, and offer some commentary on the model.



Intention RAT is focused on understanding behavioral intentions. A behavioral intention represents a person’s readiness to perform a specified action.2 In assessing behavioral intention, a questionnaire item such as shown in Figure 6.1 is commonly employed. Figure 6.1 Assessing behavioral intention (BI).



The Determinants of Intention RAT proposes that one’s intention to perform or not perform a given behavior is a function of four factors: one’s attitude toward the behavior in question, one’s injunctive norm, one’s descriptive norm, and perceived behavioral control.



Attitude Toward the Behavior The attitude toward the behavior (abbreviated AB) is the person’s general evaluation of the behavior. The expectation, of course, is that as the attitude toward the behavior becomes more positive, the intention will become more positive.3 To measure the attitude toward the behavior, several evaluative semantic differential scales can be used, as shown in Figure 6.2. Figure 6.2 Assessing attitude toward a behavior (AB)



164



Injunctive Norm The injunctive norm (abbreviated IN) is the person’s general perception of whether “important others” desire the performance or nonperformance of the behavior. As the injunctive norm becomes more positive, the intention is expected to become more positive. To obtain an index of the injunctive norm, an item such as that in Figure 6.3 is commonly employed. Figure 6.3 Assessing the injunctive norm (IN).



Descriptive Norm The descriptive norm (abbreviated DN) is the person’s perception of whether other people perform the behavior. The idea is that as people come to think a given behavior is more widely performed by others, then they themselves may be more likely to intend to perform the action. Thus as the descriptive norm becomes more positive, intentions are expected to become more positive.4 The descriptive norm can be assessed in various ways, as illustrated in Figure 6.4. Such questions can be phrased in different ways and can ask about people in general or about a specific group. That is, it is possible to assess descriptive-normative perceptions for different comparison groups. For example, a college student might give different answers to questions such as “How many people your age exercise regularly?” “How many college students exercise regularly?” “How many students at this university exercise regularly?” “How many of your friends exercise regularly?” and so forth. Formulating useful and informative descriptivenorm questions can be a considerable challenge, but the general idea is to assess the respondent’s perception of what other people do. Figure 6.4 Assessing the descriptive norm (DN).



165



Perceived Behavioral Control Perceived behavioral control (abbreviated PBC) is the person’s perception of the ease or difficulty of performing the behavior. PBC is similar to the concept of self-efficacy, which refers to a person’s perceived ability to perform or control a behavior (see Bandura, 1997). The expectation is that as PBC becomes more negative, intentions will correspondingly become more negative.5 The plausibility of this idea can perhaps be seen by considering that sometimes the obstacle to behavioral performance appears to reside not in negative attitudes or norms but rather in a perceived lack of ability to perform the action. For example, a person might have a positive attitude toward exercising regularly (“I think exercising regularly would be a good thing”), a positive injunctive norm (“Most people who are important to me think I should exercise regularly”), and a positive descriptive norm (“Most people like me exercise regularly”), but believe himself incapable of engaging in the action (“I can’t do it, I don’t have the time”—negative perceived behavioral control); as a result he does not even form the intention to exercise regularly. As another illustration, in a comparison of householders who recycled and those who did not, De Young (1989) found that recyclers and non-recyclers had similar positive attitudes about recycling; however, nonrecyclers perceived recycling as much more difficult to do than did recyclers and indicated uncertainty about exactly how to perform the behavior. That is, the barrier to recycling appeared to be a matter of perceived inability to perform the action, not a negative attitude toward the behavior. Perceived behavioral control can be assessed in various ways, but questionnaire items have often taken forms such as those in Figure 6.5. Figure 6.5 Assessing perceived behavioral control (PBC).



Weighting the Determinants 166



These four factors will not always contribute equally to the formation of intentions. In some circumstances, a person’s intentions may be determined largely by the attitude toward the behavior, and normative considerations may play little or no role; for other circumstances, a person might be strongly influenced by descriptive and injunctive normative considerations while the person’s own attitude is put aside. That is, the various influences on intention may carry varying weights in influencing intention. The RAT model expresses this algebraically, as follows: BI=AB(w1)+IN(w2)+DN(w3)+PBC(w4) Here, BI refers to behavioral intention; AB represents the attitude toward the behavior; IN represents the injunctive norm; DN represents the descriptive norm; PBC represents perceived behavioral control; and w1, w2, w3, and w4 represent the weights for each factor. One’s behavioral intentions are thus a joint function of attitude, injunctive norms, descriptive norms, and perceived behavioral control, each appropriately weighted. The relative weights of the components are determined empirically. These weights are not readily assessable for any single person. That is, there is not any satisfactory way to measure the relative weights of the components for an individual’s intention to exercise regularly. One can, however, assess the relative weights of the components (for a given behavior) across a group of respondents. For example, for a group of (say) first-year college students, one can estimate the relative influence of the various components on exercise intentions. This information is obtained through examination of the weights (the betaweights, the standardized partial regression coefficients) from a multiple regression analysis.6 In such an analysis, the four variables are used simultaneously to predict intention; the relative size of the correlation of each component with intention (in conjunction with other information, specifically, the correlations between the components) yields an indication of the relative weight of each component. (For instance, if the attitudinal component is strongly correlated with intention, and the normative components are not, the attitudinal component will receive a larger weight than either of the other two—reflecting its greater influence on intention.)



The Distinctiveness of Perceived Behavioral Control 167



There is some reason to think that PBC is not the same sort of influence on intention that AB, IN, and DN are. It makes sense that everything else being equal, a more positive AB, IN, or DN should be associated with more positive intentions. But it does not make sense that everything else being equal, greater perceived control should be associated with more positive intentions. There are many actions that I perceive to be entirely under my control—for instance, setting fire to my office—that I have no intention of performing. Just because I think I have the capability to perform an action surely does not mean that I am more likely to intend to do so. One possibility is that rather than being a straightforward influence on intention (in the ways that AB, IN, and DN are), PBC might instead be a necessary (but not sufficient) condition for the formation of intention. That is, if I do not think I have the ability to perform the behavior, then of course I will not intend to perform the action, but if I do think I have the ability to perform the behavior, then I might or might not intend to perform it (depending on my attitude and norms). This reasoning suggests an image in which AB, IN, and DN influence intention when PBC is relatively high, but when PBC is relatively low, then AB, IN, and DN will be less strongly related to intention. For example, if I think I am capable of performing the behavior of mountain climbing, then my attitude, injunctive norm, and descriptive norm can influence my intentions (if I like mountain climbing, then I’ll intend to do it; if I don’t like it, then I won’t intend to do it); but if I think the behavior is not under my control (there are no mountains where I live, it’s hard to travel to the mountains, and so forth), then my attitude, injunctive norm, and descriptive norm are irrelevant (I don’t think I can go mountain climbing, so—no matter what my attitude and norms are—I don’t intend to). That is, PBC might be thought of not as a variable that straightforwardly influences intention in the way that AB, IN, and DN do but rather as a variable that moderates the influence of AB, IN, and DN on intention; PBC might be said to enable AB, IN, and DN, in the sense that those variables will influence intention only when PBC is sufficiently high.



The Predictability of Intention Using the RAT Model 168



Various combinations of the four predictors have been explored empirically in hundreds of research studies. Behavioral intentions have proved to be rather predictable using the RAT model, across a variety of behaviors, including exercise (Brickell, Chatzisarantis, & Pretty, 2006; Everson, Daley, & Ussher, 2007; Paek, Oh, & Hove, 2012; for a review, see Hausenblas, Carron, & Mack, 1997), conservation (recycling, water conservation, and the like; Kaiser, Hübner, & Bogner, 2005; Lam, 2006; Nigbur, Lyons, & Uzzell, 2010), health screening (Mason & White, 2008; Michie, Dormandy, French, & Marteau, 2004; Sieverding, Matterne, & Ciccarello, 2010), bicycle helmet use (Lajunen & Rasanen, 2004), voting (Fishbein & Ajzen, 1981), vaccination (Dillard, 2011; Gerend & Shepherd, 2012), smoking (Hassandra et al., 2011), consumer purchases (Brinberg & Durand, 1983; Smith et al., 2008), skin cancer prevention (Branstrom, Ullen, & Brandberg, 2004; K. M. White et al., 2008), and many others. The multiple correlations (obtained using RAT model variables to predict intention) in these applications are commonly in the range of .50 to .90, with an average multiple correlation of between .65 and .70. (For some review discussions, see Albarracín, Johnson, Fishbein, & Muellerleile, 2001; Armitage & Conner, 2001; Conner & Sparks, 2005; Cooke & French, 2008; Hagger & Chatzisarantis, 2009; Hale, Householder, & Greene, 2002; McEachan, Conner, Taylor, & Lawton, 2011; Sutton, 2004; Trafimow, Sheeran, Conner, & Finlay, 2002.) This research has progressed in waves. Much early work examined only two predictors of intention: attitude and injunctive norms (Fishbein & Ajzen, 1975; for an illustrative review, see Sheppard, Hartwick, & Warshaw, 1988). A second wave of research added perceived behavioral control as a predictor (beginning with Ajzen, 1991). More recently, descriptive norms have been added as a general predictor (Fishbein & Ajzen, 2010). The rationale for this succession of additional predictors has been that each new variable has shown its value in contributing to the prediction of intentions. That is, predictions based on AB, IN, and PBC are commonly better than those based on AB and IN alone (for some relevant reviews and discussions, see Conner & Armitage, 1998; Conner & Sparks, 1996; Godin & Kok, 1996; Notani, 1998; Sutton, 1998). Similarly, adding DN has often been found to improve the prediction of intention beyond that based on AB, IN, and PBC (for reviews, see Manning, 2009; Rivis & Sheeran, 2003).7 169



Thus the four predictors described here (AB, IN, DN, and PBC) appear to be predictors of sufficiently common utility to warrant their inclusion in a single general model. That does not mean that in any given application, all four will play a significant role in influencing intention, but it does suggest that the four-predictor model is likely to be a useful starting point in trying to unravel influences on intention.



Influencing Intentions The RAT model identifies five possible avenues for changing a person’s intention to perform a given behavior: by influencing one of the four determinants of intention (AB, IN, DN, PBC)—assuming that the determinant is significantly weighted—or by changing the relative weighting of the components. (It is presumably apparent why inducing change by altering one of the four components requires that the component be significantly weighted. RAT underscores the futility of attempts to change, say, the injunctive norm in circumstances in which only the attitudinal component is significantly related to intention.) In what follows, each of those five avenues is discussed in more detail. (For some general discussion and reviews concerning RAT-based interventions, see Cappella, 2006; Fishbein & Yzer, 2003; Hackman & Knowlden, 2014; Hardeman et al., 2002; Sutton, 2002; Yzer, 2012a, 2013. For some illustrative applications, see Armitage & Talibudeen, 2010; Dillard, 2011; Elliot & Armitage, 2009; French & Cooke, 2012; Giles et al., 2014; Jemmott, 2012; Kothe, Mullan, & Amaratunga, 2011; Paek, Oh, & Hove, 2012; Stead, Tagg, MacKintosh, & Eadie, 2005.)



Influencing Attitude Toward the Behavior Presumably, a person’s attitude toward a behavior might be influenced by any number of different attitude-change mechanisms. But the RAT model provides an account of the determinants of AB that can be useful in identifying some specific ways in which it can be influenced.



The Determinants of AB An individual’s attitude toward the behavior is taken to be a function of his or her salient beliefs about the act (which commonly are beliefs concerning outcomes of the behavior). More specifically, he proposal is 170



that the evaluation of each belief (ei) and the strength with which each belief is held (bi) jointly influence one’s attitude toward the behavior, as represented in the following equation: AB=∑biei This is the same summative conception of attitude discussed in Chapter 4 (concerning belief-based attitude models). The assessment procedures are identical, a set of modally salient beliefs is usually identified, and the same sorts of belief strength scales (e.g., probable–improbable, true–false) and belief evaluation scales (e.g., good–bad, desirable–undesirable) are employed. For instance, the items in Figure 6.6 assess the respondent’s belief strength (bi) concerning a particular outcome of regular exercise. The evaluation of that outcome (ei) can be assessed with items such as those in Figure 6.7. Figure 6.6 Assessing belief strength (bi).



Figure 6.7 Assessing belief evaluation (ei).



RAT’s claims about the determinants of one’s attitude toward the act have received rather good empirical support, with correlations between ∑biei and AB commonly averaging more than .50 (for review discussions, see Albarracín, Johnson, Fishbein, & Muellerleile, 2001; Armitage & Conner, 2001; Conner & Sparks, 1996; Eagly & Chaiken, 1993, p. 176).8



Changing AB RAT thus identifies a number of possible means of changing the attitude toward the behavior (AB). Consider the case of attempting to induce an unfavorable attitude toward a given behavior such as smoking. Three broad strategies are possible. First, the evaluation of an existing salient belief might be changed. This might involve increasing the unfavorability of an existing negative belief (“You probably already know that smoking can lead to blood circulation problems—but you may not realize just how serious such problems are. Impaired circulation is very undesirable, even 171



dangerous, …”) or decreasing the favorability of an existing positive belief (“Maybe smoking does give you something to do with your hands, but that’s a pretty trivial thing”). Second, the strength (likelihood) of an existing salient belief might be changed. This might involve attempting to increase the belief strength of an existing negative belief (“You probably already realize that smoking can lead to health problems. But maybe you don’t realize just how likely it is to do so. You really are at risk …”) or to decrease the belief strength associated with an existing positive belief (“Actually, smoking won’t help you keep your weight down”). Third, the set of salient beliefs might be changed. This can be accomplished in two ways. One is to add a new salient belief (of the appropriate valence) belief about the act (“Maybe you didn’t realize that smoking leaves a bad odor on your clothes”). The other is to change the relative saliency of current beliefs such that a different set of beliefs is salient (“Have you forgotten just how expensive cigarettes are nowadays?”). Obviously, these are not mutually exclusive possibilities; a persuader might implement all these strategies. However, as discussed more extensively in Chapter 4 (concerning beliefbased attitude models), there are hidden complexities here. For example, the property of belief strength may be more categorical (“I think the behavior has the attribute”) than continuous (“I think the probability is such-and-such that the behavior has the attribute”). Hence if the persuadee already has the desired categorical judgment associating the behavior with an attribute, trying to influence the degree of association may not be useful.



Influencing the Injunctive Norm As with AB, RAT offers an account of the determinants of the injunctive norm that can be useful in identifying ways in which IN can be influenced.



The Determinants of IN An individual’s injunctive norm is taken to be based on two elements. The first is the person’s judgment of the normative expectations of specific important others (what I think my parents want me to do, what I think my best friend wants me to do, and so on). The second is the individual’s motivation to comply with each of those referents (how much I want to do what my parents think I should, etc.). Specifically, a person’s injunctive 172



norm is suggested to be a joint function of the normative beliefs that one ascribes to particular salient others (ni) and one’s motivation to comply with those others (mi). This is expressed algebraically as follows: IN=∑nimi An individual’s normative beliefs (ni) are commonly obtained through a set of items in which the normative expectation of each referent is assessed. Figure 6.8A provides one example. The motivation to comply with each referent (mi) is typically assessed through a question such as the one in Figure 6.8B. If I believe that my parents, my best friend, my physician, and others who are important to me all think that I should exercise regularly, and I am motivated to comply with each referent’s expectations, then I will surely have a positive injunctive norm regarding regular exercise. Figure 6.8 Assessing normative expectations (ni) and motivation to comply (mi).



RAT’s claims about the determinants of the injunctive norm have generally received good empirical support, with correlations between ∑nimi and IN often .50 and greater (for review discussions, see Albarracín et al., 2001; Armitage & Conner, 2001; Conner & Sparks, 1996; Eagly & Chaiken, 1993, p. 176; McEachan et al., 2011).9 Even so, there is reason for concern about the RAT’s analysis of the IN, and specifically about the motivation to comply (mi) element. These worries are of two sorts. The first is uncertainty about the most appropriate way to phrase motivation-to-comply questionnaire items. These items can be worded in a way that focuses on the specific behavior of interest (“When it comes to exercising regularly, how much do you want to do what your best friend thinks you should do?”), in a general way (“In general, how much do you want to do what your best friend thinks you should do?”), or at some intermediate level of specificity (“When it comes to health, how much …”). It is not clear how one might most appropriately choose among these.10 173



The second is some troubling empirical results concerning the role of the motivation-to-comply element. Specifically, ∑ni has often been found to be at least as good, and sometimes better, a predictor of IN than ∑nimi; that is, deleting the motivation-to-comply element does not reduce, and sometimes even improves, the prediction of IN (for some examples, see Budd, North, & Spencer, 1984; Doll & Orth, 1993; Kantola, Syme, & Campbell, 1982; Montaño, Thompson, Taylor, & Mahloch, 1997; Sayeed, Fishbein, Hornik, Cappella, & Ahern, 2005). It may be that, when a normative referent is salient, motivation to comply with that referent is already likely to be reasonably high, and hence a measure of motivationto-comply does not add useful information (Fishbein & Ajzen, 2010, p. 143).



Changing IN From the perspective of the RAT, one would influence the injunctive norm by influencing ni and mi, in ways precisely parallel to the ways in which AB is influenced through bi and ei. For example, one might attempt to reconfigure the set of salient referents by adding a new referent or by increasing the relative salience of an existing potential referent: “Have you considered what your mother would think about your doing this?” Or one might attempt to change the normative belief attributed to a current referent: “Oh, no, you’re wrong—I talked to George, and he thinks you should go ahead and do this.” Or one might try to change the motivation to comply with a current referent: “You really shouldn’t worry about what he thinks—he has no sense when it comes to things like this.” But the previously mentioned uncertainties concerning the nature and determinants of the injunctive norm make for some corresponding difficulties here. For example, if one attempts to influence a receiver’s motivation to comply with a particular referent, it is not clear whether one should attempt to change the receiver’s motivation to comply (with that referent) generally, or concerning the relevant broad behavioral domain, or regarding the specific behavior at hand. Moreover, altering motivation to comply with a given referent may not affect the receiver’s IN; given the research evidence that the IN has often been better predicted by ∑ni than by ∑nimi, perhaps changing the motivation to comply component may not affect the IN in the expected ways. So influencing injunctive norms will often be challenging for persuaders. 174



In some circumstances, some messages concerning others’ normative beliefs are likely to simply be implausible (e.g., “your friends would really be opposed to you doing this”). Yet it is plainly possible to devise successful interventions based on something like alterations of the injunctive norm. For example, Kelly et al. (1992) identified “trendsetters” who subsequently communicated HIV risk reduction information to gay men in their communities, producing substantial and sustained risk reduction behavior; one way of understanding such effects is to see them as reflecting changes in the receivers’ injunctive norms (see, relatedly, Vet, de Wit, & Das, 2011). As another example, Prince and Carey’s (2010) alcohol abuse intervention was able to affect college students’ injunctivenormative perceptions of whether the typical student approved of excessive drinking, although not parallel perceptions concerning close friends’ approval (see also Armitage & Talibudeen, 2010; Reid & Aiken, 2013). But rather than trying to change the persuadee’s injunctive norms by addressing messages to the persuadee, persuaders might sometimes consider targeting messages at the relevant important others—because if the views of those referents change, then the persuadee’s normative beliefs (the beliefs the persuadee attributes to those referents) may also change. For example, to encourage potential military recruits to enlist, recruiters might try to persuade parents to favor their child’s enlistment, thereby laying the groundwork for the potential recruit to develop the desired injunctive-normative beliefs.



Influencing the Descriptive Norm The Determinants of DN RAT does not yet provide an elaborated account of the determinants of the descriptive norm (DN). One possibility might be to conceive of the DN as arising from perceptions that parallel those determining IN (see Fishbein & Ajzen, 2010, pp. 146–148). That is, a given respondent (or set of respondents) might have a set of salient descriptive-norm referents (parallel to the salient injunctive-norm referents)—a set of individuals or groups whose behavior might be seen as a source of guidance. And the descriptive-normative beliefs about such referents might be weighted in some way (giving more weight to some referents than to others), thus yielding the person’s overall perception of the DN. But these ideas have 175



not received sustained empirical attention.



Changing DN Even without a fully explicit account of the determinants of the descriptive norm, however, it is plain that the DN might most straightforwardly be influenced by messages that convey DN information. Such messages might influence intentions either by altering the DN (e.g., in cases where people don’t know, or misperceive, the DN) or by enhancing the salience of the DN. In fact, providing descriptive-norm information has been the primary basis of quite a number of successful persuasive interventions, on such diverse subjects as conservation behavior (Goldstein, Cialdini, & Griskevicius, 2008), food choice (Burger et al., 2010), tax compliance (Wenzel, 2005), physical activity (Burger & Shelton, 2011; Slaunwhite, Smith, Fleming, & Fabrigar, 2009), and skin cancer prevention (Mahler, Kulik, Butler, Gerrard, & Gibbons, 2008). As just one illustration: People can be influenced to vote by learning that some of their Facebook friends have voted (Bond et al., 2012; see, similarly, Glynn, Huge, & Lunney, 2009). One extensively studied arena for descriptive-norm interventions has been college student alcohol consumption. Such undertakings have been motivated by the frequent observation that students commonly overestimate the frequency or amount of alcohol consumption by others (for a review discussion, see Berkowitz, 2005). On the supposition that such inaccurate descriptive-norm beliefs might lead to alcohol abuse, campaigns conveying accurate DN information have had a natural appeal. These interventions have had mixed success (for examples and discussion, see Clapp, Lange, Russell, Shillington, & Voas, 2003; Lewis & Neighbors, 2006; Mattern & Neighbors, 2004; Wechsler et al., 2003). The challenges in creating effective descriptive-norm–based interventions should not be underestimated. DN-based persuasive efforts can go off the rails in a variety of ways: The DN information in the campaign messages might not be believable (e.g., Polonec, Major, & Atwood, 2006), the messages might not provide DN information about the most appropriate referent comparison groups (e.g., Burger, LaSalvia, Hendricks, Mehdipour, & Neudeck, 2011; Larimer et al., 2009), the DN might not be strongly related to intentions or behavior (e.g., Cameron & Campo, 2006), or DN information might backfire (e.g., Campo & Cameron, 2006; see, 176



relatedly, Cialdini et al., 2006). One hopes that the accumulation of research evidence about DN-based interventions will eventuate in guidelines about how to maximize the effectiveness of such interventions (see DeJong & Smith, 2013).



Influencing Perceived Behavior Control The Determinants of PBC Perceived behavioral control is taken to be a function of the person’s beliefs about the resources and obstacles relevant to performance of the behavior. More specifically, the determinants of PBC are taken to reflect jointly the person’s perception of (a) the likelihood or frequency that a given control factor will occur and (b) the power of the control factor to inhibit or facilitate the behavior. PBC is expressed algebraically as follows: PBC=∑cipi where ci refers to the individual control belief (the perceived likelihood or frequency that the control factor will occur) and pi refers to the perceived facilitating or inhibiting power of the individual control factor. Procedures for assessing these variables are not well established, but an individual’s control beliefs (ci) might be assessed using items such as in Figure 6.9A. The perceived power of each control factor (pi) can be assessed through a question like the one in Figure 6.9B. If, for example, I think that bad weather occurs frequently where I live, that I don’t have ready access to exercise facilities, and that I don’t have much spare time, and I think that each of these conditions makes it very difficult to exercise regularly, then I will likely perceive that I have relatively little control over whether I exercise regularly.11 Figure 6.9 Assessing individual control belief (ci) and the power of each control factor (pi).



Relatively little research attention has been given to RAT’s claims about 177



the determinants of perceived behavioral control. Many RAT studies have not collected data about ci, pi, and PBC (and of those that have, some do not report the relevant correlation between ∑cipi and direct measures of PBC). The few reported results are not especially encouraging, as the correlations commonly range from roughly .10 to .35 (see, e.g., Cheung, Chan, & Wong, 1999; Elliott, Armitage, & Baughan, 2005; Parker, Manstead, & Stradling, 1995; Povey, Conner, Sparks, James, & Shepherd, 2000; Valois, Desharnais, Godin, Perron, & LeComte, 1993).12 However, stronger relationships have been reported between direct assessments of PBC and other belief-based measures, including measures based on questions about only likelihood of occurrence (i.e., ∑ci), questions about only powerfulness (∑pi), questions that appear to involve some amalgam of likelihood of occurrence and powerfulness considerations (e.g., “Which of the following reasons would be likely to stop you from exercising regularly?”), and questions about the perceived importance of various barriers. Using measures such as these, correlations with PBC measures of between roughly .25 and .60 have been obtained (Ajzen & Madden, 1986; Courneya, 1995; Elliott et al., 2005; Estabrooks & Carron, 1998; Godin, Gagné, & Sheeran, 2004; Godin, Valois, & Lepage, 1993; P. Norman & Smith, 1995; Sutton, McVey, & Glanz, 1999; Theodorakis, 1994; Trafimow & Duran, 1998).13 Interpretation of these findings is complicated by variation in the direct assessments of perceived behavioral control, in the means of establishing the set of control beliefs, and in the assessments of likelihood (ci) and powerfulness (pi).14 Taken together, however, these findings do suggest that in some fashion perceptions of behavioral control are belief-based, in the sense of being related to persons’ beliefs about resources and obstacles relevant to behavioral performance. RAT may not yet have an adequate account of exactly how beliefs combine to yield perceptions of behavioral control, but it seems plain that assessments of the resources for, and obstacles to, behavior play some role in shaping persons’ perceptions of control.



Changing PBC Influencing perceived behavioral control involves addressing the perceived barriers to and resources for behavioral performance. Unfortunately, the lack of a well-evidenced account of the determinants of PBC means that 178



there is less guidance than one might like concerning specific means of influencing PBC. Even so, there appear to be four broad alternative means by which a persuader might influence PBC. The appropriateness of each mechanism will vary depending on the particular target behavior, and combinations of these approaches may prove more effective than any one individually, but each offers an avenue to influencing perceptions of behavioral control. One means of influencing perceived behavioral control may be for the persuader to directly remove an obstacle to behavioral performance. Some such obstacles are the result of a lack of relevant information, and in such cases persuaders might find success simply by providing the information. For example, parents’ self-efficacy for lowering the temperature setting of a water heater (to prevent tap water scalding of infants) can be enhanced by a simple informational brochure describing how to perform the action (Cardenas & Simons-Morton, 1993). Similarly, better instructions may improve self-efficacy concerning do-it-yourself medical tests (Feufel, Schneider, & Berkel, 2010). Adolescents may not know how to use condoms properly, voters may not know the location of their polling places, and potential first-time home buyers may not understand the process of buying a house; in all these cases, simply providing the relevant information may remove a barrier to behavioral performance. Even when the obstacle is substantive (rather than informational), persuaders may be able to address it. For example, among low-income patients whose initial medical test results indicate a need for a return hospital visit, transportation problems might represent a significant barrier to returning; Marcus et al. (1992) found that providing such patients with free bus passes or parking permits significantly increased the likelihood of a return visit.15 Similarly, racquetball players who didn’t have eye protection equipment were willing to wear it when the recreational facility provided it at the court (Dingus, Hunn, & Wreggit, 1991). Second, a persuader might create the opportunity for successful performance of the behavior in question. The core idea is that rehearsal of a behavior—practice at performing the behavior successfully—will enhance perception of control over the action (the underlying reasoning being something such as “I’ve done it before, so I can do it again”). For instance, several studies have found that self-efficacy for condom use can be enhanced by interventions that include role-playing (or mental rehearsal) of discussions with sexual partners, practice at using condoms 179



correctly, and the like (e.g., Calsyn et al., 2010; Yzer, Fisher, Bakker, Siero, & Misovich, 1998). For other suggestions of the effect of successful performance on self-efficacy, see Duncan, Duncan, Beauchamp, Wells, and Ary (2000), Latimer and Ginis (2005a), Luzzo, Hasper, Albert, Bibby, and Martinelli (1999), and Mishra et al. (1998). Third, a persuader can provide examples of others (models) performing the action successfully; such modeling can enhance self-efficacy (by message recipients reasoning that “if they can do it, I can do it”). For example, compared with a no-treatment control group, preservice teachers who viewed a videotape that described and demonstrated various effective behavior management techniques subsequently reported enhanced selfefficacy for using such techniques (Hagen, Gutkin, Wilson, & Oats, 1998). For other examples of the potential effects of modeling on self-efficacy, see R. B. Anderson (1995, 2000), Gaston, Cramp, and Prapavessis (2012), and Ng, Tam, Yew, and Lam (1999); for some discussion of factors relevant to the choice of models, see Berry and Howe (2005), Corby, Enguidanos, and Kay (1996), and R. B. Anderson and McMillion (1995). Finally, simple encouragement may make a difference. That is, hearing a communicator say (in effect) “you can do it” may enhance a person’s perceived ability to perform an action. For instance, assuring receivers that they can successfully prevent a friend from driving while drunk can enhance receivers’ self-efficacy for that action (compared with a notreatment control condition; R. B. Anderson, 1995).16 Several studies have examined multicomponent self-efficacy interventions (i.e., interventions that combine different potential means of influencing self-efficacy, such as modeling and information; see, e.g., Luszczynska, 2004; Robinson, Turrisi, & Stapleton, 2007), and self-efficacy treatments have sometimes been included as part of a larger intervention package (as when, for instance, participants receive information designed to persuade participants of the importance of the behavior in combination with a selfefficacy treatment; for examples, see Darker, French, Eves, & Sniehotta, 2010; Fisher, Fisher, Misovich, Kimble, & Malloy, 1996; Kellar & Abraham, 2005). Such research designs can provide evidence for the influenceability of self-efficacy but obviously cannot provide information about the relative impact of different specific mechanisms of influence (although evidence is beginning to accumulate concerning that question; e.g., Anderson, 2009; Ashford, Edmunds, & French, 2010; Hyde, Hankins, Deale, & Marteau, 2008; Prestwich et al., 2014) or about the conditions 180



under which a given mechanism is most effective (although here, too, research is developing; e.g., J. K. Fleming & Ginis, 2004; Hoeken & Geurts, 2005; Luszczynska & Tryburcy, 2008; Mellor, Barclay, Bulger, & Kath, 2006).



Altering the Weights The final possible avenue of influence suggested by RAT is changing the relative weights of AB, IN, and DN.17 For instance, suppose a person has a positive attitude toward the act of attending law school but has a negative injunctive norm (believes that important others think that she should not go to law school). If the person places greater emphasis on injunctivenormative than on attitudinal considerations in making this behavioral decision, she would not intend to go to law school. A persuader who wanted to encourage the person’s attending law school might try to emphasize that insofar as a decision such as this is concerned, one’s personal feelings ought to be more important than what others think (“It’s your career choice, your life, not theirs; in situations like this, you need to do what’s right for you,” “You’re the one who has to live with the consequences, not them,” and so on). That is, the persuader might attempt to have the person place more emphasis on attitudinal than injunctivenormative considerations in forming the relevant intention. This strategy can succeed in changing intention only when the relevant components incline the person in opposite directions. For example, if a person has a positive AB, a positive IN, and a positive DN, then it won’t matter how the weights are shifted around among those three elements— the person will still have a positive intention. Intention can be changed by altering the weights of these three components only when one of those three components differs in direction from the other two.18 However, these three components are often positively correlated. More evidence is available concerning the relationship between AB and IN than concerning either the AB-DN or IN-DN relationships, but all three variables are generally reasonably positively correlated with each other, with mean correlations ranging from roughly .35 to .60 (Manning, 2009; Rivis & Sheeran, 2003).19 So, for example, persons with negative injunctive norms are likely to also have relatively unfavorable attitudes toward the behavior and relatively negative descriptive norms; as the 181



attitude toward the behavior becomes more positive, so do injunctive norms and descriptive norms; and so on. As a rule, then, it is unlikely that AB, IN, and DN will not all point in the same direction.20 The implication is that the strategy of influencing their relative weights will not find wide application.



Intentions and Behaviors Reasoned action theory focuses on factors influencing the formation of behavioral intentions, but such a focus promises to illuminate persuasion only to the extent that intentions are related to action. As it happens, there is good evidence that voluntary actions can often be successfully predicted from intentions. Several broad reviews have reported mean intentionbehavior correlations ranging from roughly .40 to .55 (Eckes & Six, 1994; M.-S. Kim & Hunter, 1993b; Sheeran, 2002; Sheppard et al., 1988), and reviews of selected subsets of relevant work have reported similar magnitudes (e.g., Cooke & French, 2008; Godin & Kok, 1996; Hagger, Chatzisarantis, & Biddle, 2002; Hausenblas et al., 1997; Ouellette & Wood, 1998; Schepers & Wetzels, 2007; Schwenk & Moser, 2009; Sheeran & Orbell, 1998).21



Factors Influencing the Intention-Behavior Relationship Given that measures of intention are thus often reasonably strongly related to behavioral assessments, the question that naturally arises is what variables influence the strength of this relationship. A variety of factors have been examined as possible influences on the intention-behavior relationship (for some examples, see Chatzisarantis & Hagger, 2007; Hall, Fong, Epp, & Elias, 2008; Prestwich, Perugini, & Hurling, 2008; for a general discussion, see Cooke & Sheeran, 2004). Three factors are discussed here as illustrative.



Correspondence of Measures First, the degree of correspondence between the measure of intention and the measure of behavior influences the strength of the observed intentionbehavior relationship (see Courneya, 1994; Fishbein & Ajzen, 2010, pp. 44–47). For instance, a questionnaire item asking about my intention to 182



buy diet cola at the grocery tonight may well be strongly related to whether I buy diet cola at the grocery tonight—but it will be less strongly related to whether I buy Diet Coke (specifically) at the grocery tonight or to whether I buy diet cola at the cafeteria tomorrow. That is, as the degree of correspondence between the two measures weakens, the intention becomes a poorer predictor of (less strongly related to) the behavior. This methodological consideration emphasizes how different means of assessing intention and behavior can affect the size of the observed association.



Temporal Stability of Intentions A second influence on the intention-behavior relationship is the temporal stability of intentions. If a person’s intentions fluctuate a good deal through time, then a measure of intention (taken at one particular time) may not necessarily be predictive of subsequent behavior (e.g., Conner, Sheeran, Norman, & Armitage, 2000; Dibonaventura & Chapman, 2005; for reviews, see Conner & Godin, 2007; Cooke & Sheeran, 2004; Rhodes & Dickau, 2013).22 In part, this is a methodological point, in the sense that if the value of a predictor variable is volatile over time, then any single assessment of it is likely to be relatively weakly related to a subsequent assessment of an outcome variable (even if the two properties are actually closely related). That is, even if behavior is entirely determined by whatever the actor’s intention is at the moment of behavioral performance, an earlier assessment of intention will be predictive of that behavioral performance only if the (earlier) assessed intention matches the (later) atthe-moment-of-action intention. Thus if people’s intentions are stable over time, then there is a good chance that their earlier intentions will match their later ones, thus yielding a strong observed relationship between the measure of intention and the measure of behavior. But if people’s intentions are variable over time, then the observed relationship will be weaker—not because intentions do not actually influence actions but because the temporal instability of intention inevitably introduces error. But there is also a substantive point here, because of the possibility that some intentions (for some people or for some types of behaviors) are generally more stable than others (see Sheeran & Abraham, 2003). There is not yet much accumulated research on this matter, but (for example) some evidence suggests that for behaviors deemed relatively important (e.g., ones taken to be closely related to one’s self-image), intentions may be more stable (compared with corresponding intentions for less important 183



behaviors) and hence more closely related to action (see Kendzierski & Whitaker, 1997; Radecki & Jaccard, 1999; Sheeran & Orbell, 2000a). In any case, the general point to notice is that to the degree that persons’ intentions are unstable, to that same degree intentions may not provide a good basis for predicting subsequent action.



Explicit Planning Third, explicit planning about behavioral performance can strengthen the relationship between intentions and actions. In a large number of studies, participants who specified when and where they would perform the action were more likely (than control group participants) to subsequently engage in the behavior. For example, Sheeran and Orbell (2000b) found that participants who specified when, where, and how they would make an appointment for a medical screening test were much more likely to subsequently attend the screening than those in a control condition. Similar effects of explicit-planning interventions have been reported for a great variety of behaviors, including exercise (e.g., Andersson & Moss, 2011), single-occupancy car use (Armitage, Reid, & Spencer, 2011), parentteacher communication (Arriaga & Longoria, 2011), smoking prevention (Conner & Higgins, 2010), contraceptive adherence (Martin, Slade, Sheeran, Wright, & Dibble, 2011), voting (Nickerson & Rogers, 2010), and many others (for some reviews, see Adriaanse, Vinkers, de Ridder, Hox, & De Wit, 2011; Gollwitzer & Sheeran, 2006; Sheeran, Milne, Webb, & Gollwitzer, 2005).23 These effects are notable because one common persuasive challenge is precisely that of encouraging people to translate their existing good intentions into action. For example, people may form an initial intention to exercise or recycle or eat a healthier diet but then fail to follow through. Obviously, encouraging receivers to engage in explicit behavioral planning is a possible mechanism for addressing such challenges. Several different explanations have been considered for these planning effects. One possibility is simply that planning makes intentions more positive, but in fact these effects do not necessarily involve enhancing intentions (see, e.g., Milne, Orbell, & Sheeran, 2002; Sheeran & Orbell, 1999b). Explicit planning appears to be able to influence the likelihood of subsequent behavior without necessarily changing intentions. A second explanation is that planning enhances PBC (self-efficacy). The 184



suggestion is that thinking through concrete action plans may convince people of their ability to successfully perform the behavior. But two considerations incline against this explanation. First, if planning enhances PBC, then (given PBC’s influence on intention) planning should also have the indirect effect of making intentions more positive; but, as just indicated, that effect seems not to occur. Second, several studies have found that planning enhances intention-behavior consistency only when PBC is already relatively high (Koring et al., 2012; Lippke, Wiedemann, Ziegelmann, Reuter, & Schwarzer, 2009; Schwarzer et al., 2010; Wieber, Odenthal, & Gollwitzer, 2010; see, relatedly, Koestner et al., 2006); that is, having high PBC appears to be a necessary condition for explicit-planning interventions to be effective. A third explanation—and it seems the best available—is that planning encourages the development of “implementation intentions,” subsidiary intentions related to the concrete realization (implementation) of a more abstract intention (for a general discussion and review, see Gollwitzer & Sheeran, 2006). A more abstract intention (“I intend to get a flu shot next week”) may be thought of as describing a goal, whereas implementation intentions specify both the concrete behavior to be performed in order to achieve the goal and the context in which that behavior will be enacted (“On Tuesday, on the way to work, I’ll stop at that pharmacy on Main Street to get my flu shot”). Interventions that encourage explicit behavioral planning thus naturally boost the development of implementation intentions. A number of factors can be expected to influence the success of explicitplanning interventions. People must already have the appropriate abstract intentions (e.g., Elliott & Armitage, 2006), the intervention must in fact lead people to plan (Michie, Dormandy, & Marteau, 2004), and, as mentioned earlier, perceived behavioral control must be sufficiently high (e.g., Koring et al., 2012). The implementation intentions explanation additionally suggests that, to be successful, explicit-planning interventions should specifically emphasize the linkage between situational cues (the context of performance) and the concrete action (Chapman, Armitage, & Norman, 2009; Van Osch, Lechner, Reubsaet, & de Vries, 2010; Webb & Sheeran, 2007; cf. Ajzen, Czasch, & Flood, 2009), so that when those contextual cues are encountered, they will naturally prompt behavioral performance. (For examples of discussion of some additional possible factors, see Adriaanse, de Ridder, & de Wit, 2009; Churchill & Jessop, 2011; Hall, Zehr, Ng, & Zanna, 2012; Knäuper, Roseman, Johnson, & 185



Krantz, 2009; Prestwich et al., 2005.)



The Sufficiency of Intention One other general aspect of the intention-behavior relationship worth considering is whether intention is a sufficient basis for the prediction of voluntary action. The RAT model proposes that intention is the only significant influence on (volitional) behavior; any additional factors that might be related to behavior are claimed to have their effect indirectly, via intention (or via the determinants of intention).24 The question at issue is this: Are there factors that have effects on behavior that are not mediated through intention? Alternatively put: Are there additional variables that might improve the prediction of behavior (over and above the predictability afforded by intention)? Among various possibilities, the most prominent and well-studied suggestion focuses on prior performance of the behavior in question. Some studies have found the prediction of behavior to be improved by taking prior behavior into account (e.g., De Wit, Stroebe, De Vroome, Sandfort, & van Griensven, 2000); specifically, persons who had performed the action in the past were more likely to perform it in the future—over and above the effects of intention on future performance. But other studies have failed to find such an effect (e.g., Brinberg & Durand, 1983). A systematic review of this research has indicated that a key differentiating factor is whether the behavior is routinized (Ouellette & Wood, 1998). Specifically, prior behavior makes an independent contribution (to the prediction of behavior) only when the behavior has become habitual and routine (and so, in a sense, automatic rather than fully intentional); where conscious decision making is required, this effect disappears, as the influence of prior behavior seems to largely be mediated through intention or its determinants. And thus, as a number of studies have found, intention-behavior correlations are smaller when habit is relatively strong than when it is relatively weak—an effect observed across such diverse behaviors as cancer screening (P. Norman & Cooper, 2011), bicycle use (de Bruijn & Gardner, 2011; de Bruijn, Kremers, Singh, van den Putte, & van Mechelen, 2009), fruit consumption (de Bruijn, 2010), exercise (de Bruijn & Rhodes, 2010), saturated fat consumption (de Bruijn, Kroeze, Oenema, & Brug, 2008), and binge drinking (P. Norman & Conner, 2006). For persuaders, these findings serve as a reminder of the persuasive difficulties created by entrenched behavioral patterns: Past behavior may 186



exert an influence on conduct that is not mediated by intention, and hence securing changes in intention may not be sufficient to yield changes in well-established behavioral routines. On the other hand, these findings also suggest the durability of persuasive effects that involve establishing such habits. (For examples and discussion concerning establishing or breaking habitual or routinized behavior, see Aarts, Paulussen, & Schaalma, 1997; Adriaanse, Gollwitzer, de Ridder, de Wit, & Kroese, 2011; Allcott & Rogers, 2012; de Vries, Aarts, & Midden, 2011; Judah, Gardner, & Aunger, 2013; Lally & Gardner, 2013.)



Adapting Persuasive Messages to Recipients Based on Reasoned Action Theory One recurring theme in theoretical analyses of persuasion is the idea that to maximize effectiveness, persuasive messages should be adapted (tailored, adjusted) to fit the audience. RAT provides an explicit treatment of two ways in which such adaptation can occur (for discussion and details, see Abraham, 2012; Ajzen, Albarracín, & Hornik, 2007; Ajzen & Manstead, 2007; Fishbein & Yzer, 2003; Sutton, 2002; Yzer, 2012a, 2013). First, messages should be adapted by addressing whichever determinants of intention are the most important influences. Concretely, this means examining the relative weights of the different RAT model elements so as to identify the ones most strongly related to intention. Generally speaking, of the four components, AB is typically the most strongly related to intention (see the reviews of Albarracín, Johnson, Fishbein, & Muellerleile, 2001; Armitage & Conner, 2001; Cooke & French, 2008; Hagger, Chatzisarantis, & Biddle, 2002; Manning, 2009; McEachan, Conner, Taylor, & Lawton, 2011; Rivis & Sheeran, 2003).25 But the relative contribution of the various components to the prediction of intention can vary from behavior to behavior (compare, e.g., Cooke and French’s 2008 findings concerning health screening behaviors and Hagger et al.’s 2002 findings concerning exercise). Indeed, sometimes AB may make a smaller contribution (to the prediction of intention) than do normative elements (e.g., Croy, Gerrans, & Speelman, 2010).26 The implication is that in any given persuasion circumstance, there will be no substitute for direct evidence about the relative importance of the various components.



187



Second, messages should be further adapted by addressing relevant beliefs underlying the component to be changed. For example, if AB is the target component, RAT-based questionnaires can be used to identify differences between those who already intend to perform the persuader’s advocated action (“intenders”) and those who do not (“nonintenders”)—differences in the strength and evaluation of salient beliefs about the behavior (see, e.g., Fishbein et al., 2002; French & Cooke, 2012; Marin, Marin, PerezStable, Sabogal, & Otero-Sabogal, 1990; Rhodes, Blanchard, Courneya, & Plotnikoff, 2009; Silk, Weiner, & Parrott, 2005; J. R. Smith & McSweeney, 2007). Such data can then be used as the basis for constructing persuasive messages—messages focused on changing those specific elements known to distinguish intenders and nonintenders (for examples of RAT-based message design, see Booth-Butterfield & Reger, 2004; Chatzisarantis & Hagger, 2005; Jordan, Piotrowski, Bleakley, & Mallya, 2012; Jung & Heald, 2009; Milton & Mullan, 2012; Stead, Tagg, MacKintosh, & Eadie, 2005).27 As attractive and useful as this general approach is, a word of caution is appropriate: The relative sizes of the weights for the four components can be misleading as a way of identifying the most important targets for persuasive messages. This is a consequence of the often-substantial correlations among the RAT model’s predictors (Albarracín et al., 2001; Hagger et al., 2002; Manning, 2009; McEachan et al., 2011; Rivis & Sheeran, 2003). Because the predictors are correlated with each other, small differences in zero-order correlations (of each predictor with intention) can produce large differences in the weights. As a concrete example: In Fishbein et al.’s (2002) study of marijuana use intentions, AB and IN had virtually identical zero-order correlations with intention, but attitude had the larger weight in the regression. Fishbein et al. reasoned that attitude was the more important target for persuasion “because attitude was the most important determinant of intention” (p. 105). But this reasoning is defective; the zero-order correlations showed that AB and IN were statistically indistinguishable (and indeed literally nearly identical) as determinants of intention. There might have been good reasons for preferring AB over IN as an intervention focus—perhaps the malleability of attitude might plausibly be expected to be greater—but the invocation of the relative sizes of the weights was not such a reason.28 In general, then, intervention designers will want to examine not only the weights but also the zero-order correlations in order to have an accurate 188



picture of the influences on intention.29 As a further illustration, consider the descriptive norm specifically: Even if the DN is not a significantly weighted predictor, a persuader might nevertheless want to present descriptive norm information, because such information could potentially affect other determinants of intention. Specifically, DN information (e.g., “everybody’s doing it”) might influence AB (“if everybody’s doing it, there must be something good about it that I haven’t seen” or “if everybody’s doing it, these negative outcomes I’ve been worried about must not be very likely, or must not be all that bad”) or PBC (“if other people are doing, it must not be as hard as I thought, so maybe I can do it too;” e.g., Fornara, Carrus, Passafaro, & Bonnes, 2011). But when DN influences intention entirely through such effects on AB and PBC, DN would not be significantly weighted; that is, DN would not make an independent contribution to the prediction of intention over and above the contributions of AB, IN, and PBC. The point here is that even when DN is not significantly weighted, presenting DN information might nevertheless be useful as a means of influencing intentions (via the effect of DN information on AB or PBC).



Commentary Three general aspects of the RAT merit some comment: consideration of additional possible predictors, some suggested revisions of the attitudinal and normative components, and the nature of the perceived control component.



Additional Possible Predictors As mentioned earlier, research on RAT has progressively added predictors to the model, with each new predictor justified by virtue of its adding to the predictability of intention. The natural question that arises is whether any other additional predictors might be generally useful. Of the possibilities that have been explored, two suggestions—anticipated affect and moral norms—are discussed here as illustrations.30



Anticipated Affect Behaviors sometimes have affective (feeling-related) consequences—they 189



can arouse regret, happiness, guilt, and so forth—and people can often foresee these consequences (as when one scans the offerings at the movie theatre, looking for a mood-brightening comedy). A good deal of research now indicates that various anticipated emotions are dependably related to intentions and behavior. A number of studies have reported such effects specifically for anticipated regret (e.g., McConnell et al., 2000; for a review, see Sandberg & Conner, 2008); for instance, Lechner, de Vries, and Offermans (1997) found that among women who had not previously undergone mammography, the best predictor of participation intentions was anticipated regret (the greater the regret anticipated from not undergoing mammography, the greater the intention to do so). Related effects have been reported for anticipated guilt (e.g., Birkimer, Johnston, & Berry, 1993; Steenhaut & Van Kenhove, 2006) and other anticipated emotions (e.g., S. P. Brown, Cron, & Slocum, 1997; Leone, Perugini, & Bagozzi, 2005). Moreover, anticipated emotions can influence intentions above and beyond the more common RAT predictors. Studies of cancer screening (McGilligan, McClenahan, & Adamson, 2009), safer sex behaviors (e.g., Conner, Graham, & Moore, 1999; Richard, de Vries, & van der Pligt, 1998), vaccination (Gallagher & Povey, 2006), lottery playing (Sheeran & Orbell, 1999a), safe driving practices (Parker et al., 1995; Simsekoglu & Lajunen, 2008), and using drugs, using alcohol, and eating fast food (Richard, van der Pligt, & de Vries, 1996a) have reported that measures of anticipated emotional states (some general, others more specific) have improved the predictability of behavioral intentions beyond that afforded by various sets of RAT predictors (for some discussion, see Conner & Armitage, 1998, pp. 1446–1448; for a review concerning anticipated regret specifically, see Sandberg & Conner, 2008). Notice that the observed independent effect of anticipated affect implies that anticipated affective reactions are distinct from AB and from IN. One might have imagined that the expected affective consequences of an action would already be included in AB (because those consequences would contribute to the overall evaluation of the action) or in IN (through recognition of the views of significant others; e.g., “My mother would want me to do this, and I’ll feel bad if I disappoint her”) and hence would not make a separate contribution to the prediction of intention. But these studies suggest otherwise. Correspondingly, such studies suggest that anticipated affective reactions 190



provide a distinctive persuasive target (see Cappella, 2007). There is good evidence that the anticipation of emotion can indeed be influenced, primarily by heightening the salience of such anticipations. Several studies have apparently influenced the salience of anticipated emotions simply by asking about such feelings, with consequent effects on intention or behavior. For example, Sheeran and Orbell (1999a, Study 4) found that persons who answered a questionnaire item about regretting not playing the lottery (and so who presumably were induced to anticipate regret) intended to buy more lottery tickets than persons who did not answer such a question. (For related manipulations, see Abraham & Sheeran, 2004, Study 2; Hetts, Boninger, Armor, Gleicher, & Nathanson, 2000; O’Carroll, Dryden, Hamilton-Barclay, & Ferguson, 2011; Richard et al., 1996b; Sandberg & Conner, 2009.) Thus it seems that one straightforward mechanism for engaging anticipated emotions is simply to invite receivers to consider how they will feel if they follow (or do not follow) a particular course of action. A more focused mechanism might involve suggesting that receivers would experience a given emotion if they followed a particular course of action— say, that they would feel guilty if they cheated on their taxes. The potential of such an approach is illustrated by Parker, Stradling, and Manstead’s (1996) research, in which participants saw one of four videos aimed at influencing intentions to speed in residential areas; the videos were meant to influence either normative beliefs, behavioral beliefs, perceived behavioral control, or anticipated regret. The anticipated regret video appeared to evoke greater anticipated regret than other videos, and only the anticipated regret video proved more successful than a control video in inducing negative attitudes toward speeding. A similar illustration is offered by an antilittering campaign in Oklahoma (employing appeals meant to make people feel guilty if they littered), which seems to have produced substantial increases in the proportion of residents who said that they would feel guilty if they littered (Grasmick, Bursik, & Kinsey, 1991). The potential influenceability of anticipated emotions is perhaps also suggested by the occurrence of certain forms of advertising. For example, guilt-based consumer advertising often seeks to engage anticipated guilt feelings. Huhmann and Brotherton (1997) examined guilt-based advertisements in popular magazines over a 2-year period and found that most of the guilt appeals were “anticipatory” appeals (offering consumers the opportunity to avoid a guilt-inducing transgression) as opposed to appeals meant to arouse guilt. Sweepstakes promoters and lottery 191



advertising often seem to seek to induce thoughts about anticipated emotions—including not just the potential positive emotional consequences of winning but also the regret of not playing (“Suppose you’ve been assigned the winning mega-prize number, but because you didn’t enter we had to give the 10 million dollars to someone else”; see Hetts et al., 2000, p. 346; Landman & Petty, 2000). All the examples thus far are ones in which a persuader seeks to encourage the anticipation of particular emotions. But sometimes a persuader might want to prevent the anticipation of certain emotions. In consumer purchases, one type of possible anticipated regret involves the prospect of finding a lower price elsewhere (“If I find a lower price at another store, I’ll regret buying the product now—hence I’ll postpone my purchase”). An appropriate price guarantee (in which the seller promises that if the buyer finds the product offered at a lower price elsewhere, the seller will match that price) can undermine the creation of that anticipated regret (as observed by McConnell et al., 2000). In short, it seems clear that persuaders can indeed effectively engage anticipated emotions, perhaps even through relatively simple mechanisms that make anticipated emotions more salient. To the extent that anticipated affect influences intentions beyond the factors identified by RAT, anticipated affect will correspondingly be an important potential target for persuaders.



Moral Norms Another possible addition to the RAT is what can be called moral norms (also sometimes termed personal norms or moral obligation), that is, a person’s conception of morally correct or required behavior. Questionnaires for assessing moral norms have included items concerning perceived obligation (e.g., “I feel a strong personal obligation to use energy-saving light bulbs,” Harland, Staats, & Wilke, 1999) or perceived moral propriety (e.g., “It would be morally wrong for me to use marijuana,” Conner & McMillan, 1999; “Not using condoms would go against my principles,” Conner, Graham, & Moore, 1999). Several studies have found that moral norms can enhance the prediction of intention above and beyond the predictors already contained in the RAT. Such increased predictability has been found, for example, in studies of marijuana use (Conner & McMillan, 1999), condom use (Kok, Hospers, 192



Harterink, & De Zwart, 2007), environmental behaviors (M. F. Chen & Tung, 2010; Harland et al., 1999), smoking cessation (Høie, Moan, Rise, & Larsen, 2012; Moan & Rise, 2005), volunteering (Warburton & Terry, 2000), driving behaviors (Conner, Smith, & McMillan, 2003; Moan & Rise, 2011), and charitable donations (J. R. Smith & McSweeney, 2007). (For review discussions concerning moral norms, see Conner & Armitage, 1998, pp. 1441–1444; Manstead, 2000.) One supposes that the inclusion of moral norms will not always contribute to the prediction of intention, but there is little firm evidence yet concerning relevant moderating factors (see Hübner & Kaiser, 2006; Manstead, 2000, pp. 27–28). As with any potential addition to RAT, the apparent influence of moral norms on intention suggests that such norms may be a distinctive focus of influence efforts. Two general influence paths are possible: One involves the creation of some new perceived moral norm, the other (surely more generally useful and plausible) involves making some existing moral norm more salient. But there is little explicit research guidance on such questions.31



The Assessment of Potential Additions The impulse to add predictors to the RAT model is a natural one. After all, earlier additions of variables proved useful, and hence pursuit of still further additions is to be expected. But in assessing possible new predictors, two criteria might be kept in mind (for related general discussion, see Ajzen, 2011; Fishbein & Ajzen, 2010, pp. 281–282; Sutton, 1998). One is the size of the improvement in predictability afforded by a given candidate addition. It will not be enough that a given variable make a dependable (statistically significant) additional contribution to the prediction of intention; a large additional contribution is what is sought. The second is the breadth of behaviors across which the proposed addition is useful. In articulating a general model of behavioral intentions, one wants evidence suggesting that a proposed addition is broadly useful. It might be the case that improved prediction results from including variable X when predicting behavioral intention Y but that result (even if frequently replicated in studies of Y) does not show that X adds to the prediction of intention sufficiently broadly (i.e., across enough different behaviors) to merit the creation of a new general model that includes X.



193



But it is important also to bear in mind that there is a natural tension between a generally useful model and accurate prediction in a given application. In studying a particular behavior, an investigator might add variables that improve the prediction of that intention, never mind whether those added variables would be helpful in improving prediction in other applications. Thus when one’s interest concerns some particular behavior of substantive interest (as opposed to concerning the elaboration of general models), RAT might be thought of as providing useful general starting points. In any particular application, there might be additional predictors (beyond AB, IN, DN, and PBC) that prove to be useful in illuminating the behavior of interest—even if those additional factors are not generally useful (that is, even if not useful in studying other behaviors). And any such additional predictor, whether general or case-specific, is another distinguishable potential target for persuaders. (An example of a specialized RAT-like model is provided by the “technology acceptance model,” which has a distinctive set of predictors of the intention to use a new technology; for discussion and reviews, see Davis, Bagozzi, & Warshaw, 1989; King & He, 2006; Schepers & Wetzels, 2007; Venkatesh & Bala, 2008. Protection motivation theory, discussed in Chapter 11, although not explicitly conceived of as a specialized RAT model, is functionally similar in focusing specifically on protective intentions and behaviors.)



Revision of the Attitudinal and Normative Components The Attitudinal Component Some commentators have suggested replacing the general attitudinal component (AB) with two more specific ones, namely, an “instrumental” attitude, reflecting the behavior’s anticipated positive or negative material consequences, and an “experiential” (or “affective”) attitude, reflecting the positive or negative experiences associated with the behavior. A number of studies have suggested the potential value of such a distinction (e.g., Elliott & Ainsworth, 2012; Keer, van den Putte, & Neijens, 2010; Lowe, Eves, & Carroll, 2002). However, rather than thinking of these as two different kinds of attitudes, one might instead think of these categories as representing two different kinds of beliefs that might contribute to attitude. So, for example, one’s 194



beliefs about exercise might include both (instrumental) beliefs about health consequences and (experiential) beliefs about how it makes one feel. So rather than distinguish two attitudes, one might instead distinguish two possible kinds of belief underpinnings to attitude. Correspondingly, persuaders will want to be attentive to the potential importance of addressing each kind of belief (see Kiviniemi, Voss-Humke, & Seifert, 2007; Lawton, Conner, & McEachan, 2009; Wang, 2009). This does, however, point to the importance of ensuring that one’s beliefelicitation procedures evoke both kinds of belief (if both kinds are salient). Some frequently used belief-elicitation procedures may be more likely to elicit instrumental beliefs than affective ones (e.g., Sutton et al., 2003). To ensure good representation of salient beliefs, then, researchers need to be attentive to such issues. (For discussion of belief elicitation procedures, see Breivik & Supphellen, 2003; Darker, French, Longdon, Morris, & Eves, 2007; Dean et al., 2006; Middlestadt, 2012.)



The Normative Components The version of RAT presented here distinguishes injunctive and descriptive norms as distinct influences on intention. Although often positively correlated, these variables are conceptually distinct.32 Still, because both concern normative influences on behavior, one might be tempted to somehow combine IN and DN into a single general normative element (e.g., Fishbein & Ajzen, 2010, pp. 130–133). But there are two good reasons to resist such a merger, at least at the moment. The first reason is that the injunctive norm and the descriptive norm appear to operate in substantively different ways. Manning’s (2009) meta-analytic review pointed to several differences between IN and DN in their relationships to intentions and behaviors (e.g., the effects of DN may be affected by the degree of social approval for the behavior in ways that IN is not). A variety of other studies have pointed to the independent operation of DN and IN (e.g., Park, Klein, Smith, & Martell, 2009; Park & Smith, 2007; Vitoria, Salgueiro, Silva, & de Vries, 2009). For example, there is evidence suggesting that whereas the effects of injunctive norms characteristically require some degree of systematic thinking, descriptive norms can operate in ways that require little cognitive effort (e.g., Göckeritz et al., 2010; Jacobson, Mortensen, & Cialdini, 2011; Melnyk, van Herpen, Fischer, & van Trijp, 2011; for a general discussion, see Goldstein & Mortensen, 2012). Taken together, such findings argue for 195



separate treatment of these two factors. The second reason is measurement-related: Whereas there are wellestablished ways of assessing attitude toward the behavior and perceived behavioral control, there are not (yet, anyway) parallel means of assessing generalized perceived norms (i.e., some assessment of the overall generalized combination of injunctive and descriptive norms). And if injunctive and descriptive norms in fact do operate substantively differently (the first reason), then such measurement challenges will eventually prove insurmountable. In sum, rather than treating IN and DN as two contributors to one general normative factor, it seems preferable— at least at present—to distinguish these as two different influences on intention.



The Nature of the Perceived Control Component PBC as a Moderator As noted above, PBC does not seem to be quite like the other determinants of intention. Instead of straightforwardly influencing intention as the attitudinal and normative components do, PBC can instead plausibly be thought to moderate the effects of those variables on intention. Specifically, PBC can be seen as a necessary, but not sufficient, condition for the formation of intentions, and hence AB, IN, and DN will influence intention only when PBC is sufficiently high. If this image is correct, persuaders would want to be alert to a possible pitfall in focusing on PBC as a persuasion target. Specifically, increasing PBC seems likely to increase behavioral intentions only when the other determinants of intention (AB, IN, DN) incline the person toward a positive intention. So, for example, persuading a person that exercising regularly is under her control would presumably lead her to intend to exercise only if she would otherwise be inclined to have a positive intention given her AB, IN, and DN. When PBC is low, persuaders ought not assume that increasing PBC will automatically mean corresponding increases in intention. If PBC operates in this sort of moderating fashion, then the usual statistical tests of RAT should reveal an interaction effect such that the relationships of intention to AB, IN, and DN would vary depending on the level of PBC 196



(and, specifically, that as PBC increases, there should be stronger relations of AB, IN, and DN to intention). There is less empirical evidence on this question than one might like, because researchers have not often conducted the appropriate analyses. When the analyses have been reported, some studies have not found the expected interaction (e.g., Crawley, 1990; Giles & Cairns, 1995), but an increasing number of studies have detected it (e.g., Bansal & Taylor, 2002; Dillard, 2011; Hukkelberg, Hagtvet, & Kovac, 2014; Kidwell & Jewell, 2003; Park, Klein, Smith, & Martell, 2009; for a review, see Yzer, 2007). There are substantial challenges to obtaining empirical evidence indicating such interactions (e.g., considerable statistical power demands; see Manning, 2009, p. 662; Yzer, 2007), so reports of nonsignificant interactions are not entirely unexpected. In any case, that PBC is not quite on all fours with AB, IN, and DN is perhaps indicated by the finding that PBC can be significantly negatively related to intentions. Wall, Hinson, and McKee (1998) observed such a relationship in a study of excessive drinking; the less control that people thought they had over excessive drinking, the more likely they were to report intending to drink to excess. (Relatedly, PBC has been found to be negatively related to binge drinking—that is, frequent binge drinkers were less likely than others to think that the behavior was under their control; P. Norman, Bennett, & Lewis, 1998.) In a similar vein, Conner and McMillan (1999) found PBC to be significantly negatively related to intentions to use marijuana. Such results are consistent with Eagly and Chaiken’s (1993, p. 189) supposition that increasing PBC might enhance intention only to the degree that the behavior is positively evaluated—and these results certainly indicate that PBC is something rather different from the other three components. Indeed, such findings invite the conclusion that PBC can be a repository for rationalization. “Why do I keep doing these bad things that I know I shouldn’t [drinking to excess, smoking, and so forth]? Because I can’t help myself; it’s not really under my control. And why do I fail to do these good things I know I should do [exercising, recycling, and so on]? Gee, I’d like to do them, I really would—but I just can’t, it’s out of my hands, not under my control.” Persuaders may need to address such rationalizations in order to lay the groundwork for subsequent behavioral change.



Refining the PBC Construct 197



Several commentators have suggested that it may be useful to distinguish different facets of perceived behavioral control (e.g., Armitage & Conner, 1999a, 1999b; Cheung et al., 1999; Estabrooks & Carron, 1998; Rhodes, Blanchard, & Matheson, 2006; Rodgers, Conner, & Murray, 2008). In good measure this suggestion has been stimulated by findings indicating that items used to measure PBC often fall into two distinct clusters (e.g., Myers & Horswill, 2006; P. Norman & Hoyle, 2004; Pertl et al., 2010; Trafimow, Sheeran, Conner, & Finlay, 2002). But the nature of these clusters is not entirely clear, and the labels used to distinguish them exhibit considerable variety (for some discussion, see Ajzen, 2002; Fishbein & Ajzen, 2010, pp. 153–178; Gagné & Godin, 2007; Yzer, 2012b). As an illustration, one possible distinction might be between internal and external aspects of PBC, that is, between internal resources and obstacles to action (motivation, personal capabilities, and the like) and external resources and obstacles (facilities, equipment, cooperation of others, and so forth). For example, I might feel as though I am personally capable of exercising regularly (internal resources) but believe that a lack of facilities prevents me from doing so (external obstacles). It is possible, however, to appreciate this distinction (and others) without replacing PBC with two separate predictors (one for internal aspects and one for external aspects); one could retain the single PBC component but recognize that different kinds of underlying beliefs might contribute to an overall sense of behavioral efficacy. For persuaders, any way of distinguishing different elements underlying PBC points to correspondingly different possible targets of persuasion. For example, internal and external control elements might need separate attention in a given persuasion setting—and these two elements might well be differentially responsive to different influence mechanisms.33



Conclusion Reasoned action theory has undergone extensive empirical examination and development over time. It is unquestionably the most influential general framework for understanding the determinants of voluntary action. And in illuminating the underpinnings of behavioral intention, RAT provides manifestly useful applications to problems of persuasion, primarily by identifying potential points of focus for persuasive efforts.



198



For Review 1. What is the most immediate determinant of voluntary action? According to reasoned action theory (RAT), what are the four primary determinants of behavioral intention? 2. What is the attitude toward the behavior (AB)? Explain the difference between attitude toward the behavior and attitude toward the object. Describe the sorts of questionnaire items commonly used for assessing the AB. What is the injunctive norm (IN)? Describe the sorts of questionnaire items commonly used for assessing the IN. What is the descriptive norm (DN)? Describe the sorts of questionnaire items commonly used for assessing the DN. Explain the difference between the injunctive norm and the descriptive norm. What is perceived behavioral control (PBC)? Describe the sorts of questionnaire items commonly used for assessing PBC. Give examples of circumstances in which PBC might plausibly be the key barrier to behavioral performance. 3. Do the components above influence intention equally? How are the relative weights of the components assessed? How is PBC different from the other three components? How predictable are intentions from the four components? 4. Describe the five possible ways of influencing intention as identified by RAT. If persuasion is attempted by changing one of the components, does that component need to be significantly weighted? Explain. 5. What are the determinants of the attitude toward the behavior (AB)? What is belief strength, and how is it assessed? What is belief evaluation, and how is it assessed? Explain how these combine to yield the AB. What does the research evidence suggest about the predictability of the AB from its determinants? Describe alternative means by which the AB might be influenced. Explain (and give examples of) changing the strength or evaluation of existing salient beliefs. Explain (and give examples of) reconfiguring the set of salient beliefs; identify two ways in which such reconfiguration might be accomplished. 6. What are the determinants of the injunctive norm (IN)? What are normative beliefs (and how are they assessed)? What is motivation to comply (and how is it assessed)? Explain how these combine to yield the IN. What does the research evidence suggest about the 199



7.



8.



9.



10.



predictability of the IN from its determinants? Identify two concerns about the motivation-to-comply element. Describe alternative means by which the IN might be influenced. Explain (and give examples of) changing the normative belief or motivation to comply that is associated with an existing salient referent. Explain (and give examples of) the two ways of reconfiguring the set of salient referents. Why is it often difficult to change the IN? Explain how directing messages to salient referents might lead to changes in the IN. Describe the current state of understanding of the determinants of the descriptive norm (DN). Explain how the DN might be changed. Give an example of a message designed to influence the DN. Describe some potential pitfalls for DN interventions. Describe RAT’s account of the determinants of perceived behavioral control (PBC). What is a control belief, and how can it be assessed? What is the perceived power of a control factor, and how can it be assessed? What is the current state of the research evidence concerning the determinants of PBC? Describe four means of influencing PBC. Explain how directly removing an obstacle to performance can influence PBC. Distinguish (and give examples of) two kinds of obstacles a persuader might try to remove. Explain how successful performance of a behavior can influence PBC; give an example. Explain how modeling can influence PBC; give an example. Explain how encouragement can influence PBC; give an example. Explain the strategy of influencing intention by changing the relative weights of the components. To which of the four components does this strategy potentially apply? In what sort of circumstance can this strategy succeed in changing intention? What is the usual pattern of association (correlation) between the AB, the IN, and the DN? What does this pattern imply about changing the weights as a means of influencing intention? What does the research evidence suggest about the predictability of behavior from intention? Identify three factors influencing the strength of the relationship between measures of intention and measures of behavior. Explain how the relationship between measures of intention and measures of behavior is affected by the degree of correspondence between the two measures. Do more specific intention measures lead to higher correlations with behavioral measures than do less specific intention measures? Explain how the relationship between measures of intention and measures of behavior is affected by the temporal stability of intentions. Explain how the 200



11.



12.



13.



14.



15.



relationship between measures of intention and measures of behavior is affected by explicit planning about behavioral performance. Give examples of circumstances in which the task facing the persuader is that of encouraging persons to act on existing intentions; describe how a persuader might approach such a task. What explains the effect of explicit-planning interventions on behavior? Does planning make intentions more positive? Does planning increase perceived behavioral control (PBC)? What are implementation intentions? Does planning encourage the development of implementation intentions? Identify four factors that might influence the effectiveness of explicitplanning interventions. What factor might improve the prediction of behavior (beyond the predictability afforded by intention)? Under what conditions does assessment of prior behavior improve the prediction of behavior? Describe two general ways in which RAT suggests persuasive messages might be adapted to recipients. Explain how the weights of the determinants of intention provide a basis for message adaptation. How can such weights be misleading? Describe how message adaptation can be guided by consideration of the beliefs that underlie the component to be changed. Describe the basis on which additional possible predictors of intention (beyond AB, IN, DN, and PBC) might be considered for inclusion in the model. Describe two criteria for assessing such additions. Identify two specific possible additional predictors. What is anticipated affect? Can anticipated affect improve the prediction of intentions beyond the predictability afforded by the four RAT components? Describe how persuaders might try to influence anticipated affect. What are moral norms? Can moral norms improve the prediction of intentions beyond the predictability afforded by the four RAT components? Describe how persuaders might try to influence moral norms. How might the attitudinal component (AB) be revised? Describe the distinction between instrumental and experiential attitudes. Explain how this distinction might reflect differences in the kinds of beliefs underlying an attitude. Describe how the injunctive norm (IN) and the descriptive norm (DN) might be revised by being merged into a single normative factor. Discuss why such a merger might not be advisable. Explain how perceived behavioral control (PBC) might moderate the effects of the other three components. Describe the current state of the 201



research evidence concerning such a moderating role. Explain how different kinds of PBC questionnaire items might represent different kinds of underlying control beliefs.



Notes 1. The viewpoint described in this chapter has appeared in a number of different forms, with a number of different labels including the “theory of reasoned action” (Ajzen & Fishbein, 1980), the “theory of planned behavior” (Ajzen, 1991), the “integrative model of behavioral prediction” (Fishbein, 2008), and the “extended theory of planned behavior” (Sieverding, Matterne, & Ciccarello, 2010). Following Fishbein and Ajzen (2010), this presentation identifies four predictors of intention: attitude toward the behavior, injunctive norms (formerly called the “subjective norm”), descriptive norms, and perceived behavioral control. However, whereas Fishbein and Ajzen’s (2010) presentation has one general “perceived norms” factor that includes both injunctive and descriptive norms, the presentation here treats those two normative elements as distinct. 2. The specification of the behavior of interest is a matter requiring close attention. For some discussion, see Fishbein and Ajzen (2010, pp. 29–32) and Middlestadt (2007). 3. The attitude of interest here is specifically the attitude toward the behavior in question. The suggestion is that, for example, the intention to buy a Ford automobile is influenced most directly by one’s attitude toward the behavior of buying a Ford automobile rather than by one’s attitude toward Ford automobiles. Attitudes toward objects may have some relationship to attitudes toward actions, but RAT suggests that attitudes toward actions are the more immediate determinant of intentions. 4. Cialdini’s (2009, p. 99) principle of “social proof”—that “we view a behavior as correct in a given situation to the degree that we see others performing it”—offers another illustration of the power of descriptivenormative perceptions. 5. RAT also expects that sometimes PBC will appear to have a direct relationship to behavior. In circumstances in which actual (not perceived) behavioral control influences performance of the behavior, then to the extent that persons’ perceptions of behavioral control are accurate (and so 202



co-vary with actual behavioral control), to that same extent PBC will be related to behavior. 6. This procedure—rather than simply asking people directly how important the attitudinal and normative considerations are to them—is used because there is reason to think such self-reports are not sufficiently accurate (Fishbein & Ajzen, 1975, pp. 159–160; for an illustration, see Nolan, Schultz, Cialdini, Goldstein, & Griskevicius, 2008). The inadequacy of these self-reports has prevented satisfactory estimation of the weights for a given individual’s intention to perform a given behavior (for exploration of some possibilities other than such self-reports, see Budd & Spencer, 1984; Hedeker, Flay, & Petraitis, 1996). 7. There is room for reasonable doubt about the generality of the contribution of descriptive norms to the prediction of intention, so some caution may be appropriate. On the one hand, DN seems to be correlated reasonably strongly with intention—sometimes as strongly as AB (see the review of Rivis & Sheeran, 2003), sometimes less strongly (see the review of Manning, 2009), but generally quite positively. And DN has often been found to make an independent contribution to the prediction of intention (i.e., it improves the predictability of intention beyond that afforded by AB, IN, and PBC). On the other hand, there is much less research evidence concerning DN compared with the other three variables; for example, Manning’s (2009) review identified 162 reports of the correlation between AB and intention, but only 17 for the correlation between DN and intention. One might reasonably anticipate that the early investigations of the role of DN could (quite sensibly) have explored behaviors for which DN was especially likely to play an important role. (A similar point about studies of anticipated regret as a possible additional predictor—discussed separately—was made by Sandberg & Conner, 2008.) It often enough happens that early research findings turn out to be overstated (Ioannidis, 2008), so the apparent contribution of DN to the prediction of behavior may well weaken as additional research results accumulate. 8. Many of the issues that have arisen in the context of belief-based models of attitude (see Chapter 4)—such as the potentially artifactual contribution of belief strength scores to the prediction of attitude—can naturally arise here as well, because the same summative model of attitude, with the same procedures, is involved (e.g., Gagné & Godin, 2000; O’Sullivan, McGee, & Keegan, 2008; Steadman, & Rutter, 2004; Trafimow, 2007). Although Armitage and Conner (2001) report a mean correlation of .50 between 203



behavioral beliefs and AB across 42 studies, it is not clear whether this represents correlations with ∑biei, ∑bi, ∑ei, or some combination of these. 9. McEachan et al.’s (2011) reported mean correlation of .53 (with IN) includes both studies using ∑nimi as the predictor and studies using ∑ni. It is not clear whether Armitage and Conner’s (2001) reported mean correlation of .50 between normative beliefs and IN (across 34 studies) is an average of correlations involving ∑nimi, ∑ni, ∑mi, or some combination of these. 10. In addition, the issues discussed in Chapter 4 (on belief-based models of attitude) concerning alternative scale scoring procedures also arise in the context of scoring normative belief and motivation-to-comply items. For some discussion of the phrasings and scorings of such items, see Fishbein and Ajzen (2010, pp. 137–138), Gagné and Godin (2000), and Kantola et al. (1982). 11. Different types of resources and obstacles will want different phrasings of questionnaire items, especially with respect to control beliefs (likelihood or frequency of occurrence). For instance, although the control belief associated with bad weather could be assessed by asking respondents how frequently bad weather occurs where they live (with scales end-anchored by phrases such as “very frequently” and “very rarely”), the control belief concerning a lack of facilities might better be assessed by asking a question such as “I have easy access to exercise facilities” (with end-anchors such as “true” and “false”). An additional complexity: Because ∑cipi is a multiplicative composite (as are ∑biei and ∑nimi), the same scale-scoring issues (e.g., unipolar vs. bipolar) can arise (see Gagné & Godin, 2000). 12. The challenge here can be illustrated by Elliott et al.’s (2005) research concerning speed limit compliance. In preliminary research, 12 possible control beliefs were identified. In the main study, a belief was retained for analysis if its control belief (ci) or its power belief (pi) or its product (cipi) was statistically significantly related to PBC. Only four beliefs survived this winnowing process. 13. Armitage and Conner (2001) reported a mean correlation of .52 (across 18 studies) between control beliefs and PBC, but it’s not clear whether this represented correlations of PBC with ∑cipi, ∑ci, ∑pi, or some combination 204



of these. McEachan et al. (2011) reported a mean correlation of .41 across 27 studies examining either the ∑cipi-PBC correlation or the ∑ci-PBC correlation, but results were not reported separately for these two sets of correlations. 14. The lack of any standardized item format for assessing control beliefs and powerfulness—necessitated by variation in the types of factors under study (see note 11 above)—has produced considerable diversity in the details of these belief-based assessments, and it is not always clear how best to characterize the measures employed. For example, P. Norman and Smith (1995) presented respondents with a list of seven barriers to physical activity (such as a lack of time or the distance from facilities) and asked respondents to indicate, “Which of the following reasons would be likely to stop you from taking regular exercise?” (with responses given on a 7-point scale anchored by “extremely likely” and “extremely unlikely”). Although based on likelihood ratings, the resulting index appears not to assess control beliefs (the perceived frequency or likelihood of occurrence of a control factor); it might better be seen as amalgamating assessment of powerfulness and likelihood of occurrence (notice that the question asks for the likelihood that the factor will prevent the behavior—not the likelihood that the factor will occur) or perhaps even as assessing simply the factor’s powerfulness (the question might be taken to mean, “Which of the following reasons, if they occurred, would be likely to stop you from taking regular exercise?”). 15. The effect of including the parking permit or bus pass was only partly attributable to its removal of transportation obstacles. Apparently, the inclusion of the permit/pass also helped convince recipients of the value of making a return visit (“This must be important, otherwise they would not send me a bus pass”), which in turn helped boost return rates (Marcus et al., 1992, p. 227). Sometimes persuasion happens in unexpected ways. 16. This list of alternative mechanisms reflects some parallels with Bandura’s (1997, pp. 79–115) analysis of sources of self-efficacy, which include “enactive attainment” (experiences of genuine mastery of the behavior), “vicarious experience” (seeing or imagining others perform successfully), and “verbal persuasion” (having others provide assurances about one’s possession of the relevant capabilities). 17. This strategy (of altering the relative weights of the components so as to influence intention) is probably in general applicable only to AB, IN, 205



and DN—not PBC. For example, if a person has a negative AB, negative IN, and negative DN, then emphasizing “you really do have the ability to do this” is unlikely to be very persuasive. This is related to the earlier point that PBC is not quite like the other determinants of intention, in that it seems more a variable that enables AB, IN, and DN (in the sense that those variables will influence intention only when PBC is sufficiently high). However, such an image does suggest that if a person has a positive AB, positive IN, and positive DN, then emphasizing “you really don’t have the ability to do this” might help to discourage formation of a positive intention. For example, imagine trying to discourage a friend from an excessively expensive purchase by saying (though perhaps not in so many words) “you really can’t afford this.” 18. More carefully: The direction of intention can be changed by altering the weights of these three components only when one of those three components differs in direction from the other two. But even if all three components incline the person in the same direction, the extremity of the intention might be changeable. For example, if AB, IN, and DN are all positive but vary in just how positive they are, then if a slightly positive component were to come to be weighted more heavily than a strongly positive component, the intention would presumably weaken (though it would still be positive). 19. As another indication that PBC may operate in a different fashion from the other three variables, the average correlations of PBC with the other three components appear to be smaller than the correlations among those three. In Rivis and Sheeran’s (2003) review, the mean correlations of PBC with AB, IN, and DN were between .05 and .20; in Manning’s (2009) review, those mean correlations ranged from roughly .20 to .45. 20. To be careful here: A positive correlation between two components does not necessarily mean that if one component is positive, the other will be as well; it means only that the two components vary directly (so that as one becomes more positive, so does the other). Imagine, for example, that in a group of respondents, each respondent has a positive attitude toward the behavior and a negative injunctive norm; those with very strongly positive attitudes, however, have only slightly negative injunctive norms, whereas those with only slightly positive attitudes have very strongly negative norms. There is a positive correlation between the two components (as the attitude becomes more positive, the norm also 206



becomes more positive—that is, less negative), although for each individual, one component is positive and the other is negative. But insofar as the persuasive strategy of altering the weights is concerned, the implication is (generally speaking) the same: Altering the weights of the components is not likely to be a broadly successful way of changing intention (because of the unusual requirements for the strategy’s working —e.g., a dramatic change in the weights may be necessary). 21. There are many potential pitfalls in interpreting intention-behavior correlations. For example, there is obviously some basis for worry about causal interpretations of cross-sectional intention-behavior correlations (i.e., correlations based on data collected at a single point in time). But even positive longitudinal correlations (where behavior is assessed subsequent to intentions) can potentially be misleading, because the correlation between intention and subsequent behavior does not indicate whether people’s intentions were realized in their behaviors. If intention and subsequent behavior are positively correlated in a given data set, that does not necessarily mean that—in any ordinary sense—people acted consistently with their intentions. For example, imagine a sample in which people said that they intended to exercise an average of 25 days over the next 30 days, but subsequent behavioral assessments indicated that in fact they exercised an average of only 5 days. This would seem to be a straightforward case of inconsistency between intention and behavior— and yet with such data intentions and behaviors could be perfectly positively correlated. To see this, imagine a small data set (N = 11) in which the intention scores (intended days of exercise) were 30, 29, 28, … 22, 21, and 20 (mean = 25.0) and the corresponding behavior scores (actual days exercised) were 10, 9, 8, … 2, 1 and zero (mean = 5.0). That is, the participant with an intention score of 30 had a behavior score of 10, the participant with an intention score of 29 had a behavior score of 9, and so on. The intention-behavior correlation is +1.00. Similarly, imagine a sample in which people said that they intended to exercise an average of 25 days over the next 30 days, and subsequent behavioral assessments indicated that in fact they did exercise an average of 25 days. One would think that this represents pretty decent intentionbehavior consistency, and yet the intention-behavior correlation in such a sample could be perfectly negative (-1.00). To see this, imagine another small data set (N = 11) in which the intention scores (intended days of exercise) were again 30, 29, 28, … 22, 21, and 20 (mean = 25.0), but with 207



corresponding behavior scores (actual days exercised) of 20, 21, 22, … 28, 29, and 30 (mean = 25.0). That is, the participant with an intention score of 30 had a behavior score of 20, the participant with an intention score of 29 had a behavior score of 21, and so on. The intention-behavior correlation is -1.00. So rather than simply examining intention-behavior correlations as indices of intention-behavior consistency, it might be more helpful to pursue something like Sheeran’s (2002) strategy. He analyzed intention-behavior relationships as a set of four cases that form a 2 × 2, where the contrasts are positive vs. negative intention and performance vs. nonperformance of the behavior. The four kinds of cases are the “inclined actor” (positive intention, performs the behavior), the “inclined abstainer” (positive intention, does not perform the behavior), the “disinclined actor” (negative attitude, but performs the behavior), and the “disinclined abstainer” (negative intention, does not perform the behavior). The first and last cases represent intention-behavior consistency, the two middle cases represent inconsistency. Sheeran reviewed several studies that collectively suggest that intention-behavior inconsistency is largely due to inclined abstainers rather than to disinclined actors. Application of Sheeran’s (2002) framework is sometimes, but not always, straightforward: It’s easy enough to create such a matrix with two-category variables (e.g., “Do you intend to get a flu shot next month—yes or no?” and “Did vs. didn’t get a shot”), but continuous variables can be a bit more challenging. Still, this framework underscores the need for closer attention to intention-behavior relationships than one would obtain from examination of correlations alone. Indeed, there are good reasons for broader concerns about the use of correlational data to assess models like RAT (Noar & Mehrotra, 2011; Weinstein, 2007). 22. It has sometimes been suggested that the strength of the intentionbehavior relationship will be affected by the time interval between the assessment of intention and the assessment of behavior (the idea being that as the time interval increases, the predictability of behavior from intention will decrease; see, e.g., Ajzen, 1985; Ajzen & Fishbein, 1980, p. 47). The supposition is that with an increased time interval between intention assessment and behavioral assessment, there would be increased opportunity for a change in intention. As it happens, it is not clear that variations in the size of the time interval have general effects on the intention-behavior relationship (for a review, see Randall & Wolff, 1994, but also see Sheeran & Orbell, 1998, pp. 234–235). But of course if (for a 208



particular behavior) persons’ intentions are relatively stable across time, then variations in the interval between assessments would not show much effect on the intention-behavior relationship. The relevant points to notice here are that (a) time interval variation is a poor proxy measure of temporal instability in intentions, and (b) an apparent absence of broad time interval effects on the strength of the intention-behavior relationship is not necessarily inconsistent with the hypothesis that temporal instability of intentions influences the strength of the intention-behavior relationship. 23. At least some explicit-planning interventions have told participants they will be more likely to perform the behavior if they write out a plan. For example, Sheeran and Orbell (2000b, p. 285) told participants “You are more likely to go for a cervical smear if you decide when and where you will go” (similarly, see Michie, Dormandy, & Marteau, 2004; Steadman & Quine, 2004). This raises the possibility that expectancyrelated effects (as have been observed with subliminal self-help audiotapes: Spangenberg, Obermiller, & Greenwald, 1992) might have contributed to some of the observed effects (but see also Chapman, Armitage, & Norman, 2009). 24. There is a slight complexity here, as RAT acknowledges that in addition to intention, PBC may also have a direct relationship with behavior by virtue of a possible correspondence between PBC and actual barriers to action. 25. At least some of these meta-analytic reviews reported confidence intervals that appear to have been based on fixed-effect analyses, which are less appropriate than random-effects analyses given interests in generalization (see, e.g., Borenstein, Hedges, Higgins, & Rothstein, 2010). Random-effects analyses would typically produce wider confidence intervals, so comparison of mean effect sizes within these meta-analyses should be undertaken cautiously. 26. Research exploring factors that might systematically affect the relative influence of the components is unfortunately scattered across a variety of factors, including self-monitoring (DeBono & Omoto, 1993), culture (AlRafee & Dashti, 2012; Bagozzi, Lee, & Van Loo, 2001), mood (Armitage, Conner, & Norman, 1999), degree of group identification (Terry & Hogg, 1996), state versus action orientation (Bagozzi, Baumgartner, & Yi, 1992), private versus collective self-concepts (Ybarra & Trafimow, 1998), and others (Latimer & Ginis, 2005b; Thuen & Rise, 1994). For some 209



discussion, see Fishbein and Ajzen (2010, pp. 193–201). 27. Nonintenders might be further subdivided on the basis of other characteristics, and those subgroups compared for potential differences (in salient referents, behavioral beliefs, and so forth) relevant to constructing persuasive messages. For instance, in designing antimarijuana messages, it may be important to recognize differences between low-risk and high-risk adolescents (Yzer et al., 2004). Of course, sometimes one size will fit all (Darker, French, Longdon, Morris, & Eves, 2007). 28. Two other examples to illustrate the importance of examining zeroorder correlations: In Dillard’s (2011) study of HPV vaccination intentions, AB and IN had relatively similar zero-order correlations with intention (.78 and .63), but very different beta-weights (.51 and .17). In Sheeran and Orbell’s (1999a) two studies of lottery playing, IN had a significant beta-weight (in regressions predicting intention) in Study 1 but not in Study 2—even though the zero-order correlation of IN with intention was identical (.41) in the two studies. 29. Some readers will recognize this as simply an example of the general point that the presence of multicollinearity conditions the interpretation of partial coefficients. 30. Of additional possible predictors not discussed here, self-identity is perhaps the most notable. For research examples and review discussions, see Astrom and Rise (2001); Booth, Norman, Harris, and Goyder, 2014; Conner and Armitage (1998); Cook, Kerr, and Moore (2002); Mannetti, Pierro, and Livi (2004); Nigbur, Lyons, and Uzzell (2010); and Rise, Sheeran, and Hukkelberg (2010). 31. At least in some circumstances, moral norms and anticipated affect may be closely intertwined; one’s moral obligations might lead to expecting to feel guilty or regretful if one fails to live up to one’s personal standards. For example, the beliefs “I feel I have a moral obligation to donate to charity” (moral norm) and “I’ll feel bad [guilty, regretful] if I fail to donate to charity” (anticipated affect) naturally hang together. In fact, some measures of moral norms (or personal norms or moral obligation) have included items concerning affective states (e.g., Conner & McMillan, 1999; Godin et al., 1996; Harland et al., 1999; Moan & Rise, 2005; Warburton & Terry, 2000). But anticipated affect and moral norms may sometimes also be usefully distinguished with respect to their influences 210



on intention. For example, a committed environmentalist’s expectations of guilt feelings from failing to recycle is probably different from a person’s anticipated regret about not playing the lottery; the former is more closely bound up with significant personal identity questions, the latter probably more with potentially forgone monetary gains. The larger point is that although there are plainly connections to be explored between moral norms and anticipated affect, one probably should not fuse these into a single element. Similarly, moral norms may be seen to be related to injunctive norms (IN). The IN has a particular referent group (“people who are important to me”) and a particular target person (the respondent): “Most people who are important to me think I should/should not engage in behavior X.” This is not far from “Most people who are important to me think it is wrong to engage in behavior X,” “Most people think it is wrong to engage in behavior X,” and “(I think) it is wrong to engage in behavior X.” But, again, running all these together into a single element is probably not advisable. 32. Rivis and Sheeran (2003) reported a mean correlation of .38 across 14 studies; Manning (2009) reported a mean correlation of .59 across 12 studies. 33. A contrast between internal and external aspects is not the only possible way of potentially distinguishing elements of PBC. For example, Fishbein and Ajzen (2010, pp. 168–177) have suggested that the key distinction (between facets of PBC) is actually that between perceived capacity (ability, capability) and perceived autonomy (degree of control), a distinction they argue is conceptually and empirically independent of a contrast between internal and external factors. For present purposes, the question of how best to interpret the various observed PBC item clusterings does not need to be settled (and there may not be only one appropriate taxonomy). The point here is that any such clusters might be seen as representing substantively different sorts of beliefs underlying PBC, with correspondingly distinct targets for persuasion.



211



Chapter 7 Stage Models The Transtheoretical Model Decisional Balance and Intervention Design Self-Efficacy and Intervention Design Broader Concerns About the Transtheoretical Model The Distinctive Claims of Stage Models Other Stage Models Conclusion For Review Notes



Persuasion characteristically has behavior change as its eventual goal. Stage models of behavioral change depict such change as involving movement through a sequence of distinct phases (stages). People at different stages are taken to need different kinds of messages (interventions, treatments, etc.) to encourage them to move to the next stage. A number of different stage models have been offered. This chapter focuses on the approach of the most-studied stage model, the “transtheoretical model.” Subsequent sections briefly discuss other stage models and the distinctiveness of stage approaches.



The Transtheoretical Model The transtheoretical model (TTM) is an approach developed from analysis of a number of different theories of psychotherapy and behavior change, in the context of changing undesirable health behaviors such as smoking. The goal was to integrate these diverse approaches by placing them in a larger (“transtheoretical”) framework. (For some general presentations, see Prochaska & DiClemente, 1984; Prochaska, Redding, & Evers, 2002.) Instead of seeing behavior change as a singular event (quitting smoking, for example), the TTM suggests that behavior change involves movement through a sequence of five distinct stages: precontemplation, contemplation, planning, action, and maintenance. In the precontemplation 212



stage, the person is not considering changing his or her behavior; for example, a smoker in the precontemplation stage is not even thinking about giving up smoking. In the contemplation stage, the person is thinking about the possibility of behavioral change; a smoker in the contemplation stage is at least considering quitting. In the planning stage, the person is making preparations for behavior change; a smoker in the planning stage is making arrangements to quit (choosing a quit date, purchasing nicotine gum, and the like). In the action stage, the person has initiated behavioral change; in this stage, the (now ex-) smoker has stopped smoking. In the maintenance stage, the person sustains that behavioral change; an ex-smoker in the maintenance stage has managed to remain an ex-smoker.1 The TTM does not claim that stage movement is always a straightforward linear process. To the contrary, it acknowledges that people may move forward, backslide, cycle back and forth between stages, and so on, in a complex and dynamic way. However, individuals are not expected to skip any stage (Prochaska, DiClemente, Velicer, & Rossi, 1992, p. 825) and hence these stages are offered as representing a general sequence through which people pass in the course of behavior change. People are said to progress through these stages using various “processes of change.” TTM presentations often list 10 such processes (e.g., Prochaska, Redding, & Evers, 2002): self-reevaluation (reconsideration of one’s self-image, such as one’s image as a smoker), environmental reevaluation (assessment of the effects of one’s behavior on others, as when a smoker consider the effects of secondhand smoke), counterconditioning (healthier behaviors that can substitute for the problem behavior, such as the use of nicotine replacement products), consciousness raising (increased awareness of causes and effects of, and cures for, the problem behavior), dramatic relief (the arousal and attenuation of emotion, as through psychodrama), self-liberation (willpower, a commitment to change), helping relationships (support for behavioral change), contingency management (creation of consequences for choices), stimulus control (removing cues that trigger the problem behavior, adding cues to trigger the new behavior), and social liberation (external policies and structures such as smoke-free zones).2 The TTM suggests that different processes of change are relevant at different stages, but research evidence on this matter is sparse. This is especially unfortunate because such evidence would provide valuable 213



guidance about how to construct effective interventions. For example, if self-reevaluation were known to be a change process distinctly associated with the movement from precontemplation to contemplation, then interventions based on that process might be targeted specifically to individuals in precontemplation. But what evidence is in hand seems to suggest that the various processes of change can often be useful across a number of different stages (see, e.g., Callaghan & Taylor, 2007; Guo, Aveyard, Fielding, & Sutton, 2009; Rosen, 2000; Segan, Borland, & Greenwood, 2004). Even so, it seems apparent that different behavior-change interventions (even if not described in terms of the 10 processes of change) should be effective for people at different stages. For example, a smoker in precontemplation presumably needs a different intervention than does a smoker in the planning stage. So the key question is how to develop interventions that are matched to the recipient’s stage. In support of the development of stage-matched interventions, the TTM has come to identify two particular mediating processes as crucial to the process of behavior change: decisional balance (the person’s assessment of the pros and cons of the new behavior) and self-efficacy (the person’s assessment of his or her ability to perform the new behavior). The following sections consider each of these.3



Decisional Balance and Intervention Design Decisional Balance One intriguing aspect of TTM research concerns “decisional balance,” the person’s assessment of the importance of the pros (advantages, gains) and cons (disadvantages, losses) associated with the behavior in question.4 The expectation is that as people progress through the stages, the importance of the pros of behavior change will come to outweigh the importance of the cons. In the relevant research, respondents are provided with a standardized list of pros and cons of the new behavior and are asked to rate the importance of each to the behavioral decision (e.g., on a scale with end-anchors such as “not important” and “extremely important”; Prochaska et al., 1994, p. 42). The relative importance of the various pros and cons can then straightforwardly be assessed. Considerable research evidence has accumulated concerning decisional 214



balance, confirming that these assessments do vary depending on the person’s stage (for some reviews, see Di Noia & Prochaska, 2010; Hall & Rossi, 2008; Prochaska et al., 1994; but see also Sutton, 2005b, pp. 228– 233). As summarized by Di Noia and Prochaska (2010, p. 619): “The balance between the pros and cons varies across stages. Because individuals in pre-contemplation are not intending to take action to change a behavior, the cons outweigh the pros in this stage. Pros increase and cons decrease from earlier to later stages. In action and maintenance stages, the pros outweigh the cons. A crossover between the pros and cons occurs between precontemplation and action stages.” This description is potentially a little misleading, however. The research in hand shows that “pros increase and cons decrease from earlier to later stages” in the specific sense that there are such changes in the perceived importance of the pros and cons. The research evidence does not concern, for example, possible changes in the number of perceived advantages (pros) and disadvantages (cons), in the desirability or perceived likelihood of the advantages and disadvantages, or in other properties of potential interest. What the evidence shows is that in pre-action stages, the perceived importance of the advantages of the new behavior is not greater than the perceived importance of the disadvantages, but that once action has been initiated, the perceived importance of the pros is greater than the perceived importance of the cons.5 In some ways this may not be too surprising. If people think that the importance of the advantages of a given behavior are not greater than the importance of its disadvantages, they may not be especially motivated to adopt that behavior. People who have adopted the behavior, on the other hand, are naturally likely to think the advantages are more important than the disadvantages.



Decisional Balance Asymmetry TTM research has identified an unusual aspect to the changes in the perceived importance of the pros and cons: they are not symmetrical. The characteristic pattern is that the size of the increase in the perceived importance of the pros is larger than the size of the decrease in the perceived importance of the cons.6 This difference has been expressed as a matter of “strong and weak principles” for progression through the stages described. These principles 215



describe the different amounts of change (in the perceived importance of the pros and the cons) in terms of the standard deviation of each. The strong principle is that “progress from precontemplation to action involves approximately one standard deviation increase in the pros of changing.” The weak principle is that “progress from precontemplation to action involves approximately .5 SD decrease in the cons of changing” (Prochaska, Redding, & Evers, 2002, pp. 105, 106). That is, the perceived importance of the new behavior’s advantages increases by about one standard deviation between precontemplation and action, while the perceived importance of the behavior’s disadvantages decreases by about half that much. (For reviews and discussion, see Di Noia & Prochaska, 2010; Hall & Rossi, 2008; Prochaska, 1994.)7



Implications of Decisional Balance Asymmetry The asymmetry in these changes suggests that the adoption of a new behavior may be less a matter of the person’s deciding that the behavior’s disadvantages are insignificant than it is a matter of deciding that the advantages make the behavior worth doing. If these two kinds of considerations (the importance of the advantages and the importance of the disadvantages) were equally influential, one might expect roughly equal amounts of change as people move through the stages. But stage progression is associated with distinctly different amounts of change in the perceived importance of the advantages and disadvantages. This asymmetry can be seen to have straightforward implications for the design of effective interventions meant to move people from the precontemplation stage to the action stage: “In progressing from precontemplation to action, tailored interventions should place primary emphasis on increasing the pros of change” (Hall & Rossi, 2008, p. 271). “For example, individuals in precontemplation could receive feedback designed to increase their pros of changing to help them progress to contemplation” (Prochaska, Redding, & Evers, 2002, p. 108). Expressed more carefully, the apparent implication is that for message recipients in precontemplation, it may be more useful for persuaders to try to increase the perceived importance of the advantages of the new behavior than to try to reduce the perceived importance of the disadvantages.8 However, such conclusions (about how to tailor interventions for preaction stages) cannot be justified by the data in hand. The characteristic research design concerning decisional balance consists of measuring both 216



people’s stages and their decisional balance assessments at one point in time. The evidence concerning decisional balance thus takes the form of a finding that (for example) for people who are in precontemplation, the perceived importance of the pros of the new behavior is not greater than the perceived importance of the cons, whereas for people who are in action or post-action stages, the perceived importance of the pros is greater than that of the cons. The trouble with such data is that one cannot tell whether these decisional balance shifts caused the movement from one stage to the next, as opposed to simply being associated with (or being a result of) stage change. For example, decisional balance might work in the following way. When the perceived importance of the pros is greater than the perceived importance of the cons, the person adopts the new behavior. This activates post-choice dissonance reduction processes (see Chapter 5), in which the perceived importance of the pros increases further and the perceived importance of the cons decreases further—and this post-action period can be the time at which the asymmetry appears (i.e., after behavioral initiation, not before). The point is: The current evidence is not sufficient to underwrite a design principle for interventions aimed at influencing decisional balance. One cannot tell, on the basis of the kinds of studies in hand, whether the asymmetry in decisional balance changes is a precursor, correlate, or consequence of such change. For supporting recommendations about intervention design, better evidence would be provided by experimental research. Such work could compare the effectiveness (in moving people from precontemplation to action) of messages that aimed either at increasing the perceived importance of the advantages of the new behavior or at reducing the perceived importance of the disadvantages of the new behavior. Evidence of this sort is not yet in hand, but would plainly be welcomed.



Self-Efficacy and Intervention Design Intervention Stage-Matching As briefly discussed above, the general idea of stage-matching of interventions (messages, treatments) is straightforward: People in different stages of change presumably need different interventions to encourage movement to the next stage. This idea is depicted in an abstract way in Figure 7.1. Intervention A is adapted to (matched to) persons who are in 217



Stage 1 and is designed to move people from Stage 1 to Stage 2. Interventions B and C are matched to persons in Stages 2 and 3, respectively, because those interventions are meant to move people to the next stage in the sequence. Thus for people in Stage 1, Intervention A should be more effective than Interventions B or C (more effective in moving people to Stage 2); for people in Stages 2 or 3, Intervention B or C, respectively, should be most effective. Figure 7.1 Matching interventions to stages.



Assessing the relative effectiveness of stage-matched and stagemismatched interventions can face some difficult challenges. Among other things, researchers need dependable ways of classifying participants into distinguishable stages, clear conceptual treatments of what counts as a matched intervention for a given stage, effective realization of those interventions, and sound assessment of relevant outcomes (such as stage progression, that is, forward movement of people through the stages). In this regard, the TTM is still a work in progress. For example, there is room for some uncertainty concerning exactly what makes for matched and mismatched interventions at various specific stages. And the difficulties in creating reliable stage assessments should not be underestimated. At the same time, it is possible to illustrate some of the relevant issues by considering one specific matter: the question of the point at which interventions should target the receiver’s self-efficacy concerning the desired behavior (their perceived ability to perform the behavior, akin to perceived behavioral control as discussed in Chapter 6).



Self-Efficacy Interventions The TTM suggests that self-efficacy interventions are not well-suited to people at earlier stages (e.g., precontemplation), because those people have not yet decided that they want to adopt the new behavior. At early stages, interventions should presumably focus on developing positive attitudes toward the new behavior (by influencing decisional balance). Self-efficacy interventions are expected to be useful only when people have already made that initial decision (e.g., are in the planning stage). The reasoning is that until people have become convinced of the desirability of an action, 218



there is little reason to worry about whether they think they can perform the behavior. These expectations have been assessed in several studies. These studies begin by identifying each participant’s stage, with an eye to distinguishing those in earlier (especially preplanning) stages and those in later (especially planning) stages. Then participants receive either an intervention designed for those in early stages (e.g., one aimed at encouraging a positive attitude toward the new behavior) or an intervention designed for those in later stages (one aimed at enhancing perceived self-efficacy); thus participants receive either a stage-matched intervention or a stage-mismatched intervention. The results of such studies have pointed to a rather more complex picture than that suggested by the TTM. Sometimes the expected pattern of results has obtained, such that interventions were more effective when matched to participants’ stages than when mismatched—and, specifically, selfefficacy interventions were effective for those at later stages but not for those at earlier stages. For example, Prentice-Dunn, McMath, and Cramer (2009) found that for encouraging sunscreen use, movement from precontemplation to contemplation was affected by the nature of threat appraisal information (about the dangers of sun exposure) but not by the nature of self-efficacy information (about the ease of using sunscreen). For individuals in contemplation, however, movement to the preparation stage was influenced by the presentation of self-efficacy information. That is, self-efficacy information was effective for influencing people in later stages but not people in earlier stages. However, a number of other studies have found that self-efficacy interventions can be useful even at relatively early stages of behavior change. For example, Schwarzer, Cao, and Lippke (2010) studied two interventions designed to encourage physical activity in adolescents. Some of the participants did not yet intend to engage in such activity (preintenders); others had formed physical activity intentions but had not yet acted upon them (intenders). One intervention, designed for pre-intenders, stressed the benefits of regular physical activity and the risks of a sedentary lifestyle (“resource communication”); the other, designed for intenders, focused on action planning to translate those intentions into behavior (“planning treatment”). As expected, the resource communication treatment was effective in increasing the frequency of physical activity for pre-intenders but not intenders—but the planning intervention was 219



effective for both pre-intenders and intenders. That is, a self-efficacy– focused intervention was effective even for people in an early (preintention) stage. As another example: Weinstein, Lyon, Sandman, and Cuite (1998) examined interventions for persuading people to undertake home radon testing. Two target audiences were distinguished: individuals who had not made up their minds about testing (the “undecided” group) and persons who had formed the intention to test but had not yet done so (the “decided to test” group). Two intervention messages were developed: The “highlikelihood” intervention aimed at convincing people that their homes were indeed vulnerable to the threat of radon; this intervention was taken to be matched to undecided receivers (because it was focused on developing the appropriate attitudes). The “low-effort” intervention aimed at convincing people that the testing process was simple and inexpensive; this intervention was taken to be matched to decided-to-test receivers (because it was focused on self-efficacy). As expected, for decided-to-test receivers, the low-effort intervention was dependably more successful (than the highvulnerability intervention) in moving people to a subsequent stage. But for undecided receivers, the two interventions were equally effective. That is, the intervention aimed at influencing self-efficacy was successful in producing stage progression (movement to a more advanced stage) even for people at early stages—people for whom the intervention was putatively mismatched. (For similar indications of the effectiveness of selfefficacy interventions at early stages, see Malotte et al., 2000; Quinlan & McCaul, 2000.) In short, several studies have failed to confirm the expected matching effects for self-efficacy interventions (for a review offering a similar conclusion concerning smoking specifically, see Cahill, Lancaster, & Green, 2010). Such failures might be explained in any number of ways. For example, participants might not have been correctly classified with respect to stages; if some participants who were classified as being in early stages (pre-intending, precontemplation, etc.) were actually in later stages (intending, planning, etc.), then the self-efficacy–focused intervention would in fact have been matched to those early-stage participants. Or perhaps the self-efficacy interventions inadvertently included some elements not focused on self-efficacy (but instead focused on developing more positive attitudes), which made those interventions effective for early-stage participants.



220



However, there is another reason to have doubts that self-efficacy–focused interventions should be deployed only for persons in later stages of change: research on threat appeals (discussed more extensively in Chapter 11). A threat appeal message has two components, one designed to arouse fear or anxiety about possible negative events or consequences associated with a possible threat, and one that offers a recommended course of action to avert or reduce those negative outcomes. So, for instance, a message might depict the dangers of not wearing a seat belt (the threat component) as a way of encouraging seat belt use (the recommended action). One much-studied threat-appeal message variation concerns differences in the depicted ease with which the recommended action can be adopted. Perhaps unsurprisingly, when the advocated action is depicted as easy to do, people are (ceteris paribus) subsequently more likely to intend to perform it. But the finding relevant to the present discussion is this: When the advocated action is depicted as easy to do, people are also subsequently more likely to have positive attitudes about the behavior (see the meta-analytic review of Witte & Allen, 2000).9 That is, what looks like a self-efficacy intervention can have effects on people’s attitudes. In the context of the TTM, then, the plain implication is that persuaders should not withhold self-efficacy interventions until people have already developed positive attitudes. On the contrary, self-efficacy interventions might be useful even when—or especially when—people do not yet have positive attitudes about the new behavior.10 All told, then, there is good reason to think that self-efficacy–oriented interventions can potentially be useful at earlier stages than are contemplated by the transtheoretical model. That is to say, with respect to this one aspect, the TTM’s expectations about stage-matching appear not to be well-founded.11



Broader Concerns About the Transtheoretical Model As just indicated, self-efficacy interventions may be useful at earlier stages than one would expect on the basis of the transtheoretical model. Of course, even if the TTM is mistaken about this one particular aspect of stage-matching (concerning the timing of self-efficacy interventions), that does not by itself somehow invalidate or undermine the TTM. However, two broader concerns have been raised about the TTM.



221



One continuing concern is the transtheoretical model’s description of, and procedures for assessing, the various stages. In TTM research, an individual’s stage is most commonly assessed based on answers to a small number of yes-no questions about current behavior, intentions to change, and the like. But some of the resulting stage classifications can appear artificial; for example, in some classification systems, an individual planning to stop smoking in the next 30 days is placed in the preparation stage, but an individual planning to quit in the next 31 days is described as being in the contemplation stage (see Sutton, 2000, 2005b, pp. 238–242). Moreover, the questions used (and the criteria used to subsequently classify respondents) have varied from study to study, making for difficulties in assessing the validity of such measures (Littell & Girvin, 2002). The challenge of creating reliable and valid stage assessments has received considerable attention, with a variety of alternative stagemeasurement procedures being explored, but no easy resolution is in hand (for some illustrative discussions, see Balmford, Borland, & Burney, 2008; Bamberg, 2007; de Nooijer, van Assema, de Vet, & Brug, 2005; Lippke, Ziegelmann, Schwarzer, & Velicer, 2009; Marttila & Nupponen, 2003; Napper et al., 2008; Richert, Schüz, & Schüz, 2013).12 A second general concern is the paucity of empirical support for TTMbased stage-matched interventions as compared with nonmatched interventions. As indicated earlier, the expectations of the transtheoretical model have not been confirmed concerning the specific question of the timing of self-efficacy interventions. More generally, however, several broader reviews have raised questions about the effectiveness of TTMbased stage-matched interventions as compared to nonmatched interventions (e.g., Adams & White, 2005; Bridle et al., 2005; Cahill, Lancaster, & Green, 2010; Riemsma et al., 2003; Tuah et al., 2011). To be sure, there are many ways in which doubt can creep in about such conclusions (e.g., measurement error in assessing stages, or inadequate realization of interventions; see, e.g., Hutchison, Breckon, & Johnston, 2009). But at present there is insufficient evidence to permit one to confidently conclude that TTM-based stage-matched interventions are in general more successful than nonmatched interventions.13 Other questions have been raised about, for example, whether the transtheoretical model’s stages in fact constitute mutually exclusive categories, whether there is good evidence of sequential movement through the stages, whether the model is sufficiently well-specified to permit useful empirical examination, and so forth. (For a particularly nice 222



review, see Sutton, 2005b. For other general critical discussions of the TTM, see Armitage, 2009; Herzog, 2008; Littell & Girvin, 2002; Sutton, 2000, 2005a; R. West, 2005; Whitelaw, Baldwin, Bunton, & Flynn, 2000.) These doubts about the adequacy of the transtheoretical model, however, do not necessarily undermine the general idea of a stage model of behavior change. Any failings of the TTM might reflect its particular realization of the stage approach rather than some defect with the very idea of a stage model. But before considering some alternative (non-TTM) stage models, it will be useful to reflect on the distinctive claims advanced by stage (as opposed to nonstage) approaches.



The Distinctive Claims of Stage Models Stage models such as the TTM take on a set of distinctive empirical commitments. A stage model must have a clear set of categories (with principles for assigning people to categories), must show that those categories are temporally ordered and so constitute stages, must define those stages in a way that people in the same stage face common behaviorchange challenges (and so would profit from the same interventions), and must define those stages in a way that people in different stages face different challenges (and so need different interventions). (For a careful elaboration of these evidentiary requirements, see Weinstein, Rothman, & Sutton, 1998.) As a consequence, stage models are quite different from nonstage models of behavior such as reasoned action theory (RAT; Chapter 6). Nonstage models might be called “continuum” approaches, in the sense that people are seen as arrayed along a continuum reflecting the likelihood of their engaging in the behavior (Weinstein, Rothman, & Sutton, 1998). In RAT, for instance, people’s placements on the intention continuum are seen to be influenced by four broad factors (attitude toward the behavior, injunctive norms, descriptive norms, and perceived behavioral control). Where continuum models see a range of propensities to act, with the same general factors potentially influencing everyone, stage models see qualitative differences between people on the basis of action preparation or propensity (i.e., different stages), with different factors influencing people in those different stages. If one were to arbitrarily divide the intention continuum into distinct parts, a stage-like classification might seem to result. For example, on an 11223



point intention scale, one could group together persons with scores of 1 and 2, those with scores of 3 and 4, those with scores of 5, 6, and 7, those with scores of 8 and 9, and those with scores of 10 and 11—and then conceive of these as five distinct “stages” along the way to action. But these would more appropriately be called “pseudo-stages” (Weinstein, Rothman, & Sutton, 1998). After all, there is no reason to suppose that people need to pass through these various “stages” in succession; a person might have a weak intention at one point in time (say, an intention score of 2) but subsequently be convinced to have a much stronger intention (say, a score of 10) without having to pass through the points in between. Moreover, two people in the same region of the continuum might need different kinds of treatments; for example, two people might have equally negative intentions, but (expressed in terms of RAT) one person could have a negative attitude while the other person had low perceived behavioral control—and hence those two people would need different kinds of interventions. To come at this same point from a different angle: Many theoretical perspectives on persuasion and behavior change recognize that recipients might be in different states that are relevant to behavioral performance. Correspondingly, many different approaches emphasize the importance of matching messages (interventions, treatments) to the current state of the recipient—adapting the intervention to the recipient’s current beliefs, attitude function, cultural values, personality characteristics, elaboration likelihood, and so forth. For example, although reasoned action theory identifies four abstract general influences on intention (attitude toward the behavior, injunctive norms, descriptive norms, and perceived behavioral control), RAT stresses that interventions should target only those determinants of intention that significantly influence the intention in question. That is, RAT emphasizes that interventions should be adapted to the recipients’ current state; for instance, if injunctive norms are not significantly related to the relevant intention, then injunctive-norm interventions are presumably inappropriate. Stage models share this emphasis on adapting interventions to recipients’ states—but differ by virtue of suggesting that the various different recipient states form a temporal sequence, that is, a set of stages. It’s not just that different people might be in different states (and so need different interventions); it’s that people pass through those states in a definite temporal order. The claim that the relevant states are temporally ordered (and so form a set of stages) brings with it additional burdens of proof. For 224



example, the appropriate empirical evidence for stage models must consist of something more than showing that state-matched interventions are more effective than non-matched (mismatched or unmatched) interventions. After all, any number of continuum (that is, non-stage) approaches also contain the idea that state-matched interventions will enjoy greater success than non-matched interventions. A finding that state-matched interventions were more effective than non-matched interventions does not show that the states form a sequence, that is, does not show that the states are stages.14 One kind of evidence especially relevant to the distinctive claims of stage models concerns intervention sequencing. If different recipient states represent genuine stages (i.e., are temporally sequenced), then some intervention sequences should be more effective than others. To express this point in terms of the abstract representation in Figure 7.1: If Intervention A is needed to move people from Stage 1 to Stage 2, and intervention B is needed to move people from Stage 2 to Stage 3, then it should be more effective to deliver Intervention A first followed by Intervention B, rather than the reverse. Again, approaches such as reasoned action theory provide a useful contrast. From a RAT perspective, if both attitude (AB) and descriptive norms (DN) are significant predictors of intention, then each would be an appropriate target for influence—and the sequence of such changes would be irrelevant (that is, there’d be no necessary reason to change AB before DN or vice versa). By contrast, with a stage model, the sequence of treatments is expected to be crucial. So for comparing the utility of stage models against non-stage (continuum) models such as reasoned action theory, studies of the relative effectiveness of different intervention sequences would be very helpful. In the case of the TTM (and stage models more generally), research attention has unfortunately focused on examining the relative effectiveness of stagematched and non-matched interventions without yet taking up the more complex question of intervention sequencing. This focus is understandable, especially in the early stages of research, but a genuine stage model will contain clear predictions about intervention sequencing that deserve empirical attention.



Other Stage Models 225



The transtheoretical model (TTM) is the most-studied stage model of behavior change, but several others have also been developed. For example, a number of stage models have concerned consumer purchasing behavior. These models propose that a sequence of distinct stages precedes product purchase. The description of these stages varies (in number and composition), but common elements include awareness of the brand, knowledge about the brand, attitude toward the brand, brand purchase intention, purchase, and brand loyalty. These models are often described as “hierarchy of advertising effects” models, because they putatively identify a sequence of desired effects of advertising—to make the consumer aware of the brand, to ensure the consumer has knowledge about the brand, and so forth. (For one classic treatment, see Lavidge & Steiner, 1961.) Although there is variation in the details of these hierarchy-of-effects models, they share a common shortcoming, namely, that there is not good evidence for the expected temporal sequence of advertising effects (Fennis & Stroebe, 2010, pp. 29–34; Vakratsas & Ambler, 1999; Weilbacher, 2001).15 Health-related behaviors have been the focus of several non-TTM stage models (e.g., the precaution adoption model of Weinstein & Sandman, 1992, 2002; for a general review of health-related stage models, see Sutton, 2005b). Of these, perhaps the best developed is the Health Action Process Approach (HAPA; Schwarzer & Fuchs, 1996). The HAPA identifies three stages in behavioral change: non-intention, intention, and action. Those in the non-intention stage (“non-intenders”) have not formed the relevant intention; those in the intention stage (“intenders”) have formed the intention but have not acted on it; those in the action stage (“actors”) have engaged in the new behavior.16 The HAPA suggests that different kinds of messages are appropriate at different stages. For example, information about the disadvantages of the current behavior or the advantages of the new behavior (e.g., the benefits of regular exercise) is likely to be most relevant for non-intenders, whereas intenders may need explicit planning advice (about how to translate their intentions into action). But the HAPA suggests that self-efficacy information (about one’s capability to perform the new behavior) is potentially valuable at all stages. (For a general review of HAPA research, see Schwarzer, 2008a. For some illustrative specific applications, see Craciun, Schüz, Lippke, & Schwarzer, 2012; Parschau et al., 2012; Schüz,, Sniehotta, Mallach, Wiedemann, & Schwarzer, 2009; Schüz, Sniehotta, & Schwarzer, 2007. For some commentaries on the HAPA, see Abraham, 226



2008; Conner, 2008; Leventhal & Mora, 2008; Sutton, 2008.) However, there is room for doubt about whether the HAPA is in fact a stage model, as opposed to being something more like a continuum model (see especially Sutton, 2005b). Notice that the HAPA stages correspond to different portions of an intention continuum: Pre-intenders presumably have negative (or insufficiently positive) intentions, and intenders and actors presumably have positive intentions (but are distinguished by whether they have acted on those intentions). That is, HAPA seems to divide the propensity-to-act (intention) continuum into two general categories: those with negative intentions and those with positive intentions. This differentiation, however, is arguably insufficiently sharp to create a genuine stage distinction. The intention continuum has no natural boundary (the mean? the median? the scale midpoint?), but rather offers only a blurry distinction between intenders and non-intenders (Abraham, 2008). The implication is that the HAPA’s demarcation of intenders and non-intenders seems an arbitrary division rather than a genuine stage distinction. At the same time, some such differentiation does seem potentially useful. In this context it may be illuminating to consider the HAPA against the backdrop of reasoned action theory. From the perspective of RAT, one set of factors underlies the formation of positive intentions (intention formation is influenced by attitude toward the behavior, injunctive norms, descriptive norms, and perceived behavioral control), but the realization of intentions in action involves different processes.17 In this regard, RAT and the HAPA look rather similar, in that each (implicitly or explicitly) recognizes a distinction between getting people to have the desired intention and getting people to act on that intention.18 In any event, the HAPA is not a straightforward stage model. It might be thought of instead as a continuum theory, akin to—and indeed an alternative to—reasoned action theory (see Sutton, 2005b), or as an attempt to somehow blend stage and continuum approaches (see Schwarzer, 2008b; cf. Conner, 2008).



Conclusion Stage models of behavioral change are naturally quite appealing. The idea that behavior change requires movement through a sequence of stages 227



sounds like a very plausible idea, but it turns out to be surprisingly difficult to redeem that abstract idea in empirically and conceptually sound ways. The emphasis that stage models place on matching treatments to recipients is valuable but not unique. What makes stage models distinctive is the suggestion of a set of distinct stages—temporally ordered states through which people pass in the course of behavior change. However, defining these stages clearly, developing sound assessments of stages, identifying stage-appropriate interventions, displaying the greater effectiveness of stage-matched treatments over nonmatched treatments, showing the expected effects of variation in intervention sequencing—these have proved difficult challenges to surmount.



For Review 1. How do stage models describe the process of behavioral change? In the transtheoretical model (TTM), what stages are distinguished? Describe each stage: precontemplation, contemplation, planning, action, and maintenance. Is stage movement a linear process? Explain why different kinds of treatments (messages) might be needed for people at different stages of change. 2. What is decisional balance? How does decisional balance change as people progress through the stages? Is the size of the change in the perceived importance of the pros the same as the size of the decrease in the perceived importance of the cons? Which is larger? Describe the implications of such a difference for the design of effective interventions. Explain why the existing research evidence about decisional balance does not necessarily support recommendations about intervention design. 3. Explain the general idea of stage-matching, and identify challenges in comparing the effectiveness of stage-matched and stage-mismatched interventions. According to the TTM, when should self-efficacy interventions be most effective, at early stages or later ones? What does the research evidence indicate about the appropriate timing of self-efficacy interventions? 4. Describe some concerns that have been raised about the TTM’s procedures for assessing stages. How extensive is the empirical evidence for the superiority of TTM-based stage-matched interventions over nonmatched interventions? 5. Describe the distinctive claims of stage models. Explain the 228



difference between stage models and continuum models of behavior. How can dividing an intention continuum into segments produce the appearance of a stage-like set of categories? Why is such a set of categories not a genuine stage model? Describe the difference between thinking of recipients as being in different states and thinking of them as being in different stages. Why would finding that state-matched interventions are more effective than nonmatched interventions not necessarily show that the states are stages? Explain why, according to stage models, some intervention sequences should be more effective than others. 6. Describe some stage models other than the transtheoretical model (the TTM). Sketch the kinds of stages included in a hierarchy-ofadvertising-effects model. Describe the stages identified by the Health Action Process Approach (HAPA). Is the HAPA a straightforward stage model? Explain.



Notes 1. The number and description of the stages can vary, especially depending on the application area. For example, some TTM presentations focused on smoking cessation have described the maintenance stage as one in which the ex-smoker is actively working to prevent relapse, and have added a subsequent “termination” stage in which the ex-smoker has achieved complete self-control (Prochaska, Redding, & Evers, 2002). 2. The various processes of change appear to form a haphazard collection of activities and circumstances that some smokers have sometimes found useful in quitting rather than a principled or systematically organized set of fundamental processes. This naturally impairs their value in guiding research or intervention design. 3. Viewed from the perspective of reasoned action theory (RAT; see Chapter 6), decisional balance can be seen to be related to the person’s attitude toward the new behavior (AB), with self-efficacy reflecting perceived behavioral control (PBC). But the TTM’s treatment of these two elements is rather less careful than that of RAT, and TTM does not emphasize normative considerations (injunctive or descriptive) in the way that RAT does. 4. Decisional balance was originally developed as part of Janis and Mann’s (1968) account of decision making. 229



5. The TTM’s research evidence focuses exclusively on the perceived importance of pros and cons, but—as suggested by belief-based models of attitude (see Chapter 4)—other properties of perceived advantages and disadvantages (such as evaluation or perceived likelihood) might arguably be of at least as much interest to persuaders. It should not pass unnoticed that, as discussed in Chapter 4, when the evaluation and strength (likelihood) of salient beliefs are used to predict attitude, belief importance (the property apparently of key interest to the TTM) does not add to the predictability of attitude. 6. Again, common ways of expressing this finding can be misleading. Consider, for example: “the magnitude of the increase in the pros was greater than the magnitude of the decrease in the cons” (Di Noia & Prochaska, 2010, p. 629). This might more carefully be expressed as “the magnitude of the increase in the perceived importance of the pros was greater than the magnitude of the decrease in the perceived importance of the cons,” so as not to invite misapprehension of the property under discussion. 7. A divergent effect was reported for organ donation decisions by McIntyre et al. (1987). Organ donors and nondonors generally did not differ in their ratings of the importance of various reasons for donating (pros) but did often differ in their ratings of the importance of reasons for not donating (cons). 8. TTM presentations of these principles are not always as clear as might be wanted. For instance, Prochaska, Redding, and Evers (2002, p. 106) misleadingly assert that one of the “practical implications” of the strong and weak principles is that increasing the number of perceived advantages (pros) should be emphasized more than decreasing the number of perceived disadvantages (cons). The research evidence concerning decisional balance, however, does not address variations in the number of pros or cons, only in their perceived importance. 9. In Witte and Allen’s (2000) meta-analysis, the mean effect of variation in depicted self-efficacy on attitudes, expressed as a correlation, was .12 (across eight cases). For comparison, the mean effect of variation in depicted self-efficacy on perceived self-efficacy was .36 (across 17 cases; Witte & Allen, 2000, p. 599, Table 2). This latter effect, however, was based only on cases with a successful “manipulation check” (that is, studies in which the experimental variation in depicted self-efficacy 230



produced statistically significant differences in perceived self-efficacy); if studies with “failed” manipulation checks were to have been included, the observed mean effect would probably be smaller. 10. Expressed in the framework of reasoned action theory (Chapter 6), these findings indicate that an intervention meant to influence perceived behavioral control (PBC) can also influence behavioral attitudes (AB). Hence if AB is negative, an intervention focused on PBC (self-efficacy) might nevertheless be useful, because that intervention might both enhance PBC and make AB more positive. 11. One might also suppose that an assumption of some strict temporal segregation between attitudinal considerations (the person’s evaluation of the new behavior) and self-efficacy considerations (the person’s perceptions of their ability to perform the action) is implausible. The implicit suggestion of the TTM is that if people have a negative attitude toward the new behavior, then they won’t even think about self-efficacy considerations; the only time people think about self-efficacy (according to this view) is when they already have a positive attitude (and so have reached a stage at which self-efficacy becomes relevant). But if this were the case, then no one would ever simultaneously express a negative attitude and self-efficacy doubts—no one would ever say, for example, “I don’t think exercise is all that valuable, and besides I don’t have time for it anyway.” And yet this seems like a perfectly natural state of mind. In fact, in a circumstance in which people have both a negative attitude toward the desired behavior and doubts about their ability to perform it, persuaders might well sometimes want to address those self-efficacy concerns first. For instance, low perceived self-efficacy might sometimes be a rationalization device, a way of justifying failing to do something that people know they should do (“the recycling rules are so hard to understand”). And when low perceived self-efficacy does function this way, then removal of that rationalization could be crucially important for encouraging behavioral change. 12. One potential issue with stage models is that stages can be defined in ways that evade (or submerge) certain empirical questions. As one illustration, stage definitions can guarantee a certain sequencing of stages. For example, suppose a “precontemplation” stage is defined in such a way that once a person has thought about adopting the new behavior, by definition that person cannot ever be in precontemplation again. In that case, a “contemplation” stage would by definition follow precontemplation 231



—which would mean that no empirical evidence would be needed to confirm that temporal relation. Similarly, stage definitions can guarantee that a given stage cannot possibly be skipped. For example, if part of the definition of a “planning” stage is that the person must have already at least been thinking about adopting the new behavior (i.e., must already have been in a “contemplation” stage), then it would be conceptually impossible for a person to reach the “planning” stage without passing through the “contemplation” stage. The larger point here is that the development of reliable and valid stage assessments is bound up with definitional questions—and certain ways of defining stages can transform what would look to be empirical questions (“Does stage Y always follow stage X?”) into definitional matters (“Stage Y, by definition, follows stage X”). 13. This conclusion is not inconsistent with a belief that state-matched interventions are generally more effective than nonmatched interventions (e.g., Noar, Benac, & Harris, 2007), or with a belief than non-TTM-based stage-matched interventions are more effective than nonmatched interventions, or with a belief that TTM-based stage-matched interventions are more effective than a no-treatment control condition. 14. As a parallel case: One could imagine devising different messages that were matched to differing political dispositions (say, Republicans and Democrats in the U.S.). Finding that Republicans were more persuaded by messages matched to Republicans than by messages matched to Democrats (with the reverse true for Democrats) would not show that the two categories (Democrat and Republican) formed any sort of temporal sequence. Similarly here: Finding that people in the contemplation category were more persuaded by interventions matched to contemplation than by messages matched to planning (with the reverse true for people in the planning category) would not show that the two categories form any sort of temporal sequence. Showing that stage-matched interventions are more successful than nonmatched interventions is a minimum requirement for a valid stage theory but does not satisfy all the burdens of proof incurred by stage models. 15. Notice, however, that even if the sequence of stages posited by a hierarchy-of-advertising-effects model is not accurate, those different “stages” might nevertheless be useful as a conceptualization of different possible purposes of advertising. So, for example, the challenges of an advertiser may be different if consumers don’t know the product exists 232



than if they know it exists but believe it’s not very good. 16. Some presentations of the HAPA have described just two stages—a motivational stage (in which persons need to develop the appropriate motivations to change) and a volitional stage (in which people already have the relevant motivations)—but with the volitional stage further divided between people who merely intend to adopt the new behavior and those who have already adopted that behavior. Because such an analysis eventuates in three relevant categories, it seems better to describe the HAPA as proposing three (not two) stages. However, the “two-stage” language does draw attention to the similarity between those in the intention stage and those in the action stage, namely, that they all have the requisite intention (which distinguishes them from those in the nonintention stage). 17. On this latter subject (the intention-behavior relationship), RAT has focused on the question of what influences the intention-behavior correlation, so as to be able to better specify the conditions under which intentions will predict behavior. RAT itself has not given so much attention to the question of how to influence the intention-behavior relationship. 18. This distinction, or something like it, has a long intellectual history, dating back at least to 18th-century faculty psychology. A distinction between “conviction” (influencing belief, associated with the faculty of understanding) and “persuasion” (influencing action, associated with the faculty of the will) can be found in George Campbell’s 1776 The Philosophy of Rhetoric and even more sharply formulated in Bishop Richard Whately’s 1828 The Elements of Rhetoric. However, this conviction-persuasion distinction needlessly confused distinctions between communicative purposes and distinctions between communicative means, with unhappy consequences (O’Keefe, 2012a).



233



Chapter 8 Elaboration Likelihood Model Variations in the Degree of Elaboration: Central Versus Peripheral Routes to Persuasion The Nature of Elaboration Central and Peripheral Routes to Persuasion Consequences of Different Routes to Persuasion Factors Affecting the Degree of Elaboration Factors Affecting Elaboration Motivation Factors Affecting Elaboration Ability Summary Influences on Persuasive Effects Under Conditions of High Elaboration: Central Routes to Persuasion The Critical Role of Elaboration Valence Influences on Elaboration Valence Summary: Central Routes to Persuasion Influences on Persuasive Effects Under Conditions of Low Elaboration: Peripheral Routes to Persuasion The Critical Role of Heuristic Principles Varieties of Heuristic Principles Summary: Peripheral Routes to Persuasion Multiple Roles for Persuasion Variables Adapting Persuasive Messages to Recipients Based on the ELM Commentary The Nature of Involvement Argument Strength One Persuasion Process? Conclusion For Review Notes



The elaboration likelihood model (ELM) of persuasion is an approach developed by Richard Petty, John Cacioppo, and their associates (the most comprehensive single treatment of the ELM is provided by Petty & Cacioppo, 1986a; for briefer presentations, see Petty & Briñol, 2010, 2012a; Petty, Cacioppo, Strathman, & Priester, 2005). The ELM suggests that important variations in the nature of persuasion are a function of the likelihood that receivers will engage in elaboration of (that is, thinking about) information relevant to the persuasive issue. Depending on the 234



degree of elaboration, two types of persuasion process can be engaged (one involving systematic thinking and the other involving cognitive shortcuts)—with different factors influencing persuasive outcomes in each. In the sections that follow, the nature of variations in the degree of elaboration are described, factors influencing the degree of elaboration are discussed, the two persuasion processes are described, and then various complexities of persuasion processes are considered. The ELM is an example of a dual-process approach to social informationprocessing phenomena (see Chaiken & Trope, 1999), an example focused specifically on persuasion phenomena. An alternative dual-process image of persuasion has been provided by the heuristic-systematic model (HSM; see Chaiken, 1987; S. Chen & Chaiken, 1999; Todorov, Chaiken, & Henderson, 2002). Although the two models differ in some important ways, the ELM and HSM share the broad idea that persuasion can be achieved through two general avenues that vary in the amount of careful thinking involved.



Variations in the Degree of Elaboration: Central Versus Peripheral Routes to Persuasion The Nature of Elaboration The ELM is based on the idea that under different conditions, receivers will vary in the degree to which they are likely to engage in elaboration of information relevant to the persuasive issue. Elaboration here refers to engaging in issue-relevant thinking. Thus sometimes receivers will engage in extensive issue-relevant thinking: They will attend closely to a presented message, carefully scrutinize the arguments it contains, reflect on other issue-relevant considerations (e.g., other arguments recalled from memory or arguments they devise), and so on. But sometimes receivers will not undertake so much issue-relevant thinking; no one can engage in such effort for every persuasive topic or message, and hence sometimes receivers will display relatively little elaboration.1 A number of means have been developed for assessing variations in the degree of elaboration that occurs in a given circumstance (for discussion, see Petty & Cacioppo, 1986a, pp. 35–47). Perhaps the most straightforward of these is the thought-listing technique: Immediately following the receipt of a persuasive message, receivers are simply asked 235



to list the thoughts that occurred to them during the communication (for a more detailed description, see Cacioppo, Harkins, & Petty, 1981, pp. 38– 47; for a broad review of such techniques, see Cacioppo, von Hippel, & Ernst, 1997). The number of issue-relevant thoughts reported is presumably at least a rough index of the amount of issue-relevant thinking.2 Of course, the reported thoughts can also be classified in any number of ways (e.g., according to their substantive content or according to what appeared to provoke them); one classification obviously relevant to the illumination of persuasive effects is one that categorizes thoughts according to their favorability to the position being advocated by the message. As is probably already apparent, the degree to which receivers engage in issue-relevant thinking forms a continuum, from cases of extremely high elaboration to cases of little or no elaboration. One might be tempted to think that in circumstances in which little or no elaboration occurs, little or no persuasion will occur; after all, the receiver has not really engaged the message. But the ELM suggests that persuasion can take place at any point along the elaboration continuum, although the nature of the persuasion processes will be different as the degree of elaboration varies. To bring out the differences in these persuasion processes, the ELM offers a broad distinction between two routes to persuasion: a central and a peripheral route.



Central and Peripheral Routes to Persuasion The central route to persuasion represents the persuasion processes involved when elaboration is relatively high. When persuasion is achieved through the central route, it commonly comes about through extensive issue-relevant thinking: careful examination of the information contained in the message, close scrutiny of the message’s arguments, consideration of other issue-relevant material (e.g., arguments recalled from memory, arguments devised by the receiver), and so on. In short, persuasion through the central route is achieved through the receiver’s thoughtful examination of issue-relevant considerations. The peripheral route represents the persuasion processes involved when elaboration is relatively low. When persuasion is achieved through peripheral routes, it commonly comes about because the receiver employs some simple decision rule (some heuristic principle) to evaluate the advocated position. For example, receivers might be guided by whether 236



they like the communicator or by whether they find the communicator credible. That is, receivers may rely on various peripheral cues (such as communicator credibility) as guides to attitude and belief, rather than engaging in extensive issue-relevant thinking. Thus as elaboration decreases, peripheral cues presumably become progressively more important determinants of persuasive effects, but as elaboration increases, peripheral cues should have relatively smaller effects on persuasive outcomes. Indeed, one indirect marker of the amount of elaboration (in a given circumstance) is precisely the extent to which observed persuasive effects are a function of available peripheral cues as opposed to (for example) the quality of the message’s arguments. If, in a given experimental condition, variations in peripheral cues have more influence on persuasive outcomes than do variations in the strength of the message’s arguments, then presumably relatively little elaboration occurred. That is, the persuasive outcomes were presumably achieved through a peripheral, not a central, route (but see Bless & Schwarz, 1999). This distinction between the two routes to persuasion should not be permitted to obscure the underlying elaboration continuum. The central and peripheral routes to persuasion are not two exhaustive and mutually exclusive categories or kinds of persuasion; they simply represent prototypical extremes on the high-to-low elaboration continuum. The ELM recognizes, for example, that at moderate levels of elaboration, persuasion involves a mixture of central route and peripheral route processes, with correspondingly complex patterns of effects (see, e.g., Petty & Wegener, 1999, pp. 44–48). Thus, in considering the differing character of persuasion achieved through central and peripheral routes, it is important to bear in mind that these routes are offered as convenient idealized cases representing different points on the elaboration continuum. A classic illustration of the distinction between central and peripheral routes to persuasion is provided by Petty, Cacioppo, and Goldman’s (1981) study of the effects of argument strength and communicator expertise on persuasive effectiveness. In this investigation, the personal relevance of the message topic for receivers was varied, such that for some receivers the topic was quite relevant (and so presumably disposed receivers to engage in high elaboration), whereas for other receivers the topic was much less relevant (and hence these receivers would presumably be less likely to engage in elaboration). The design also varied the quality of the message’s arguments (strong vs. weak arguments) and the expertise 237



of the communicator (high vs. low). High-topic-relevance receivers were significantly affected by the quality of the arguments contained in the message (being more persuaded by strong arguments than by weak arguments) but were not significantly influenced by the communicator’s degree of expertise. By contrast, low-topicrelevance receivers were more affected by expertise variations (being more persuaded by the high-expertise source than by the low) than by variations in argument quality. That is, when receivers were inclined (by virtue of topic relevance) to engage in extensive elaboration, the results of their examination of the message’s arguments were much more influential than was the peripheral cue of the communicator’s expertise. But when receivers were not inclined to invest the cognitive effort in argument scrutiny, the peripheral cue of expertise had more influence. As this investigation indicates, persuasion can be obtained either through a central route (involving relatively high elaboration) or through a peripheral route (in which little elaboration occurs). But the factors influencing persuasive success are different in the two cases. Moreover, as indicated in the next section, the consequences of persuasion are not identical for the two routes.



Consequences of Different Routes to Persuasion Although persuasion can be accomplished at any point along the elaboration continuum, the persuasive effects obtained are not necessarily identical. The ELM suggests that with variations in the amount of elaboration (i.e., variations in the route to persuasion), there are corresponding variations in the character of the persuasive outcomes effected. Specifically, the ELM suggests that attitudes shaped under conditions of high elaboration will (compared with attitudes shaped under conditions of low elaboration) display greater temporal persistence, be more predictive of intentions and subsequent behavior, and be more resistant to counterpersuasion. Each of these claims enjoys both some supportive direct research evidence and some previous research that can be interpreted as indicating such effects. For example, Petty, Cacioppo, and Schumann (1983) reported that attitudes were more strongly correlated with intentions when the attitudes were formed under conditions of high (as opposed to low) personal relevance of the topic. Cacioppo, Petty, Kao, and Rodriguez (1986) found 238



that persons high in need for cognition (and so presumably higher in elaboration motivation) displayed greater attitude-intention and attitudebehavior consistency than did persons lower in need for cognition. Verplanken (1991) reported greater persistence of attitudes and greater attitude-intention consistency under conditions of high (rather than low) elaboration likelihood (as indicated by topic relevance and need for cognition). MacKenzie and Spreng (1992) experimentally varied elaboration motivation and found stronger attitude-intention relationships under conditions of higher (as opposed to lower) elaboration motivation. (For other illustrations, see Gasco, Briñol, & Horcajo, 2010; Haugtvedt, Schumann, Schneier, & Warren, 1994. For some general reviews and discussions, see Petty & Cacioppo, 1986a, pp. 173–195; Petty, Haugtvedt, & Smith, 1995; Petty & Wegener, 1999, pp. 61–63.) These effects may seem intuitively plausible (in the sense that the greater issue-relevant thinking affiliated with central route processes might well be expected to yield attitudes that are stronger in these ways), but the mechanism by which these outcomes arise is not entirely well understood (for some discussion, see Petty, Haugtvedt, & Smith, 1995, pp. 119–123).3 Nevertheless, there is good reason for persuaders to presume that persuasion accomplished through high elaboration is likely to be more enduring (less likely to decay through time, less likely to succumb to counterpersuasion) and to be more directive of behavior than is persuasion accomplished through low elaboration. Given that the underlying persuasion process varies depending on the level of elaboration, and given that the different routes to persuasion have these different consequences, it becomes important to consider what factors influence the degree of elaboration that receivers are likely to undertake.



Factors Affecting the Degree of Elaboration Two broad classes of factors influence the degree of elaboration that a receiver will likely undertake in any given circumstance. One concerns the receiver’s motivation for engaging in elaboration, the other the receiver’s ability to engage in such elaboration. For extensive elaboration to occur, both ability and motivation must be present. High elaboration will not occur if the receiver is motivated to undertake issue-relevant thinking but is unable to do so, nor will it occur if the receiver is able to engage in elaboration but is unmotivated to do so. 239



Factors Affecting Elaboration Motivation A variety of factors have received research attention as influences on receivers’ motivation to engage in issue-relevant thinking, including the receiver’s mood (e.g., Banas, Turner, & Shulman, 2012; Bless, Mackie, & Schwarz, 1992; Bohner & Weinerth, 2001; Côté, 2005; Ziegler, 2010; but also see Bless & Schwarz, 1999),4 attitudinal ambivalence (i.e., the degree to which the attitude is based on a mixture of positive and negative elements; e.g., Hänze, 2001; Jonas, Diehl, & Brömer, 1997; Maio, Esses, & Bell, 2000), and perceived information sufficiency (B. B. Johnson, 2005; Trumbo, 1999). Two influences are discussed here as illustrative: the personal relevance of the topic to the receiver and the receiver’s degree of need for cognition.



Personal Relevance (Involvement) The most studied influence on the receiver’s motivation for engaging in issue-relevant thinking is the personal relevance of the topic to the receiver. As a given issue becomes increasingly personally relevant to a receiver, the receiver’s motivation for engaging in thoughtful consideration of that issue presumably increases—and indeed a number of investigations have reported findings confirming this expectation (e.g., Petty & Cacioppo, 1979b, 1981; Petty, Cacioppo, & Goldman, 1981; Petty, Cacioppo, & Schumann, 1983).5 The ELM’s research evidence on this matter has employed a clever methodological innovation (introduced by Apsler & Sears, 1968). In many earlier studies of the effect of topic relevance variations on persuasive processes, researchers commonly employed two message topics, one presumably quite relevant for the population from which receivers were drawn and one not so relevant. This obviously creates difficulties in interpreting experimental results because any observed differences between high- and low-relevance conditions might be due not to the relevance differences but to some factor connected to the topic differences (e.g., the necessarily different arguments used in the messages on the two topics). The procedure followed by ELM researchers is exemplified in a study in which the participants were college undergraduates. The persuasive messages advocated the adoption of senior comprehensive examinations as 240



a graduation requirement—either at the receivers’ college (the highrelevance condition) or at a different, distant college (the low-relevance condition). With this form of manipulation, receivers in parallel high- and low-relevance conditions could hear messages identical in every respect (e.g., with the same arguments and evidence) save for the name of the college involved, thus simplifying interpretation of experimental findings (Petty & Cacioppo, 1979b). A note about terminology: In ELM research reports, these variations in personal relevance have often been labeled as variations in the receiver’s level of “involvement” with the message topic (and so, for instance, in the high-relevance condition, receivers would be said to be “highly involved” with the topic). But in persuasion research, the term involvement has also been used to cover other variations in the sort of relationship that message recipients have to the topic of advocacy, including the person’s judgment of the importance of the issue, the degree to which the person is strongly committed to a stand on the issue, the extent to which the person’s sense of self is connected to the stand taken, and ego-involvement (see Chapter 2), an omnibus concept meant to encompass a number of such elements. But these are different properties, so it is important to bear in mind that the involvement manipulations in ELM research are specifically ones that induce variation in the personal relevance of the topic.6



Need for Cognition A second factor influencing elaboration motivation is the receiver’s level of need for cognition (NFC), which refers to “the tendency for an individual to engage in and enjoy thinking” (Cacioppo & Petty, 1982, p. 116). This tendency varies among people; some persons are generally disposed to enjoy and engage in effortful cognitive undertakings, whereas others are not. Need-for-cognition scales have been developed to assess this individual difference (e.g., Cacioppo & Petty, 1982). Persons high in NFC tend to agree with statements such as “I really enjoy a task that involves coming up with new solutions to problems” and “I like to have the responsibility of handling a situation that requires a lot of thinking,” whereas individuals low in NFC are more likely to agree with statements such as “I like tasks that require little thought once I’ve learned them” and “I think only as hard as I have to.” (For general reviews of research on need for cognition, see Cacioppo, Petty, Feinstein, & Jarvis, 1996; Petty, Briñol, Loersch, & McCaslin, 2009.)



241



As one might suppose, a good deal of research suggests that need for cognition influences elaboration likelihood. Persons high in NFC are likely to report a larger number of issue-relevant thoughts (following message exposure) than are persons low in need for cognition (e.g., S. M. Smith, Haugtvedt, & Petty, 1994; for a review, see Cacioppo, Petty, et al., 1996, pp. 230–231). Relatedly, those high in NFC are more influenced by the quality of the message’s arguments than are those low in need for cognition (e.g., Axsom, Yates, & Chaiken, 1987; Green, Garst, Brock, & Chung, 2006; for a review, see Cacioppo, Petty, et al., 1996, pp. 229– 230).7 Such findings, of course, are consistent with the supposition that persons high in need for cognition have generally greater motivation for engaging in issue-relevant thinking than do persons low in need for cognition.8



Factors Affecting Elaboration Ability Several possible influences on receivers’ ability to engage in issue-relevant thinking have been investigated, including such variables as message repetition (Claypool, Mackie, Garcia-Marques, McIntosh, & Udall, 2004) and the receiver’s body posture (Petty, Wells, Heesacker, Brock, & Cacioppo, 1983). Two factors with relatively more extensive research support are discussed here: the presence of distraction in the persuasive setting and the receiver’s prior knowledge about the persuasive topic.



Distraction In this context, distraction refers to the presence of some distracting stimulus or task accompanying a persuasive message. Research concerning the effects of such distractions has used a variety of forms of distraction, including having an audio message be accompanied by static or beep sounds and having receivers monitor a bank of flashing lights, copy a list of two-digit numbers, or record the location of an X flashing from time to time on a screen in front of them (for a general discussion of such manipulations, see Petty & Brock, 1981). The theoretical importance of distraction effects to the ELM should be plain. Under conditions that would otherwise produce relatively high elaboration, distraction should interfere with such issue-relevant thinking. Such interference should enhance persuasion in some circumstances and reduce it in others. Specifically, if a receiver would ordinarily be inclined 242



to engage in favorable elaboration (that is, to predominantly have thoughts favoring the advocated position), then distraction, by interfering with such elaboration, would presumably reduce persuasive effectiveness. But if a receiver would ordinarily be inclined to predominantly have thoughts unfavorable to the position advocated, then distraction should presumably enhance the success of the message (by interfering with the having of those unfavorable thoughts).9 Distraction’s effects on persuasion have been extensively studied, although regrettably little of this research is completely suitable for assessing the predictions of the ELM (for some general discussions of this literature, see Baron, Baron, & Miller, 1973; Buller & Hall, 1998; Petty & Brock, 1981). But what relevant evidence exists does seem largely compatible with the ELM. For example, studies reporting that distraction enhances persuasive effects have commonly relied on circumstances in which elaboration likelihood was high and predominantly unfavorable thoughts would be expected (see Petty & Brock, 1981, p. 65). More direct tests of the ELM’s predictions have also been generally supportive (for a review, see Petty & Cacioppo, 1986a, pp. 61–68). For instance, one study found that increasing distraction increased the effectiveness of a counterattitudinal message containing weak arguments but decreased the effectiveness of a counterattitudinal message containing strong arguments. The weakargument message ordinarily evoked predominantly unfavorable thoughts, and hence distraction—by interfering with such thoughts—enhanced persuasion for that message, but the strong-argument message ordinarily evoked predominantly favorable thoughts, and thus distraction inhibited persuasion for that message (Petty, Wells, & Brock, 1976, Experiment 1; see, relatedly, Albarracín & Wyer, 2001; Jeong & Hwang, 2012; Miarmi & DeBono, 2007).



Prior Knowledge A second factor influencing elaboration ability is the receiver’s prior knowledge about the persuasive topic: The more extensive such prior knowledge, the better able the receiver is to engage in issue-relevant thinking. Several studies have indicated that as the extent of receivers’ prior knowledge increases, more issue-relevant thoughts occur, the influence of argument strength on persuasive effects increases, and the influence of peripheral cues (such as source likability and message length) decreases (e.g., Averbeck, Jones, & Robertson, 2011; Laczniak, Muehling, & Carlson, 1991; W. Wood, 1982; W. Wood & Kallgren, 1988; W. Wood, 243



Kallgren, & Preisler, 1985).10 As one might expect, this suggests that when receivers with extensive prior knowledge encounter a counterattitudinal message, such receivers are better able to generate counterarguments and hence are less likely to be persuaded (in comparison with receivers with less extensive topic knowledge). But receivers with extensive prior knowledge are also more affected by variations in message argument strength; hence increasing the strength of a counterattitudinal message’s arguments will presumably enhance persuasion for receivers with extensive knowledge but will have little effect on receivers with less extensive knowledge.11



Summary As should be apparent, a variety of factors can influence the likelihood of elaboration in a given circumstance by affecting the motivation or the ability to engage in issue-relevant thinking. With variations in elaboration likelihood, of course, different sorts of persuasion processes are engaged: As elaboration increases, peripheral cues have diminished effects on persuasive outcomes, and central route processes play correspondingly greater roles. But the factors influencing persuasive effects are different, depending on whether central or peripheral routes to persuasion are followed. Thus the next two sections consider what factors influence persuasive outcomes when elaboration likelihood is relatively high and when it is relatively low.



Influences on Persuasive Effects Under Conditions of High Elaboration: Central Routes to Persuasion The Critical Role of Elaboration Valence Under conditions of relatively high elaboration, the outcomes of persuasive efforts will largely depend on the outcomes of the receiver’s thoughtful consideration of issue-relevant arguments (as opposed to simple decision principles activated by peripheral cues). Broadly put, when elaboration is high, persuasive effects will depend on the predominant valence (positive or negative) of the receiver’s issue-relevant thoughts: To the extent that the receiver is led to have predominantly favorable thoughts 244



about the advocated position, the message will presumably be relatively successful in eliciting attitude change in the desired direction; but if the receiver has predominantly unfavorable thoughts, then the message will presumably be relatively unsuccessful. Thus the question becomes: Given relatively high elaboration, what influences the predominant valence (the overall evaluative direction) of elaboration?



Influences on Elaboration Valence Of the many influences on the evaluative direction of receivers’ issuerelevant thinking, two factors merit attention here: whether the message’s advocated position is proattitudinal or counterattitudinal and the strength (quality) of the message’s arguments.



Proattitudinal Versus Counterattitudinal Messages The receiver’s initial attitude and the message’s advocated position, considered jointly, will surely influence the valence of elaboration. When the advocated position is one toward which the receiver is already favorably inclined—that is, when the message advocates a pro-attitudinal position—the receiver will presumably ordinarily be inclined to have favorable thoughts about the position advocated. By contrast, when the message advocates a counter-attitudinal position, receivers will ordinarily be inclined to have unfavorable thoughts about the view being advocated. That is, everything else being equal, one expects proattitudinal messages to evoke predominantly favorable thoughts and counterattitudinal messages to evoke predominantly unfavorable thoughts. But of course, this cannot be the whole story—otherwise nobody would ever be persuaded by a counterattitudinal message. At least sometimes people are persuaded by the arguments contained in counterattitudinal communications, and hence the ELM suggests that a second influence on elaboration valence is the strength of the message’s arguments.



Argument Strength Recall that under conditions of high elaboration, receivers are motivated (and able) to engage in extensive issue-relevant thinking, including careful examination of the message’s arguments. Presumably, then, the valence of receivers’ elaboration will depend (at least in part) on the results of such 245



scrutiny: The more favorable the reactions evoked by that scrutiny of message material, the more effective the message should be. If a receiver’s examination of the message’s arguments reveals shoddy arguments and bad evidence, one presumably expects little persuasion; but a different outcome would be expected if the message contains powerful arguments, sound reasoning, good evidence, and the like. That is, under conditions of high elaboration, the strength (quality) of the message’s arguments should influence the evaluative direction of elaboration (and hence should influence persuasive success). Many investigations have reported results indicating just such effects (e.g., Lee, 2008; Levitan & Visser, 2008; Petty & Cacioppo, 1979b; Petty, Cacioppo, & Schumann, 1983; for complexities, see Park, Levine, Westermann, Orfgen, & Foregger, 2007).



Other Influences on Elaboration Valence Some research has examined other possible influences on elaboration valence. For example, some evidence indicates that when elaboration likelihood is high, warning receivers of an impending counterattitudinal message can encourage receivers to have more unfavorable thoughts about the advocated position than they otherwise would have had (e.g., Petty & Cacioppo, 1979a; for a review, see W. Wood & Quinn, 2003). As another example, when elaboration is high, the receiver’s mood can incline the receiver to have mood-congruent thoughts, so that positive moods encourage positive thoughts (e.g., Wegener & Petty, 2001; for complexities, see, e.g., Agrawal, Menon, & Aaker, 2007; Mitchell, Brown, Morris-Villagran, & Villagran, 2001). But the greatest research attention concerning influences on elaboration valence has been given to variations in argument strength.



Summary: Central Routes to Persuasion Under conditions of high elaboration (e.g., high personal relevance of the topic to the receiver), the outcome of persuasive efforts depends on the valence of receivers’ elaboration: When a persuasive message leads receivers to have predominantly favorable thoughts about the position being advocated, persuasive success is correspondingly more likely. And the valence of receivers’ elaboration will depend (at least in part) on the character of the message’s arguments.12 246



Influences on Persuasive Effects Under Conditions of Low Elaboration: Peripheral Routes to Persuasion The Critical Role of Heuristic Principles The ELM suggests that under conditions of relatively low elaboration, the outcomes of persuasive efforts will not generally turn on the results of the receiver’s thoughtful consideration of the message’s arguments or other issue-relevant information. Instead, persuasive effects will be much more influenced by the receiver’s use of simple decision rules or heuristic principles.13 These heuristic principles (or heuristics, for short) represent simple decision procedures requiring little information processing. The principles are activated by peripheral cues, that is, by extrinsic features of the communication situation such as the characteristics of the communicator (e.g., credibility). For example, in a circumstance in which elaboration likelihood is low, receivers may display agreement with a liked communicator because a simplifying decision rule (“If I like the source, I’ll agree”) has been invoked. Heuristic principles have ordinarily not been studied in a completely direct fashion—and for good reason. One would not expect (for instance) that self-report indices of heuristic use would be valuable; presumably, these heuristics are commonly used in a tacit, nonconscious way, and thus receivers may well not be in a good position to report on their use of such principles (S. Chen & Chaiken, 1999, pp. 86–87; Petty & Cacioppo, 1986a, p. 35). Instead, the operation of heuristic principles has been inferred from the observable influence of peripheral cues on persuasive outcomes. The ELM expects particular patterns of cue effects on persuasion: The influence of peripheral cues should be greater under conditions of relatively low elaboration likelihood (e.g., lower topic relevance) or under conditions in which the cue is relatively more salient. The primary evidence for the operation of heuristic principles consists of research results conforming to just such patterns of effect (for some discussion, see Bless & Schwarz, 1999).



Varieties of Heuristic Principles Although a number of heuristic principles have been suggested, three 247



heuristics have received relatively more extensive research attention: the credibility, liking, and consensus heuristics.14



Credibility Heuristic One heuristic principle is based on the apparent credibility of the communicator and amounts to a belief that “statements by credible sources can be trusted” (for alternative expressions of related ideas, see Chaiken, 1987, p. 4; Cialdini, 1987, p. 175). As discussed in Chapter 10, studies have indicated that as the personal relevance of the topic to the receiver increases, the effects of communicator credibility diminish (e.g., Byrne, Guillory, Mathios, Avery, & Hart, 2012; H. H. Johnson & Scileppi, 1969; Petty, Cacioppo, & Goldman, 1981; Rhine & Severance, 1970). Similar results have been obtained when elaboration likelihood has been varied in other ways (e.g., Janssen, Fennis, Pruyn, & Vohs, 2008; Kumkale, Albarracín, & Seignourel, 2010). Thus, consistent with ELM expectations, the peripheral cue of credibility has been found to have greater impact on persuasive outcomes when elaboration likelihood is relatively low. Moreover, some research suggests that variations in the salience of credibility cues lead to corresponding variations in credibility’s effects (e.g., Andreoli & Worchel, 1978). All told, there looks to be good evidence for the existence of a credibility heuristic in persuasion.



Liking Heuristic A second heuristic principle is based on how well the receiver likes the communicator and might be expressed by beliefs such as these: “People should agree with people they like” and “People I like usually have correct opinions” (for alternative formulations of this heuristic, see Chaiken, 1987, p. 4; Cialdini, 1987, p. 178). When this heuristic is invoked, liked sources should prove more persuasive than disliked sources. As discussed in more detail in Chapter 10, the research evidence does suggest that the ordinary advantage of liked communicators over disliked communicators diminishes as the personal relevance of the topic to the receiver increases (e.g., Chaiken, 1980, Experiment 1; Petty, Cacioppo, & Schumann, 1983). Confirming findings have been obtained in studies in which elaboration likelihood varied in other ways (e.g., Kang & Kerr, 2006; W. Wood & Kallgren, 1988) and in studies varying the salience of liking cues (e.g., Chaiken & Eagly, 1983): As elaboration likelihood declines or cue saliency increases, the impact of liking cues on persuasion increases. 248



Taken together, then, these studies point to the operation of a liking heuristic that can influence persuasive effects.



Consensus Heuristic A third heuristic principle is based on the reactions of other people to the message and could be expressed as a belief that “if other people believe it, then it’s probably true” (for variant phrasings of such a heuristic, see Chaiken, 1987, p. 4; Cialdini, 1987, p. 174). When this heuristic is employed, the approving reactions of others should enhance message effectiveness (and disapproving reactions should impair effectiveness). A number of studies now indicate the operation of such a consensus heuristic in persuasion (for a more careful review, see Axsom et al., 1987). For example, several investigations have found that receivers are less persuaded when they overhear an audience expressing disapproval (versus approval) of the communicator’s message (e.g., Landy, 1972; Silverthorne & Mazmanian, 1975). (For some related work, see Darke et al., 1998. For complexities, see Beatty & Kruger, 1978; Hilmert, Kulik, & Christenfeld, 2006; Hodson, Maio, & Esses, 2001; Mercier & Strickland, 2012.)



Other Heuristics Various other principles have been suggested as heuristics that receivers may employ in reacting to persuasive messages (e.g., Chang, 2004; Forehand, Gastil, & Smith, 2004). For example, it may be that the number of arguments in the message (Chaiken, 1980, Experiment 2) or the sheer length of the message (W. Wood et al., 1985) can serve as cues that engage corresponding heuristic principles (“the more arguments, the better” or “the longer the message, the better its position must be”). But for the most part, relatively little research evidence concerns such heuristics, and hence confident conclusions are perhaps premature.



Summary: Peripheral Routes to Persuasion Under conditions of low elaboration likelihood, the outcome of persuasive efforts depends less on the valence of receivers’ issue-relevant thinking than on the operation of heuristic principles, simple decision rules activated by peripheral cues in the persuasion setting. When receivers are unable or unmotivated to engage in extensive issue-relevant thinking, their reactions to persuasive communications will be guided by simpler 249



principles such as the credibility, liking, and consensus heuristics.



Multiple Roles for Persuasion Variables One important contribution of the ELM to the general understanding of persuasion is its emphasizing that a given variable might play different roles in persuasion under different conditions. Viewed through the lens of the ELM, a variable might influence persuasion in three broad ways.15 First, it might influence the degree of elaboration (and thus influence the degree to which central route or peripheral route processes are engaged). Second, it might serve as a peripheral cue (and so influence persuasive outcomes when peripheral route persuasion is occurring). Third, it might influence the valence of elaboration (and so influence persuasive outcomes when central route persuasion is occurring), by being an argument or by otherwise biasing (that is, encouraging one or another valence of) elaboration.16 The ELM emphasizes that a given variable need not play only one of these roles (e.g., Petty & Cacioppo, 1986a, pp. 204–215; Petty & Wegener, 1998a, 1999). In different circumstances, a variable might affect persuasion through different mechanisms. For example, consider the variable of message length (the simple length of a written message). This might serve as a peripheral cue that activates a length-based heuristic (such as “longer messages probably have lots of good reasons for the advocated view”; see W. Wood et al., 1985). When message length operates this way, longer messages will be more persuasive than shorter ones. But message length might also (or instead) influence elaboration motivation. For example, on a highly technical subject, the length of the message might serve as a sign of whether the message was likely to be worth close examination. Shorter messages might get little attention (because receivers would think that the message could not possibly contain the necessary amount of technical information), whereas longer messages would be examined more carefully. (For some evidence of such a phenomenon, see Soley, 1986.) In such a circumstance, obviously, a longer message would not necessarily be more persuasive than a shorter one; the persuasiveness of the longer message would turn on the outcome of the closer scrutiny engendered by the message’s length. For example, lengthening a message by adding weak arguments might enhance persuasion for recipients who were not examining the message carefully 250



but diminish persuasion for recipients who were engaged in close scrutiny (e.g., Friedrich, Fetherstonhaugh, Casey, & Gallagher, 1996). Similarly, communicator attractiveness might operate as a peripheral cue (engaging some version of the liking heuristic), might influence the amount of elaboration (a communicator’s attractiveness might draw attention toward or away from the message), or might serve as an argument (e.g., in advertisements for beauty products) and hence influence elaboration valence (see, e.g., Puckett, Petty, Cacioppo, & Fischer, 1983). Another example: The articulation of justification (supporting argumentation and evidence) in a message might influence the amount of elaboration (as when the presence of such support leads receivers to think that paying close attention to the message’s arguments will be worthwhile), might serve as a peripheral cue (by suggesting the credibility of the communicator or by activating a heuristic such as “if the message cites information sources, the position must be worthy of belief”), or might influence elaboration valence by encouraging more positive thoughts about the advocated view (see O’Keefe, 1998). The possibility of different persuasion roles for a single variable implies considerable complexity in persuasion. Consider, for instance: What will be the effect (on persuasive outcomes) of varying the communicator’s attractiveness? The ELM’s analysis implies that no simple prediction can be made; instead, the effects will be expected to vary depending on (among other things) whether attractiveness operates as an influence on the extent of elaboration, as an influence on the valence of elaboration, or as a peripheral cue. So, for instance, increasing the communicator’s attractiveness might enhance persuasion (e.g., if attractiveness operates as a peripheral cue that activates a liking-implies-correctness heuristic, if attractiveness enhances message scrutiny and the message contains strong arguments, if attractiveness reduces message scrutiny and the message contains weak arguments, or if greater attractiveness encourages positive elaboration by serving as an argument) or inhibit persuasion (e.g., if attractiveness enhances message scrutiny and the message contains weak arguments, or if attractiveness reduces message scrutiny and the message contains strong arguments). (For some illustrations of multiple persuasion roles for variables, see J. K. Clark, Wegener, & Evans, 2011; Mondak, 1990; Tormala, Briñol, & Petty, 2007.) Obviously, the key question that arises concerns specifying exactly when a variable is likely to play one or another role. The ELM offers a general 251



rule of thumb for anticipating the likely function for a given variable, based on the overall likelihood of elaboration (Petty, Wegener, Fabrigar, Priester, & Cacioppo, 1993, p. 354). When elaboration likelihood is low, then if a variable affects attitude change, it most likely does so by serving as a peripheral cue. When elaboration likelihood is high, then any effects of a variable on attitude change probably come about through influencing elaboration valence. When elaboration likelihood is moderate, then the effects of a variable on attitude change are likely to arise from affecting the degree of elaboration (e.g., when some aspect of the persuasive situation suggests that closer scrutiny of the message will be worthwhile). There is reason to doubt that this ELM rule of thumb is genuinely informative, because it amounts to little more than a restatement of the distinction between the two routes to persuasion. For instance, the proffered principle says in effect that “when elaboration is high, attitude change happens through elaboration valence and so anything that affects attitude change under such conditions does so by influencing elaboration valence.” This verges on a tautology, in which by definition something that influences attitude change under conditions of high elaboration must be affecting elaboration valence. The value of this rule of thumb thus turns on the degree to which one can independently assess whether peripheral or central processes are engaged, and such independent assessments are not easily had (as acknowledged by Petty & Briñol, 2006, p. 217). However, the ELM’s analysis does point to distinctive predictions about the different roles of a given variable, based on the operation of moderating variables. For example, if the physical attractiveness of a communicator in an advertisement is processed as a peripheral cue (which activates the liking heuristic), then the nature of the advertised product is unlikely to influence the cue’s effects. By contrast, if attractiveness influences elaboration valence because of being processed as an argument, then the nature of the product should be a moderating variable: The effect of attractiveness should occur for some products (namely, those for which attractiveness is a plausible argument, such as beauty products) but not for others (Petty & Briñol, 2006, p. 218). The implication is that by examining the effects of a moderator variable, one can distinguish whether a given property is activating a heuristic or influencing elaboration valence. The larger point is that the ELM draws attention to the mistake of thinking that a given variable can influence persuasive outcomes through only one pathway. Even in the absence of some well-articulated account of the 252



circumstances under which a given variable will serve in this or that persuasion role, persuaders will be well-advised to be alert to such complexities—and the ELM’s underscoring of these intricacies of persuasion represents an especially important contribution.



Adapting Persuasive Messages to Recipients Based on the ELM Given that variations in the likelihood of the message recipient’s elaboration make for very different underlying persuasion processes, the ELM would presumably recommend that persuaders tailor their persuasive efforts to the audience’s likely level of elaboration. For high-elaboration recipients, the key to effective persuasion will be having strong arguments. Under conditions of high elaboration, message receivers will be engaged in close scrutiny of the message’s arguments, so having high-quality arguments will be important. For low-elaboration recipients, on the other hand, argument quality will presumably be less of an influence on persuasive outcomes than will whatever heuristics the receiver is led to engage. Thus when elaboration is likely to be low, the ELM would recommend considering closely just what sorts of heuristics might potentially be available to be activated in the persuasion situation; for example, if the communicator’s expertise can easily be displayed, then perhaps one might try to encourage use of the credibility heuristic.17 But sometimes persuaders will have opportunities to influence the recipient’s degree of elaboration. For persuaders, achieving persuasion through central route processes is clearly advantageous, so one might think that persuaders will always want to encourage elaboration. But successful persuasion using counterattitudinal messages under conditions of high elaboration is likely to require having strong arguments, and at least sometimes persuaders might worry that the quality of their arguments will not be sufficient to overcome recipients’ inclination to counterargue. So where a persuader expects considerable counterarguing (negative elaboration), that persuader might well prefer to interfere with the audience’s ability to elaborate, perhaps by ensuring that something distracting is at hand. For instance, a scam artist selling dubious financial products might not want people to give his pitch their undivided attention —and so makes that pitch while people are eating the free dinner he has provided. 253



Commentary The ELM has stimulated a great deal of research. It is noteworthy that the ELM provides a framework that offers the prospect of reconciling apparently competing findings about the role played in persuasion by various factors. For example, why might the receiver’s liking for the communicator sometimes exert a large influence on persuasive outcomes and sometimes little? One possibility is simply that as elaboration varies, so will the impact of a simple decision rule such as the liking heuristic. Indeed, the ELM’s capacity to account for conflicting findings from earlier research makes it an especially important theoretical framework and unquestionably the most influential recent theoretical development in persuasion research. Even so, several facets of ELM theory and research require some commentary.



The Nature of Involvement In persuasion research, the concept of involvement has been used by a variety of theoretical frameworks to describe variations in the relationship that receivers have to the message topic (e.g., social judgment theory’s use of ego-involvement; see Chapter 2). In the ELM, as noted earlier, involvement refers specifically to the personal relevance of the message topic to the recipient. But several commentators have recommended distinguishing different forms of involvement, arguing that different varieties of involvement have different effects on persuasion. For example, B. T. Johnson and Eagly (1989) distinguished outcomerelevant involvement (in which concrete short-term outcomes or goals are involved) and value-relevant involvement (in which abstract values are engaged). Their meta-analytic evidence suggested that high outcomerelevant involvement produces the pattern of effects expected by the ELM, in which variations in argument strength produce corresponding variations in persuasive effects, but that high value-relevant involvement leads receivers to defend their opinions when exposed to counterattitudinal messages, regardless of whether the message contains strong or weak arguments. Petty and Cacioppo (1990), however, have argued that the same process might underlie these apparently divergent patterns of effect (for some further discussion, see B. T. Johnson & Eagly, 1990; Levin, Nichols, & Johnson, 2000; Petty & Cacioppo, 1990; Petty, Cacioppo, & Haugtvedt, 1992; see also Park, Levine, Westermann, Orfgen, & Foregger, 254



2007). As another example, Slater (2002) approached the task of clarifying involvement’s role in persuasion not by identifying different kinds of involvement but by identifying different kinds of message processing— and then working backward to consider how different kinds of involvement (and other factors) might influence different kinds of processing. Slater’s analysis distinguished outcome-based processing (motivated by the goal of self-interest assessment), value-affirmative processing (motivated by the goal of value reinforcement), and hedonic processing (motivated by the goal of entertainment)—with these influenced by, respectively, outcome relevance (akin to “outcome-relevant involvement”), value centrality (akin to “value-relevant involvement”), and narrative interest. Slater (2002, p. 179) thus argued that “simply distinguishing value-relevant involvement from the issue-or outcomerelevant involvement manipulated in ELM research does not go far enough.” In the present context, the point to be borne in mind is that the ELM conception of involvement is a specific one, namely, personal relevance— and other kinds of “involvement” might not have the same pattern of effects as is associated with personal relevance.



Argument Strength In ELM research, argument strength (argument quality) variations have been defined in an unusual way: in terms of persuasive effects under conditions of high elaboration. To obtain experimental messages containing strong or weak arguments, ELM researchers commonly pretest various messages: A strong-argument message is defined as “one containing arguments such that when subjects are instructed to think about the message, the thoughts that they generate are predominantly favorable,” and a weak-argument message is defined as one in which the arguments “are such that when subjects are instructed to think about them, the thoughts that they generate are predominantly unfavorable” (Petty & Cacioppo, 1986a, p. 32). That is, a high-quality argument is one that, in pretesting, is relatively more persuasive (compared to a low-quality argument) under conditions of high elaboration. By definition, then, high-quality arguments lead to greater persuasion under conditions of higher elaboration than do low-quality arguments. 255



Thus to say, “Under conditions of high elaboration, strong arguments have been found to be more effective than weak arguments” is rather like saying, “Bachelors have been found to be unmarried.” No empirical research is needed to confirm this claim (and indeed, there would be something wrong with any empirical research that seemed to disconfirm such claims). Notice, thus, how misleading the following statement might be: “A message with strong arguments should tend to produce more agreement when it is scrutinized carefully than when scrutiny is low, but a message with weak arguments should tend to produce less overall agreement when scrutiny is high rather than low” (Petty & Cacioppo, 1986a, p. 44). Appearances to the contrary, these are not empirical predictions; these are not expectations that might be disconfirmed by empirical results. If a message does not produce more agreement when scrutinized carefully than when scrutiny is low, then (by definition) it cannot possibly be a message with strong arguments.18 This way of defining argument quality reflects the role that argument quality has played in ELM research designs. In ELM research, argument quality variations have been used “primarily as a methodological tool to examine whether some other variable increases or decreases message scrutiny, not to examine the determinants of argument cogency per se” (Petty & Wegener, 1998a, p. 352). When argument quality is operationalized as the ELM has defined it, argument quality variations provide simply a means of indirectly assessing the amount of elaboration that has occurred. Thus to see whether a given factor influences elaboration, one can examine the difference in the relative persuasiveness of high- and low-quality arguments as that factor varies: High- and lowquality arguments will be most different in persuasiveness precisely when message scrutiny is high, and hence examining the size of the difference in persuasiveness between high- and low-quality arguments provides a means of assessing the degree of message scrutiny. For instance, one might detect the effect of distraction on elaboration by noticing that when distraction is present, there is relatively little difference in the persuasiveness of highquality arguments and low-quality arguments, but that without distraction, there is a relatively large difference in persuasiveness. Such a pattern of effects presumably reflects distraction’s effect on elaboration, because— by definition—high- and low-quality arguments differ in persuasiveness when elaboration is high. But this way of defining argument quality means the ELM has a curious lacuna. Consider the plight of a persuader who seeks advice about how to 256



construct an effective counterattitudinal persuasive message under conditions of high elaboration. Presumably the ELM’s advice would be “use strong arguments.” But because argument strength has been defined in terms of effects (a strong argument is one that persuades under conditions of high elaboration), this advice amounts to saying “to be persuasive under conditions of high elaboration, use arguments that will be persuasive”—which is obviously unhelpful (for some elaboration of this line of reasoning, see O’Keefe, 2003). Avoiding this shortcoming will require identification of the particular argument features that give rise to the observed effects of “argument quality” variations. Unfortunately, the experimental messages used in ELM experiments appear to have confounded a great many different appeal variations, making it challenging to identify just which features might have been responsible for the observed effects. However, at least one active ingredient in ELM messages has been identified, namely, variation in the perceived desirability of the outcomes associated with the advocated view (Areni & Lutz, 1988; Hustinx, van Enschot, & Hoeken, 2007; van Enschot-van Dijk, Hustinx, & Hoeken, 2003; see also B. T. Johnson, Smith-McLallen, Killeya, & Levin, 2004). That is, one key way in which ELM “strong argument” and “weak argument” messages have varied is in the desirability of the consequences of the advocated action or policy. As an example: One recurring message topic in ELM research has been a proposal to mandate senior comprehensive examinations as a university graduation requirement. In studies with undergraduates as research participants, the “strong argument” messages used arguments such as “with mandatory senior comprehensive exams at our university, graduates would have better employment opportunities and higher starting salaries,” whereas the “weak argument” messages had arguments such as “with mandatory senior comprehensive exams at our university, enrollment would increase” (see Petty & Cacioppo, 1986, pp. 54–59, for other examples of such arguments). Obviously, for these message recipients, these different outcomes almost certainly varied in desirability. And, correspondingly, it’s not surprising that, at least under conditions of relatively close attention to message content, the “strong argument” messages would be more persuasive than the “weak argument” messages for these receivers.



257



The identification of this key ingredient permits a redescription of the research findings concerning the effects of argument strength in terms of outcome desirability, as follows: When elaboration is low, the persuasiveness of a message is relatively unaffected by variation in the perceived desirability of the outcomes, whereas when elaboration is high, persuasive success is significantly influenced by the perceived desirability of the outcomes. That is, under conditions of high elaboration, receivers are led to have more positive thoughts about the advocated view when the message’s arguments indicate that the advocated view will have outcomes that the receivers think are relatively desirable than they do when the arguments point to outcomes that are not so desirable—but this difference is muted under conditions of low elaboration.19 What is not yet clear is whether there are other message variations that might function in a way similar to outcome desirability, that is, ones that serve as an influence on elaboration valence. Put somewhat differently, the question is: Are there other quality-related features of persuasive appeals whose variation makes relatively little difference to persuasive outcomes under conditions of low elaboration, but whose variation makes a more substantial difference under conditions of high elaboration?20 One natural candidate is outcome likelihood. A general expectancy-value treatment of attitudes (as in Fishbein’s, 1967a, belief-based model, described in Chapter 4) suggests that attitudes are a joint function of evaluations (the perceived desirability of the attitude object’s various characteristics) and likelihood (the likelihood that the object has each of those different characteristics). So one might expect that messages varying in the depicted likelihood of outcomes might have effects parallel to those of messages varying in the depicted desirability of outcomes, such that variations in outcome likelihood would make a greater difference to persuasiveness under conditions of high elaboration than under conditions of low elaboration. Little direct evidence bears on this expectation. However, the indirect evidence in hand is not very encouraging. A number of studies have reported that messages varying in the depicted likelihood of consequences did not differentially influence persuasive outcomes, but messages varying in the desirability of depicted outcomes did vary correspondingly in persuasiveness (e.g., Hass, Bagley, & Rogers, 1975; B. T. Johnson et al., 2004; for a review, see O’Keefe, 2013a). That is, under conditions in which elaboration was presumably sufficiently high to permit 258



consequence-desirability variations to produce differential effects, consequence-likelihood variations did not produce parallel effects.21 In any case, the general question remains open: There may be other quality-related message characteristics (in addition to outcome desirability) that enhance message persuasiveness under conditions of high elaboration. Identification of such message properties would represent an important advance in the understanding of persuasion generally and argument quality specifically.



One Persuasion Process? The Unimodel of Persuasion The two persuasion routes sketched by the ELM can be seen to be similar in a key way: In each route, people are trying to reach conclusions about what views to hold, and they do so on the basis of evidence that is available to them. Different sorts of evidence might be relied on in the two cases (peripheral cues in the peripheral route, the carefully scrutinized message arguments in the central route), but—it has been argued—there are not really two fundamentally different underlying processes here. Instead, there is just one process—the process of reasoning to conclusions based on evidence. Hence (this analysis suggests) in place of a dualprocess analysis, all that is needed is a unimodel of persuasion. (For some presentations of the unimodel approach, see Bohner, Erb, & Siebler, 2008; Kruglanski et al., 2006; Kruglanski & Thompson, 1999a, 1999b; E. P. Thompson, Kruglanski, & Spiegel, 2000.) It is important to be clear about exactly how the unimodel approach differs from a framework such as the ELM. The unimodel approach does not deny, for example, the roles played by motivational and ability variables in influencing the degree to which evidence is processed. The key difference is that the unimodel denies, whereas the ELM is said to assert, that a qualitative difference in persuasion processes arises as a consequence of whether persuasion occurs through the processing of message contents as opposed to the processing of extrinsic information (peripheral cues). The unimodel claims that there is an underlying uniformity to the persuasion process, no matter which type of information is processed. That is, the unimodel proposes that there is a “functional equivalence of cues and message arguments” (E. P. Thompson et al., 2000, p. 91), in the sense that 259



cues and message arguments simply serve as evidence bearing on the receiver’s conclusion about whether to accept the advocated view. One way of expressing this equivalence is to see that both peripheral cues and message arguments can be understood as supplying premises that permit the receiver to complete a conditional (“if-then”) form of reasoning. In the case of peripheral cues, the reasoning can be exemplified by a receiver who believes that “if a statement comes from an expert, the statement is correct.” A message from a source that the receiver recognizes as an expert, then, satisfies the antecedent condition (a statement coming from an expert), and hence the receiver reasons to the appropriate conclusion (that the statement is correct). In the case of message arguments, the reasoning can be exemplified by a receiver who believes (for instance) that “if a public policy has the effect of reducing crime, it is a good policy.” Accepting a message argument indicating that current gun control policies have the effect of reducing crime, then, satisfies the antecedent condition (that the policy reduces crime), and hence the receiver reasons to the indicated conclusion (that current gun control policies are good ones). Thus the unimodel proposes that there is really only one type of persuasion process, a process that accommodates different (but functionally equivalent) sources of evidence (viz., cues and message arguments).



Explaining ELM Findings One question that immediately arises is how the unimodel might explain the substantial accumulated evidence supporting the ELM. For instance, there appears to be considerable evidence showing that receivers vary in their relative reliance on peripheral cues or message arguments depending on such factors as the personal relevance of the topic; such evidence seems to imply that cues and message arguments are not actually functionally equivalent evidence sources. The unimodel’s analysis of such research begins with the point that both peripheral cues and message arguments can vary in their complexity, ease of processing, brevity, and so forth. The unimodel acknowledges that peripheral cues are often the sorts of things that are easily processed (and message arguments are commonly the sorts of things that require more processing), but that need not be so: “Cue and heuristic information need not be briefer, less complex, or easier to process than message information” (Kruglanski & Thompson, 1999b, p. 96). 260



But (the unimodel suggests) ELM research has commonly confounded the cue-versus-message contrast with other contrasts—in particular, with complexity and temporal location. That is, in ELM research, receivers are offered a simple source of evidence at the beginning of the message in the form of a peripheral cue (e.g., information about source credibility) and then later are given a complex source of evidence (in the form of message arguments). The unimodel analysis suggests that in such a research design, under conditions of low personal relevance (low motivation to process), receivers will naturally be more influenced by the brief, easily processed, initially presented peripheral cue than by the subsequent difficult-toprocess argumentative material; when the message arguments appear later in the sequence of information—and require more processing than do the cues—they will likely affect only those receivers who (by virtue of higher topic relevance) have greater motivation to process. Thus the unimodel approach argues that the apparent differences between peripheral cues and message arguments (in their relative impacts on persuasion as personal relevance varies) do not reflect some general difference between cues and arguments (as the ELM is taken to assert) but rather a confounding of evidence type (peripheral cue vs. message argument) and other features of evidence (brevity, ease of processing, and temporal location). If (the unimodel suggests) peripheral cues and message arguments were equalized on these other dimensions, then the putative dual-process differences between them would evaporate. As an illustration of research supporting the unimodel’s view, Kruglanski and Thompson (1999b, Study 1) found that when source expertise information was relatively lengthy, source expertise influenced the attitudes of receivers for whom the topic was personally relevant but not the attitudes of receivers for whom the topic was not relevant. In other words, source expertise information and topic relevance interacted in just the way that argument quality and topic relevance did in earlier ELM studies. The apparent implication is that cues (such as expertise) and arguments function identically in persuasion, once the complexity and temporal location of each is equalized. (For related research, see Erb, Pierro, Mannetti, Spiegel, & Kruglanski, 2007; Pierro, Mannetti, Erb, Spiegel, & Kruglanski, 2005.)



Comparing the Two Models The unimodel raises both empirical and conceptual issues concerning the ELM, and these issues are sufficiently complicated that it will take some 261



time to sort them out. (For some discussion of these and related issues, see, e.g., Chaiken, Duckworth, & Darke, 1999; Petty & Briñol, 2006; Petty, Wheeler, & Bizer, 1999; Wegener & Claypool, 1999.) Empirically, there is room for some uncertainty about exactly when (or, indeed, whether) the ELM and the unimodel make genuinely different predictions. The description given here of the unimodel has stressed its putative contrasts with the ELM, but those contrasts may be less substantial than is supposed. For instance, presentations of the unimodel depict the distinction between cues and arguments as crucially important to the ELM, in that the ELM is seen to treat these as functionally different influences on persuasive outcomes (as opposed to the unimodel view, in which cue and argument are simply two content categories and are not functionally different as sources of evidence in the receiver’s reasoning processes). It is certainly true that ELM theorists have sometimes used the terms cue and argument in ways that make these into opposed categories (e.g., Petty & Wegener, 1999), which invites some misunderstanding. However, the key distinction for the ELM is not the contrast between peripheral cues and message arguments but variation along the elaboration continuum that yields a general trade-off between peripheral processes (e.g., as represented by the influence of peripheral cues) and central processes (as represented by elaboration valence, not message arguments specifically) as influences on persuasive outcomes.22 Indeed, one important source of confusion can be a failure to grasp the ELM’s insistence that a given variable can play different roles in persuasion—and hence (for example) it is inappropriate to treat source characteristics as necessarily always and only serving as peripheral cues (Petty & Briñol, 2006, p. 217). Consider, for instance, the previously mentioned finding that complex information about source expertise had more influence on persuasive outcomes when the topic was personally relevant to receivers than when it was not (Kruglanski & Thompson, 1999b, Study 1). From a unimodel perspective, this is taken to be inconsistent with the ELM, because the ELM is assumed to expect that source cues will have a smaller influence on persuasion as topic relevance increases. But—bearing in mind that a given variable might affect persuasion through various pathways—the ELM might explain this result in several ways, including the possibility that expertise information was processed as an argument or provoked elaboration of self-generated (as opposed to message) arguments (Petty, Wheeler, & Bizer, 1999, pp. 159– 160). 262



The general point is that it is not yet clear whether (or exactly how) the ELM and the unimodel can be made to offer contrasting empirical predictions. Research findings indicating that communicator characteristics and message arguments can both function either as peripheral cues or as influences on elaboration valence—the kinds of research findings offered as support for the unimodel—are in fact not necessarily inconsistent with the ELM. Conceptually, the unimodel does implicitly point to some unclarities in the ELM. Consider, for example, the question of whether it is true by definition that peripheral cues are easy to process. If part of the very concept of a peripheral cue is that it is easy to process, then it does not make sense to speak of there being any “confounding” of cues and simplicity—and so the unimodel’s suggestion that there might be complex cues would be conceptually malformed. On the other hand, if peripheral cues are not necessarily (definitionally) easy to process, then it makes sense to explore the effects of hard-to-process peripheral cues. Thus the unimodel has raised some important questions concerning the ELM. One may hope that continuing attention to these issues will lead to more focused empirical predictions and better-articulated conceptual equipment.



Conclusion The elaboration likelihood model may be seen to contribute two key insights about persuasion. One is the recognition of the variable character of topic-related thinking engaged in by message recipients. Because the extensiveness of topic-relevant thinking varies (from person to person, from situation to situation, etc.), the central factors influencing persuasive success vary: Simple heuristic principles may prevail when little elaboration occurs, but when extensive elaboration is undertaken, then the character of the message’s contents takes on greater importance. The second is the recognition that a given variable may play different roles in the persuasion process. The same variable (in different circumstances) might influence the degree of elaboration, might influence the valence of elaboration, and might serve as a peripheral cue—and so might have different effects on persuasive outcomes depending on the situation. Taken together, these two ideas offer the prospect of reconciling apparently conflicting findings in the research literature concerning the role played by various factors in influencing persuasive effects and mark the ELM as an 263



important step forward in the understanding of persuasion.



For Review 1. What is elaboration? How can the degree of elaboration be assessed? Do variations in the amount of elaboration form a continuum or discrete categories? Describe the general difference between central and peripheral routes to persuasion. Explain how persuasion can occur even under conditions of low elaboration. 2. Are the consequences of central route persuasion and peripheral route persuasion identical? Identify three differences in the consequences of persuasion’s being achieved through one or the other route. How is the persistence of persuasion different? How is the strength of the relationship of attitudes to intentions and behaviors different? How is resistance to counterpersuasion different? 3. Identify two broad categories of factors that influence the amount of elaboration undertaken. What is elaboration motivation? Identify two factors influencing elaboration motivation. Explain how the personal relevance of the topic (involvement) influences elaboration motivation. What is “need for cognition”? Explain how need for cognition influences elaboration motivation. What is elaboration ability? Identify two factors influencing elaboration ability. What is “distraction”? Explain how distraction influences elaboration ability. Explain how prior knowledge influences elaboration ability. 4. In central route persuasion, what is the key determinant of persuasive outcomes? Explain. Identify two factors that influence elaboration direction (valence). Explain how the message’s proattitudinal or counterattitudinal position influences elaboration direction. What is argument strength (quality)? Explain how argument strength influences elaboration direction. 5. In peripheral route persuasion, what influences the outcomes of persuasive efforts? What is a heuristic principle? What activates heuristic principles? Give three examples of heuristic principles. What is the credibility heuristic? Explain how it works. Under what conditions does credibility have relatively greater influence on persuasive outcomes? What is the liking heuristic? Explain how it works. Under what conditions does liking have relatively greater influence on persuasive outcomes? What is the consensus heuristic? Explain how it works. Under what conditions will the consensus heuristic have relatively greater influence on persuasive outcomes? 264



6. Explain the idea that a given variable might play different roles in persuasion in different circumstances. Describe three different roles identified by the ELM. Give examples of how a variable might serve in different roles. What is the ELM’s rule of thumb for expecting what role a given variable will play? How useful is that rule of thumb? 7. Describe how persuaders might adapt messages to recipients using the ELM. How should messages be adapted to high-elaboration recipients? How should messages be adapted to low-elaboration recipients? Why might persuaders want to influence the likely amount of elaboration? 8. How is involvement defined in ELM research? How is this sense of involvement different from social judgment theory’s egoinvolvement? What various kinds of involvement might be distinguished? 9. How is argument strength defined in ELM research? Explain how this definition does not specify the message features that underlie argument quality variations. Identify one active argument-quality ingredient in the messages used in ELM research. 10. What is the unimodel of persuasion? Describe how the unimodel suggests that one process, not two, underlies persuasion effects. How does the unimodel attempt to explain ELM findings (about, e.g., the relative influence of peripheral cues and argument quality)? Do the unimodel and the ELM make different predictions? Explain how the unimodel raises questions about the clarity of some ELM concepts. 11. Identify and explain two key insights about persuasion contributed by the ELM.



Notes 1. There has been some variation in the ELM’s definition of elaboration. Elaboration has sometimes been conceived in broad terms (as here), namely, engaging in issue-relevant thinking (e.g., Petty & Wegener, 1999, p. 46). But elaboration has also been defined more narrowly as issuerelevant thinking undertaken with the motivation of impartially determining the merits of the arguments (e.g., Cacioppo, Petty, & Stoltenberg, 1985, p. 229) or as message scrutiny (e.g., Petty & Cacioppo, 1986a, p. 7). But the broadest definition is the most common. 2. Variations in the conceptualization of elaboration (mentioned in note 1) 265



have produced corresponding variations in proposed assessments of elaboration (see, e.g., Cacioppo et al., 1985, p. 229). But most procedures for the assessment of elaboration (as discussed by Petty & Cacioppo, 1986a, pp. 35–47) appear to represent indices of the amount of issuerelevant thinking generally. 3. It can be tempting to say that central route persuasion makes for “stronger” attitudes, but it is not plain that this is an entirely satisfactory account, for several reasons. First, dangers of tautology lurk; if a “strong” attitude is one that by definition is resistant to persuasion, then attitude strength cannot possibly be an explanation of resistance. Second, a number of different strength-related attitude properties are distinguishable, and there is reason to suppose that these are best treated as distinct attributes rather than being bundled together in an omnibus concept (see Visser, Bizer, & Krosnick, 2006). 4. Although it seems common for this research literature to be glossed as showing that positive moods lead to less elaboration than do negative moods, this characterization seems not to quite capture the complexity in the literature. In Hullett’s (2005) meta-analysis, the mean correlation of argument strength with attitude (with this correlation serving as a proxy for elaboration, because it presumably reflects sensitivity to argumentstrength variation) in positive-mood conditions (.29 across 12 effect sizes) was not significantly different from that in negative-mood conditions (.39 across 21 effect sizes). (For additional discussion of this meta-analysis, see Chapter 12, note 4.) And several studies have reported significantly greater elaboration (or elaboration proxies) in positive moods than in negative moods, at least under some conditions (e.g., Das & Fennis, 2008; Das, Vonkeman, & Hartmann, 2012; Sinclair, Moore, Mark, Soldat, & Lavis, 2010; Ziegler & Diehl, 2011). 5. This research support largely consists of evidence showing that as personal relevance increases, the effects of argument quality increase and the effects of peripheral cues decrease. 6. Other properties captured under the term involvement may not have the same effects as does personal relevance. As a simple illustration, the effects on message scrutiny (that is, close attention to the message’s contents) may not be the same for increasing personal relevance and for increasing commitment to a position. As personal relevance increases, message scrutiny increases, but as position commitment increases, one can 266



imagine message scrutiny either increasing or decreasing (e.g., increasing when there are cues that message scrutiny will yield position-bolstering material but decreasing when scrutiny looks to yield position-threatening material). For some general discussions of involvement, see B. T. Johnson and Eagly (1989, 1990), K. D. Levin et al. (2000), Petty and Cacioppo (1990), Slater (2002), and Thomsen, Borgida, and Lavine (1995). 7. Across the 11 cases reviewed by Cacioppo, Petty, et al. (1996, p. 229), the mean effect corresponds to a correlation of roughly .15. There is not a corresponding difference in the influence of peripheral cues. That is, persons low in need for cognition are not dependably more influenced by peripheral persuasion cues than are those high in need for cognition (for a review and discussion, see Cacioppo, Petty, et al., 1996, p. 230). 8. Because need-for-cognition indices are positively correlated with various measures of intellectual ability (mean correlations are roughly in the range of .15 to .30; for a review, see Cacioppo, Petty, et al., 1996, p. 214), one might wonder whether the apparent effects of need for cognition on elaboration likelihood should be ascribed to differences in elaboration motivation or differences in elaboration ability (Chaiken, 1987, pp. 16– 17). The evidence in hand appears to favor a motivational difference explanation rather than an ability difference explanation (e.g., Cacioppo, Petty, Kao, & Rodriguez, 1986, Study 1; See, Petty, & Evans, 2009); for instance, the presence of additional motivational incentives (to engage in elaboration) can minimize these effects of need for cognition (e.g., Priester & Petty, 1995), suggesting that a difference in dispositional motivation (not an ability difference) underlies the effects. 9. The ELM’s analysis of distraction effects is actually a bit more complex than this. For instance, the ELM acknowledges that when the distraction is so intense as to become the focus of attention, thus interfering with even minimal message reception, one does not expect to find the otherwise predicted distraction effects. For a more careful discussion, see Petty and Cacioppo (1986a, pp. 61–68). 10. The studies by W. Wood (1982), W. Wood and Kallgren (1988), and W. Wood et al. (1985) all use the same message topic with (it appears) similar messages, which means that this research evidence does not underwrite generalizations as confident as one might prefer. 11. As a further complexity, however, consider that prior knowledge might have still other effects. For instance, although prior knowledge may 267



enhance elaboration ability, it could also diminish elaboration motivation —as might happen if receivers think that they have sufficient information and so expect that there would be little gained from close processing of the message (see B. T. Johnson, 1994; Trumbo, 1999). (For another example of diverse effects of receiver knowledge, see Biek, Wood, & Chaiken, 1996; for a general discussion, see W. Wood, Rhodes, & Biek, 1995.) 12. The organization of this description of the ELM has separated influences on the amount of elaboration (factors affecting elaboration motivation and/or elaboration ability) and influences on the valence (evaluative direction) of elaboration (following, e.g., Petty & Wegener, 1999, p. 43, Figure 3.1). Alternatively (see Petty & Wegener, 1999, pp. 52–59), one might distinguish (a) variables that affect message processing “in a relatively objective manner” (i.e., that influence elaboration motivation and/or ability in such a way as to affect positive and negative thoughts more or less equally; e.g., distraction interferes with elaboration ability generally) and (b) variables that affect message processing “in a relatively biased manner” (i.e., that influence elaboration motivation and/or ability in selective ways that encourage a particular evaluative direction to thinking; e.g., a message’s counterattitudinal stance might enhance motivation to engage in specifically negative elaboration, that is, counterarguing). 13. The ELM suggests that there are other peripheral route processes in addition to heuristic principles—specifically, “simple affective processes” (Petty & Cacioppo, 1986a, p. 8) in which attitudes change “as a result of rather primitive affective and associational processes” (p. 9) such as classical conditioning. Indeed, this additional element is one important difference between the ELM and the HSM. The HSM’s systematic processing mode corresponds to the ELM’s central route, and the HSM’s heuristic mode refers specifically to the use of heuristic principles of the sort discussed here. Although the ELM’s peripheral route is thus broader than is the HSM’s heuristic mode, here the peripheral route is treated in a way that makes it look like the heuristic mode. That is, the present treatment focuses on the simple rules/inferences (the heuristic principles) rather than on the primitive affective processes that are taken to also represent peripheral routes to persuasion. There are several reasons for this. First, the nonheuristic peripheral route processes have not gotten much attention in ELM research. Second, the ELM could abandon a belief in any particular 268



nonheuristic peripheral process (say, classical conditioning) with little consequence for the model, which suggests that the ELM’s commitment to any specific such process is inessential to the model. Third, it may be possible to translate some apparently nonheuristic peripheral processes into heuristic principle form (e.g., mood effects might reflect a tacit heuristic such as “if it makes me feel good, it must be right”). 14. As Chaiken (1987, p. 5, n. 1) pointed out, a number of heuristic principles appear to be represented in the various compliance principles identified by Cialdini (1984; Cialdini & Trost, 1998); see, similarly, Petty and Briñol (2012b). 15. Presentations of the ELM have expressed this “multiple roles” idea in various ways (e.g., Petty & Briñol, 2008, 2010), but these have not always been as clear as one might like. For instance, one formulation is that “the ELM notes that a variable can influence attitudes in four ways: (1) by serving as an argument, (2) by serving as a cue, (3) by determining the extent of elaboration, and (4) by producing a bias in elaboration” (Petty & Wegener, 1999, p. 51). But the ways in which (what would conventionally be called) arguments can influence attitudes, from the perspective of the ELM, seem to be (a) by serving as a cue (e.g., when the number of arguments activates a heuristic such as “there are many supporting arguments so the position must be correct”), (b) by influencing the extent of elaboration (as when a receiver thinks that “there seem to be a lot of arguments here so maybe it’s worth looking at them closely”), and (c) by producing a bias in elaboration (i.e., by influencing the evaluative direction of elaboration). That is, the roles of arguments appear already subsumed in the other three roles (peripheral cue, influence on degree of elaboration, and influence on valence of elaboration); it is not clear how arguments might otherwise function in persuasion within an ELM framework. Hence the presentation here does not distinguish “serving as an argument” as a distinct role for a persuasion variable. At least part of the confusion appears to concern the ELM’s use of the word argument, about which three points might be noted. First, “arguments” are sometimes conceived of as “bits of information contained in a communication that are relevant to a person’s subjective determination of the true merits of an advocated position” (Petty & Cacioppo, 1986b, p. 133); but taken at face value, such a definition would accommodate at least some peripheral cues as arguments (after all, from the perspective of the heuristic processor, a cue is a bit of information relevant to assessing 269



the true merits of the advocated view—it just happens to provide a shortcut to such assessment), which seems a unimodel-like view (Kruglanski & Thompson, 1999b), surely to be resisted by the ELM. Second, “argument” and “cue” sometimes appear to be used as shorthand to cover anything that affects, respectively, central route and peripheral route persuasion (e.g., Petty & Wegener, 1999, p. 49). But when they are used this way, it is not clear why argument-based persuasion roles are to be distinguished from persuasion roles involving influencing elaboration valence (given that elaboration valence is presumably the engine of persuasion within central route processes). Third, distinguishing “serving as an argument” does at least underscore the broad possible application of “argument” within an ELM perspective. For example, the communicator’s physical attractiveness is recognized by the ELM as potentially not simply a peripheral cue but also an argument (as in advertisements for beauty products). Still, when attractiveness serves this argumentative role, it presumably influences persuasive effects by influencing elaboration valence (just as arguments of more conventional form do). 16. One additional possible role for variables is as influences on metacognitive states (e.g., Petty & Briñol, 2010, pp. 230–232) such as thought confidence (i.e., confidence in the accuracy or correctness of one’s thoughts or beliefs) or attitude confidence (i.e., confidence in the accuracy or correctness of one’s attitude). These metacognitive effects appear to represent an expansion of the sorts of persuasion-relevant outcomes that researchers might examine. The implicit suggestion is that in addition to assessing attitude (valence and extremity of a general evaluation), one might also assess (a) putative determinants of attitude (such as thought confidence; Briñol & Petty, 2009a, 2009b; Petty & Briñol, 2010, pp. 230– 231) and (b) properties of attitude other than valence and extremity (e.g., attitudinal certainty or confidence; Petty, Briñol, Tormala, & Wegener, 2007, pp. 260–262). With respect to metacognitive states that might be determinants of attitude: Having a sound account of the determinants of attitude is very much to be hoped for. Considerable work has already explored the roles of belief evaluation and belief likelihood as possible determinants (see Chapter 4). It may be that additional factors, such as belief confidence (thought confidence), will also have some role to play, but this is not yet entirely clear. (For example, it is not plain that belief confidence is so straightforwardly an influence on attitude in the way that belief evaluation is: e.g., Bennett & Harrell, 1975; Feldman, 1974.) In principle, at least, 270



metacognitive properties that influence attitude provide persuaders with additional avenues to attitude change. However, such metacognitive states presumably—though this is not entirely clear—influence attitude change only under conditions of relatively high elaboration (i.e., only under conditions in which attributes of thoughts, such as their valence or confidence, influence attitudes). So one might treat the roles of “influencing elaboration valence” and “influencing metacognitive states that affect attitudes” as representing two concrete realizations of a single more abstract role, namely, “influencing attitude-relevant thought properties” (valence, confidence, likelihood, and so forth). But having a clear picture of all this will require a careful enumeration of attituderelevant thought properties and a description of their relationships and effects, and this task is not to be underestimated (for one such sketch, see Petty, Briñol, Tormala, & Wegener, 2007). With respect to metacognitive states that are properties of attitude: Additional attitudinal properties, such as certainty or importance, represent additional possible targets for persuasive efforts. For example, a persuader might not want (or need) to make an existing attitude any more positive but might want to make people more confident in those positive attitudes. Pursuing research on these lines will want a clear conceptual treatment of these various attitudinal properties (for some discussion, see Fabrigar, MacDonald, & Wegener, 2007; Visser, Bizer, & Krosnick, 2006). Increasing research attention is being given especially to attitude certainty (e.g., Druckman & Bolsen, 2012; Nan, 2009; Tormala & Rucker, 2007) and attitudinal ambivalence (e.g., Conner & Armitage, 2008; van Harreveld, Schneider, Nohlen, & van der Pligt, 2012). Research on the determinants and consequences of these additional attitudinal properties is not yet as extensive as that concerning attitude valence and extremity, but it is an extraordinarily promising new line of development. 17. Not a few investigators have tried to design message formats or appeals so as to adapt messages to different levels of need for cognition (NFC)—without much success. For example, Steward, Schneider, Pizarro, and Salovey (2003) compared the effects of a simple message (with framed cartoons, grammatically simple sentences, and so on) and a complex message (no pictures, detailed statistical information, and so on) on recipients varying in NFC, expecting the simple message to be more persuasive than the complex message for those low in NFC, with the reverse effect expected for those high in NFC; no such effect obtained. For similar failures to find expected interaction effects involving NFC and 271



message variations, see McKay-Nesbitt, Manchanda, Smith, and Huhmann (2011), Rosen and Haaga (1998), and Williams-Piehota, Pizarro, Silvera, Mowad, and Salovey (2006); for an exception, see Bakker (1999). 18. Another example: “Subjects led to believe that the message topic (e.g., comprehensive exams) will (vs. won’t) impact on their own lives have also been shown to be less persuaded by weak messages but more persuaded by strong ones” (Chaiken & Stangor, 1987, p. 594). Despite the statement’s appearance, this is not a discovery. It is not an empirical result or finding, something that research “shows” to be true, or something that could have been otherwise (given the effect of topic relevance on elaboration). The described relationship is true by definition. 19. Indeed, when elaboration is low as a consequence of low personal relevance of the topic (low receiver involvement), then an even more pointed redescription is possible: When a message recipient is directly affected by the advocated policy, the desirability of the claimed outcomes matters a great deal—but when the policy isn’t personally relevant, receivers are less affected by the apparent desirability of the outcomes (after all, the outcomes aren’t going to happen to them). Unsurprising, really: When the outcomes affect the message recipient directly, the desirability of those outcomes becomes especially important as an influence on the message’s persuasiveness. 20. The importance of this undertaking (identifying quality-related message characteristics other than outcome desirability that influence persuasiveness under conditions of high elaboration) is magnified when one recognizes the commonality with which personal relevance (“involvement”) variations have been used as an experimental means of inducing variations in elaboration likelihood. The ELM’s claim is that “argument quality variations make a larger difference to persuasive outcomes under conditions of high elaboration than under conditions of low elaboration”—but this is different from the narrower claim that “outcome desirability variations make a larger difference to persuasive outcomes under conditions of high elaboration than under conditions of low elaboration,” and still more different from the even narrower claim that “outcome desirability variations make a larger difference to persuasive outcomes under conditions of high personal relevance than under conditions of low personal relevance.” This last claim is amply supported by ELM research, but that research is not necessarily evidence for any broader claims. 272



As a first step in seeing what additional evidence is needed, consider whether it is possible to have high elaboration without high personal relevance (high involvement). It is certainly possible to have low elaboration without low personal relevance (e.g., when personal relevance —and so elaboration motivation—is high, distraction might interfere with elaboration ability and so produce low elaboration despite high personal relevance), but it is difficult to imagine conditions under which elaboration would be high while personal relevance was low. If high personal relevance is indeed a necessary condition for high elaboration, then ELM-related descriptions ought to reflect that relationship. For example, high elaboration should not be described simply as something that might be influenced by (among other things) personal relevance but rather as something that is entirely dependent on the level of personal relevance. If, on the other hand, high personal relevance is not required for high elaboration, then additional research evidence is needed to fill the gap between the narrower (and well-supported) claim that “outcome desirability variations make a larger difference to persuasive outcomes under conditions of high personal relevance than under conditions of low personal relevance” and the broader claim that “argument quality variations make a larger difference to persuasive outcomes under conditions of high elaboration than under conditions of low elaboration.” What is needed is evidence that parallel effects (parallel to those observed with outcome desirability variations and personal relevance variations) can be obtained when (a) argument quality is varied but outcome desirability is held constant and (b) elaboration is varied but personal relevance is low. It is not plain that such evidence is in hand. 21. It may be that outcome likelihood variations just don’t have the same effects that outcome desirability variations do, or it might be that the amount of message scrutiny required to yield outcome likelihood effects is higher than that needed to produce outcome desirability effects (so that a given level of elaboration might be high enough for recipients to be affected by the apparent desirability of the depicted outcomes but not yet high enough for recipients to be affected by the relative likelihood of those outcomes). 22. In unimodel presentations and research, the “cue versus argument” distinction sometimes seems to be cast as a “source versus message” 273



distinction (e.g., Kruglanski & Thompson, 1999b, p. 84). But this also does not seem to capture the ELM’s assertions. The ELM does not partition source variables and message variables as having intrinsically different roles to play in persuasion but, on the contrary, emphasizes that each category of variable can serve different persuasion roles in different circumstances (Petty & Briñol, 2006, p. 217; Petty et al., 1999, p. 157; Wegener & Claypool, 1999, pp. 176–177).



274



Chapter 9 The Study of Persuasive Effects Experimental Design and Causal Inference The Basic Design Variations on the Basic Design Persuasiveness and Relative Persuasiveness Two General Challenges in Studying Persuasive Effects Generalizing About Messages Variable Definition Conclusion For Review Notes



The research to be discussed in the next three chapters is, overwhelmingly, experimental research that systematically investigates the influence that various factors (communicator characteristics, message variations, and so on) have on persuasive outcomes. This chapter first provides some general background on the underlying logic of such experimental research and then discusses some challenges that arise in the study of persuasive effects.



Experimental Design and Causal Inference Various experimental arrangements are used in persuasion effects research, but these can usefully be thought of as variations on a basic design.



The Basic Design The simplest sort of research design employed in the work to be discussed is an experimental design in which the researcher manipulates a single factor (the independent variable) to see its effects on persuasive outcomes (the dependent variable). For instance, an investigator who wishes to investigate the effects of explicit conclusion drawing on attitude change might design a laboratory investigation of the following sort. The researcher prepares two persuasive messages identical in every respect except that in one message, the persuader’s conclusion is drawn explicitly 275



at the end of the message (the explicit conclusion message), whereas in the other message, the persuader’s conclusion is left implicit (the implicit conclusion message). When participants in this experiment arrive at the laboratory, their attitudes on the persuasive topic are assessed, and then they receive one of the two messages; which message a given participant receives is a matter of chance, perhaps determined by flipping a coin. After exposure to the persuasive message, receivers’ attitudes are assessed again to ascertain the degree of attitude change produced by the message. Suppose that (following conventional statistical procedures) the results indicate reliably greater attitude change for those receiving the explicit conclusion message than for those receiving the implicit conclusion message. In considering how such a result might be explained, one can rule out systematic bias in assigning participants to hear one or the other of the messages because participants were randomly assigned to hear messages. For instance, one can confidently say that it is unlikely that those hearing the explicit conclusion message were people who are just generally more easily persuaded than those hearing the implicit conclusion message because participants were randomly distributed across the two groups. The obvious explanation for the obtained results, of course, is the presence or absence of an explicit conclusion. Indeed, because this is the only factor that varies between the two messages, it presumably must be the locus of the observed differences. This is the general logic of experimental designs such as this: These designs are intended to permit unambiguous causal attribution precisely by virtue of experimental control over factors other than the independent variable.



Variations on the Basic Design There are innumerable ways in which this basic experimental arrangement might be varied. For example, one might dispense with the initial attitude assessment, reasoning that the random assignment of participants to the two experimental conditions would likely ensure that the two groups would have roughly comparable initial attitudes; this is commonly called a posttest-only design (because there would be only a postmessage assessment of attitude). Or an investigator might create an independent variable with more than two conditions (more than two levels). For instance, one might compare the persuasive effects of communicators who are high, moderate, or low in credibility. 276



The most common and important variation, however, is the inclusion of more than one independent variable in a single experiment. Thus (for instance) rather than doing one experiment to study implicit versus explicit conclusions and a second study to examine high versus moderate versus low credibility, a researcher could design a single investigation to study these two variables simultaneously. This would involve creating all six possible combinations of conclusion type and credibility level (3 credibility conditions × 2 conclusion type conditions = 6 combinations). Experimental designs with more than one independent variable permit the detection of interaction effects involving those variables. An interaction effect is said to occur if the effect of one independent variable depends on the level of another independent variable; conversely, if the effect of one variable does not depend on the level of another variable, then no interaction effect exists. For example, if the effect of having an implicit or explicit conclusion is constant, no matter what the credibility of the source, then no interaction effect exists between credibility and conclusion type. But if the effect of having an implicit or explicit conclusion varies depending on the credibility of the source (say, if high-credibility sources are most effective with explicit conclusions, and low-credibility sources most effective with implicit conclusions), then an interaction effect (involving credibility and conclusion type) exists; the effect of one variable (conclusion type) depends on the level of another (credibility).1



Persuasiveness and Relative Persuasiveness These experimental designs are meant to provide information about the relative persuasiveness of two or more messages—not the absolute persuasiveness of any one message. So, for example, an experiment that found that explicit conclusion messages are more persuasive than implicit conclusion messages would not necessarily show that explicit conclusion messages are highly persuasive. That is, the research question in these studies is characteristically not “How persuasive are messages of kind X?” but rather “Which is more persuasive, kind X or kind Y?” So the question is not “How persuasive are messages with explicit conclusions?” but “Which is more persuasive, messages with explicit conclusions or messages with implicit conclusions? ”2 There is much potential for misunderstanding here. Imagine, for example, that research underwrites a general conclusion that messages with explicit 277



conclusions are generally more persuasive than those with implicit conclusions. Message designers should not think that having explicit conclusions will somehow automatically make their messages highly persuasive. Messages with explicit conclusions may be relatively persuasive compared with those with implicit conclusions, but that does not mean that explicit conclusion messages are inevitably highly persuasive in absolute terms. The larger point is that the research under discussion here can certainly provide useful information to message designers about how to enhance message persuasiveness (by creating messages of one sort rather than another), but it does not offer evidence bearing on the absolute persuasiveness of any given kind of message.3



Two General Challenges in Studying Persuasive Effects Two noteworthy general challenges arise in investigating factors influencing the effectiveness of persuasive messages. One of these concerns the difficulty in making reliable generalizations about the effects of message types; the other concerns the task of defining independent variables in studies of persuasive effects.



Generalizing About Messages The earlier description of experimental design might make it seem easy to arrive at generalizations about factors influencing persuasive effects. To compare the persuasive effects of (for example) explicit and implicit conclusions, one simply does an experiment of the sort previously described: Create two versions of a given message (one of each conclusion type), and see whether there is any difference in persuasive effect. Indeed, this is overwhelmingly the most common sort of experimental design used in studies of persuasive effects. But this experimental design has important weaknesses, at least if one is interested in arriving at dependable generalizations about the persuasive effects of variations in message features (such as implicit vs. explicit conclusions). This design uses a single message to represent each general category (level, type) of the message variable. That is, the experiment compares one particular instance of an explicit conclusion message and one particular instance of an implicit conclusion message. Such singlemessage designs create two important barriers to generalization: One is 278



that the design does not permit unambiguous causal attribution; the other is that the design is blind to the possibility that the effects of a given message factor may not be constant (uniform) across different messages (Jackson & Jacobs, 1983).



Ambiguous Causal Attribution Although the logic of experimental research is designed to permit clear and unambiguous causal attribution, single-message designs inevitably create some ambiguity concerning the cause of any observed differences. This ambiguity arises because manipulating the variable of interest (say, implicit vs. explicit conclusion) inevitably means also concomitantly manipulating other variables that are not of interest. For example, suppose a researcher created the implicit and explicit conclusion messages in the following way. First, the explicit conclusion message was written. Then, to create the implicit conclusion message, the researcher simply eliminated the final paragraph (which contained the explicit conclusion). These two messages differ in conclusion type, but that is not the only thing that distinguishes the two messages. For one thing, the explicit conclusion message is now longer than the implicit conclusion message. It is probably apparent what difficulty this poses for arriving at generalizations. If the persuasiveness of the two messages differ, one’s initial inclination might well be to explain that difference as resulting from the type of conclusion used. But one could equally well suppose (given the evidence) that it was message length, not conclusion type, that created the difference. Worse, these are not the only two possibilities. The explicit conclusion message might be more repetitive than the implicit conclusion message, it might seem more insulting (because it says the obvious), or it might be more coherent or better organized, and so on. The problem is that one does not know whether conclusion type or some other variable leads to the observed difference in persuasiveness. There is another way of expressing this same problem: In a single-message design, the manipulation of a given message variable can be described in any number of ways, and, consequently, problems of causal attribution and generalization arise. Consider, for example, the following experimental manipulation. Two persuasive messages are prepared arguing in favor of making the sale of cigarettes illegal. Both messages emphasize the harmful 279



physical consequences of smoking and indeed are generally similar, except for the following sort of variation. Message A reads, “There can be no doubt that cigarette smoking produces harmful physical effects,” whereas Message B reads, “Only an ignorant person would doubt that cigarette smoking produces harmful physical effects.” The statement in Message A, “It is therefore readily apparent that the country should pass legislation to make the sale of cigarettes illegal,” is replaced in Message B by the statement, “Only the stupid or morally corrupt would oppose passage of legislation to make the sale of cigarettes illegal” (with four such alterations in the messages). What is the independent variable under investigation here? That is, how should this experimental manipulation be described? Framing some causal generalization will require that the difference between the two messages be expressed somehow—but exactly how? Several different answers have been offered. The original investigators described this manipulation as a matter of “opinionated” as opposed to “nonopinionated” language (G. R. Miller & Baseheart, 1969), but others have characterized it varying “language intensity” (Bradac, Bowers, & Courtright, 1980, pp. 200–201), having a “confident style in debating” (Abelson, 1986, p. 227), or using a “more dynamic” rather than “subdued” style (McGuire, 1985, p. 270). Of course, not even these exhaust the possibilities. For instance, one could describe this as a contrast between extreme and mild (or nonexistent) denigration of those holding opposing views. These different descriptions of the experimental manipulation could all be correct, but of course they are not identical. Unfortunately, if one wishes to frame a causal generalization from research using this concrete experimental manipulation, one must choose some particular description. But which one? Given this single-message design, all the various interpretations are equally good—which is to say, researchers cannot make the desired sorts of unambiguous causal attributions.



Nonuniform Effects of Message Variables A second barrier to generalization created by single-message designs arises from the possibility (or probability) that the effect of a given message variable will not be uniform (constant) across all messages. Consider again the example of a single-message study examining the effects of having implicit versus explicit conclusions. Suppose that this 280



study found the explicit conclusion message to be significantly more persuasive than the implicit conclusion message (and let’s overlook the problem of deciding that it was conclusion type, not some other factor, that was responsible for the difference). One should not necessarily conclude that having explicit conclusions will always improve the effectiveness of a message. After all, there might have been something peculiar about the particular message that was studied (remember that only one message was used). Perhaps something (unnoticed by the researchers) made that particular message especially hospitable to having an explicit conclusion— maybe the topic, maybe the way the rest of the message was organized, or maybe the nature of the arguments that were made. Other messages might not be so receptive to explicit conclusions. To put that point more abstractly: The effect of a given message variable may not be uniform across messages. Some messages might be helped a lot by having an explicit conclusion, some helped only a little, and some even hurt by it. But if that is true, then looking at the effects of conclusion type on a single message does not really provide a good basis for drawing a general conclusion. So, once again, the typical single-message design used in persuasion effects research creates an obstacle to dependable message generalization because it overlooks the possibility of nonuniform effects across messages.4



Designing Future Persuasion Research It has probably already occurred to the reader that there is a straightforward way of dealing with these two obstacles to dependable message generalization. Because those two obstacles arise from the use of a single message to represent an entire category of messages, the straightforward solution is to use multiple messages to represent each category (Jackson & Jacobs, 1983; for discussion of parallel approaches in other contexts, see Highhouse, 2009; Wells & Windschitl, 1999). For example, a study of implicit versus explicit conclusions would want to have many instances of each message type with as much variation within a category as one could achieve (variation in topic, organization, length, etc.). With such a multiple-message design, the possibility of nonuniform effects across messages is acknowledged. There is no presumption that the effect of conclusion type will be constant across messages; on the contrary, the design may permit the detection of variation in the effect that conclusion 281



type has across messages. And the chances for unambiguous causal attribution are improved by such a design: Given the variation within the set of messages for a given conclusion type, the researcher can rule out alternative explanations and be more confident in attributing observed differences to the sort of conclusion used.5 Beyond the desirability of using multiple-message designs, the generalization problems associated with single-message designs also have implications for how experimental messages are constructed in persuasion research. A number of complex considerations bear on the question of how to construct (or obtain) experimental messages (for discussion, see Bradac, 1986; Jackson, 1992, pp. 131–149, 1993; Jackson & Jacobs, 1983). But a sense of the relevant implications can be obtained by focusing on one particular research practice: the practice of using the same experimental messages more than once. The problem of generalizing about message types from individual messages would be serious enough if each investigation of persuasive message effects used only one message to represent each message type (category) but with a different concrete message used in each study (so that, e.g., every investigation of implicit versus explicit conclusions used only one instance of each type, but every investigation created a new instance of each type). But the problem is worse because sometimes the same messages are used more than once in persuasion research (consider, e.g., Pratkanis, Greenwald, Ronis, Leippe, & Baumgardner, 1986). For example, related messages were used by Burnkrant and Howard (1984), Petty, Cacioppo, and Heesacker (1981), and Swasy and Munch (1985); by Holbrook (1978) and Venkatraman, Marlino, Kardes, and Sklar (1990); by B. T. Johnson (1994), Petty and Cacioppo (1979b, 1984), Solomon, Greenberg, Psyczynski, and Pryzbylinski (1995), and Sorrentino, Bobocel, Gitta, Olson, and Hewitt (1988); by Kamins and Assael (1987a, 1987b) and Kamins and Marks (1987); by Lalor and Hailey (1990) and Meyerowitz and Chaiken (1987); by Mann, Sherman, and Updegraff (2004), Sherman, Mann, and Updegraff (2006), and Updegraff, Sherman, Luyster, and Mann (2007); by Shiv, Edell, and Payne (1997, 2004); and by W. Wood (1982), W. Wood and Kallgren (1988), and W. Wood et al. (1985). This practice is readily understandable. First, the task of creating satisfactory experimental materials is difficult and time-consuming, and if existing messages already represent the variables of interest, then it can be 282



awfully tempting to employ those materials. Second, in a continuing line of research, a desire for tight experimental control may suggest the reuse of earlier messages. But in the end this way of proceeding is unsatisfactory, precisely because it complicates, rather than eases, the task of obtaining sound causal generalizations. It complicates this task because single-message instantiations are an unsatisfactory basis for generalizations about message types.6 Even a very large number of studies of one message (or message pair) cannot provide evidence for generalizations beyond that message.



Interpreting Past Persuasion Research Employing multiple-message designs (and avoiding reusing experimental messages) may help future researchers avoid the message generalization problems created by single-message designs, but a great deal of earlier research relied on single-message designs. Obviously, any such singlemessage study should be interpreted very cautiously. The interpretive difficulties created by single-message designs are such that one cannot confidently make broad generalizations from a single study using such a design. But if several single-message studies address the same research question, then some greater confidence may be warranted. If 10 investigations compare explicit and implicit conclusions, and each one has a different single example of each message category, a review that considers the body of studies taken as a whole can transcend this limitation of the individual investigations and provide a sounder basis for generalization. Meta-analytic statistical techniques can be particularly helpful here. Broadly, meta-analysis is a family of quantitative techniques for summarizing the results obtained in a number of separate research studies. Using meta-analytic techniques, a researcher can systematically examine the different results obtained in separate investigations, combine these separate studies to yield a picture of the overall effect obtained, look for variations among the results of different studies, and so on. (For general treatments, see Borenstein, Hedges, Higgins, & Rothstein, 2009; Card, 2012; Field & Gillett, 2010. For a single comprehensive source, see Cooper, Hedges, & Valentine, 2009.) Obviously, meta-analysis offers the possibility of overcoming some of the limitations of existing research using single-message designs.7 283



However, two aspects of meta-analytic practice deserve some notice. The first concerns which studies to include in a review, specifically whether to attempt to locate unpublished studies (conference papers, dissertations, and so forth). There are good reasons to want to include unpublished studies whenever possible, because the published research literature may produce a misleading picture. For example, studies that find statistically significant differences may be more likely to be published than those with nonsignificant results. (For some discussions of publication biases and related questionable research practices, see Bakker, van Dijk, & Wicherts, 2012; Ioannidis, 2005, 2008; John, Loewenstein, & Prelec, 2012; Levine, Asada, & Carpenter, 2009.) The second concerns how to analyze meta-analytic data, specifically the choice between fixed-effect and random-effects models. The technical details need not detain us here, but results derived from fixed-effect analyses characteristically overstate the precision of meta-analytic findings —and only random-effects models permit generalization beyond the cases in hand, which is usually the goal of undertaking meta-analytic reviews (see Borenstein, Hedges, Higgins, & Rothstein, 2010; Card, 2012, pp. 233–234).8 Even so, the use of fixed-effect analyses is distressingly common (for some discussion, see Cafri, Kromrey, & Brannick, 2010).



Beyond Message Variables This discussion has focused on the message-generalizing problems associated with single-message designs. These problems are especially salient for persuasion researchers because—despite widespread interest in generalizing across messages—single-message designs have been the norm for studies of persuasive effects (but see Brashers & Jackson, 1999). Of course, the same general considerations apply not just to message factors but to anything; dependable generalizations about a collection of things (messages, people, tables, and so on) commonly require multiple instances of the class. There is, that is, nothing unique about these problems of message generalization. But some focused attention to matters of message generalization is important, if only because single-message designs are so common.



Variable Definition The other noteworthy challenge that arises in studying persuasive effects 284



concerns how independent variables are defined in research practice. Because this issue arises most clearly in the context of defining message variables (i.e., message variations or message types), the following discussion focuses on such variables; as will be seen, however, the difficulties that ensue are not limited to message variables.



Message Features Versus Observed Effects Broadly put, a message variable can be defined in one of two ways: on the basis of intrinsic message features or on the basis of observed effects on recipients. Most message variables have been defined using message features (as one might expect), but occasionally investigators have defined message types using engendered recipient responses (i.e., the effects observed in message receivers). Because these two ways of defining message variations are consequentially different, the distinction is an important one. A useful example is provided by the extensive research on threat appeals (also called fear appeals) in persuasive messages. A threat appeal is a particular type of persuasive message, but it has been defined in varying ways. Some investigators define a threat appeal as a message that contains certain sorts of message content (e.g., graphic depictions of consequences of not following the communicator’s recommendations, as in gruesome films of traffic accidents in driver education classes). But for other investigators, a threat appeal message is one that arouses fear or anxiety in message recipients (that is, threat appeal is defined by the responses of message receivers). Obviously, these two definitions will not necessarily correspond. That is, a message that contains gruesome content (a threat appeal by the first definition) might not arouse fear or anxiety in message recipients (i.e., might not be a threat appeal by the second definition). Similarly, a message might succeed in arousing fear without containing graphic message content.



The Importance of the Distinction It is important to be clear about the different ways of defining message variables (by reference to message features or by reference to effects) because the distinction is consequential. (For discussion, see O’Keefe, 2003; Tao & Bucy, 2007.) 285



First, generalizations about message types can only cautiously lump together investigations that employ different ways of defining a given variable. Two studies might call themselves studies of threat appeals, but if one of them defines threat appeal by message content whereas the other defines it by recipient response, it may be difficult to draw reliable generalizations that encompass the two studies. Second, the different ways of defining message types raise different evidentiary issues concerning the soundness of experimental manipulations. Consider that to construct defensible examples of threat appeal messages for use in research, an investigator who defines threat appeal by the presence of certain sorts of message content need only ensure that the messages do contain the requisite sort of content. By contrast, an investigator who defines threat appeal by audience responses must show that the messages engender the required responses.9 Third, and most important, feature-based definitions offer several advantages over effect-based definitions. In particular, research using feature-based definitions can give obvious direct advice for persuaders concerning the construction of persuasive messages (“Put features X, Y, and Z in your message”), whereas effect-based definitions are likely to be much less helpful (“Do something that engenders such-and-such effects”).10 Similarly, exploration of the role of mediating states (intervening between messages and effects) can be obscured with effectbased definitions (O’Keefe, 2003). The extended example in this discussion of the problem of variable definition has been that of a particular message variable (threat appeals), although other message variables—most notably, the variable of argument strength that figures prominently in elaboration likelihood model research (discussed in Chapter 8)—could have served as well. But these issues of variable definition are not limited to message factors. For example, communicator credibility could be defined by observed persuasive effects (so that, by definition, higher credibility would be associated with greater persuasion) or by other criteria (such as receivers’ impressions of the source’s believability, expertise, honesty, and the like; see Chapter 10).



Conclusion Experimental research examining the influence of various factors on 286



persuasive outcomes offers the prospect of useful insights into persuasive processes and effects, but the task of creating dependable generalizations from such research can be more challenging than might appear at first look.



For Review 1. Describe the simplest experimental design used in the study of persuasive message effects. What is an independent variable? A dependent variable? Explain how experimental designs are meant to permit unambiguous causal attribution for observed effects. Describe some variations on the basic design. What is an interaction effect? Explain how experimental designs with more than one independent variable permit detection of interaction effects. Explain the difference between a conclusion about the relative persuasiveness of two messages and a conclusion about the absolute persuasiveness of one message. Are experimental designs meant to provide evidence about the absolute persuasiveness of a given message? 2. What is a single-message experimental design? Explain why such designs do not provide good evidence for generalizations. How do such designs undermine unambiguous causal attribution for effects? How do such designs overlook the potential variability (nonuniformity) of a given message factor across different concrete messages? What is a multiple-message design? How do multiplemessage designs address some concerns about generalization? Why is re-using messages (from earlier studies) a generally undesirable research practice? How can generalizations be obtained from previous studies using single-message designs? What is meta-analysis? 3. Describe two ways in which a message variable can be defined in research. Explain the difference between defining a message variable on the basis of message features and defining it on the basis of recipient responses (effects). Describe how the criteria for satisfactory experimental manipulations differ depending on whether a featurebased or an effect-based definition is used. Explain why feature-based definitions provide a better basis for advice to message designers.



Notes 1. An interaction effect can also be described as a “moderator” effect, in the sense that one variable moderates (influences) the effect of another 287



variable. To continue the example: If the effect of implicit-versus-explicit conclusions varies depending on the level of credibility, then credibility would be said to moderate the effect of conclusion type (credibility would be a moderator variable). Moderator effects, in which variable X influences the relationship of variables A and B, are different from mediator effects, in which X mediates the relationship of A and B by being between them in a causal chain (A influences X, which in turn influences B). The classic treatment of this distinction is Baron and Kenny (1986). For some subsequent discussion, see Fairchild and MacKinnon (2009), Green, Ha, and Bullock (2010), Kraemer, Kiernan, Essex, and Kupfer (2008), Preacher and Hayes (2008), Spencer, Zanna, and Fong (2005), and Zhao, Lynch, and Chen (2010). 2. In experimental research concerning relative persuasiveness, the most common persuasive outcome assessments have been attitudes, intentions, and behaviors. (Attitude assessment is briefly discussed in Chapter 1. For some treatment of intention and behavior assessments, see Fishbein & Ajzen, 2010, pp. 29–43.) Although attitudes, intentions, and behaviors are commonly generally positively correlated (see Chapter 6 concerning reasoned action theory), it is also the case that the persuasiveness of a given message might vary across these outcomes (e.g., in response to a given message, people might change their attitudes a lot, their intentions only a little, and their behaviors not at all). However, where the research question concerns the relative persuasiveness of two message kinds (are messages of type A more persuasive than messages of type B?), those different outcomes yield substantively identical research conclusions: If message type A is more persuasive than message type B with attitudinal outcomes, it is also more persuasive (and equally more persuasive) with intention outcomes and with behavioral outcomes. More carefully: The mean effect sizes (describing the difference in persuasiveness between two message types) for attitudinal outcomes, for intention outcomes, and for behavior outcomes are statistically indistinguishable (O’Keefe, 2013b). 3. Such potential confusions should also be borne in mind when experimental results are described (as they should be) using effect sizes. In experimental studies comparing the persuasiveness of two message types, the effect size of interest describes the size and direction of the difference between the two message conditions—not the persuasiveness of either message. And when effect sizes are compared (as when, e.g., the effect size for a given message variable is examined under two different conditions), the same cautions are relevant. If the effect size for a given 288



message variable is larger under condition X than under condition Y, that does not mean that messages are more persuasive in condition X than in condition Y; it means that the difference in persuasiveness between the two messages is larger in condition X than in condition Y. As an example: O’Keefe and Jensen’s (2011) meta-analysis of gain-loss message framing effects concerning obesity-related behaviors reported that the mean effect size (expressed as a correlation, with positive values indicating a persuasive advantage for gain-framed appeals) was .17 for physical activity messages and .02 for healthy eating messages, a statistically significant difference. This does not mean that physical activity messages were more persuasive than healthy eating messages, or that gain-framed physical activity messages were more persuasive than gain-framed healthy eating messages. It means only that the difference in persuasiveness between gain-framed and loss-framed appeals was larger for physical activity messages than for healthy eating messages. 4. Indeed, heterogeneity (variability, nonuniformity) in persuasive message effects is not just an abstract possibility. Meta-analytic reviews of persuasive message variables can provide straightforward evidence here, because such reviews contain multiple estimates of the size of a given variable’s effect. In such reviews, effect size variability is common and substantial; indeed, it is rare for the mean effect size to be larger than the standard deviation and common for the variability to be twice or three times as large as might be expected given human sampling error (for a review, see O’Keefe, 1999c). For examples of primary research studies reporting such variability from multiple-message designs, see Greene, Krcmar, Rubin, Walters, and Hale (2002), Reichert, Heckler, and Jackson (2001), and Siegel et al. (2008). 5. It’s not enough just to have multiple messages; an appropriate statistical analysis (a random-effects analysis) is also required. For a general treatment of these matters, see Jackson (1992). For additional discussion, see M. Burgoon, Hall, and Pfau (1991), Jackson and Brashers (1994), Jackson, O’Keefe, and Brashers (1994), and Slater (1991). For discussion of such issues in other research contexts, see Baayen, Davidson, and Bates (2008), H. H. Clark (1973), Fontenelle, Phillips, and Lane (1985), Judd, Westfall, and Kenny (2012), Raaijmakers, Schrijnemakers, and Gremmen (1999), Rietveld and van Hout (2007), and Siemer and Joormann (2003). 6. When the same message is used repeatedly as the instantiation of a message type, meta-analytic techniques (discussed shortly) are less helpful 289



than they might be as a means of coping with message generalization problems. Indeed, where messages have been reused in primary research and generalizations about message types are wanted, the appropriate metaanalytic procedure is to collapse the results (across studies) for a given message pair; that is, the appropriate unit of analysis is the message pair, not the study. To concretize this: Imagine two data sets in which 20 studies have provided the experimental contrast of interest (comparing a message of kind A versus a message of kind B). In one data set, each study used a different message pair. In the other, 10 studies used one specific message pair and the other 10 studies used a second message pair. Plainly, one’s confidence in any generalizations would be greater in the first data set than in the second, and the meta-analytic procedure should reflect that (making the number of cases 20 in the first data set and 2 in the second). 7. Meta-analytic methods can be especially attractive because they naturally shift the focus away from statistical significance (was there a statistically significant difference in persuasiveness between the two message kinds?) and toward effect sizes (how large was the difference in persuasiveness between the two message kinds?) and confidence intervals (what is the likely range of population values for the effect, given the data in hand?). In addition, meta-analysis can encourage conceiving of message effects in terms of the effect size distribution (with some mean and variance) associated with a given variable (see Brashers & Jackson, 1999). 8. These questions (about the statistical treatment of meta-analytic data) arise in parallel form in the context of multiple-message primary research designs (see note 5 above). 9. This can create some confusion about “manipulation checks” concerning message variables. Where message variations are defined on the basis of intrinsic features, customary manipulation checks (under that description) are unnecessary and inappropriate (O’Keefe, 2003). 10. As an illustration, consider Stephenson et al.’s (2005) research in which hearing protection messages were pretested to identify messages evoking positive, negative, and neutral affect; in a subsequent study, the relative persuasiveness of these messages was then tested. The finding that the affectively neutral messages were most successful gives only limited help to future message designers because the message categories were defined not by some intrinsic properties of the messages but by the reactions the messages evoked. 290



Chapter 10 Communicator Factors Communicator Credibility The Dimensions of Credibility Factors Influencing Credibility Judgments Effects of Credibility Liking The General Rule Some Exceptions and Limiting Conditions Other Communicator Factors Similarity Physical Attractiveness About Additional Communicator Characteristics Conclusion The Nature of Communication Sources Multiple Roles for Communicator Variables For Review Notes



Persuasion researchers have quite naturally focused considerable research attention on the question of how various characteristics of the communicator influence the outcomes of the communicator’s persuasive efforts. This chapter’s review of such research is focused on two particular communicator characteristics—the communicator’s credibility and likability—but also considers other source factors, including the communicator’s similarity to the audience.



Communicator Credibility The Dimensions of Credibility Credibility (or, more carefully expressed, perceived credibility) consists of the judgments made by a perceiver (e.g., a message recipient) concerning the believability of a communicator. Communicator credibility is thus not an intrinsic property of a communicator; a message source may be thought highly credible by one perceiver and not at all credible by another. But this general notion of credibility has been given a somewhat more careful 291



specification in investigations aimed at identifying the basic underlying dimensions of credibility.



Factor-Analytic Research There have been quite a few factor-analytic studies of the dimensions underlying credibility judgments (e.g., Andersen, 1961; Applbaum & Anatol, 1972, 1973; Baudhuin & Davis, 1972; Berlo, Lemert, & Mertz, 1969; Bowers & Phillips, 1967; Falcione, 1974; McCroskey, 1966; Schweitzer & Ginsburg, 1966). In the most common research design in these investigations, respondents rate communication sources on a large number of scales. The ratings given of the sources are then submitted to factor analysis, a statistical procedure that (broadly put) groups the scales on the basis of their intercorrelations: Scales that are comparatively highly intercorrelated will be grouped together as indicating some underlying “factor” or dimension.



Expertise and Trustworthiness as Dimensions of Credibility Without overlooking potential weaknesses in this research (see, e.g., Delia, 1976; McCroskey & Young, 1981) or the variations in obtained factor structures (compare, e.g., Berlo et al., 1969, with Schweitzer & Ginsburg, 1966), one may nevertheless say that with some frequency, two broad (and sensible) dimensions have commonly emerged in factor-analytic investigations of communicator credibility. These are variously labeled in the literature, but two useful terms are expertise and trustworthiness. The expertise dimension (sometimes called competence, expertness, authoritativeness, or qualification) is commonly represented by scales such as experienced-inexperienced, informed-uninformed, trained-untrained, qualified-unqualified, skilled-unskilled, intelligent-unintelligent, and expert–not expert. These items all seem directed at the assessment of (roughly) whether the communicator is in a position to know the truth, to know what is right or correct. Three or more of these scales are reported as loading on a common factor in investigations by Applbaum and Anatol (1972), Baudhuin and Davis (1972), Beatty and Behnke (1980), Beatty and Kruger (1978), Berlo et al. (1969), Bowers and Phillips (1967), Falcione (1974), McCroskey (1966), Pearce and Brommel (1972), Schweitzer and Ginsburg (1966), and Tuppen (1974). And (as these factor-analytic results would indicate) measures of perceived expertise that are composed of 292



several such items commonly exhibit high internal reliability (e.g., reliability coefficients of .85 or greater have been reported by Beatty & Behnke, 1980; McCroskey, 1966). The trustworthiness dimension (sometimes called character, safety, or personal integrity) is commonly represented by scales such as honestdishonest, trustworthy-untrustworthy, open-minded–closed-minded, justunjust, fair-unfair, and unselfish-selfish. These items all appear to be related to the assessment of (roughly) whether the communicator will likely be inclined to tell the truth as he or she sees it. Three or more of these scales are reported as loading on a common factor in investigations by Applbaum and Anatol (1972), Baudhuin and Davis (1972), Berlo et al. (1969), Falcione (1974), Schweitzer and Ginsburg (1966), Tuppen (1974), and Whitehead (1968). Correspondingly, indices of perceived trustworthiness that are composed of several such items have displayed high internal reliability (e.g., reliabilities of .80 or better have been reported by Bradley, 1981; Tuppen, 1974).1 These two dimensions parallel what have been described as the two types of communicator bias that message recipients might perceive: knowledge bias and reporting bias. “Knowledge bias refers to a recipient’s belief that a communicator’s knowledge about external reality is nonveridical, and reporting bias refers to the belief that a communicator’s willingness to convey an accurate version of external reality is compromised” (Eagly, Wood, & Chaiken, 1978, p. 424; see also Eagly, Wood, & Chaiken, 1981). A communicator perceived as having a knowledge bias will presumably be viewed as relatively less expert; a communicator viewed as having a reporting bias will presumably be seen as comparatively less trustworthy. Perhaps it is not surprising that both expertise and trustworthiness emerge as basic dimensions of credibility because as a rule, only the conjunction of expertise and trustworthiness makes for reliable communications. A communicator who knows what is correct (has expertise) but who nevertheless misleads the audience (is untrustworthy, has a reporting bias) produces messages that are unreliable guides to belief and action, just as does the sincere (trustworthy) but uninformed (low-expertise, knowledgebiased) communicator. These two dimensions, however, represent only the most general sorts of credibilityrelevant judgments made by recipients about communicators. The particular judgments underlying credibility may vary from 293



circumstance to circumstance, as can the emphasis placed on one or another dimension of judgment. Thus it may be useful to develop credibility assessments tailored to particular situations (for some examples, see Frewer, Howard, Hedderley, & Shepherd, 1996; Gaziano & McGrath, 1986; Hilligoss & Rieh, 2008; Ohanian, 1990; M. D. West, 1994). Notably, however, even such situation-specific assessments commonly identify expertise and trustworthiness as key credibility dimensions, as in studies of expert courtroom witnesses (Brodsky, Griffin, & Cramer, 2010), risk communication (Siegrist, Earle, & Gutscher, 2003; Twyman, Harvey, & Harries, 2008), and corporations (Newell & Goldsmith, 2001).



Factors Influencing Credibility Judgments Judgments of a communicator’s expertise and trustworthiness are surely influenced by a great many factors, and it is fair to say that research to date leaves us rather far from a comprehensive picture of the determinants of these judgments. What follows is a selective review of some features that have received relatively more attention.



Education, Occupation, and Experience Although little systematic research investigates exactly how credibility judgments are influenced by information about the communicator’s training, experience, and occupation, these characteristics are precisely the ones most frequently manipulated by investigators in experimental studies of the effects of variations in communicator credibility. That is, a researcher who wishes to compare the effects of a high-credibility source with those of a low-credibility source will most commonly manipulate the receiver’s perception of the communicator’s credibility by varying the information given about the communicator’s credentials. For instance, a classic study of messages about nuclear radiation described the highcredibility communicator as “a professor of nuclear research, recognized as a national authority on the biological effects of radioactivity,” whereas the low-credibility introduction described the source as “a high school sophomore, whose information was based on a term paper prepared for a social studies class” (Hewgill & Miller, 1965, p. 96). Similar manipulations are commonplace, and researchers commonly confirm the success of these manipulations by assessing respondents’ judgments of the communicators’ expertise and trustworthiness. As one might expect, such high-credibility introductions do indeed generally lead 294



receivers to perceive the source as more trustworthy and (particularly) more expert than do low-credibility introductions. What systematic research exists on this matter is (perhaps not surprisingly) consistent with these effects (e.g., Falomir-Pichastor, Butera, & Mugny, 2002; Hurwitz, Miron, & Johnson, 1992; Tormala, Briñol, & Petty, 2006).



Nonfluencies in Delivery There have been a number of investigations of how variations in delivery can influence the credibility judgments made of a speaker. Unfortunately, several of these studies have investigated conceptions of delivery that embrace a number of behavioral features (e.g., Pearce & Brommel, 1972). But one delivery characteristic that has been studied in isolation is the occurrence of nonfluencies in the delivery of oral communications. Nonfluencies include vocalized pauses (“uh, uh”), the superfluous repetition of words or sounds, corrections of slips of the tongue, articulation difficulties, and the like. Several investigations have found that with increasing numbers of nonfluencies, speakers are rated significantly lower on expertise, with judgments of trustworthiness typically unaffected (e.g., Engstrom, 1994; for a review of effects on expertise judgments, see Carpenter, 2012a).



Citation of Evidence Sources Persuaders commonly include evidence in their persuasive messages, that is, relevant facts, opinions, information, and the like, intended to support the persuader’s claims. Several investigations have studied how citing the sources of such evidence—as opposed to providing only vague documentation (“Studies show that”) or no documentation at all— influences perceived communicator credibility. On the whole, a communicator’s citation of the sources of evidence appears to enhance perceptions of the communicator’s expertise and trustworthiness, although these effects are sometimes small (e.g., Whitehead, 1971; for reviews and discussion, see O’Keefe, 1998; Reinard, 1998). These investigations employed relevant supporting materials that were attributed (when source citations were provided) to high-credibility sources. One should not expect enhanced communicator credibility to result from citations to lowcredibility evidence sources or from citations for poor or irrelevant evidence (Luchok & McCroskey, 1978). But the citation of expert and trustworthy sources of evidence in the message appears to influence the communicator’s perceived expertise and trustworthiness; in a sense, the 295



high credibility of the cited sources seems to rub off on the communicator.2



Position Advocated The position that the communicator advocates on the persuasive issue can influence perceptions of the communicator’s expertise and trustworthiness. Specifically, a communicator is likely to be perceived as more expert and more trustworthy if the advocated position disconfirms the audience’s expectations about the communicator’s views (when such expectations derive from knowledge of the source’s characteristics or circumstances), although certain sorts of trustworthiness judgments (concerning objectivity, openmindedness, and unbiasedness) appear to be more affected than others (such as sincerity and honesty). The most straightforward examples of this phenomenon are communicators who argue for positions that are apparently opposed to their own self-interest. Ordinarily, of course, we expect persons to take positions that forward their own interests; sources who support views opposed to their interests thus disconfirm our expectations. If we wonder why a source is taking this (apparently unusual) position, we may well be led to conclude that the communicator must be especially well-informed (expert) and honest (trustworthy): The source must really know the truth and must really be willing to tell the truth, otherwise why would the source be advocating that position? Conversely, where communicators advocate views that forward their self-interest, credibility is likely to suffer by comparison. So, for example, online contributors who donate their revenue shares to charity (as compared with those who retain the revenue; Hsieh, Hudson, & Kraut, 2011), salespeople whose compensation is salary-based (rather than commission-based; Straughan & Lynn, 2002) or who flatter the customer after the sale (rather than before; M. C. Campbell & Kirmani, 2000), prosecutors who argue that prosecutorial powers should be decreased (rather than increased; Walster, Aronson, & Abrahams, 1966), politicians who praise their opponents (as opposed to denigrating them; Combs & Keller, 2010)—all are likely to enjoy relatively enhanced credibility. Similarly, physicians find their colleagues more believable sources of drug information than they do advertisements or salespeople (Beltramini & Sirsi, 1992), and they are more likely to prescribe drugs that have been studied in government-funded research than in industry-funded research 296



(Kesselheim et al., 2012). Of course, receivers’ expectations about the position that a communicator will express can derive from sources other than the ordinary presumption that people will favor viewpoints that are in their own interest. A general analysis of the bases of premessage expectancies (and their effects on perceived credibility and persuasive outcomes) has been provided by Eagly et al. (1981). As briefly mentioned earlier, Eagly et al. (1981) distinguished two sorts of perceived communicator bias that receivers can use to form premessage expectancies about the communicator’s position. One is knowledge bias, which refers to the receiver’s belief that the communicator’s knowledge of relevant information is somehow biased (perhaps because of the source’s background or experience) and thus that the source’s message will not accurately reflect reality. The other is reporting bias, which refers to the receiver’s belief that a communicator may not be willing to accurately convey relevant information (for instance, because situational pressures might lead the source to withhold or distort information). A receiver’s perception of either sort of communicator bias will lead the receiver to have certain expectations about the position that a communicator will express on the issue. When communicators confirm those expectations (e.g., the lifelong Democrat speaking in favor of a Democratic political candidate, or a speaker opposing gun control legislation when addressing the National Rifle Association), we have ready explanations for why the communicators acted as they did. But when a communicator advocates a position that violates an expectancy based on knowledge or reporting bias, the receiver faces the task of explaining why the communicator is defending the advocated position— why the lifelong Democrat is speaking in support of a Republican candidate or why the speaker addressing the National Rifle Association is urging stricter gun control legislation. The most plausible explanation at least sometimes will be that the facts of the matter were so compelling that the communicator was led to override those personal or situational pressures (that had generated the receiver’s expectations) and thus defend the advocated position. Correspondingly, the receiver may be led to perceive the communicator as especially expert and trustworthy, precisely because the communicator’s expressed position violates the receiver’s expectations (for relevant research, see L. Anderson, 1970; Eagly & Chaiken, 1975; Eagly et al., 1978; Peters, Covello, & McCallum, 1997; W. Wood & Eagly, 1981).3



297



A related expectancy disconfirmation effect has been observed in studies of advertisements for consumer products. Ordinarily, consumers expect advertisements to tout the advertised product or brand as “the best” on every feature or characteristic that is mentioned. Thus an advertisement for exterior house paint that claimed that the product was superior to its competitors on only three mentioned product features (durability, number of coats needed, and ease of cleanup) while being equal on two others (number of colors available and nonspill lip on container) would disconfirm receivers’ expectations about the message’s contents (particularly by contrast to an advertisement claiming that the product was superior on each of these five features; R. E. Smith & Hunt, 1978). There have been several experimental comparisons of these two types of advertisements—an advertisement suggesting superiority for all the mentioned features of the product (a one-sided advertisement), and an advertisement that acknowledges (and does not refute or deny) some ways in which the product is not superior (a nonrefutational two-sided advertisement). As one might suppose, when an advertisement acknowledges ways in which competing products are just as good as the advertised product (or acknowledges weaknesses of the advertised product), the ad is commonly perceived as more credible than when the ad claims superiority on every product feature that is mentioned (e.g., Alden & Crowley, 1995; Eisend, 2010; Pechmann, 1992; for reviews, see Eisend, 2006; O’Keefe, 1999a).4



Liking for the Communicator Some indirect evidence indicates that the receiver’s liking for the communicator can influence judgments of the communicator’s trustworthiness, although not judgments of the communicator’s expertise. This evidence, derived from factor-analytic investigations of credibility judgments, is the finding that various general evaluation items often load on the same factor as do trustworthiness scales. For example, items such as friendly-unfriendly, pleasant-unpleasant, nice–not nice, and valuableworthless have been reported as loading on a common factor with such trustworthiness items as honest-dishonest, trustworthyuntrustworthy, unselfish-selfish, and just-unjust (see, e.g., Applbaum & Anatol, 1972; Bowers & Phillips, 1967; Falcione, 1974; McCroskey, 1966; Pearce & Brommel, 1972). This suggests that liking and trustworthiness judgments are probably more likely to co-vary than are liking and expertise judgments. Such a pattern of results surely makes good sense: One’s general liking for a communicator is much more likely to influence one’s 298



judgments about the communicator’s dispositional trustworthiness (the communicator’s general honesty, fairness, open-mindedness, and the like) than the communicator’s expertise.5



Humor Including humor in persuasive messages has been found to have rather varied effects on perceptions of the communicator. When positive effects of humor are found, they tend to most directly involve enhancement of the audience’s liking for the communicator—and thus occasionally the trustworthiness of the communicator (because liking and trustworthiness are associated)—but rarely judgments of expertise (e.g., Chang & Gruner, 1981; Gruner, 1967, 1970; Gruner & Lampton, 1972; Skalski, Tamborini, Glazer, & Smith, 2009). The use of humor, however, can also decrease the audience’s liking for the communicator, the perceived trustworthiness of the communicator, and even the perceived expertise of the source (e.g., Bryant, Brown, Silberberg, & Elliott, 1981; Munn & Gruner, 1981). These negative effects seem most likely when the humor is perceived as excessive or inappropriate for the context. Small amounts of appropriate humor thus may have small enhancing effects on perceived trustworthiness but are unlikely to affect assessments of the communicator’s expertise.



Summary This selective review has touched on several broad factors that can influence credibility judgments. However, different specific influences may be at work in different persuasion circumstances. So, for example, one might expect that different factors will affect the perceived credibility of courtroom witnesses (e.g., Dahl et al., 2007), blogs (e.g., Armstrong & McAdams, 2009), journalists (e.g., Jensen, 2008), online reviews (e.g., Pan & Chiou, 2011), and so forth.6



Effects of Credibility What effects do variations in communicator credibility have on persuasive outcomes? It might be thought that the answer to this question is pretty simple: As one’s credibility increases, so will one’s effectiveness. But the answer turns out to be much more complicated.



Two Initial Clarifications 299



Two preliminary clarifications need to be made concerning the research on the effects of communicator credibility. The first is that in this research, the two primary dimensions of credibility (expertise and trustworthiness) are usually not separately manipulated. That is, the research commonly compares a source that is relatively high in both expertise and trustworthiness (the high-credibility source) with a source that is relatively low in both (the low-credibility source). Obviously, because expertise and trustworthiness are conceptually distinct aspects of credibility, it would be possible to manipulate these separately and so examine their separate effects on persuasive outcomes. One could, for instance, compare the effectiveness of a source high in expertise but low in trustworthiness with that of a source low in expertise but high in trustworthiness. Overwhelmingly, however, expertise and trustworthiness have not been independently manipulated in investigations of credibility’s effects. There have been a few efforts at disentangling the effects of expertise and trustworthiness (e.g., Mowen, Wiener, & Joag, 1987; O’Hara, Netemeyer, & Burton, 1991; Terwel, Harinck, Ellemers, & Daamen, 2009), but to date no clear generalizations seem possible.7 The point, thus, of this first clarification is to emphasize the limits of current research on credibility’s effects: This research concerns credibility generally, rather than different dimensions of credibility individually.8 The second preliminary clarification concerns the nature of the lowcredibility sources in this research: The low-credibility sources are not low in absolute terms but simply relatively low in credibility. In absolute terms, the low-credibility communicators are probably accurately described as no better than moderate in credibility.9 Several researchers have remarked that it is difficult to create believable experimental manipulations that will consistently yield credibility ratings that are low in absolute terms.10 Thus although this discussion (like most in the literature) will be cast as a matter of the differential persuasive effectiveness of highas opposed to low-credibility communicators, the comparison made in the relevant research is nearly always between a relatively higher credibility communicator and a relatively lower one, not necessarily between two sources that are in absolute terms high and low in credibility. With these preliminaries out of the way, we can now turn to a consideration of just how variations in communicator credibility influence 300



persuasive effectiveness. The effects of credibility on persuasive outcomes are not completely straightforward but depend centrally on other factors. These factors can be usefully divided into two general categories: factors that influence the magnitude of credibility’s effects and factors that influence the direction of credibility’s effects.



Influences on the Magnitude of Effect The size of the effect that communicator credibility has on persuasive outcomes is not constant but varies from one circumstance to another. Researchers have identified at least two factors that affect just how consequential a role communicator credibility plays in persuasion. The first is the degree of direct personal relevance that the issue has for the receiver. As the issue becomes more personally relevant for the receiver, variations in the source’s credibility make less difference; under conditions of low personal relevance, the communicator’s credibility may make a great deal of difference to the outcome, whereas on highly relevant topics, the source’s credibility may have little impact (for a classic illustration, see Petty, Cacioppo, & Goldman, 1981; for a review, see E. J. Wilson & Sherrell, 1993).11 In some ways, it may seem paradoxical that as an issue becomes more personally relevant for a receiver, the source’s expertise and trustworthiness become less important. But this relationship may be more understandable when viewed from the perspective of the elaboration likelihood model (ELM; see Chapter 8). For issues of little personal relevance, receivers may be content to let their opinions be shaped by the communicator’s apparent credibility; for such an issue, it is not worth the effort to follow the details of the arguments. But for highly relevant topics, receivers will be more likely to attend closely to the details of the message, to scrutinize the communicator’s arguments and evidence, and to invest the effort involved in thinking closely about the contents of the message— and this comparatively greater importance of the message contents means that the communicator’s credibility will play a smaller role than it otherwise might have.12 Indeed, this first factor might be cast in a more general framework, by suggesting that as the amount of issuerelevant thinking varies (whether because of personal relevance or other factors), so will the effect of credibility variations. For example, Kumkale, Albarracín, and Seignourel’s (2010) meta-analysis indicated that the effect of credibility variations is greatest when recipients have poorly formed 301



attitudes and little background knowledge—conditions likely to be conducive to relatively low elaboration. The second factor influencing the magnitude of credibility’s impact is the timing of the identification of the communicator. Often, of course, the communicator’s identity is known before the message is received by the audience (e.g., because the source is well-known and can be seen by the audience or because another person introduces the communicator). But in some circumstances, it can be possible to delay identification of the source until after the audience has been exposed to the message (e.g., in television advertisements, in which the source’s identity may be withheld until the end of the ad, or in multipage web or print articles, in which information about the writer may not appear on the first page but instead only at the end). The timing of the identification of the source does make a substantial difference in the role that source credibility plays in persuasion. Specifically, the impact of communicator credibility appears to be minimized when the identity of the source is withheld from the audience until after the message has been presented (e.g., Allen et al., 2002; O’Keefe, 1987; see, relatedly, Nan, 2009; Tormala, Briñol, & Petty, 2007). When the communicator’s identity is delayed until after the audience has received the message, the message is apparently heard more nearly on its own terms, without the influence of the communicator’s credibility. It might be thought that this finding implies that high-credibility communicators should be sure not to delay their identification (but instead should be sure to identify themselves before the message), whereas lowcredibility communicators should strive, where circumstances permit, to have their messages received before the audience is given information about their credibility. But that is a mistaken conclusion because it is based on an unsound (although natural) assumption that the direction of credibility’s effect is constant, with higher credibility always yielding greater persuasion. As discussed in the next section, sometimes lowercredibility communicators can be more successful persuaders than highercredibility sources.



Influences on the Direction of Effect One might plausibly suppose that the direction of credibility’s effect would be constant—specifically, that increases in credibility would yield only increases in persuasive effectiveness. Perhaps sometimes only small 302



increases would occur, and sometimes (e.g., when the topic is personally relevant to the receiver) no increase at all, but at least whenever credibility had an effect, it would be in a constant direction, with high-credibility sources being more effective than low-credibility sources. However plausible such a supposition may seem, it is not consistent with the empirical evidence. The direction of credibility’s effect is not constant: Several investigations have found that at least sometimes low-credibility communicators are significantly more effective than high-credibility communicators (e.g., Bock & Saine, 1975; Bohner, Ruder, & Erb, 2002; Chebat, Filiatrault, Laroche, & Watson, 1988; Dholakia, 1987; FalomirPichastor, Butera, & Mugny, 2002; Harmon & Coney, 1982; Sternthal, Dholakia, & Leavitt, 1978; Tormala, Briñol, & Petty, 2006; for some discussion, see Pornpitakpan, 2004). This finding is not easily impeached, as these results have been obtained by different investigators, using various topics, with different participant populations, and with good evidence for the success of the credibility manipulations employed. An entirely clear picture is not yet in hand, but one factor that appears critical in determining the direction of credibility’s effects appears to be the nature of the position advocated by the message—specifically, whether the message advocates a position initially opposed by the receiver (a counterattitudinal message) or advocates a position toward which the receiver initially feels at least somewhat favorable (a proattitudinal message). With a counterattitudinal message, the high-credibility communicator will tend to have a persuasive advantage over the lowcredibility source; with a proattitudinal message, however, the lowcredibility communicator appears to enjoy greater persuasive success than the high-credibility source. The most direct evidence of this relationship comes from investigations that have varied the counter- or proattitudinal stance of the message (under conditions of low topic relevance and with communicators identified prior to messages). Under these conditions, high-credibility communicators are more effective than low-credibility communicators with counterattitudinal messages, but this advantage diminishes as the advocated position gets closer to the receiver’s position, to the point that with proattitudinal messages, the low-credibility communicator is often more effective than the high-credibility source (Bergin, 1962; Bochner & Insko, 1966; Chebat et al., 1988; J. K. Clark, Wegener, Habashi, & Evans, 2012; Harmon & Coney, 1982; McGinnies, 1973; Sternthal et al., 1978, Study 2). 303



Perhaps one way of understanding this effect is to consider the degree to which, given a proattitudinal message, receivers might be stimulated to think about arguments and evidence supporting the advocated view. When receivers hear their views defended by a high-credibility source, they may well be inclined to presume that the communicator will do a perfectly good job of advocacy, will defend the viewpoint adequately, will present the best arguments, and so forth—and so they sit back and let the source do the work. But when the source is low in credibility, receivers might be more inclined to help the communicator in defending their common viewpoint, and hence they might be led to think more extensively about supporting arguments—thereby ending up being more persuaded than if they had listened to a higher-credibility source. Expressed in ELM terms, a proattitudinal message may provoke more elaboration, and more favorable elaboration, when it comes from a low-credibility communicator than when it comes from a high-credibility communicator (for some evidence consistent with such an account, see Clark et al., 2012; Sternthal et al., 1978).13 However, greater success of low-credibility communicators should not be expected in every case of proattitudinal messages, nor should one expect that high-credibility communicators will have an edge whenever counterattitudinal messages are employed. Rather, one should find such effects only when the conditions promote credibility’s having a substantial effect (e.g., only when the topic is not especially personally relevant and the communicator is identified prior to the message).



Liking The General Rule Perhaps it comes as no surprise that a number of investigations have found support for the general principle that on the whole, liked communicators are more effective influence agents than are disliked communicators (e.g., Eagly & Chaiken, 1975; Sampson & Insko, 1964).14 But the general principle that liked persuaders are more successful can be misleading. Important exceptions and limiting conditions on that principle are discussed in the following section.



Some Exceptions and Limiting Conditions 304



Extant research evidence suggests at least three important caveats concerning the effects of liking for the communicator on persuasive outcomes: The effects of liking can apparently be overridden by credibility, the superiority of liked over disliked communicators is minimized as the topic becomes more personally relevant to the receiver, and disliked communicators can at least sometimes be significantly more effective persuaders than can liked communicators. (For indications of additional possible limiting conditions, see Chebat, Laroche, Baddoura, & Filiatrault, 1992; Roskos-Ewoldsen & Fazio, 1992.)



Liking and Credibility The effects of liking on persuasive outcomes appear to be weaker than the effects of credibility (e.g., Lupia & McCubbins, 1998, pp. 196–199; Simons, Berkowitz, & Moyer, 1970; see, relatedly, Eisend & Langner, 2010). Thus when the receiver’s judgment of the source’s credibility conflicts with the receiver’s liking for the source, the effects of liking may be overridden by the effects of credibility. This may be exemplified by the results of an investigation in which participants were asked to make a judgment about the size of the monetary award to be given in a personal injury damage suit. Each participant heard a persuasive message from a source who advocated either a relatively small or a relatively large monetary award; the source was portrayed either as cold and stingy or as warm and generous. Although the warm, generous source was liked better than was the cold, stingy communicator, the stingy source was nevertheless sometimes a more effective persuader, namely, when the stingy source was arguing for a relatively large award. Indeed, of the four source-message combinations, the two most effective combinations were the stingy source arguing for a large award and the generous source arguing for a small award (Wachtler & Counselman, 1981). Both these combinations, of course, represent sources who are (given their personalities) advocating an unexpected position and who thus may well have been perceived as relatively higher in credibility. Of particular interest is that the communicator who was disliked and (presumably) high in credibility (the stingy source advocating the large award) was significantly more effective than the communicator who was liked and (presumably) low in credibility (the generous source advocating the large award), thus suggesting that the effects of liking for the communicator are weaker than the effects of communicator credibility.



305



Liking and Topic Relevance The effects of liking on persuasive outcomes are minimized as the topic becomes more personally relevant to the receiver. Thus, although betterliked sources may enjoy some general persuasive advantage, that advantage is reduced when the issue is personally relevant to the receiver (Chaiken, 1980; see, relatedly, Kang & Kerr, 2006). This result is, of course, compatible with the image offered by the ELM (discussed in Chapter 8). When receivers find the topic personally relevant, they are more likely to engage in systematic active processing of message contents and to minimize reliance on peripheral cues such as whether they happen to like the communication source. But when personal relevance is low, receivers are more likely to rely on simplifying heuristics emphasizing cues such as liking (“I like this person, so I’ll agree”).



Greater Effectiveness of Disliked Communicators At least sometimes disliked communicators can be more effective persuaders than liked communicators—even when the communicators are comparable in other characteristics such as credibility. A demonstration of this possibility was provided by a classic investigation in which participants were induced to eat fried grasshoppers. In one condition, the communicator acted snobbish, cold, bossy, tactless, and hostile (the disliked communicator); the liked communicator displayed none of these characteristics. The two communicators were roughly equally successful in inducing participants to eat the fried grasshoppers, but that is not the result of interest. What is of interest is how, among participants who did eat the grasshoppers, their attitudes toward eating grasshoppers changed. As one might predict from dissonance-theoretic considerations, among those who ate the grasshoppers, the disliked communicator was much more effective in changing attitudes in the desired direction than was the liked communicator: The person who ate the grasshoppers under the influence of the disliked communicator presumably experienced more dissonance— and thus exhibited more attitude change—than did the person induced to eat by the liked source (Zimbardo, Weisenberg, Firestone, & Levy, 1965). This, of course, is the familiar induced compliance counterattitudinal action circumstance (as discussed in Chapter 5). But similar results have been obtained in straightforward persuasive communication situations (J. Cooper, Darley, & Henderson, 1974; Himmelfarb & Arazi, 1974; R. A. Jones & Brehm, 1967; cf. Eagly & Chaiken, 1975). That is, disliked 306



communicators can indeed potentially be more persuasive than liked communicators. However, when disliked communicators have been found to be more successful persuaders than liked communicators, the circumstances appear to typically have involved the receivers’ having freely chosen to listen to the message. For example, in J. Cooper et al.’s (1974) investigation, suburban householders received a counterattitudinal communication (i.e., one opposed to the receiver’s views) from either a deviant-appearing communicator (a long-haired hippie) or a conventional-appearing communicator. The deviant-appearing communicator was significantly more effective than the conventionally dressed communicator in persuading these suburbanites, but the message recipients all had freely chosen to receive the communication (and indeed had had two opportunities to decline to receive the communication). If one remembers that dissonance effects are expected in induced compliance circumstances only when the person has freely chosen to engage in the discrepant action, the finding that disliked communicators can be more successful than liked communicators only under conditions of choice is perhaps not surprising. Receivers who freely choose to listen to (what turns out to be) an unlikable communicator presumably face a dissonance reduction task that is not faced by receivers who find themselves (through no fault of their own) listening to an unlikable source. Hence the greater success of disliked (as opposed to liked) communicators is, as the research evidence suggests, obtained only when the receiver has chosen to listen to the message. (For an experimental illustration of this moderating factor, see R. A. Jones & Brehm, 1967.)



Other Communicator Factors Beyond credibility and liking, a large number of other communicator factors have received at least some research attention as possible influences on persuasive outcomes. This section focuses on two such factors—similarity and physical attractiveness—but concludes with a more general discussion of other communicator characteristics.



Similarity It seems common and natural to assume that to the degree that receivers 307



perceive similarities between themselves and a persuader, to that same degree the persuader’s effectiveness will be enhanced. The belief that “greater similarity means greater effectiveness” is an attractive one and is commonly reflected in recommendations that persuaders emphasize commonalities between themselves and the audience. But the relationship of similarity to persuasive effectiveness is much more complex than this common assumption indicates. Indeed, while some research findings indicate that persuasive effectiveness can be enhanced by similarity (e.g., Brock, 1965; Woodside & Davenport, 1974), other findings suggest that persuasive effectiveness can be reduced by similarity (e.g., Infante, 1978; S. W. King & Sereno, 1973; Leavitt & Kaigler-Evans, 1975) or that similarity has little or no effect on persuasive outcomes (e.g., Klock & Traylor, 1983; Wagner, 1984). Two initial clarifications will be helpful in untangling these complexities. First, there is “an infinite number of possible dimensions” of similaritydissimilarity (Simons et al., 1970, p. 3). One might perceive oneself to be similar or dissimilar to another person in age, occupation, attitudes, physique, income, education, speech dialect, personality, ethnicity, political affiliation, interpersonal style, clothing preferences, and on and on. Thus there is not likely to be any truly general relationship between similarity and persuasive effectiveness, or indeed between similarity and any other variable. Different particular similarities or dissimilarities will have different effects, making impossible any sound generalization about similarity. Second, similarities most likely do not influence persuasive effectiveness directly (R. G. Hass, 1981; Simons et al., 1970). Rather, similarities influence persuasive outcomes indirectly, especially by affecting the receiver’s liking for the communicator and the receiver’s perception of the communicator’s credibility (expertise and trustworthiness). Because the effects of similarities may not be identical for liking, perceived expertise, and perceived trustworthiness, the relationship of similarities to each of these needs separate attention.



Similarity and Liking Given the infinite varieties of possible similarities, any general relationship is unlikely between perceived similarity and liking for another person. That is, “there is no singular ‘similarity’ effect” on liking but rather “a 308



multiplicity of effects that depend on both content and context” (Huston & Levinger, 1978, p. 126). However, the effect on liking of one particular sort of similarity—attitudinal similarity—has received a good deal of empirical attention. Attitudinal similarity is having similar attitudes (similar evaluations of attitude objects), as opposed to, say, having similar traits, abilities, occupations, or backgrounds. A fair amount of evidence now indicates that as a general rule, perceived attitudinal similarity engenders greater liking, at least among previously unacquainted persons (for reviews, see Berscheid, 1985; Byrne, 1969; Sunnafrank, 1991). Thus to the extent that message recipients perceive that the communicator has attitudes (on matters other than the topic of the influence attempt) that are similar to theirs, those recipients are likely to come to like the communicator more. Hence, even when not especially relevant to the topic of the influence attempt, perceived attitudinal similarities (between source and audience) can enhance the audience’s liking of the source and so can potentially influence persuasive effectiveness. The hypothesis that attitudinal similarities can influence persuasive effectiveness by influencing the receiver’s liking for the communicator is bolstered by the results of investigations that have varied both communicator credibility (specifically, expertise) and communicatorreceiver attitudinal similarity. As discussed previously, the effects of liking on persuasive effectiveness appear to be weaker than the effects of credibility. Thus, if attitudinal similarities influence persuasive effects by influencing liking for the communicator, then the effect of attitudinal similarities on persuasive effectiveness should be smaller than the effect of credibility. Indeed, several studies have found persuasive success to be more influenced by the communicator’s expertise than by the communicator’s attitudinal similarity (Wagner, 1984; Woodside & Davenport, 1974). But enhanced liking of a communicator will not always mean enhanced persuasive effectiveness; as discussed earlier, greater liking for a communicator may enhance, reduce, or have no effect on persuasive effectiveness. Correspondingly, greater perceived attitudinal similarities may (through their influence on the receiver’s liking for the communicator) enhance, reduce, or have no influence on persuasive effectiveness. Thus one should not assume that with greater perceived attitudinal similarity comes greater persuasive effectiveness. Rather, with 309



greater perceived attitudinal similarity comes greater liking, which may or may not mean greater effectiveness.15



Similarity and Credibility: Expertise Judgments Perceived similarities (or dissimilarities) between source and audience can certainly influence the audience’s judgment of the source’s expertise. But there are two noteworthy features of this relationship. First, the similarity or dissimilarity must be relevant to the influence attempt if it is likely to influence judgments of expertise. For example, a communicator seeking to influence a receiver’s judgment of the president’s budget policy will probably not obtain enhanced expertise judgments by pointing out that the communicator and recipient are wearing the same color shirt. In a study that varied the communicator’s occupational similarity (student vs. nonstudent, for an audience of students) in advertisements for several consumer products, receivers’ judgments of the source’s expertise were found to be unrelated to judgments of perceived similarity (Swartz, 1984); presumably, the variations in similarity were not relevant to the persuasive issues involved. Only relevant similarities (or dissimilarities) are likely to influence judgments of the communicator’s expertise. Second, not all relevant similarities will enhance perceived expertise, and not all relevant dissimilarities will damage perceived expertise. For example, a perceived similarity in relevant training and experience may reduce the perceived expertise of a communicator (because the receiver may be thinking, “I know as much about this topic as the speaker does”). A perceived dissimilarity in relevant training and experience, on the other hand, might either enhance or damage perceived expertise, depending on the direction of the dissimilarity: If the receiver thinks that the communicator is dissimilar because the communicator has better training and experience, then presumably enhanced judgments of the communicator’s expertise will be likely, but if the receiver thinks that the communicator is dissimilar because the communicator has poorer training and experience, then most likely the communicator’s perceived expertise will suffer. A demonstration of this sort of complexity was provided by a study of speech dialect similarity, in which persons who spoke a general American dialect heard one of two versions of a message from a speaker using either 310



a general American dialect or a Southern dialect; the message concerned a well-known Southern governor (who enjoyed some popularity in the South but not elsewhere), with one version offering a favorable view of the governor and the other an unfavorable view. Regardless of the position advocated, the speaker with the Southern (dissimilar) speech dialect was perceived as more expert than the speaker with the general American (similar) dialect, presumably because the Southern speaker could be assumed to have better access to relevant information than would the general American speaker (Delia, 1975). Thus similarities should have varying effects on perceived expertise, depending on the particulars of the circumstances. One should not be surprised that the research literature indicates that similar others are sometimes seen as more expert than are dissimilar others (e.g., Mills & Kimble, 1973), sometimes as less expert (e.g., Delia, 1975), and sometimes as not significantly differing in expertise (e.g., Atkinson, Winzelberg, & Holland, 1985; Swartz, 1984). The effects of perceived similarities and dissimilarities on judgments of communicator expertise depend on whether, and how, the receiver perceives these as relevant to the issue at hand.



Similarity and Credibility: Trustworthiness Judgments The relationship between similarities and judgments of the communicator’s trustworthiness appears to be complex as well. As previously mentioned, certain sorts of similarities—specifically, perceived attitudinal similarities—can influence the receiver’s liking for the communicator, and enhanced liking for the communicator is commonly accompanied by enhanced judgments of the communicator’s trustworthiness. One would thus expect that perceived attitudinal similarities might (through their influence on liking) exert some effect on perceptions of the communicator’s trustworthiness. This interplay of attitudinal similarity, liking, and trustworthiness judgments is nicely illustrated in research by Meijinders et al. (2009). The perceived trustworthiness of a journalist writing about genetically modified food was greater when the journalist was described as having similar (as opposed to dissimilar) attitudes about an unrelated subject. One can see this effect as straightforwardly reflecting how attitudinal similarity can enhance liking, which in turn enhances perceived trustworthiness.



311



However, there are intricacies here. In the previously described speech dialect investigation, greater trustworthiness was ascribed to the progovernor speaker using the similar (general American) dialect and to the antigovernor speaker using the dissimilar (Southern) dialect (Delia, 1975). This effect is, of course, readily understandable: The Southern speaker arguing against the Southern governor and the non-Southern speaker supporting that governor could each have been seen as offering views that ran against the tide of regional opinion—and hence seen as speakers who must be especially sincere and honest in their expressions of their opinions. But notice the complexity of these results regarding similarity: Sometimes similarity enhanced perceptions of trustworthiness, but sometimes it diminished such perceptions, depending on the position advocated. And (to round things out) other investigators have found that sometimes similarities have no significant effect on trustworthiness judgments (e.g., Atkinson et al., 1985).



Summary: The Effects of Similarity Perhaps it is now clear just how inadequate is a generalization such as “greater similarity leads to greater persuasive effectiveness.” The effects of similarity on persuasive outcomes are complex and indirect, and no single easy generalization will encompass those varied effects. Indeed, in several instances, similarities have been found in a single investigation to enhance persuasive effectiveness under some conditions but to inhibit persuasive effectiveness under other circumstances (e.g., Goethals & Nelson, 1973; S. W. King & Sereno, 1973; Mills & Kimble, 1973). So consider, as an example, the effects of salient similarities or dissimilarities in group membership (“Ah, this communicator is a student at my university”) on persuasion. Such group membership similarities might provide bases for inferences about likely attitudinal similarities between receiver and communicator or more generally might provide bases for inferences about likability or credibility. Hence (following ELMlike reasoning), on topics that are not especially personally relevant to the receiver, such similarities might serve as peripheral cues (that engage corresponding heuristics) and thus enhance the persuasiveness of messages from similar communicators (those sharing group membership with the receiver). By contrast, on topics of greater personal relevance (for which greater message scrutiny is likely), these peripheral cue effects of group membership similarities may be diminished. On such personally relevant topics, however, group membership similarity might encourage closer 312



scrutiny of messages from similar communicators; such closer scrutiny might enhance or inhibit persuasion, depending on such factors as the quality of the message’s arguments. (For some relevant empirical work and general discussions, see M. A. Fleming & Petty, 2000; Mackie & Queller, 2000; N. Wyer, 2010.) Such complexities might lead one to wonder about the common practice of using peers (of the target audience) in health education programs on such topics as smoking and unsafe sex; this practice can be seen to reflect a generalized belief in the persuasive power of similarity. Given the observed complexities of similarity’s roles and effects in persuasion, however, perhaps it should not be surprising that several reviews have concluded that peer-based health interventions are not dependably more successful—and sometimes are significantly less successful—than programs without such peer bases (Durantini, Albarracín, Mitchell, Earl, & Gillette, 2006; Posavac, Kattapong, & Dew, 1999; cf. Cuijpers, 2002).16 In sum, if there is a general conclusion to be drawn about source-receiver similarities in persuasion, it surely is that simple generalizations will not do. To say, for example, that “receivers are more likely to be persuaded by communicators they perceive as similar to themselves” is to overlook the complexities of the effects that similarities have on persuasive outcomes.



Physical Attractiveness The effects of physical attractiveness on persuasive outcomes—like the effects of similarity—are rather varied. For the most part, “existing research does indicate that heightened physical attractiveness generally enhances one’s effectiveness as a social influence agent” (Chaiken, 1986, p. 150; for some illustrative examples, see Horai, Naccari, & Fatoullah, 1974; Micu, Coulter, & Price, 2009; Widgery & Ruch, 1981). But physical attractiveness appears to commonly operate in persuasion in a fashion akin to similarity; that is, physical attractiveness affects persuasive outcomes indirectly, by means of its influence on the receiver’s liking for the communicator and the receiver’s assessment of the communicator’s credibility.



Physical Attractiveness and Liking Unsurprisingly, greater physical attractiveness tends to lead to greater 313



liking (for a review, see Berscheid & Walster, 1974). And, as discussed previously, there is good evidence for the general proposition that on the whole, liked communicators will be more effective persuaders than disliked communicators. Hence the observed effects of physical attractiveness on persuasive success might straightforwardly be explained as arising from the recipient’s liking for the communicator (for a careful elaboration of this idea, see Chaiken, 1986; for some illustrative results, see Horai et al., 1974; Snyder & Rothbart, 1971). In addition to the parallel overall general effect on persuasion (i.e., the parallelism of the generally positive effect of attractiveness on persuasion and the generally positive effect of liking on persuasion), there is also at least some evidence that the complexities attendant to liking’s persuasive effects also attach to attractiveness’s effects. Specifically, there are indications that (a) credibility can be a more important determinant of persuasion than physical attractiveness (e.g., Maddux & Rogers, 1980), (b) the effect of physical attractiveness on persuasion is reduced as elaboration increases (e.g., Kang & Kerr, 2006; see Chaiken, 1986, for a discussion of the preponderance of low-relevance topics in studies reporting significant effects of communicator physical attractiveness on persuasion), and (c) an unattractive communicator can under some circumstances be a more successful persuader than an attractive one (Buunk & Dijkstra, 2011; J. Cooper et al., 1974; Kang & Kerr, 2006). These parallel effects thus further strengthen the case for supposing that many of the effects of communicator physical attractiveness on persuasive outcomes can best be explained by the hypothesis that “physical attractiveness affects social influence via its more direct impact on liking for the social influence agent” (Chaiken, 1986, p. 151).



Physical Attractiveness and Credibility It is, however, also conceivable that physical attractiveness can affect persuasive outcomes through its effects on perceived credibility. But a clear treatment of this possibility requires separate consideration of the expertise and trustworthiness dimensions of credibility. Concerning the effects of attractiveness on expertise: Investigations that have found physically attractive persuaders to be more successful than unattractive persuaders have typically not found the attractive communicators to be rated higher in expertise (e.g., Chaiken, 1979; Horai et al., 1974; Snyder & Rothbart, 1971; see also R. Norman, 1976; cf. 314



Patzer, 1983; and Praxmarer, 2011). Thus it is not plausible to suppose that differential judgments of the communicator’s expertise generally mediate the effect of communicator physical attractiveness on persuasive outcomes. To be sure, in certain specific circumstances, the communicator’s physical attractiveness might influence judgments of expertise—namely, when the topic of influence is related to physical attractiveness in relevant ways. For example, physically attractive sources might enjoy greater perceived expertise in the realm of beauty products. But generally speaking, the effect of the source’s physical attractiveness on persuasive outcomes appears not to be achieved through enhanced perceptions of the source’s expertise. Concerning the effects of attractiveness on trustworthiness: Physical attractiveness may (at least indirectly) influence judgments of the communicator’s trustworthiness. Physical attractiveness influences liking for the communicator; and, as discussed earlier, there is at least some indirect evidence that the receiver’s liking for the communicator can influence the receiver’s judgment of communicator trustworthiness. But this roundabout path of influence is likely to mean that physical attractiveness will have only weak effects on trustworthiness judgments: If the effect of communicator physical attractiveness on trustworthiness judgments is mediated by the receiver’s liking for the communicator, then (given that liking for the communicator can be influenced by many other things besides physical attractiveness) one should expect that the effect of attractiveness on trustworthiness will be less strong than the effect of attractiveness on liking (as was found by Patzer, 1983) and indeed will typically be comparatively small, even negligible; Maddux and Rogers (1980), for example, found that physically attractive persuaders were indeed better liked but were not rated as significantly more sincere or honest than were their physically unattractive counterparts (for related results, see Snyder & Rothbart, 1971).



Summary Understanding the role that communicator physical attractiveness plays in influencing persuasive outcomes seems to require that central emphasis be given to the influence of physical attractiveness on liking. Physical attractiveness appears to affect persuasive outcomes not directly but rather indirectly, especially (though not exclusively) by means of its influence on the receiver’s liking for the communicator.



315



About Additional Communicator Characteristics This discussion of the persuasive effects of communicator-receiver similarity and communicator physical attractiveness has focused on how those factors might influence credibility and liking, because the research evidence seems to indicate that similarity and physical attractiveness influence persuasive outcomes indirectly, through their effects on credibility and liking. Indeed, in thinking about the effects of any given additional source characteristic on persuasion, one useful avenue to illuminating that characteristic’s effects can be a consideration of how that characteristic might influence credibility or liking (and thereby indirectly influence persuasive outcomes). Such an approach will reveal considerable complexities in how a given factor might eventually influence persuasive effects. Consider, as an example, the question of the comparative persuasive success of communicators varying in ethnicity (e.g., a Latino communicator and an Anglo communicator) in influencing receivers who also vary in ethnicity. The answer to this question almost certainly varies from case to case, depending on the particulars involved. With one topic, the Latino communicator may be perceived (by all receivers, no matter their ethnicity) to be more credible than the Anglo communicator; with a different topic, the Anglo communicator may be perceived as more credible (no matter the receiver’s ethnicity); with yet another topic, there may be no credibility differences associated with the ethnicity of the communicator; or the credibility judgments may depend not just on the topic addressed but also on the position advocated (as in the previously discussed study of speech dialects, which had a topic on which regional differences in knowledge and attitude were likely). But (to add to the complexity here) these credibility judgments may not influence persuasive outcomes substantially because variations in credibility are not always associated with variations in effects; even when variations in these credibility judgments are associated with variations in outcomes, sometimes the lower-credibility communicator will be more effective than the higher-credibility source. When one adds the complex relationship between ethnicity and credibility to the complex relationship between credibility and persuasive outcomes, the result is a rococo set of possibilities for the relationship between ethnicity and persuasive effects. (And notice: The discussion of this example has focused only on the direct ethnicity-credibility relationship; the picture of ethnicity’s effects becomes even more complex when one considers in addition the ethnicity-liking 316



relationship or the role of perceived ethnic similarity.)



Conclusion As perhaps is apparent, communicator characteristics can have complicated relationships with each other and can have various direct and indirect effects on persuasive outcomes. But two further complexities deserve mention: the nature of communication sources and the multiple roles that communicator variables might play in persuasion.



The Nature of Communication Sources The treatment of communicator factors here—and, for the most part, in the research literature—rests implicitly on an image of the communicator as an individual person communicating through linguistic means (speaking or writing). In such cases, it is straightforward to conceptualize properties such as the recipient’s liking for the source or the source’s perceived credibility. But when persuasive messages take other forms, the nature of the communication “source” undergoes some transformation. Consider, for example, consumer advertisements, in which often there is no identifiable individual person who is the communicator. Advertisements themselves, however, are recognizable objects toward which people might have attitudes (evaluations) and—given their communicative function—they are objects whose credibility might appropriately be assessed. That is, just as message recipients might have attitudes toward, and credibility perceptions concerning, some identifiable person delivering a message, so they can have attitudes toward advertisements and perceptions of advertisement credibility. And thus corresponding research questions can arise about the conceptualization and assessment of ad liking (attitudes toward the advertisement) and ad credibility (e.g., Eisend, 2002), about the factors that influence ad liking and ad credibility (e.g., Kelly, Slater, & Karan, 2002; Lord, Lee, & Sauer, 1995), and about the persuasive effects of variations in ad liking and ad credibility (e.g., S. P. Brown & Stayman, 1992; Nan & Zhao, 2010; Smit, van Meurs, & Neijens, 2006). Even in the context of advertisements, however, there are further complications introduced by the possibility of advertisements’ using spokespeople or endorsers. For example, in a consumer product ad that 317



uses a celebrity endorser, message recipients might have relevant attitudes about the endorser (how well liked the endorser is), perceptions of the endorser’s credibility (expertise, trustworthiness), attitudes toward the ad as a whole, and perceptions of the ad’s overall credibility. (For some illustrative research on endorsers, see Amos, Holmes, & Strutton, 2008; Austin, de Vord, Pinkleton, & Epstein, 2008; Biswas, Biswas, & Das, 2006; Eisend & Langner, 2010; Lafferty, Goldsmith, & Newell, 2002; Magnini, Garcia, & Honeycutt, 2010; Ohanian, 1990.)17 Websites (of all sorts, including personal or institutional webpages, wikis, blogs, etc.) provide a similar case. As with advertisements, people might have attitudes toward, and credibility perceptions concerning, websites. And thus corresponding research questions can arise about the conceptualization and assessment of website credibility and liking (e.g., Hu & Sundar, 2010; Metzger, Flanagin, Eyal, Lemus, & McCann, 2003; Walther, Wang, & Loh, 2004) and about the antecedents and consequences of variations in website credibility and liking (e.g., Cheung, Sia, & Kuan, 2012; Flanagin & Metzger, 2007; Hong, 2006; Lim, 2013; Metzger, Flanagin, & Medders, 2010; Robins, Holmes, & Stansbury, 2010; Yi, Yoon, Davis, & Lee, 2013). The general point is that message “sources” can take a variety of forms— people, advertisements, websites, and so forth—and consequently the nature and operation of source characteristics (such as credibility) naturally may vary across these different communication formats. Parallel research questions will arise (about the nature, antecedents, and effects of source characteristics), but the answers can be expected to differ.



Multiple Roles for Communicator Variables A still larger complexity is also to be borne in mind here, namely, that a variable can play many roles in persuasion (as suggested by the elaboration likelihood model; see Chapter 8). For example, when a highercredibility communicator is observed to be more persuasive than a lowercredibility communicator, this might have occurred because the communicator’s apparent credibility served as a cue (and so engaged a credibility-based heuristic), because the higher-credibility communicator engendered greater message scrutiny than did the lower-credibility communicator (in a circumstance in which the message had strong arguments), because the higher-credibility communicator engendered less message scrutiny than did the lower-credibility communicator (in a 318



circumstance in which the message had weak arguments), or because the higher-credibility communicator more or less directly biased (influenced the evaluative direction of) elaboration in a way favorable to the advocated view. Moreover—apart from whatever influence credibility might otherwise have on the persuasiveness of a message—the communicator’s credibility may affect whether the communicator has access to the audience (e.g., editors may provide space in the op-ed section of a news outlet only to persons who appear to have relevant expertise) and whether the audience pays much attention to the message (i.e., credibility may influence message exposure or scrutiny). (For some examples of research illustrating such varied roles for communicator characteristics, see J. K. Clark, Wegener, & Evans, 2011; Howard & Kerin, 2011; Sinclair, Moore, Mark, Soldat, & Lavis, 2010; Tormala, Briñol, & Petty, 2007; Ziegler & Diehl, 2001.)18



For Review 1. What is credibility? What are the primary dimensions of credibility? What is expertise? Describe the questionnaire items commonly used to assess expertise. What is trustworthiness? Describe the questionnaire items commonly used to asses trustworthiness. Describe the research used to identify the primary dimensions of credibility. What is factor analysis? What is knowledge bias? Reporting bias? Explain the relationships of knowledge bias, reporting bias, expertise, and trustworthiness. 2. Identify factors influencing credibility. Which of these influence expertise and which trustworthiness? Describe the effect of knowledge of the communicator’s education, occupation, experience, and training on expertise and on trustworthiness. Describe the effect of nonfluencies in delivery on expertise and on trustworthiness. Describe the effect of citation of evidence sources on expertise and on trustworthiness. Describe the effect of the advocated position on expertise and on trustworthiness; explain the roles of knowledge bias and reporting bias in this phenomenon. Describe the effect of liking for the communicator on expertise and on trustworthiness. Describe the effects of humor on expertise and on trustworthiness. 3. In research on the effects of credibility variations, are expertise and trustworthiness usually manipulated separately? Explain. In this research, are the low-credibility communicators low in absolute terms or only relatively low? Explain. 319



4. Explain the idea that the magnitude of credibility’s effect on persuasive outcomes might vary. Identify two factors that influence the magnitude of credibility’s effect. Describe how the personal relevance of the topic influences the magnitude of credibility’s effect. Under what sort of relevance condition (high or low) will the effect of credibility be relatively larger? Describe how the timing of identification of the communicator influences the magnitude of credibility’s effect. What timing of identification leads to relatively larger effects of credibility? 5. Explain the idea that the direction of credibility’s effect on persuasive outcomes might vary. Identify a factor that influences the direction of credibility’s effect. Under what conditions will higher-credibility sources be more persuasive than lower-credibility sources? And under what conditions will the opposite effect occur? Describe a possible explanation for the latter effect. 6. What is the general rule of thumb concerning the effect of variations in liking (of the communicator) on persuasive outcomes? Explain how that general principle can be misleading (e.g., identify a limiting condition). Describe the relative strength of the effects of credibility and the effects of liking (on persuasive outcomes). Describe how variations in the personal relevance of the topic influence the effects of liking. What relevance conditions (high or low) lead to relatively larger effects of liking? Can a disliked communicator be more persuasive than a liked communicator? Can a disliked communicator be more persuasive than a liked communicator even when the two communicators are equivalent with respect to other characteristics (e.g., credibility)? Identify a necessary condition for an otherwise equivalent disliked communicator’s being more persuasive than a liked communicator. 7. Does perceived similarity influence persuasive outcomes directly or indirectly? Explain. Through what avenues does perceived similarity influence persuasive outcomes? What is attitudinal similarity? How does perceived attitudinal similarity influence liking? Can liking be influenced by perceived similarities that are not relevant to the message topic? Can perceived similarities influence judgments of communicator expertise? Identify a necessary condition for a perceived similarity to influence expertise judgments. Will all relevant perceived similarities enhance expertise? Will all relevant perceived dissimilarities diminish expertise? Explain. Can perceived similarities influence judgments of communicator trustworthiness? Explain. Why is it misleading to assume that greater perceived 320



similarity enhances persuasive effectiveness? 8. Does the physical attractiveness of the communicator influence persuasive outcomes directly or indirectly? Explain. Through what avenues does physical attractiveness influence persuasive outcomes? How does physical attractiveness influence liking? Can physical attractiveness enhance perceived expertise? Give an example. Can physical attractiveness enhance perceived trustworthiness? 9. Explain how other communicator characteristics (i.e., other than credibility, liking, similarity, and physical attractiveness) influence persuasive outcomes indirectly. 10. Explain how a communication “source” might not be an identifiable individual; give examples. Describe how questions abut the nature, antecedents, and effects of source characteristics can arise concerning such sources. Explain how communicator variables might play multiple roles in persuasion; give examples.



Notes 1. Not all the factors that in the research literature have been labeled “trustworthiness” (or “character,” “safety,” or the like) contain many of the items that here are identified as assessing trustworthiness (e.g., McCroskey, 1966). An important source of confusion is the apparent empirical association between a receiver’s liking for a communicator and the receiver’s judgment of the communicator’s trustworthiness; this covariation is reflected in factor analyses that have found items such as honest-dishonest, trustworthyuntrustworthy, and fair-unfair to load on the same factor with items such as friendly-unfriendly, pleasantunpleasant, nice–not nice, and valuable-worthless (see, e.g., Applbaum & Anatol, 1972; Bowers & Phillips, 1967; Falcione, 1974; McCroskey, 1966; Pearce & Brommel, 1972). This pattern can plausibly be interpreted as reflecting the effects of liking on trustworthiness judgments (receivers being inclined to ascribe greater trustworthiness to persons they like). But such empirical association should not obscure the conceptual distinction between trustworthiness and liking, especially because the empirical association is imperfect; see Delia’s (1976, pp. 374–375) discussion of Whitehead’s (1968) results, or consider the stereotypical used car salesman who is likable but untrustworthy. In this chapter, investigations are treated as bearing on judgments of trustworthiness only when it appears that trustworthiness (and not liking) has been assessed.



321



2. This finding (that citation of evidence sources can enhance perceptions of the communicator’s expertise and trustworthiness) may be seen to have implications for the elaboration likelihood model (ELM; see Chapter 8). Although source and message variables are not partitioned by the ELM as having intrinsically different roles to play in persuasion, it is clear that message materials might have implications for perceptions of source characteristics (as when advocacy of an unexpected position enhances perceptions of communicator trustworthiness and thereby engenders reduced message scrutiny; Priester & Petty, 1995). The finding under discussion points specifically to the possibility that variations in argumentative message content may alter impressions of the communicator’s credibility (Slater & Rouner, 1996). (As an aside: Compared with premessage identification, postmessage identification of communicators has sometimes been seen to yield more positive impressions of credibility [Ward & McGinnies, 1973]. This result might easily be understood as a consequence of participants’ larger reliance on message materials—which commonly appear to have been of good quality —as a basis for credibility judgments in conditions in which identification follows the message as compared with those in which it precedes the message.) In the context of the ELM, this implies that variations in argument strength might affect persuasive outcomes by providing what amounts to credibility-related cue information (for some relevant evidence, see Reimer, 2003). The existence of such a pathway, in turn, invites reconsideration of the commonly observed enhanced effect that argument strength manipulations have (on persuasive outcomes) under conditions of high personal relevance; that effect could come about through the use of a credibility-related heuristic (and not through anything such as genuinely thoughtful consideration of substantive arguments), as long as there was sufficiently close message scrutiny to permit receivers to notice whatever message elements are used as a basis for inferences about credibility. The point here is not that this pathway provides an entirely satisfactory account of the accumulated findings on this matter but only that the possibility of this pathway points to some complexities in untangling what lies behind the effects observed in ELM research. 3. Although expectancy disconfirmation can enhance perceptions of the communicator’s expertise and trustworthiness, the communicator likely to be perceived as the most expert and trustworthy may be a qualified source about whom the audience has no expectations so far as message position is concerned (see, e.g., Arnold & McCroskey, 1967; Weinberger & Dillon, 1980). 322



4. Estimates of the mean effect on credibility perceptions, expressed as a correlation, range from .16 to .22 (Eisend, 2006; O’Keefe, 1999a). Notably, the credibility-enhancing effect (of mentioning opposing considerations without refuting them) that obtains in consumer advertising messages is not found in other messages (e.g., those concerning public policy issues; O’Keefe, 1999a). It may simply be that skepticism about consumer advertising is substantially greater than that about public policy advocacy—and hence nonrefutational acknowledgment of potential counterarguments is more surprising when it occurs in consumer advertisements than when it occurs in other messages. 5. For results consistent with this expectation, see Marquart, O’Keefe, and Gunther (1995), who found that perceived attitudinal similarity (which can influence liking; see, e.g., Berscheid, 1985) influenced ratings of sources’ trustworthiness but not expertise. 6. For some reviews of research concerning factors affecting credibility in specific persuasion contexts, see Hoyt (1996), G. R. Miller and Burgoon (1982), and Wathen and Burkell (2002). 7. Some experimental manipulations do appear to be especially targeted to influencing perceptions of expertise (e.g., by providing background information about occupation and training), and it can be tempting to interpret such studies as speaking specifically to questions about the effects of variations in perceived expertise. But a manipulation of apparent expertise may also affect perceptions of trustworthiness, and thus the results of such a study should not necessarily be interpreted as reflecting distinctly expertise effects. (Indeed, at least in some domains, perceptions of expertise, trustworthiness, and attractiveness are sufficiently highly correlated that some recommend collecting these under a single global credibility construct; see Hoyt, 1996, concerning therapist credibility.) One way of protecting against this problem can be to assess perceptions of both expertise and trustworthiness and to examine the effects of the manipulation on each. (It will not be sufficient to show that the experimental manipulation significantly influenced expertise perceptions but not trustworthiness perceptions. Such a finding would not necessarily mean that the manipulation influenced expertise perceptions significantly more than it influenced trustworthiness perceptions.) But when the research question of interest concerns the relationship between perceived expertise (or perceived trustworthiness) and persuasive outcomes, analyzing data by examining the relationship between the experimental 323



manipulation and persuasive outcomes fails to provide relevant information; with such a research question, the relationship between the perceptual state (e.g., perceived expertise) and persuasive outcomes should be examined directly. 8. E. J. Wilson and Sherrell’s (1993) review reported that expertise manipulations appear to have larger effects on persuasive outcomes than do trustworthiness manipulations. But (as discussed in the preceding note) this does not speak to the question of the relative influence (on persuasive effects) of perceived expertise and perceived trustworthiness. Concerning perceived therapist credibility specifically, Hoyt’s (1996) review found (with a relatively small number of effect sizes) no reason to suppose that perceived expertise and perceived trustworthiness were differentially related to therapist influence; but the generality of such a result is an open question. 9. For example, in Terwel, Harinck, Ellemers, & Daamen’s (2009) study, the means on the trustworthiness index (average of three 7-point scales, with a midpoint of 4) were 5.24 for the high-credibility source and 3.52 for the low-credibility source. On Nan’s (2009) 7-point trustworthiness scale (midpoint of 4), the mean ratings were 4.41 for the high-credibility source and 4.00 for the low-credibility source. Such results are common (e.g., Bochner & Insko, 1966; Falomir-Pichastor, Butera, & Mugny, 2002; Greenberg & Miller, 1966, Experiment 1; H. H. Johnson & Scileppi, 1969; Sternthal et al., 1978; Tormala, Briñol, & Petty, 2006). 10. See, for example, Greenberg and Miller (1966) and Sternthal, Dholakia, and Leavitt (1978). This difficulty is consistent with studies of the ratings given to “ideal” high- and low-credibility communicators, which have found that when respondents are asked to indicate where a perfectly credible and a perfectly noncredible communicator would be rated on expertise and trustworthiness scales, the ratings are not at the absolute extremes (R. A. Clark, Stewart, & Marston, 1972; see also J. K. Burgoon, 1976). 11. As mentioned in Chapter 8 (concerning the elaboration likelihood model), a good deal of research concerning variations in personal relevance has used the term involvement as a label for this variable (and so the point under discussion could be phrased as a matter of credibility’s impact declining as the receiver’s involvement with the issue increases). But the term involvement has also been used to cover other variations in 324



the relationship that receivers have to the message topic and so, in interests of clarity, is avoided here. 12. As discussed in Chapter 8, the elaboration likelihood model (ELM) actually proposes a slightly more specific version of this generalization (about the decline of credibility’s effect on persuasive outcomes as personal relevance increases); it stresses that as elaboration declines (e.g., as topic relevance declines), credibility plays less of a role as a peripheral cue but may still influence persuasion through other mechanisms. But when (for instance) credibility influences persuasion through influencing the degree of message scrutiny, then (given that increasing or reducing message scrutiny might lead to either greater or lesser persuasion, depending on the outcomes of greater message scrutiny) credibility’s apparent relationship with persuasive outcomes will presumably also appear to weaken. On the one hand, there is the observed empirical regularity (as personal relevance increases, the simple apparent relationship between credibility variations and persuasive effects weakens), and on the other, an ELM-based explanation of that observed regularity (viz., that credibility has a lessened role as a peripheral cue but might serve in other roles that also would naturally make for a weaker apparent relationship); one presumably does not want to confuse these. 13. For a suggestion that parallel effects (i.e., generation of supportive advocacy for proattitudinal messages) may be more likely to occur when manifest argument quality is weak than when it is strong, see Akhar, Paunesku, & Tormala (2013). 14. Because variation in liking can affect persuasive outcomes, anything that influences the message recipient’s liking for the communicator thus might potentially influence persuasive outcomes (via its effects on liking). For example, doing a favor for the recipient can influence liking and thereby persuasion (e.g., Goei, Lindsey, Boster, Skalski, & Bowman, 2003). Or, as discussed in a subsequent section, perceived attitudinal similarities can influence liking and hence persuasion. But a word of caution: Something that enhances both liking and persuasion might do so without liking being the mechanism that produces the persuasive effects. For example, compliments (given to the recipient by the communicator) can enhance both liking and persuasion—but the increased liking might not be responsible for the increased persuasion (Grant, Fabrigar, & Lim, 2010).



325



15. Other kinds of perceived similarities (beyond attitudinal similarities) might also enhance liking and thereby potentially influence persuasive outcomes. For example, incidental similarities in first names, birthdays, or birthplaces appear capable of producing such effects (see, e.g., Burger, Messian, Patel, del Prado, & Anderson, 2004; Garner, 2005; Guéguen, Pichot, & Le Dreff, 2005; Jiang, Hoegg, Dahl, & Chattopadhyay, 2010; Silvia, 2005; for some complexities, see Howard & Kerin, 2011). 16. Several other reviews have offered evidence concerning the effectiveness of peer-based health interventions (e.g., Kennedy, O’Reilly, & Sweat, 2009; Maticka-Tyndale & Barnett, 2010; Simoni, Nelson, Franks, Yard, & Lehavot, 2011; Webel, Okonsky, Trompeta, & Holzemer, 2010)—but the question of interest here is the relative effectiveness of peer-based and non-peer-based interventions. 17. And this does not exhaust the potentially relevant endorser-related perceptions. For example, there is reason to think that the effectiveness of endorser ads can be driven not so much by liking or credibility as by perceptions of the fit between other attributes of the endorser and attributes of the product (see, e.g., Kamins, 1990; Misra & Beatty, 1990; Mittelstaedt, Riesz, & Burns, 2000; Till & Busier, 2000; Törn, 2012). 18. As discussed in Chapter 8 (concerning the elaboration likelihood model), there is not yet a completely well-articulated account of the circumstances under which a given variable will serve in one or another persuasion role. For the moment, then, the point is to be attentive to the mistake of assuming that a given communicator characteristic can function in only one way in persuasion.



326



Chapter 11 Message Factors Message Structure and Format Conclusion Omission Recommendation Specificity Narratives Prompts Message Content Consequence Desirability One-Sided Versus Two-Sided Messages Gain-Loss Framing Threat Appeals Sequential Request Strategies Foot-in-the-Door Door-in-the-Face Conclusion For Review Notes



This chapter reviews research concerning the effects that selected message variations have on persuasion. The message factors discussed are grouped into three broad categories: message structure and format, message content, and sequential-request strategies.



Message Structure and Format This section discusses four aspects of the structure and format of persuasive messages that have been investigated for their possible effects on persuasive outcomes: whether the message’s conclusion is explicitly stated, the degree of specificity with which the communicator’s advocated action is described, the use of narratives as vehicles for persuasive messages, and the use of simple prompts.



Conclusion Omission Obviously, persuasive messages have some point—some opinion or belief that the communicator hopes the audience will accept, some recommended 327



action that the communicator wishes to have adopted. But should the message explicitly make that point—explicitly state the conclusion or recommendation—or should the message omit the conclusion and so leave the point unstated?1 Intuitively, there look to be good reasons for each alternative. For instance, one might think that making the conclusion explicit would be superior because receivers would then be less likely to misunderstand the point of the message. On the other hand, it might be that if the communicator simply supplies the premises, and the audience reasons its own way to the conclusion, then perhaps the audience will be more persuaded than if the communicator had presented the desired conclusion (more persuaded, because they reached the conclusion on their own). There have been a number of investigations of this question, in which an explicit conclusion is either included in or omitted from the message.2 For example, Struckman-Johnson and Struckman-Johnson (1996) compared AIDS public service announcements with and without an explicit recommendation to use condoms. In such studies, the overwhelmingly predominant finding is that messages that include explicit conclusions or recommendations are more persuasive than messages without such elements (for a review, see O’Keefe, 2002b; see, relatedly, Moyer-Gusé, Jain, & Chung, 2012).3 There has often been speculation that the apparent advantage of explicit conclusions may be moderated by factors involving the hearer’s ability and willingness to draw the appropriate conclusion when left unstated. Hence variables such as the receiver’s intelligence (which bears on ability) and initial opinion (which bears on willingness) have often been mentioned as possible moderators (e.g., McGuire, 1985). The expectation has been that explicit conclusions may not be necessary to, and might even impair, persuasive success for intellectually more capable audiences and for audiences initially favorable to the advocated view (because such audiences should be able and willing to reason to the advocated conclusion). What little relevant empirical evidence exists, however, gives no support to these speculations. For example, in several studies, the audience was comparatively intelligent and well-educated (college students), and even so, there was a significant advantage for messages with explicit recommendations or conclusions (e.g., Fine, 1957). One possible explanation for the persuasive advantage of explicitly stated 328



conclusions is that when the conclusion is omitted, assimilation and contrast effects are encouraged. As discussed in Chapter 2 (concerning social judgment theory), assimilation and contrast effects are perceptual distortions concerning what position is being advocated by a message (C. W. Sherif et al., 1965; M. Sherif & Hovland, 1961): An assimilation effect occurs when a receiver perceives the message to advocate a view closer to his or her own than it actually does; a contrast effect occurs when a receiver perceives the message to advocate a position more discrepant from his or her own than it actually does. Both assimilation and contrast effects reduce persuasive effectiveness—contrast effects because they make the message appear to urge an even more unacceptable viewpoint, assimilation effects because they reduce the amount of change apparently sought by the message. Notably, relatively ambiguous messages (i.e., messages ambiguous about what position is being advocated) appear especially susceptible to assimilation and contrast effects (Granberg & Campbell, 1977). Thus the reduced persuasive success of messages omitting explicit conclusions may arise because such messages are relatively more subject to assimilation and contrast effects. In any case, the research evidence suggests that persuaders commonly have little to gain (and much to lose) by leaving the message’s conclusion implicit. Ordinarily, messages containing explicit statements of the conclusion will be more effective than messages that omit such statements.4



Recommendation Specificity When a communicator is urging some particular action, the message can vary in the specificity with which the advocated action is described. The contrast here is between messages that provide only a general description of the advocate’s recommended action and messages that provide a more specific (detailed) recommendation. Both messages contain an explicitly stated conclusion (in the form of an explicitly identified desired action), but one conclusion is more detailed than the other. For example, Leventhal, Jones, and Trembly (1966) compared persuasive messages recommending that students get tetanus shots at the student health clinic with messages providing a more detailed description of the recommended action (e.g., mentioning the location and hours of the clinic—although students were already familiar with such information). Similarly, Evans, Rozelle, Lasater, Dembroski, and Allen (1970) compared messages giving 329



relatively general and unelaborated dental care recommendations with messages giving more detailed, specific recommendations. Such studies have commonly found that messages with more specific descriptions of the recommended action are more persuasive than those providing general, nonspecific recommendations (for a review, see O’Keefe, 2002b).5 It is not yet clear what might explain this effect. One possibility is that more specific descriptions of the recommended action enhance the receiver’s behavioral self-efficacy (perceived behavioral control). As discussed in Chapter 6, reasoned action theory suggests that one factor influencing a person’s behavioral intention is the individual’s belief in his or her ability to engage in the behavior (perceived behavioral control). For example, people who do not think that they have the ability to engage in a regular exercise program (because they lack the time, the equipment, and so forth) are unlikely to undertake such behavior, even if they have positive attitudes toward exercising. It may be that—akin to the enhanced self-efficacy that can arise from seeing another person perform the action —receivers who encounter a detailed description of the recommended action may become more convinced of their ability to perform the behavior. A second (not necessarily competing) possible explanation is that a specific action description encourages people to plan their behavioral performance and thus develop implementation intentions (subsidiary intentions related to the concrete realization of a more abstract intention, as discussed in Chapter 6; see Gollwitzer & Sheeran, 2006), which in turn make behavioral performance more likely.6



Narratives Broadly conceived, a narrative is a story, that is, a depiction of a sequence of related events. Much research attention has recently been given to studying narrative as a distinctive message format for persuasion. For example, instead of trying to persuade by making explicit arguments, instead one might use a story as the vehicle for persuasive information.



Complexities in Studying Narrative and Persuasion Studying narratives and persuasion is complicated for at least three reasons. First, narratives can take many forms. A narrative might figure as a brief illustration (e.g., when one part of an advertisement for a diet plan is a before-and-after testimonial), as a more extensive story that forms the 330



bulk of the message (e.g., when one example serves as a “case study”), or as an even more extended story (e.g., when a daytime television drama has a multiepisode story arc concerning some health topic). Narratives might be fictional or factual, might have a simple natural order or a more complex structure (e.g., with flashbacks), might be delivered in first person (“I did X”) or third person (“He did X”) forms, and so forth. Second, there is no single clearly conceived message form against which to contrast narrative messages. Researchers have naturally wanted to compare the effects of narrative messages against nonnarrative messages, but a great variety of nonnarrative forms have been explored. For example, some research has compared a narrative form in which a single illustrative case is presented against a nonnarrative form in which statistical information is presented (e.g., Uribe, Manzur, & Hidalgo, 2013). Other studies have compared the effects of a narrative message against a “didactic” or “argument-based” message (e.g., Prati, Pietrantoni, & Zani, 2011).7 Third, narratives might conceivably play any number of different roles in persuasion. As examples: A narrative might engage the audience’s attention and so enhance message exposure or message elaboration. A narrative might provide evidence for a claim and hence enhance belief in that claim (“you actually can integrate exercise into your daily routine— here’s the story of someone like you who did just that”). A narrative might induce a persuasion-relevant psychological state, such as a (positive or negative) mood (e.g., Cerully & Klein, 2010) or a counterfactual mind-set (e.g., Nan, 2008), which in turn might affect the amount or evaluative direction of subsequent message processing. (This surely does not exhaust the possible roles that narrative might play in persuasion.)8



The Persuasive Power of Narratives For all that studying narrative persuasion is a complex challenge, and despite the scattered and incomplete research to date, the potential persuasiveness of narratives is undeniable. In a number of studies, (various kinds of) narrative messages have been found to be more persuasive than (various kinds of) nonnarrative messages. As examples: Adamval and Wyer (1998) found that vacations were evaluated more favorably when a travel brochure presented a narrative rather than a list of features of the vacation. Dillard, Fagerlin, Dal Cin, Zikmund-Fisher, and Ubel (2010) reported that adding a narrative to a traditional educational message 331



increased interest in colorectal cancer screening. In Murphy, Frank, Chatterjee, and Baezconde-Garbanati’s (2013) study, a fictional narrative film was more effective than a nonnarrative message in enhancing knowledge and intentions concerning cervical cancer. Polyorat, Alden, and Kim (2007) reported that narrative ads produced more favorable evaluations of several consumer products than did “factual” ads. Prati et al. (2011) found that a narrative message was more effective than a didactic message in influencing various risk and efficacy perceptions concerning flu shots. (For some other illustrations, see Appel & Richter, 2007; H. S. Kim, Bigman, Leader, Lerman, & Cappella, 2012; Larkey & Gonzalez, 2007; Masser & France, 2010; Morgan, Cole, Struttmann, & Piercy, 2002; Morman, 2000; Niederdeppe, Shapiro, & Porticella, 2011; Ricketts, Shanteau, McSpadden, & Fernandez-Medina, 2010.) In short, there is little room for doubt that narratives can be more persuasive than nonnarrative messages. This research evidence does not show that narratives are generally more persuasive than nonnarrative messages—only that it is possible for narratives to have a persuasive advantage. It remains to be seen exactly under what circumstances a given narrative form will be more or less persuasive than some specific nonnarrative message form.9



Factors Influencing Narrative Persuasiveness The persuasiveness of narratives (in absolute terms or as compared with nonnarrative messages) is likely to vary from one instance to another. A number of different possible moderating factors have received at least some research attention, but two specific factors warrant mention here. One is the degree to which the recipient identifies with the narrative’s characters.10 In a number of studies, greater identification with characters has been found to be associated with greater persuasive effects of narratives. As examples: In Moyer-Gusé, Chung, and Jain’s (2011) study of narratives in which safer-sex conversations were modeled, greater identification with the characters enhanced recipients’ self-efficacy for having such conversations themselves. Igartua (2010) found that character identification influenced the degree to which a fictional film affected story-relevant beliefs and attitudes (see, relatedly, Igartua & Barrios, 2012). De Graaf, Hoeken, Sanders, and Beentjes (2012) manipulated character identification by varying the perspective from which the story was told, which yielded corresponding effects on story-consistent attitudes 332



(recipients were more inclined to have attitudes consistent with the perspective of the character from whose perspective the story was told). (For other illustrations, see Sestir & Green, 2010; van den Hende, Dahl, Schoormans, & Snelders, 2012. For a review, see Tukachinsky & Tokunaga, 2013).11 A second factor influencing the persuasiveness of narratives is the degree to which the recipient is “transported” by the narrative—caught up in, or carried away by, the story. Narratives are potentially capable of inducing a state of “transportation” in message recipients, a state in which the recipients are so immersed in (transported by) the story that they become completely focused on the world depicted in the story.12 (For discussions of the assessment of transportation, see Busselle & Bilandzic, 2009; Green & Brock, 2000.) The suggestion has been that when people are thusly transported, their real-world beliefs are more likely to be affected by narrative content. For example, Green and Brock (2000) had participants read a story about a student whose sister is stabbed to death by a psychiatric patient at a shopping mall. Those participants who were relatively highly transported into the story were (compared with lesstransported participants) more likely to subsequently report storyconsistent beliefs (e.g., enhanced beliefs in the likelihood of violence, reduced beliefs that the world is just, etc.). That is, transportation made people more susceptible to narrative influence. (For related findings, see, e.g., Banerjee & Greene, 2012; Dunlop, Wakefield, & Kashima, 2010; Escalas, 2004, 2007; Reinhart & Anker, 2012; Vaughn, Hesse, Petkova, & Trudeau, 2009; Zwarun & Hall, 2012. For reviews, see Tukachinsky & Tokunaga, 2013; van Laer, de Ruyter, Visconti, & Wetzels, 2014.)13 Transportation might lead to enhanced persuasion through any number of different mechanisms (see Carpenter & Green, 2012)—reduced counterarguing, enhanced imagery, and so on—but there is relatively little evidence bearing on these possibilities. (For discussion, see Slater & Rouner, 2002; Vaughn, Childs, Maschinski, Niño, & Ellsworth, 2010. For some illustrative research findings, see Dunlop, Wakefield, & Kashima, 2010; McQueen, Kreuter, Kalesan, & Alcaraz, 2011; Moyer-Gusé & Nabi, 2009; Niederdeppe, Kim, Lundell, Fazili, & Frazier, 2012.) Research is also only beginning to explore the factors that might affect the likelihood of transportation, such as features of the story (e.g., de Graaf & Hustinx, 2011), characteristics of the recipient (e.g., Appel & Richter, 2010; Dal Cin, Zanna, & Fong, 2004; Thompson & Haddock, 2012), and properties of the medium (e.g., Braverman, 2008); for a general discussion, see Green 333



(2008). The relationship of these two influences on narrative persuasion (transportation and character identification) is not entirely clear. One might treat each of these as a more specific realization of a broader concept of narrative “involvement” (e.g., Tukachinsky & Tokunaga, 2013). Or one might treat them as distinct influences on (i.e., alternative pathways to) narrative persuasion. Or these might be related in some fashion; in particular, character identification might foster transportation (see, e.g., van Laer et al., 2014). But to date there is too little research evidence to permit confident conclusions about the relationships of these factors (for some relevant work, see Murphy, Frank, Moran, & Patnoe-Woodley, 2011; Sestir & Green, 2010; Tal-Or & Cohen, 2010). And, of course, character identification and transportation are not the only possible influences on narrative persuasiveness. A variety of other factors have also been explored, including the nature of the communication source (e.g., Hopfer, 2012) and various properties of the narrative material (see, e.g., Appel & Mara, 2013; Dahlstrom, 2010; H. S. Kim et al., 2012; Moyer-Gusé, Jain, & Chung, 2012; Tal-Or, Boninger, Poran, & Gleicher, 2004). However, obtaining dependable generalizations about any such factors—and identifying exactly how and why such factors influence narrative persuasiveness (keeping in mind that the effects might be obtained by affecting character identification or transportation)—remains some distance in the future.



Entertainment-Education One particular application of persuasive narratives is worth mention: entertainmenteducation. Entertainment-education (EE) is the purposeful design of entertainment media specifically as vehicles for educating—and thereby influencing behavior. A classic example is provided by the South African dramatic television series Soul City, which was initially created for the purpose of conveying HIV prevention information. The series proved both enormously popular and effective in providing the desired information. In subsequent years the program expanded to address other subjects (e.g., tobacco control and domestic violence) and to include other media (e.g., a radio series). (For an overview of Soul City development, see Usdin, Singhal, Shongwe, Goldstein, & Shabalala, 2004.) Similar programs have been created in a number of developing countries (see, e.g., Abdulla, 2004; Kuhlmann et al., 2008; Ryerson & Teffera, 2004; Smith, 334



Downs, & Witte, 2007) and, less commonly, in the developed world (e.g., van Leeuwen, Renes, & Leeuwis, 2013; Wilkin et al., 2007). The challenge in creating EE programs is striking the right balance between entertainment (which attracts the audience) and education (which is the reason for creating the program in the first place)—and this can be difficult to manage (for some discussion, see Renes, Mutsaers, & van Woerkum, 2012). A variant approach is to use some existing entertainment program as the vehicle for conveying health information. That is, rather than creating an entertainment program from scratch, instead one can weave the relevant information into some existing program (ideally, a relatively popular one that already attracts the desired audience). For example, storylines about breast cancer, immunization, and safer sex have been embedded in various popular television shows with the explicit purpose of influencing viewers (for examples, see Bouman, 2004; Glik et al., 1998; Hether, Huang, Beck, Murphy, & Valente, 2008; Kennedy, O’Leary, Beck, Pollard, & Simpson, 2004; Whittier, Kennedy, St. Lawrence, Seeley, & Beck, 2005). In EE applications of narrative persuasion, persuasive effects can come about in two ways. First, the program can have the usual sort of direct effects on viewers. For example, consistent with the findings discussed earlier, several studies have reported that the effectiveness of EE programming is enhanced when recipients identify with a narrative character (e.g., Kuhlmann et al., 2008; Smith, Downs, & Witte, 2007; Wilkin et al., 2007). Second, indirect effects can arise from communication stimulated by the entertainment programming. For example, exposure to a Nepalese radio drama serial concerning family planning enhanced the likelihood that women would discuss family planning with their spouse (Sharan & Valente, 2002; for related findings, see Love, Mouttapa, & Tanjasiri, 2009; Pappas-DeLuca et al., 2008).14



Summary Narratives can be powerful vehicles for persuasion, but many open questions remain about how and why narratives persuade. At the moment not much is securely known about exactly when any given narrative form will be more persuasive than some specifiable nonnarrative form, what factors influence the relative persuasiveness of different narrative forms, or how such moderating factors are related to the mechanisms underlying narrative effects. Continuing attention to these questions will be 335



welcomed. (For some general discussions of narrative persuasion, see Bilandzic & Busselle, 2013; Carpenter & Green, 2012; Green & Clark, 2013; Hinyard & Kreuter, 2007; Larkey & Hecht, 2010; Larkey & Hill, 2012; Moyer-Gusé, 2008; Slater & Rouner, 2002; Vaughn et al., 2010; Winterbottom, Bekker, Conner, & Mooney, 2008.)



Prompts A prompt (reminder) is a simple cue that that makes behavioral performance salient and hence can trigger the behavior. Depending on the context, a prompt might be delivered by a small sign or poster, a text message, an automated phone call, an email, regular mail, and so on. The message can be variously phrased—as a explicit reminder (“Don’t forget to …”), as an invitation (“Have you considered … ?”), as a simple rationale for the behavior (“Taking the stairs burns calories”), and so on— but is characteristically relatively brief. A large number of experiments have demonstrated the potential of such prompts to influence behavior.15 As examples: Simple signs can increase stair use in shopping malls, train stations, and workplaces, especially when the alternative is an escalator (e.g., Andersen, Franckowiak, Snyder, Bartlett, & Fontaine, 1998; Boen, Maurissen, & Opdenacker, 2010; Kwak, Kremers, van Baak, & Brug, 2007; for a review, see Nocon, MüllerRiemenschneider, Nitzschke, & Willich, 2010). Text message reminders can encourage sunscreen use (Armstrong et al., 2009), voting (Dale & Strauss, 2009), and saving money (Karlan, McConnell, Mullainathan, & Zinman, 2010). Reminders can increase cancer screenings (Burack & Gimotty, 1997), medication initiation (Waalen, Bruning, Peters, & Blau, 2009), medication adherence (Lester et al., 2010; Vervloet et al., 2012), diabetes monitoring (Derose, Nakahiro, & Ziel, 2009), and newborn immunization rates (Alemi et al., 1996). Prompts can increase organ donor registration (A. J. King, Williams, Harrison, Morgan, & Havermahl, 2012), seat belt use (Austin, Alvero, & Olson, 1998; Austin, Sigurdsson, & Rubin, 2006; Cox, Cox, & Cox, 2000), and the frequency with which physicians recommend preventive care measures (such as immunization and cancer screening) to their patients (Dexheimer, Talbot, Sanders, Rosenbloom, & Aronsky, 2008; Shea, DuMouchel, & Bahamonde, 1996). (For some reviews, see Head, Noar, Iannarino, & Harrington, 2013; Fry & Neff, 2009; Szilagyi et al., 2002; Tseng, Cox, Plane, & Hia, 2001. For some examples of unsuccessful prompting, see Amass, Bickel, Higgins, 336



Budney, & Foerg, 1993; Blake, Lee, Stanton, & Gorely, 2008.)16 There are probably at least two necessary conditions for a prompt to be effective in inducing behavior. First, the recipients must presumably already have the appropriate positive attitude; reminding people to do something they don’t want to do is unlikely to be very effective. Second, the recipients must likely already believe themselves capable of performing the behavior (expressed in terms of reasoned action theory, perceived behavioral control must be sufficiently high); prompting people to do something they don’t think they can do is presumably unlikely to be very effective. So when people are willing to perform a behavior (positive attitude) and think themselves able to perform the behavior (high perceived behavioral control), but nevertheless aren’t doing the behavior, then perhaps all that’s needed is a simple prompt.17



Message Content This section reviews research concerning the persuasive effects of certain variations in the contents of messages. Literally dozens of content variables have received at least some empirical attention; this review focuses mainly on selected message content factors for which the empirical evidence is relatively more extensive.



Consequence Desirability One common way of trying to persuade people is by appealing to the consequences of the advocated action. The general abstract form is “If the advocated action A is undertaken, then desirable consequence D will occur.” A good deal of research has addressed questions about the relative persuasiveness of various forms of consequence-based arguments. Of specific interest here is the comparison of appeals invoking more and less desirable consequences of compliance with the advocated view. Abstractly put, the experimental contrast is between arguments of the form “If advocated action A is undertaken, then very desirable consequence D1 will occur” and “If advocated action A is undertaken, then slightly desirable consequence D2 will occur.” Now one might—with some reason—think that this research question is hardly worth investigating, because the answer is obvious. And, perhaps understandably, the overt research question has not been expressed quite 337



this way. Even so, substantial research evidence, collected in other guises, has accumulated on this matter. For example, many studies have examined a question of the form “do people who differ with respect to characteristic X differ in their responsiveness to corresponding kinds of persuasive appeals?”—where characteristic X is actually a proxy for variations in what people value. For example (as discussed in Chapter 3 concerning functional approaches to attitude), a number of studies have examined how people varying in self-monitoring (concern about the image one projects to others) differ in their responsiveness of different kinds of persuasive appeals. Specifically, high self-monitors are more persuaded by consumer ads that use imagebased appeals than by ads that use product-quality-oriented appeals, whereas the reverse effect is found for low self-monitors. As suggested in Chapter 3, this effect can be seen to reflect differences in the values of high and low self-monitors—high self-monitors value the image-related attributes of products in ways low self-monitors do not. Naturally, then, each kind of person is more persuaded by messages invoking consequences that (from their perspective) are relatively more desirable. A similar example is offered by research on the individual-difference variable called “consideration of future consequences” (CFC; Strathman, Gleicher, Boninger, & Edwards, 1994). CFC refers to differences in the degree to which people consider longer-term as opposed to shorter-term behavioral consequences. As one might expect, persons differing in CFC respond differently to persuasive messages depending on whether the message’s arguments emphasize immediate consequences (more persuasive for those low in CFC) or long-term consequences (more persuasive for those high in CFC; see, e.g., Orbell & Hagger, 2006; Orbell & Kyriakaki, 2008). Yet another example is provided by research on cultural differences in “individualismcollectivism,” which refers to the degree to which individualist values (e.g., independence) are prioritized as opposed to collectivist values (e.g., interdependence; Hofstede, 2001). Persons from cultures differing in individualism-collectivism respond differently to persuasive messages depending on whether the message’s appeals emphasize individualistic or collectivistic outcomes. For example, advertisements for consumer goods have been more persuasive for American audiences when the ads emphasize individualistic outcomes (“this watch will help you stand out”) rather than collectivistic ones (“this 338



watch will help you fit in”), with the reverse being true for Chinese audiences (e.g., Aaker & Schmitt, 2001; for a review, see Hornikx & O’Keefe, 2009). Plainly, this effect reflects underlying differences in the perceived desirability of various product attributes. Several other lines of research similarly converge on the unsurprising general conclusion: Consequence-based appeals are more persuasive when they invoke outcomes of the advocated action that are (taken by the audience to be) relatively more desirable than when they invoke outcomes that are not valued so highly (for a review, see O’Keefe, 2013a).18 As obvious as this point might be, it nevertheless underscores the importance of fashioning appeals that are tailored to the audience’s wants. And it is an empirical question which appeals will be most persuasive for a given audience. For instance, one might think that skin protection behaviors (sunscreen use, tanning avoidance, and so on) might be successfully influenced by invoking health-related consequences such as skin cancer—but at least some people are more persuaded by appeals concerning appearance-related consequences (J. L. Jones & Leary, 1994; Thomas et al., 2011). Similarly, HPV (human papillomavirus) vaccine can prevent both cancer and sexually transmitted infections, but these two consequences may not be equally persuasive for all recipients (Krieger & Sarge, 2013; Leader, Weiner, Kelly, Hornik, & Cappella, 2009). (See, relatedly, Gollust, Niederdeppe, & Barry, 2013; Segar, Updegraff, Zikmund-Fisher, & Richardson, 2012.)19



One-Sided Versus Two-Sided Messages In many circumstances, a persuader will be aware of potential arguments supporting an opposing view. How should a persuader handle such arguments? One possibility is simply to ignore the opposing arguments, to not mention or acknowledge them in any way; the persuader would offer only constructive (supporting) arguments. Alternatively, the persuader might not ignore opposing arguments but rather discuss them explicitly (while also presenting supporting arguments). This basic contrast— ignoring or discussing opposing arguments—has commonly been captured in persuasion research as the difference between a one-sided message (which presents supporting arguments but ignores opposing ones) and a two-sided message (which discusses both supporting and opposing arguments). (The broadest review is O’Keefe, 1999a; for other reviews and 339



discussions, see Allen, 1991, 1993, 1998; Crowley & Hoyer, 1994; Eisend, 2006, 2007; O’Keefe, 1993; Pechmann, 1990.)20 There appears to be no general difference in persuasiveness between onesided and two-sided messages. Previous discussions of sidedness effects have mentioned many possible moderating factors that might influence the relative persuasiveness of one-sided and two-sided messages, including the audience’s level of education, the availability of counterarguments to the audience (sometimes represented as the audience’s familiarity with the topic), and the audience’s initial opinion on the topic (see, e.g., Chu, 1967; R. G. Hass & Linder, 1972). None of these factors, however, appears to moderate this effect; for example, the relative persuasiveness of one- and two-sided messages does not vary as a function of whether the audience initially favors or opposes the advocated view (O’Keefe, 1999a). A more complex picture emerges, however, if one distinguishes two varieties of twosided messages. A refutational two-sided message attempts to refute opposing arguments; this might involve criticizing the reasoning of an opposing argument, offering evidence to undermine an opposing claim, challenging the relevance of an opposing argument, and so forth. A nonrefutational two-sided message acknowledges opposing considerations but does not attempt to refute them directly; the message might suggest that the supporting arguments outweigh the opposing ones, but it does not directly attack the opposing considerations. One way of expressing the difference is to say that refutational two-sided messages characteristically attempt to undermine opposing arguments (by refuting them), whereas nonrefutational two-sided messages characteristically attempt to overwhelm opposing arguments (by deploying supportive ones). These two types of two-sided message have dramatically different persuasive effects when compared with one-sided messages. Specifically, refutational two-sided messages are dependably more persuasive than onesided messages; nonrefutational two-sided messages, on the other hand, are slightly less persuasive than their one-sided counterparts.21 That is, acknowledging opposing arguments without refuting them generally makes messages somewhat less persuasive (compared with ignoring opposing arguments), whereas refuting opposing arguments enhances persuasiveness.22 As with the overall contrast between one-sided and twosided messages, this pattern of effects seems largely unaffected by commonly proposed moderator factors such as audience education or initial attitude, although the research evidence is often sketchy (O’Keefe, 340



1999a). However, there is almost certainly a limiting condition on the apparent persuasive advantage of refutational two-sided messages. Examination of the messages in these studies suggests that the refuted counterarguments were ones that might well have been entertained by the audience as potentially significant objections. One ought not necessarily expect the same results if implausible or trivial objections were to be refuted. Notice that, at least when designed in recognition of this limiting condition, the use of refutational two-sided messages represents a form of audience adaptation. Two-sided messages that attempt to undermine the audience’s active objections to the recommended action are tailored to the audience’s psychological state (the audience’s potential counterarguments). Perhaps it is only to be expected that meeting such objections head-on will be more persuasive than ignoring them. As an example of the potential importance of addressing active objections, consider smoking cessation. Smokers often express concern that if they quit smoking, they will gain weight. One possible way of defusing this objection would be to couple a smoking cessation intervention with a weight control intervention. And, indeed, such combined treatments produce significantly greater abstinence, at least in the short term, than do smoking cessation treatments alone (for a review, see Spring et al., 2009).23 One additional complexity is worth noting: Nonrefutational two-sided messages appear to have different effects in consumer advertising messages than in other persuasive messages (messages concerning political questions, public policy issues, and the like). In nonadvertising messages, nonrefutational two-sided messages are dependably less persuasive than their one-sided counterparts, but in consumer advertisements, nonrefutational two-sided messages are neither more nor less persuasive than one-sided advertisements (O’Keefe, 1999a).24 Put differently: Nonrefutational two-sided advertising messages do not seem to suffer the same negative persuasion consequences that parallel nonadvertising messages do. Some light might be shed on this by considering the effects of nonrefutational twosided messages on credibility perceptions. The credibility of consumer advertising is boosted by the use of nonrefutational two-sided messages, but nonrefutational two-sided messages on other 341



persuasive topics do not produce the same enhancement of credibility (O’Keefe, 1999a). It may be that receivers’ initial skepticism about consumer advertising leads receivers to expect that advertisers will provide a one-sided depiction of the advertised product—and thus when an advertisement freely acknowledges (and does not refute) opposing considerations, the advertiser’s credibility is enhanced (akin to the credibility enhancement effects obtained when communicators advocate positions opposed to their apparent self-interest, as discussed in Chapter 8). This enhanced credibility for advertisements, in turn, could have varying effects. It might boost the believability of both the supportive arguments and the acknowledged counterarguments (with these effects canceling each other out), it might enhance the counterarguments more than the supportive arguments (making the ad less persuasive), or it might enhance the supportive arguments more than the counterarguments (making the ad more persuasive). Across a number of nonrefutational two-sided ads, then, one might expect to find no dependable overall difference in persuasiveness—which is precisely the observed effect.25 In sum, persuaders are best advised to meet opposing arguments head-on, by refuting them, rather than ignoring or (worse still) merely mentioning such counterarguments—save, perhaps, in the case of consumer advertising, where nonrefutational acknowledgment of opposing arguments promises to be about as persuasive as ignoring opposing considerations.



Gain-Loss Framing One especially well-studied persuasive message variation is gain-loss message framing.26 A gain-framed message emphasizes the advantages of undertaking the advocated action; a loss-framed message emphasizes the disadvantages of not engaging in the advocated action. So, for example, “If you wear sunscreen you’ll have attractive skin when you’re older” is a gain-framed appeal, whereas “If you don’t wear sunscreen you’ll have unattractive skin when you’re older” is a loss-framed appeal.



Overall Effects The phenomenon of negativity bias provides a reason for expecting that loss-framed appeals might have a general persuasive advantage over gainframed appeals. Negativity bias refers to the greater sensitivity to, and 342



impact of, negative information compared with equally extreme positive information (for a review, see Cacioppo, Gardner, & Berntson, 1997). For example, negative information has a disproportionate impact on evaluations or decisions compared with otherwise equivalent positive information (for a review, see Rozin & Royzman, 2001); learning one new negative thing about a person often has a much larger effect than learning one new positive thing. The phenomenon of negativity bias naturally suggests that loss-framed messages, which emphasize the negative consequences of noncompliance with the recommended action, should be more persuasive than gain-framed appeals. However, there is no such general advantage for loss-framed appeals. Gain-framed and loss-framed appeals do not generally differ in persuasiveness (for a review, see O’Keefe & Jensen, 2006).27



Disease Prevention Versus Disease Detection Even though gain-framed and loss-framed appeals do not generally differ in persuasiveness, some moderating factor might be at work—a factor that makes gain-framed appeals more advantageous in some circumstances and loss-framed appeals more advantageous in others (thus yielding no overall difference). One especially well-studied moderator concerns the nature of the advocated behavior, specifically whether the advocated action is a disease detection behavior (such as cancer screening) or a disease prevention behavior (such as regular flossing). The hypothesis was that loss-framed messages will be more persuasive than gain-framed messages for disease detection behaviors, with the reverse pattern expected for disease prevention behaviors (see, e.g., Salovey, Schneider, & Apanovitch, 2002). This hypothesis was motivated by researchers having noticed what looked like a pattern in some early framing studies. For example, a lossframed appeal was more persuasive than a gain-framed appeal in a study of breast cancer screening (Meyerowitz & Chaiken, 1987), but gainframed appeals were more persuasive than loss-framed appeals in a study of sunscreen use (Detweiler, Bedell, Salovey, Pronin, & Rothman, 1999).28 The accumulated research evidence, however, indicates that this hypothesis is also faulty. Concerning disease detection topics, several reviews have concluded that there is no overall persuasive advantage for loss-framed appeals (Gallagher & Updegraff, 2012; O’Keefe & Jensen, 2009).29 With respect to disease prevention topics, the evidence is more 343



complex, but here too there does not appear to be any general persuasive advantage for gain-framed appeals (Gallagher & Updegraff, 2012; O’Keefe & Jensen, 2007).30



Other Possible Moderating Factors In the search for other potential moderators of gain-loss message framing effects, a number of different possibilities have been explored, including the temporal proximity of the consequences (Gerend & Cullen, 2008), the type of supporting evidence (Dardis & Shen, 2008), and the colors used in the message (Chien, 2011). But the greatest attention has focused on individual-level variables (see Latimer, Salovey, & Rothman, 2007; Updegraff & Rothman, 2013), including self-efficacy (van ’t Riet, Ruiter, Werrij, & de Vries, 2010), attitudinal ambivalence (Broemer, 2002), perceived threat susceptibility (Gallagher, Updegraff, Rothman, & Sims, 2011), need for cognition (Steward, Schneider, Pizarro, & Salovey, 2003), mood (Chang, 2007), and so on. However, the evidence is not yet sufficiently extensive to permit confident conclusions about any such candidates (see, e.g., Covey, 2014). Moreover, identifying a genuine moderator of gain-loss message framing effects can be a challenging task, because isolating the effect of this message variation is potentially difficult. The difficulty arises from the nature of the appeal variation. Both gain-framed and loss-framed appeals are conditional arguments (“if-then” arguments) that depict some consequence as being a result of some antecedent. Gain- and loss-framed appeals differ in the antecedent condition: in gain-framed appeals, the antecedent is compliance with the communicator’s recommended course of action (“If you wear sunscreen, then …”), whereas in loss-framed appeals the antecedent is noncompliance (“If you don’t wear sunscreen, then …”).31 Because the communicative purpose is persuasion, there is correspondingly a difference in the valence of the claimed consequences: persuaders naturally argue that compliance produces desirable consequences and that noncompliance produces undesirable consequences.32 However, even though the valence of the consequences is necessarily dictated by the nature of the antecedent, the substance of the consequences can potentially vary. And here experimenters can face quite a challenge. If the consequences invoked by the two appeals are substantively different, then any difference in persuasiveness might reflect the difference in the consequences, not the difference in the antecedent. 344



As an example: One suggested individual-level moderating factor is the recipient’s approach/avoidance motivation (BAS/BIS; Carver & White, 1994). Individuals vary in their general sensitivity to reward (desirable outcome) or punishment (undesirable outcome) cues, and the hypothesis has been that approach-oriented individuals will be more persuaded by gain-framed appeals than by loss-framed appeals, with the reverse pattern holding for avoidance-oriented individuals (e.g., Jeong et al., 2011; Latimer, Salovey, & Rothman, 2007). A related motivational difference is the recipient’s regulatory focus (Higgins, 1998); regulatory focus variations reflect a broad motivational difference between a promotion focus, which emphasizes obtaining desirable outcomes, and a prevention focus, which emphasizes avoiding undesirable outcomes. Correspondingly, this hypothesis has sometimes been phrased in terms of regulatory focus: Promotion-oriented individuals should be more persuaded by gain-framed appeals than by loss-framed appeals, but prevention-oriented individuals should be more persuaded by loss-framed appeals than by gain-framed appeals. But these motivational differences are presumably reflected in differing evaluations of substantively different consequences; for example, approach-oriented people will presumably prefer approach-oriented consequences over avoidance-oriented consequences. And the distinction between these two kinds of consequences (approach-oriented vs. avoidance-oriented) is different from the distinction between the two kinds of antecedents (compliance vs. noncompliance). Thus the conjunction of gain-loss variations (different kinds of antecedents) and these motivational variations (different kinds of consequences) yields four possible appeal types: (1) gain-framed appeals that emphasize avoidanceoriented consequences (e.g., “if you exercise, you’ll reduce your heart attack risk”), (2) gain-framed appeals that emphasize approach-oriented consequences (e.g., “if you exercise, you’ll increase your energy”), (3) loss-framed appeals that emphasize avoidance- oriented consequences (e.g., “if you don’t exercise, you won’t reduce your heart attack risk”), and (4) lossframed appeals that emphasize approach-oriented consequences (e.g., “if you don’t exercise, you won’t increase your energy”). In order to show that approach/avoidance orientation affects the relative persuasiveness of gain-framed and loss-framed appeals, investigators must take care to ensure that the gain-loss framing manipulation (compliance vs. noncompliance) is not confounded with variation in the kinds of consequences invoked. Consider, for example, a study of messages aimed 345



at encouraging flossing in which the gain-framed appeal emphasized having healthy gums (an approach-oriented consequence) and the lossframed appeal emphasized avoiding gum disease (an avoidance-oriented consequence). The results indicated that the former appeal was more persuasive than the latter for approach-oriented participants, with the reverse result for avoidance-oriented participants (Sherman, Mann, & Updegraff, 2006). But, given the confounding of the type of antecedent and the type of consequence, such results might more plausibly be said to reflect differences in the consequences invoked (approach vs. avoidance) than in the antecedent (compliance vs. noncompliance). (For some relevant evidence, see Chang, 2010, Experiment 2; for discussion, see Cesario, Corker, & Jelinek, 2013; O’Keefe, 2013a.) Broadly speaking, individual differences in motivational orientation (approach vs. avoidance, promotion vs. prevention) map easily onto a contrast between different kinds of consequences (approach/promotionoriented consequences and avoidance/preventionoriented consequences), but do not match up so well with a contrast between gain-framed (compliance-focused) and loss-framed (noncompliance-focused) appeals. For that reason, one might be justifiably skeptical that individual motivational-orientation differences will turn out to moderate the effects of gain-loss message framing variations independent of the kinds of consequences invoked.33 But this illustrates the potential difficulty of identifying moderating factors for this message variation—difficulties arising from the need to distinguish the gain-loss framing variation (variation in the antecedent of the appeal) from variation in the substantive consequences invoked.34



Summary Gain-framed and loss-framed appeals do not differ much in persuasiveness. Research has not yet identified moderating factors that yield substantial differences in persuasiveness between gain-framed and loss-framed appeals—and identifying such factors will be challenging.



Threat Appeals Threat appeals (also called fear appeals) are messages designed to encourage the adoption of behaviors aimed at protecting against a potential threat. Threat appeals have two components. One is material depicting the 346



threatening event or consequences; the other is material describing the recommended protective action. So, for example, driver education programs may show films depicting gruesome traffic accidents (in an effort to reduce dangerous driving practices such as drinking and driving), antismoking messages may display the horrors of lung cancer (so as to discourage smoking initiation), and dental hygiene messages may emphasize the ravages of gum disease (in an effort to encourage regular flossing).



Protection Motivation Theory Because threat appeals concern specifically protective behaviors, an analysis of the factors underlying such behaviors can be useful. Protection motivation theory (PMT; Rogers & Prentice-Dunn, 1997) identifies two broad processes that influence the adoption of protective behaviors, namely, threat appraisal and coping appraisal (appraisal of the recommended action). That is, whether a person adopts a given protective behavior will be a function of the person’s assessment of the threat and the person’s assessment of the suggested way of coping with the threat (the advocated action). Each of these two elements, however, is further unpacked by PMT. Threat appraisal is said to depend on perceived threat severity (the perceived severity of the problem) and perceived threat vulnerability (one’s perception of the likelihood that one will encounter or be susceptible to the threat); as persons perceive the threat to be more severe and as they perceive themselves to be more vulnerable to the threat, persons are expected to have higher protection motivation. Coping appraisal is said to depend on perceived response efficacy (perceived recommendation effectiveness, the degree to which the recommended action is perceived to be effective in dealing with the problem) and on perceived self-efficacy (one’s perceived ability to adopt or perform the protective action); as the perceived efficaciousness of the response increases and as perceived ability to perform the action increases, persons are expected to have greater protection motivation.35 For example, imagine someone contemplating adopting an exercise program to prevent heart disease who thinks that heart disease is not all that serious a problem (low perceived threat severity) and, in any case, perceives herself to be relatively unlikely to have heart disease because no one in her family has had it (low perceived threat vulnerability); moreover, 347



she is not convinced that exercise will really prevent heart disease (low perceived recommendation effectiveness), and she does not think that she has the discipline to stick with an exercise program (low perceived selfefficacy). Such a person presumably will have relatively low protection motivation, as reflected in corresponding actions (namely, not exercising) and intentions (not intending to exercise). A number of studies have indicated that protective intentions and behaviors are indeed affected by these four underlying factors (for a review, see Milne, Sheeran, & Orbell, 2000).36 That is, PMT appears to have identified four influences on protective intentions and actions—and thereby also identified four possible foci for persuasive messages concerning protective behaviors. Such messages might focus on the severity of the threat, on the recipient’s vulnerability to the threat, on the effectiveness of the advocated action, or on the recipient’s ability to adopt that action. (Of course these are not mutually exclusive.) And, plainly, this framework provides a basis for adapting (tailoring) messages to recipients. A person who acknowledges that a threat has really bad consequences (high threat severity) but thinks it won’t happen to them (low perceived vulnerability) presumably should be sent messages emphasizing susceptibility, whereas a person who acknowledges their vulnerability but doesn’t think the threat’s consequences are all that terrible presumably should be sent messages focused on threat severity. Considerable research evidence has shown that persuasive messages can affect each of those four factors. The relevant experimental designs involve manipulating some message feature in an effort to influence the theoretically important mediating state. For instance, to influence perceived threat severity, participants would receive either a message suggesting that the consequences are extremely negative or one suggesting that they are minor; to influence perceived recommendation effectiveness, the message either would depict the recommended action as highly effective in dealing with the threat or would describe it as an inconsistent or unreliable means of coping with the threat; and so on. (For some examples, see Brouwers & Sorrentino, 1993; Smerecnik & Ruiter, 2010.)37 Such experimental message variations have been found to have the anticipated effects on the corresponding mediating states (e.g., messages varying in their depictions of recommendation effectiveness produce corresponding variations in perceived recommendation effectiveness; for reviews, see Milne, Sheeran, & Orbell, 2000; Witte & Allen, 2000; see also Mongeau, 1998, p. 62, note 4) and have been found 348



to have parallel (but weaker) effects on relevant persuasive outcome variables such as attitudes, intentions, and behavior (for reviews, see de Hoog, Stroebe, & de Wit, 2007; Floyd, Prentice-Dunn, & Rogers, 2000; Witte & Allen, 2000).38



Threat Appeals, Fear Arousal, and Persuasion Threat appeals plainly have the potential to arouse fear or anxiety about the threatening events. For that reason, much research attention has focused on variations in how threat appeals depict the negative consequences associated with the threat—and on how such message variations are associated with fear arousal and persuasion. The specific contrast of interest is between a message containing explicit, intense, vivid depictions of those negative consequences (a strong threat appeal) and a message with a tamer, toned-down version (a weak threat appeal).39 Research concerning the evocation of fear by threat appeals is extensive and complex, but an overview of the key findings can be expressed in five broad conclusions. First, messages with more intense content do generally arouse greater fear. To be sure, the relationship between threat appeal message variations and aroused fear is not perfect (expressed as a correlation, estimates of the mean effect range between .25 and .40), suggesting that influencing the audience’s level of fear is not necessarily something easily accomplished (for reviews, see Boster & Mongeau, 1984; de Hoog et al., 2007; Mongeau, 1998; Witte & Allen, 2000).40 On reflection, of course, this may not be too surprising. A persuader may be mistaken about what will be fearful, and what one person finds extremely fearful may be only mildly worrisome to another person (for examples and discussion, see Botta, Dunker, Fenson-Hood, Maltarich, & McDonald, 2008; Ditto, Druley, Moore, Danks, & Smucker, 1996; Henley & Donovan, 2003; Murray-Johnson et al., 2001). Still, in general, stronger threat appeal contents do arouse greater fear.41 Second, threat messages with more intense content also are more persuasive than those with less intense content, although this effect is smaller than the effect of threat appeal variations on aroused fear (expressed as correlations, the effects average between .10 and .20; for reviews, see Boster & Mongeau, 1984; Mongeau, 1998; Sutton, 1982; Witte & Allen, 2000). This weaker effect on persuasive outcomes is consistent with the idea that fear (the aroused emotional state) mediates the 349



effect of threat appeal message manipulations on persuasive outcomes. That is, the invited image is that varying the message contents produces variations in aroused fear, which in turn are related to changes in attitudes, intentions, and actions (and thus the relationship between message manipulations and persuasive outcomes would be expected to be weaker than the relationship between message manipulations and fear).42 Third (and a natural corollary of the first two), messages that successfully arouse greater fear are also generally more persuasive. That is, in studies with messages that have been shown to arouse dependably different amounts of fear, the messages that arouse greater fear are more persuasive than the messages that arouse lesser fear (for reviews, see Sutton, 1982; Witte & Allen, 2000).43 Fourth, these relationships are roughly linear. That is, generally speaking, as message content becomes more intense, greater fear is aroused and greater persuasion occurs. It has sometimes been thought that very intense message materials will (compared with less intense materials) produce less persuasion (because recipients tune out the message and so do not come to accept the recommendations). That is, the thought has been that a persuader might go “too far” in threat appeal intensity, producing a curvilinear relationship between intensity and persuasion (and specifically an inverted-U-shaped relationship, such that the highest levels of message intensity are associated with relatively lower levels of persuasion). But the evidence in hand gives little indication of any such curvilinear effects (see Boster & Mongeau, 1984; Mongeau, 1998; Sutton, 1992; Witte & Allen, 2000).44 Fifth, there are at least two conditions under which more intense threat appeals are unlikely to be more persuasive than less intense ones. One is circumstances in which the recipients’ fear level is already relatively high. If message recipients are already experiencing sufficiently high levels of concern, it may not be necessary—or even possible—to increase it further. In such circumstances, messages might more appropriately focus on barriers to adopting the recommended action—perhaps receivers’ concerns about whether the action is effective or their doubts about their ability to perform the action. For example, Earl and Albarracín’s (2007)’s review of HIV prevention interventions found that HIV counseling was more effective than fear-inducing arguments in encouraging condom use, arguably because HIV-related anxiety was already relatively high, and counseling provided information about how to address such concerns. 350



(See, relatedly, Kessels & Ruiter, 2012; Muthusamy, Levine, & Weber, 2009.) The second is circumstances in which the message recipients do not have a sufficiently positive assessment of the recommended action (e.g., when the advocated action is perceived as ineffective or difficult to perform). Increasing the intensity of the depiction of the threat is unlikely to be helpful when the key barrier to recommendation acceptance is a negative appraisal of the recommended action. In the absence of a sufficiently positive assessment of the advocated behavior, more intense threat appeals are unlikely to be more persuasive than less intense appeals. Correspondingly, greater fear arousal is likely to be associated with greater persuasion only when a workable, effective protective action is (perceived to be) available. (For a review, see Peters, Ruiter, & Kok, 2012.)45 One way of understanding these two limiting conditions is to see that a threat appeal has what amounts to a problem-solution message format: It identifies a potential (threatening, fear-inducing) problem and offers a solution (the recommended action). If people are already convinced of the problem (already experiencing high fear) or if they do not see a good solution available (insufficiently positive assessment of the recommended action), then it doesn’t much matter whether the message has more or less intense depictions of the problem.



The Extended Parallel Process Model The extended parallel process model (EPPM) provides a useful framework for organizing these findings concerning threat appeals, fear, and persuasive effects (Witte, 1992, 1998). Like protection motivation theory (PMT), the EPPM is concerned with the factors influencing protective actions. Although the two models have different terminology, they incorporate the same key components. EPPM identifies “perceived threat” (threat appraisal) and “perceived efficacy” (coping appraisal, that is, recommendation appraisal) as the immediate influences on protective behaviors and offers the same elements underlying each of those: Perceived threat is influenced by perceived threat severity and perceived threat susceptibility (vulnerability), and perceived efficacy is influenced by response efficacy (i.e., perceived recommendation effectiveness) and perceived self-efficacy. But the EPPM additionally identifies two different (parallel) processes that 351



can be activated by threat appeals. People may want to control the apparent danger posed by the threat (danger control processes), and they may want to control their feelings of fear (fear control processes). The activation of these processes varies depending on variations in the combination of perceived threat and perceived efficacy, as follows.46 In a circumstance in which message recipients have high perceived threat (high perceived threat severity and susceptibility) and high perceived efficacy (high perceived recommendation effectiveness and high perceived self-efficacy), then danger control processes will be engaged. People will understand the danger posed by the threat and will see themselves as in a position to deal with that danger by adopting the recommended action. By contrast, in a circumstance in which message recipients have high perceived threat but low perceived efficacy, then fear control processes will be engaged. People will be facing a significant threat, which arouses fear, but they do not believe they have a suitable way to control that threat —and hence people look for ways to control their fear. For example, people might avoid thinking about the threat (defensive avoidance), or they might reassess their threat perceptions so as to diminish their feelings of fear (e.g., they might decide that the threat really isn’t all that severe or that they’re not really vulnerable to it).47 Where perceived threat is low (because the threat is seen as trivial or because people think themselves invulnerable to it), then people are unlikely to adopt the protective action. After all, there’s no apparent threat. So under conditions of low perceived threat, variations in perceived efficacy will not affect behavioral adoption. Briefly, then, from the perspective of the EPPM, the role of threat-related perceptions (perceived severity and vulnerability) is contingent on efficacy-related perceptions (perceived recommendation efficacy and selfefficacy). High perceived threat alone is insufficient to motivate protective action; only the combination of high perceived threat and high perceived efficacy activates the danger control processes that encourage protective behavior. Correspondingly, the effect of threat-related message materials is expected to vary depending on the recipient’s efficacy-related perceptions. In particular, the EPPM emphasizes that efforts to increase threat perceptions may be unhelpful or unwise in circumstances in which efficacy 352



perceptions are not sufficiently high. Thus the EPPM provides a more nuanced basis for message design than simply suggesting that persuaders focus on whichever of the four perceptual determinants of protective action (threat severity, threat vulnerability, recommendation effectiveness, self-efficacy) needs attention. (For some examples of EPPM applications, see Campo, Askelson, Carter, & Losch, 2012; Kotowski, Smith, Johnstone, & Pritt, 2011; Krieger & Sarge, 2013; Murray-Johnson et al., 2004. For general discussions of EPPM-based message design, see Basil & Witte, 2012; Cho & Witte, 2005.) Other theoretical frameworks for understanding threat appeals have also offered distinctive message design recommendations (see, e.g., de Hoog, Stroebe, & de Wit, 2007, 2008; Rimal, Bose, Brown, Mkandawire, & Folda, 2009; Rimal & Juon, 2010; Rimal & Real, 2003; Ruiter & Kok, 2012), although the EPPM is currently the most widely applied framework. At the same time, for all that the EPPM provides an attractive housing for many threat-appeal research findings, at least some of the model’s more detailed claims have not received much research attention— and of those that have, the research evidence may not be univocal (for a useful review, see Popova, 2012; see also Mongeau, 2013, pp. 191–192; cf. Maloney, Lapinski, & Witte, 2011).48



Summary Although some aspects of threat appeals have come into focus, many unanswered questions remain, including exactly how threat-related perceptions and efficacy-related perceptions combine to influence protective intentions and actions (see, e.g., Goei et al., 2010), the appropriate structure of the threat and efficacy components in threat appeals (e.g., Carcioppolo et al., 2013; Wong & Cappella, 2009), whether the within-individual dynamics of fear over time conform to theoretical expectations (e.g., Algie & Rossiter, 2010; Dillard & Anderson, 2004; for a useful discussion, see Shen & Dillard, 2014), and the potential role of individual differences in reactions to threat appeals (e.g., Nestler & Egloff, 2012; Ruiter, Verplanken, De Cremer, & Kok, 2004; Schlehofer & Thompson, 2011; van ’t Riet, Ruiter, & de Vries, 2012; for a general treatment, see Cho & Witte, 2004). In short, there is much more to be learned about threat appeals. (For some general discussions of threat appeal research, see Dilliplane, 2010; Mongeau, 2013; Ruiter, Abraham, & Kok, 2001; Ruiter, Kessels, Peters, & Kok, 2014; Yzer, Southwell, & Stephenson, 2013.) 353



Beyond Fear Arousal Fear is perhaps the best studied of the various emotions that persuasive appeals might try to engage, although there has also been some work on other emotions such as anger (Moons & Mackie, 2007; Quick, Bates, & Quinlan, 2009), disgust (Leshner, Bolls, & Thomas, 2009; Nabi, 1998; Porzig-Drummond, Stevenson, Case, & Oaten, 2009), and especially guilt (e.g., Cotte, Coulter, & Moore, 2005; Hibbert, Smith, Davies, & Ireland, 2007; Turner & Underhill, 2012; for some reviews, see O’Keefe, 2000, 2002a). These lines of work have a common underlying idea, namely, that one avenue to persuasion involves the arousal of an emotional state (such as fear or guilt), with the advocated action providing a means for the receiver to deal with those aroused feelings.49 (For some general treatments of emotions and persuasion, see Dillard & Nabi, 2006; Dillard & Seo, 2013; Nabi, 2002, 2007; Turner, 2012.) There is another way, however, in which emotions might figure in persuasion, namely, through the anticipation of emotional states. Anticipated emotions plainly shape intentions and actions (e.g., people often avoid actions that they think would make them feel guilty; Birkimer et al., 1993), and such anticipations can be influenced (e.g., O’Carroll, Dryden, Hamilton-Barclay, & Ferguson, 2011). This research is discussed more extensively in Chapter 6 (concerning reasoned action theory) in the context of considering how inclusion of anticipated affective states might enhance the predictability of intentions. The point to be noticed here is simply that emotional considerations might play a role in persuasive messages either through the arousal of emotions or through the invocation of expected emotional states.



Sequential Request Strategies Substantial research has been conducted concerning the effectiveness of two sequential request influence strategies: the foot-in-the-door (FITD) strategy and the door-in-the-face (DITF) strategy. In each strategy, the request of primary interest to the communicator (the target request) is preceded by some other request; the question is how compliance with the target request is affected by the presence of the preceding request.50



Foot-in-the-Door 354



The Strategy The FITD strategy consists of initially making a small request of the receiver, which the receiver grants, and then making the (larger) target request. The hope is that having gotten one’s foot in the door, the second (target) request will be looked on more favorably by the receiver. The question thus is whether receivers will be more likely to grant a second request if they have already granted an initial, smaller request.51



The Research Evidence The research evidence suggests that this FITD strategy can enhance compliance with the second (target) request. For example, in Freedman and Fraser’s (1966, Experiment 2) FITD condition, homeowners were initially approached by a member of the “Community Committee for Traffic Safety” or the “Keep California Beautiful Committee.” The requester either asked that the receiver display a small sign in their front window (“Be a Safe Driver” or “Keep California Beautiful”) or asked that the person sign a petition supporting appropriate legislation (legislation that would promote either safer driving or keeping California beautiful). Two weeks later, a different requester (from “Citizens for Safe Driving”) approached the receiver, asking if the receiver would be willing to have a large, unattractive “Drive Carefully” sign installed in the front yard for a week. In the control condition, in which receivers heard only the large request, fewer than 20% agreed to put the sign in the yard. But in the FITD conditions, more than 55% agreed.52 This effect was obtained no matter whether the same topic was involved in the two requests (safe driving or beautification), and no matter whether the same action was involved (displaying a sign or signing a petition): Even those who initially signed the “Keep California Beautiful” petition were more likely to agree to display the large safe driving yard sign. As these results suggest, the FITD strategy can be quite successful. Several factors that influence the strategy’s effectiveness have been identified. First, if the FITD strategy is to be successful, there must be no obvious external justification for complying with the initial request (for reviews, see Burger, 1999; Dillard, Hunter, & Burgoon, 1984). For example, if receivers are given some financial reward in exchange for complying with the first request, then the FITD strategy is not very successful. Second, the larger the first request (presuming it is agreed to by 355



the receiver), the more successful the FITD strategy (see the review by Fern, Monroe, & Avila, 1986). Third, the FITD strategy appears to be more successful if the receiver actually performs the action requested in the initial request, as opposed to simply agreeing to perform the action (for reviews, see Beaman, Cole, Preston, Klentz, & Steblay, 1983; Burger, 1999; Fern et al., 1986; cf. Dillard et al., 1984). Fourth, the FITD strategy is more effective when the requests are prosocial requests (that is, requests from institutions that might provide some benefit to the community at large, such as civic or environmental groups) as opposed to nonprosocial requests (from profit-seeking organizations such as marketing firms; Dillard et al., 1984).53 Notably, several factors apparently do not affect the success of the FITD strategy. The time interval between the two requests does not make a difference (Beaman et al., 1983; Burger, 1999; Dillard et al., 1984; Fern et al., 1986); for example, Cann, Sherman, and Elkes (1975) obtained equivalent FITD effects with no delay between the two requests and with a delay of 7–10 days. Similarly, it does not appear to matter whether the same person makes the two requests (Fern et al., 1986).54



Explaining FITD Effects The most widely accepted explanation for FITD effects is based on selfperception processes (for a brief statement, see Freedman & Fraser, 1966; a more extensive discussion is provided by DeJong, 1979). Briefly, the explanation is that compliance with the first request leads receivers to make inferences about themselves; in particular, initial compliance is taken to enhance receivers’ conceptions of their helpfulness, cooperativeness, and the like. These enhanced self-perceptions, in turn, are thought to increase the probability of the receiver’s agreeing to the second request. The observed moderating factors are consistent with this explanation. For example, the presence of an external justification for initial compliance obviously undermines enhancement of the relevant self-perceptions: If one is paid money in exchange for agreeing to the initial request, it is more difficult to conclude that one is especially cooperative and helpful just because one agreed. Similarly, the larger the request initially agreed to, the more one’s self-perceptions of helpfulness and cooperativeness should be enhanced (“If I’m going along with this big request, without any obvious external justification, then I must really be a pretty nice person, the kind of person who does this sort of thing”). And it’s easier to think of oneself as a 356



helpful, socially minded person when one agrees to requests from civic groups (as opposed to marketing firms) or when one actually performs the requested action (as opposed to merely agreeing to perform it).55 Only a few studies have included direct assessments of participants’ selfperceptions of helpfulness, but recent results have been supportive. For example, Burger and Caldwell (2003) found that first-request compliance produced the expected changes in selfperceived helpfulness—and such self-perceptions were in turn related to second-request compliance. (For similar results, see Burger & Guadagno, 2003; cf. Gorassini & Olson, 1995; Rittle, 1981. For some discussion, see Cialdini & Goldstein, 2004, pp. 602–604.) So the direct evidence to date is slim but encouraging. As attractive as the self-perception account is, there is one nagging concern. As several commentators have remarked, it seems implausible to suppose that self-perceptions of helpfulness would be deeply affected by compliance with small requests of the sort used in FITD research (Rittle, 1981, p. 435; Gorassini & Olson, 1995, p. 102). Presumably, persons’ beliefs about their helpfulness rest on some large number of relevant experiences, and it seems unlikely that such beliefs would be significantly altered by consenting to a single small request. However, the selfperception explanation’s ability to accommodate the observed moderatorvariable effects, and the emerging direct evidence of the role played by self-perceptions in FITD effects, recommend it as the best available explanation.56



Door-in-the-Face The Strategy The DITF strategy turns the FITD strategy on its head. The DITF strategy consists of initially making a large request, which the receiver turns down, and then making the smaller target request. The question is whether initially having (metaphorically) closed the door in the requester’s face will enhance the receiver’s compliance with the second request.



The Research Evidence Studies of the DITF strategy have found that it can indeed enhance compliance. That is, receivers will at least sometimes be more likely to 357



agree to a second smaller request if they have initially turned down a larger first request. For example, in a study reported by Cialdini et al. (1975, Experiment 1), individuals on campus sidewalks were approached by a student who indicated that he or she represented the county youth counseling program. In the DITF condition, persons were initially asked to volunteer to spend 2 hours a week for a minimum of 2 years as an unpaid counselor at a local juvenile detention center; no one agreed to this request. The requester then asked if the person would volunteer to chaperone a group of juveniles from the detention center on a 2-hour trip to the zoo. Among those in the control condition, who received only the target request, only 17% agreed to chaperone the zoo trip; but among those in the DITF condition, who initially turned down the large request, 50% agreed.57 The research evidence also suggests that various factors moderate the success of the DITF strategy. DITF effects are larger if the two requests are made by the same person as opposed to by different persons (for relevant reviews, see Feeley, Anker, & Aloe, 2012; Fern et al., 1986; O’Keefe & Hale, 1998, 2001), if the two requests have the same beneficiary as opposed to benefiting different persons (Feeley et al., 2012; O’Keefe & Hale, 1998, 2001), if there is no delay between the requests (Dillard et al., 1984; Feeley et al., 2012; Fern et al., 1986; O’Keefe & Hale, 1998, 2001), and if the requests come from prosocial rather than nonprosocial organizations (Dillard et al., 1984; Feeley et al., 2012; O’Keefe & Hale, 1998, 2001).58



Explaining DITF Effects Several alternative explanations of DITF effects have been offered. One is the reciprocal concessions explanation (see Cialdini et al., 1975). This explanation proposes that the successive requests make the situation appear to be one involving bargaining or negotiation—that is, a situation in which a concession by one side is supposed to be reciprocated by the other. The smaller second request represents a concession by the requester —and so the receiver reciprocates (“Okay, you gave in a little bit by making a smaller request, so I’ll also make a concession and agree with that request”). Some of the observed moderator variable effects are nicely explained by this account. For example, given this analysis, it makes sense that DITF effects should be smaller if different persons make the requests; if different persons make the requests, then no concession has been made (and hence there is no pressure to reciprocate a concession). 358



But for several reasons, one might doubt whether the reciprocal concessions explanation is entirely satisfactory. First, some moderator variable effects are not so obviously accommodated by the explanation. For example, it is not clear why the strategy should work better for prosocial requests than for nonprosocial requests. Second, several meta-analytic reviews have found that DITF effects are not influenced by the size of the concession made (Fern et al., 1986; O’Keefe & Hale, 1998, 2001), and this seems inconsistent with the reciprocal concessions account. The reciprocal concessions account appears to predict that larger concessions will make the DITF strategy more effective (by putting greater pressure on the receiver), and hence the failure to find such an effect seemingly indicates some weakness in the explanation.59 Third, DITF effects do not appear to be influenced by emphasizing or deemphasizing the concession. For example, stressing that the second request represents a concession does not enhance the strategy’s effectiveness, and deemphasizing the fact of a concession does not weaken the strategy (e.g., Goldman, McVeigh, & Richterkessing, 1984; for a review, see O’Keefe, 1999b). A second explanation suggests that guilt arousal underlies DITF effects. The general idea is that first-request refusal potentially represents a violation of one’s own standards for socially responsible conduct, which gives rise to feelings of guilt. Those guilt feelings then motivate targetrequest compliance (for discussion of such explanations, see O’Keefe & Figgé, 1999; Tusing & Dillard, 2000). A guilt-based account appears capable of accommodating many of the observed moderator variable effects. For example, declining prosocial requests probably generates greater guilt than does declining nonprosocial requests, thus making the strategy more effective for prosocial organizations. Similarly, because guilt feeling presumably dissipate over time, delaying the second request will diminish the strategy’s effectiveness. However, the guilt-based account does not provide a good explanation of why the DITF strategy is more effective when the same person makes the two requests (see Cialdini & Goldstein, 2004, p. 601). One might think 359



that feelings of guilt would be better reduced by “making it up to” the requester (as opposed to agreeing to a request from someone else)—but that’s not the way guilt works. There is considerable evidence that guiltreduction behaviors do not necessarily have to involve making amends to the victim of the guiltinducing behavior. For example, when people are feeling guilty about having committed a transgression (e.g., telling a lie) that harmed another person, they are more likely to comply with a subsequent request (than are people in a no-transgression control condition)—but this effect is the same no matter whether the request comes from the injured party or someone else (for a review, see O’Keefe, 2000). The implication of this finding is that DITF compliance cannot be explained by guilt arousal alone; if DITF compliance arose purely from guilt, then the strategy’s effectiveness would not be influenced by whether the same person made both requests. So it may be that DITF effects arise from a combination of reciprocitybased processes and guilt-based processes. The reciprocity-based explanation can explain the effect of having the requests come from the same person, and the guilt-based account can accommodate other observed moderating factors. Guilt-based and reciprocity-based motivations can presumably operate simultaneously, and the evidence in hand points to just such a combination.



Conclusion Researchers have investigated a large number of message characteristics as possible influences on persuasive effectiveness. These message factors are varied, ranging from the details of message components (the phrasing of the message’s conclusion) to the sequencing of multiple messages (as in the FITD and DITF strategies). Indeed, this discussion can do no more than provide a sampling of the message features that have been studied. (For other general discussions of persuasive message variations, see Perloff, 2014; Pratkanis, 2007; Shen & Bigsby, 2013; Stiff & Mongeau, 2003.)



For Review 1. What does the research evidence suggest about the relative persuasiveness of stating the message’s conclusion explicitly as opposed to omitting the conclusion (leaving the conclusion implicit)? 360



2.



3.



4.



5.



6.



Does this difference vary depending on the audience’s educational level? Does it vary depending on the audience’s initial favorability toward the advocated view? Describe a possible explanation for the observed effect. What does the research evidence suggest about the relative persuasiveness of providing a general (as opposed to a more specific) description of the advocated action? Describe two possible explanations for the observed effect. What is a narrative? Explain why studying the role of narratives in persuasion can be challenging. Can narratives be more persuasive than nonnarrative messages? Are narratives generally more persuasive than nonnarrative messages? Identify two factors that influence the persuasiveness of narratives. What is character identification? How does character identification influence narrative persuasiveness? What is transportation? How does transportation influence narrative persuasiveness? What is entertainment-education? Describe two ways in which entertainment-education programs can produce persuasive effects. What is a prompt? Give examples. Can prompts influence behavior? Identify two necessary conditions for prompts to be effective in influencing behavior. Why is an existing positive attitude such a condition? Why is sufficiently high perceived behavioral control (PBC, self-efficacy) such a condition? What is a consequence-based argument? How do variations in the perceived desirability of the consequences affect the persuasiveness of such arguments? Give examples. What is a one-sided message? A two-sided message? Which is more persuasive? Distinguish two varieties of two-sided messages. What is a refutational two-sided message? What is a nonrefutational twosided message? Comparing one-sided messages and refutational twosided messages, which generally is more persuasive? Identify an implicit limiting condition on the occurrence of these differences. What general differences, if any, are there in persuasiveness between one-sided messages and nonrefutational two-sided messages? In advertising contexts, how do one-sided messages and nonrefutational two-sided messages differ in persuasiveness? Outside advertising contexts (that is, in “nonadvertising” messages), how do one-sided messages and nonrefutational two-sided messages differ in persuasiveness? What might explain the observed differences between advertising messages and other persuasive messages in how nonrefutational two-sided messages work? Are the effects of 361



nonrefutational two-sided messages (compared with one-sided messages) on credibility perceptions the same for advertising messages and for nonadvertising messages? Explain how skepticism about advertising might underlie the different effects of nonrefutational two-sided messages in advertising contexts as opposed to nonadvertising contexts. 7. What is a gain-framed message? A loss-framed message? Describe a reason for hypothesizing that loss-framed appeals might generally be more persuasive than gain-framed appeals. Which kind of appeal is generally more persuasive? Which kind of appeal is more persuasive when the message topic concerns disease prevention? Which kind of appeal is more persuasive when the message topic concerns disease detection? Explain why it is difficult to identify a factor that moderates the effects of gain- and loss-framed appeals. Describe the hypothesis that the relative persuasiveness of gain-framed and lossframed appeals will vary depending on whether the recipient is relatively approach/promotion–oriented or avoidance/prevention– oriented; explain how such motivational differences are related to different kinds of behavioral consequences. 8. What is a threat appeal? Describe the two parts of a threat appeal. What is protection motivation theory (PMT)? What is protection motivation? Identify the two processes underlying protection motivation. What is threat appraisal? Identify two factors that influence threat appraisal. What is perceived threat severity? What is perceived vulnerability to threat? What is coping appraisal? Identify two factors that influence coping appraisal. What is perceived response efficacy? What is perceived self-efficacy? What is the relationship between the intensity of threat appeal content and the degree of fear aroused in receivers? Are messages with more intense content generally more persuasive than those with less intense content? Are messages that arouse greater fear generally more persuasive than those that arouse lesser amounts of fear? Does the relationship between the intensity of message contents and the amount of aroused fear take the shape of an inverted U? Explain. Does the relationship between the intensity of message contents and persuasive outcomes take the shape of an inverted U? Explain. Identify two conditions under which more intense threat appeals are unlikely to be more persuasive than less intense appeals. Describe the extended parallel process model (EPPM). What does the EPPM add to protection motivation theory (PMT)? What is danger control? What is fear control? Describe how the activation of fear control and 362



danger control processes varies as a function of variations in perceived threat and perceived efficacy. From the perspective of the EPPM, is high perceived threat sufficient to motivate protective action? Explain. What emotions other than fear might be involved in persuasion? Describe how the anticipation of emotional states can play a role in persuasion. 9. Describe the foot-in-the-door (FITD) strategy. Identify four factors that influence the success of the FITD strategy (four moderating factors). How does the presence of an obvious external justification (for initial-request compliance) influence the effectiveness of the strategy? How does the size of the initial request influence the effectiveness of the strategy? How is the strategy’s effectiveness influenced by whether the initially requested behavior is actually performed? How is the strategy’s effectiveness influenced by whether the requests come from prosocial or nonprosocial organizations? Does the time interval between the two requests influence the strategy’s success? Is the strategy’s success affected by whether the same person makes both requests? What is the selfperception explanation of FITD effects? Describe how that explanation accounts for the observed moderating factors. Identify a potential problem with the self-perception explanation. 10. Describe the door-in-the-face (DITF) strategy. Identify four factors that influence the success of the DITF strategy (four moderating factors). How is the success of the strategy affected by whether the same person makes the two requests? How is the success of the strategy affected by whether the two requests have the same beneficiary? How is the success of the strategy affected by the presence of a delay between the requests? How is the strategy’s effectiveness influenced by whether the requests come from prosocial or nonprosocial organizations? Describe the reciprocalconcessions explanation of DITF effects. Describe how that explanation accounts for some of the observed moderating factors; describe how that explanation has a difficult time accounting for other moderating factors. Does the size of the concession (the reduction in request size from the first to the second request) influence the success of the strategy? Are DITF effects influenced by emphasizing or deemphasizing the concession? Describe the guilt-based explanation of DITF effects. Describe how that explanation accounts for some of the observed moderating factors; describe how that explanation has a difficult time accounting for other moderating factors. Do guiltreduction behaviors necessarily involve making amends to the person 363



injured by the guilt-producing behavior? Explain how DITF effects might reflect a combination of reciprocity-based and guilt-based processes.



Notes 1. This variation (stating or omitting the message’s overall conclusion) thus is different from varying whether the message states the conclusions to its individual supporting arguments (e.g., Kao, 2007). Both have been glossed as “conclusion omission” manipulations but are plainly distinguishable variations. 2. The experimental contrast of interest here is between a message in which the overall conclusion is stated explicitly and one in which that conclusion is simply omitted. A slightly different contrast compares a message with an explicit overall conclusion and a message in which recipients are explicitly urged to “decide for yourself” (Martin, Lang, & Wong, 2003; Sawyer & Howard, 1991). The latter contrast might yield effects different from the former; for example, a “decide for yourself” message might minimize reactance and thereby diminish any relative advantage of an explicit conclusion message. There is good evidence that adding “but you are free to refuse” can enhance request compliance (e.g., Guéguen & Pascual, 2005; for a review, see Carpenter, 2013), so one ought not assume that an omitted conclusion message and a decide-foryourself conclusion message will produce identical effects when compared with an explicit conclusion message. 3. The observed mean effect size (across 17 such studies) corresponds to a correlation of .10 (O’Keefe, 2002b; for some earlier reviews, see Cruz, 1998; O’Keefe, 1997). 4. It may be that in social influence settings such as psychotherapy, or perhaps in unusual persuasive message circumstances (see, e.g., Linder & Worchel, 1970), there can be some benefit to letting the receiver draw the conclusion. But in ordinary persuasive message contexts, the evidence indicates that such benefits are unlikely to accrue. And it might be that in consumer advertising, any advantage of explicit conclusions (over omitted conclusions) would be diminished because recipients are unlikely to misperceive what position is being advocated. It would plainly be useful for future research to obtain a principled description of the circumstances in which explicit conclusions might impair persuasive success relative to 364



omitted conclusions. 5. Across 18 such studies, the mean effect size, expressed as a correlation, was .10 (O’Keefe, 2002b). 6. Either of these explanations might be related to the possibility that a message with a more specific description of the recommended action makes it easier for receivers to imagine themselves performing that action. This possibility is worth noticing because at least under some conditions, imagining performing a hypothetical future behavior can lead to increased (perceived and actual) likelihood of performing that behavior (e.g., C. A. Anderson, 1983; Armitage & Reidy, 2008; Gregory, Cialdini, & Carpenter, 1982; Libby, Shaeffer, Eibach, & Slemmer, 2007; R. T. Sherman & Anderson, 1987). But the mechanism underlying such imagined behavior effects is not yet clear. It might be that imagined behavioral performance enhances perceived behavioral control (selfefficacy), enhances the development of implementation intentions (because it is akin to explicit planning; see Chapter 6), makes reasons for performing the behavior more salient (see R. T. Sherman & Anderson, 1987), or increases behavioral performance through some other mechanism. (For related work, see Escalas & Luce, 2004; Knäuper et al., 2011; Petrova & Cialdini, 2005; Ratcliff et al., 1999.) The larger point is that there are several strands of research (concerning specific action descriptions, explicit planning, imagining behavioral performance, and developing implementation intentions) that appear to be related but whose connections have not been fully explored. 7. Taken together, these first two complexities should make plain the challenges in reaching dependable generalizations about the persuasive effects of narrative. Given that there are many different narrative forms (the first point) and many different nonnarrative forms against which a narrative form might be compared (the second point), quite a diverse set of message contrasts is naturally possible. 8. This analysis is not quite of the sort recommended by the elaboration likelihood model’s idea of different persuasion roles—or at least this is not an analysis that relies on the ELM’s description of such roles (see Chapter 8). But this general line of thinking has obvious affinities with an ELMbased analysis. For a general discussion of the capabilities of narrative in communication (not only persuasive communication) in the context of cancer prevention and control, see Kreuter et al. (2007). 365



9. One much-studied specific realization of the contrast between narrative and nonnarrative messages is the contrast between a message that provides a single example (in a form amounting to a narrative) and a message that provides corresponding statistical information about many cases (nonnarrative). In primary research, one can find results indicating a statistically significant advantage for examples (e.g., Uribe, Manzur, & Hidalgo, 2013), results indicating a statistically significant for statistical summaries (e.g., Lindsey & Ah Yun, 2003), and results reporting no significant difference (e.g., Hong & Park, 2012; Mazor et al., 2007; Schulz & Meuffels, 2012). A thorough review of the relevant research is not yet in hand. Allen and Preiss’s (1997) review included studies that did not compare the persuasive effectiveness of examples and parallel quantitative summaries (e.g., Harte, 1976). The review by Zebregs, van den Putte, Neijens, and de Graaf (2015) was careful to avoid such problems but did not include unpublished studies or several seemingly relevant published studies (e.g., Dardis & Shen, 2008; Studts, Ruberg, McGuffin, & Roetzer, 2010). 10. Character identification has been conceptualized and assessed in a number of different ways, including perceived similarity to a character, liking for a character, adopting the role or perspective of a character, wishful identification with a character (wanting to be like the character in some way), and “parasocial interaction” with a character (a sense of having a relationship with the character). The present treatment bundles these together, but there is reason to think that, in the long run, a more differentiated analysis will be useful (see Moyer-Gusé, 2008; Tukachinsky & Tokunaga, 2013). 11. Tukachinsky and Tokunaga’s (2013) meta-analysis distinguished, inter alia, perceived character similarity (homophily), empathic character identification, and parasocial relationships. In a random-effects analysis, perceived similarity and empathic identification produced statistically significant effects on storyconsistent attitudes, beliefs, and behaviors, but parasocial relationships did not. The mean effects were r = .30 for perceived similarity (31 studies), r = .26 for empathic identification (17 studies), and r = .12 for parasocial relationships (five studies). 12. A number of different terms have been used to describe this phenomenon, including transportation, engagement, immersion, and absorption (Slater & Rouner, 2002, p. 179).



366



13. In Tukachinsky & Tokunaga’s (2013) random-effects meta-analysis, greater transportation was significantly associated with story-consistent attitudes, beliefs, and behaviors; the mean effect was r = .29 (across 31 cases). In van Laer, de Ruyter, Visconti, and Wetzels’s (2014) metaanalysis, greater transportation was significantly associated with stronger persuasive effects on beliefs (the random-effects unadjusted mean effect, across 31 studies, corresponded to a correlation of .23), attitudes (31 studies, mean r = .41), and intentions (nine studies, mean r = .29). 14. Video games may provide another vehicle for EE narrative persuasion. Games can easily encourage transportation (immersion in the game world) and character identification (as when the player is a character). When exposure to a game can be mandated (e.g., when school children are required to play a health-oriented game as part of their instruction), then the intrinsic appeal of the game may not matter so much; however, where voluntary game playing is concerned, then (just as with, say, voluntary exposure to EE television programming), the challenge arises of making the game sufficiently entertaining (so people will want to play it) while also ensuring delivery of the desired persuasive contents. (For some review discussions of games as persuasive vehicles, see Lieberman, 2012, 2013; Lu, Baranowski, Thompson, & Buday, 2012; Peng, Crouse, & Lin, 2013; Primack et al, 2012.) 15. In these studies, the comparison condition is commonly a no-message control (or the evidence of effectiveness is drawn from a pre- versus postintervention comparison). The consequence is that conclusions are not yet possible about the relative effectiveness of prompts and other (e.g., more complex) messages. 16. Head et al.’s (2013) reported meta-analytic results were based on fixed-effect analyses and hence cannot appropriately be used as a basis for generalizing about text messaging interventions other than the ones already studied. A reanalysis of Head et al.’s (2013) effect sizes, converted from standardized mean differences (ds) to correlations (rs) and analyzed using Borenstein and Rothstein’s (2005) random-effects analysis, suggests that some, but not all, of Head et al.’s conclusions can appropriately be generalized. For example, the overall comparison between text messaging interventions and comparison conditions yields a statistically significant effect with both analyses (across 19 cases, the random-effects mean r was .13, with 95% confidence interval limits of .08 and .19). By contrast, the claimed effect of personalization (where personalized text messaging was 367



reported as more effective than nonpersonalized text messaging) does not generalize: in a random-effects analysis, the mean effects for personalized interventions (four cases, mean r = .17, 95% CI limits of .06 and .28) and for nonpersonalized interventions (15 cases, mean r = .12, 95% CI limits of .06 and .18) were not significantly different, Q(1) = .625, p = .43. 17. Of course, if behavioral performance is already at a practical maximum in the target population (such that nearly all the people with appropriately positive attitudes and perceived behavioral control are already engaging in the behavior), then prompts may have little or no effect. For example, Clack, Pitts, and Kellermann’s (2000) finding that parking deck prompts did not increase safety belt use might have been a consequence of the high baseline performance rate (83%). 18. A parallel generalization can be drawn concerning messages that invoke undesirable consequences of noncompliance with the advocated action—for example, “If you don’t do the advocated action [wear your seat belt, use sunscreen, etc.] then this bad thing can happen to you [serious injury, skin cancer, etc.].” Appeals invoking consequences of noncompliance are more persuasive when they invoke consequences that are (taken by the audience to be) relatively more undesirable than when they invoke outcomes that are relatively less undesirable. The research evidence here is drawn from work on threat appeals (fear appeals), where the message variation of interest is commonly described as variations in “threat severity”—and the relevant finding is that threats perceived as more severe (i.e., more undesirable) make for more effective persuasive appeals than do threats perceived as less severe (less undesirable); see the meta-analytic reviews of de Hoog, Stroebe, and de Wit (2007), Floyd, Prentice-Dunn, and Rogers (2000), and Witte and Allen (2000). For discussion, see O’Keefe (2013a). 19. Persuaders might usefully be reminded here that their reasons for wanting a behavior performed are not necessarily the reasons that will be most persuasive to message recipients. The public health official may want to encourage sunscreen use so as to reduce skin cancer, but appeals to health-related consequences might be less persuasive than appeals to appearance-related consequences. 20. Keller & Lehmann’s (2008) meta-analysis also offered conclusions about message sidedness (and a number of other message variations), but their procedures included nonexperimental data. For example, a study in 368



which all the messages were two-sided had its results included in the analysis of the effects of twosided messages. That is, Keller and Lehmann’s conclusions about a given message variable were not based exclusively on experiments (randomized trials) in which levels of that variable were manipulated. In fact, they reported, “we had relatively few manipulated levels for many of the variables” whose effects they reviewed (p. 120). There are, of course, very good and familiar reasons to prefer conclusions based on randomized trials (“this experiment compared the effectiveness of one-sided and two-sided messages and found …”) over those based on observational studies (“in this study all the messages were two-sided, and people were really persuaded, so therefore …”). Correspondingly, there are good reasons to prefer meta-analytic conclusions based exclusively on randomized-trial data over those based largely on observational studies. 21. The average persuasive advantage of refutational two-sided messages over one-sided messages (across 42 studies) corresponds to a correlation of .08. The average persuasive disadvantage of nonrefutational two-sided messages compared with one-sided messages (across 65 studies) corresponds to a correlation of –.05 (O’Keefe, 1999a). 22. Both kinds of two-sided messages are perceived as more credible than one-sided messages. For refutational two-sided messages, the effect corresponds to a correlation of .11 (across 20 studies); for nonrefutational two-sided messages, the correlation is .08 (across 36 cases; O’Keefe, 1999a). 23. In Spring et al.’s (2009) review, the (statistically significant) mean effect size for short-term abstinence (under three months) for this comparison, expressed as an odds ratio, was 1.29, which corresponds to a correlation of .07. The (nonsignificant) mean effect size for long-term abstinence (over six months) was 1.23, which corresponds to a correlation of .06. (For some discussion, see Parsons, Lycett, & Aveyard, 2011; Spring, Rademaker, McFadden, & Hitsman, 2011.) 24. Refutational two-sided messages appear to enjoy a persuasive advantage (over one-sided messages) in both advertising (mean effect corresponding to a correlation of .07, across nine cases) and nonadvertising (mean effect corresponding to a correlation of .08, across 33 cases) messages. However, there are too few studies of refutational two-sided advertisements to permit one to be confident of this effect (see 369



O’Keefe, 1999a). 25. Eisend’s (2006, 2007) meta-analysis of the effects of sidedness variations, which was restricted to studies of consumer advertising, appears to provide only partial confirmation of this picture. Concerning effects on credibility perceptions, O’Keefe’s (1999a) results concerning nonrefutational two-sided advertising messages and Eisend’s (2006) results for (refutational and nonrefutational) two-sided advertising messages were consistent: Such messages were more credible than onesided messages (with mean effect sizes, expressed as correlations, of .16 and .22, respectively). However, the two meta-analyses produced dramatically different results concerning persuasion outcomes: Where O’Keefe (1999a) reported no significant different in persuasiveness between nonrefutational two-sided advertisements and one-sided advertisements (mean effect size of –.02), Eisend (2006) reported significant advantages for two-sided advertising messages (mean effect sizes of .12 and .14 for brand attitude and purchase intention, respectively). However, these two meta-analyses differed in several ways. When the same message pair was used in multiple studies, O’Keefe’s procedure combined results across studies (thus treating the message pair as the unit of analysis), whereas Eisend’s procedures recorded separate effect sizes when a message pair was reused. When multiple measures of a given outcome were available in a given study (e.g., three measures of attitude), O’Keefe’s procedures created a single composite effect size for that outcome, whereas Eisend’s procedures treated these as contributing separate effect sizes. These and other procedural differences made for a large apparent difference in the size of the meta-analytic databases: For advertising messages, O’Keefe’s (1999a) database had 35 persuasionoutcome and 22 credibility-outcome effect sizes. Eisend’s (2006) database had 65 brand-attitude, 37 purchase-intention, and 32 credibility-outcome effect sizes—despite not including a number of cases included in O’Keefe’s (1999a) data set. [Comparison of O’Keefe’s (1999a) Tables 6.1 and 6.4 with Eisend’s (2006) Table 2 indicates that 18 effect sizes concerning persuasion outcomes (brand attitude and purchase intention) and 12 effect sizes concerning credibility outcomes that appeared in the earlier data set were not included in the later one.] If Eisend’s (2006) procedures had treated multiple measures of a given outcome in a study as contributing a single effect size, and if brand-attitude and intention outcomes had been analyzed together as representing persuasion outcomes (as in O’Keefe, 1999a; see also O’Keefe, 2013b), then the data set would have consisted of 21 persuasion effect sizes—of which 17 were included 370



in O’Keefe’s (1999) data set of 35 advertising persuasion effect sizes. For credibility outcomes, if multiple measures in a study had been treated as contributing a single effect size, Eisend’s (2006) data set would have consisted of 10 effect sizes—of which seven were included in O’Keefe’s (1999) data set of 22 advertising credibility effect sizes. As an indication of the potential of such differences to influence the results: The 18 advertising persuasion-outcome effect sizes included in O’Keefe’s (1999) data set but not in Eisend’s (2006) data set had a random-effects mean effect size, expressed as a correlation, of –.04 (N = 4,148; 95% confidence interval limits of –.12 and .05). This effect was not different for refutational (six cases, mean r = .02, 95% CI limits of –.07 and .12) and nonrefutational (12 cases, mean r = –.08, 95% CI limits of –.19 and .04) advertising messages, Q(1) = 1.7, p = .19. There may have been good reasons for the observed procedural variations (e.g., for some cases in the earlier data set to have been excluded from the later one). On the face of things, however, one might reasonably be cautious about supposing that nonrefutational two-sided advertising messages generally enjoy the size of persuasive advantage over one-sided messages that might be implied by Eisend’s (2006) results. 26. The phrase message framing has been used to cover a variety of different message variations. Messages have been described as differently framed when they have varied in the substantive consequences invoked (as when HPV vaccine is described either as preventing genital warts or as preventing both genital warts and cancer; McRee, Reiter, Chantala, & Brewer, 2010) or in the description of a property of the attitude object (as when ground beef is described as “75% lean” or “25% fat”; Levin & Gaeth, 1988). Messages have also been labeled as framed differently when a given outcome of the advocated action is described in various ways; for example, the results of a surgical procedure might be described in terms of the probability of living or the probability of dying (e.g., McNeil, Pauker, Sox, & Tversky, 1982), or price might be characterized in terms of daily expense (“pennies a day”) as opposed to aggregate cost (e.g., Gourville, 1999; see, relatedly, Chandran & Menon, 2004). All of these are plainly distinguishable message variations, and not much is gained by lumping all of them together as “message framing.” For one effort at sorting out such matters, see Levin, Schneider, and Gaeth (1998). 27. In O’Keefe and Jensen’s (2006) review, the mean effect size, expressed as a correlation, was .02 (not significantly different from zero) across 165 cases; that review covered a variety of advocacy topics and 371



included both published and unpublished research. O’Keefe’s (2011) review analyzed O’Keefe and Jensen’s (2006) cases, the studies of disease prevention topics reviewed by O’Keefe and Jensen (2007), and the studies of disease detection topics reviewed by O’Keefe and Jensen (2009); it produced a similar result (mean effect size of r = .01 across 219 cases). Other meta-analytic reviews of gain-loss message framing studies have commonly had more limited scope by virtue of examining only certain kinds of advocacy subjects (say, only health behaviors) or excluding unpublished studies (e.g., Akl et al., 2011; Gallagher & Updegraff, 2012; O’Keefe & Jensen, 2011). 28. This hypothesis was seemingly buttressed by Kahneman and Tversky’s (1979) prospect theory, which was taken to suggest that preventive actions would be motivated more by gains than by losses, with disease detection behaviors more motivated by potential losses than by gains. But this reading of prospect theory arguably misapplied the theory (for discussion, see O’Keefe & Jensen, 2006, p. 23; O’Keefe, 2012b, pp. 12–15). Moreover, direct examination of the mechanism hypothesized to underlie putative framing differences between prevention and detection topics (namely, perceived risk) has not yielded confirming evidence (van ’t Riet et al., 2014). 29. O’Keefe and Jensen’s (2009) review reported a small but statistically significant advantage for lossframed appeals for disease detection topics (mean r = –.04 across 53 cases), but this effect reflected the results for breast cancer detection (a statistically significant mean r of –.06 across 17 cases) and did not obtain for other kinds of detection (a nonsignificant mean r of –.03 across 36 cases). Gallagher and Updegraff’s (2012) review undertook separate analyses for different persuasion outcomes (attitude, intention, and behavior); for disease detection topics, they reported no significant differences between framing conditions (no statistically significant mean effect size) for attitude outcomes (mean r = –.03 across 16 cases), intention outcomes (mean r = –.03 across 32 cases), or behavior outcomes (mean r = –.04 across 18 cases). The population effect (for gain-loss framing for disease detection topics) is almost certainly not literally zero, but, taken together, these meta-analytic results suggest that any such effect is likely to be quite small. For example, O’Keefe & Jensen (2009, p. 306) pointed out that their results were consistent with a belief that the population effect is –.02 both overall and for each of the different detection topics they distinguished; that is, that 372



value falls within the 95% confidence interval around each of the various mean effects. And that value also falls within the 95% confidence interval for the three separate effects computed over (a corrected version of) Gallagher and Updegraff’s (2012) data set (O’Keefe, 2013b). So the gainloss message framing population effect for disease detection topics may not be zero, but it is not very distant from zero. 30. O’Keefe and Jensen’s (2007) review, which included both published and unpublished studies, reported a small but statistically significant advantage for gain-framed appeals for disease prevention topics (mean r = .03 across 93 cases). But this effect reflected the results for dental hygiene messages (a statistically significant mean r of .15 across nine cases) and did not obtain for other kinds of prevention topics (a nonsignificant mean r of .02 across 84 cases). Gallagher and Updegraff’s (2012) review, which was restricted to published studies, undertook separate analyses for different persuasion outcomes (attitude, intention, and behavior); for disease prevention topics, they reported no significant difference between framing conditions for attitude (mean r = .04 across 45 cases) and intention (mean r = .03 across 46 cases) outcomes but did find a significant mean effect size for behavioral outcomes (mean r = .08 across 32 cases). But a closer analysis of (a corrected version of) Gallagher and Updegraff’s data set indicates that those three mean effect sizes were not significantly different from each other (for details and discussion, see O’Keefe, 2013b); that is, for prevention topics, the mean effect size for behavior outcomes was not significantly larger than the mean effect sizes for attitude or intention. Expressed differently: There is no evidence that gain-loss framing effects on prevention topics vary as a consequence of the kind of outcome examined; the observed effects on the different outcomes were statistically indistinguishable. Of course, the population effect (for gain-loss framing for disease prevention topics) is almost certainly not literally zero; but, taken together, these meta-analytic results suggest that any such effect is likely to be quite small. For example, O’Keefe and Jensen’s (2007) results are consistent with a belief that the population effect is .03 both overall and for each of the different prevention topics they distinguished; that is, that value falls within the 95% confidence interval around of each the various mean effects. And a value of .04 falls within the 95% confidence intervals for the three separate effects computed over Gallagher and Updegraff’s (2012) corrected data set (O’Keefe, 2013b). So, for disease prevention topics, as with disease detection topics, the gain-loss message framing population 373



effect may not literally be zero, but it is not very far from that value. 31. It might have been more transparent to have labeled this appeal variation as the difference between “compliance-focused” and “noncompliance-focused” appeals. But the terminology of “gain-framed” and “loss-framed” is too well established to hope for any better labeling to take hold. 32. Nonrefutational two-sided messages are an exception. In such messages, the persuader as much as acknowledges some disadvantage of compliance. 33. It is worth noticing, however, that an appropriately chosen gain-loss frame might amplify differences (in persuasive effectiveness) arising from whether the invoked consequences match the recipient’s motivational focus. That is, the persuasive advantage that accrues to a message’s having motivationally matched consequences (e.g., promotion-oriented consequences for promotion-oriented people) could potentially be enhanced or diminished depending on whether the appeal was phrased in terms of consequences of compliance (gain-framed) or consequences of noncompliance (loss-framed). For some related work, see Cesario, Corker, and Jelinek (2013). 34. To put this problem a bit more abstractly: If a given individual-level variable is associated with value differences (differences in the evaluations of various substantively different consequences), then special care must be taken in constructing experimental messages so as to ensure that such value differences are not confounded with the gain-loss framing manipulation. A sufficiently clever experimenter could show that promotion-oriented people are more persuaded by gain-framed appeals than by loss-framed appeals (by having the gain-framed appeal invoke promotion consequences and the loss-framed appeal invoke prevention consequences)—or could show the exact opposite (by having the gainframed appeal invoke prevention consequences and the loss-framed appeal invoke promotion consequences). Or (see Chapter 3 concerning functional approaches to attitude) an experimenter could show that high self-monitors are more persuaded by gain-framed appeals than by loss-framed appeals (by having the gain-framed appeal invoke symbolic consequences and the loss-framed appeal invoke instrumental consequences)—or could show the exact opposite. Or an experimenter could show that people in individualist cultures are more persuaded by gainframed than by loss-framed appeals 374



(by having the gain-framed appeal invoke individualist consequences and the loss-framed appeal invoke collectivistic consequences)—or could show the exact opposite. In general, any individual-difference variable that goes proxy for, or straightforwardly represents, value variations makes the task of experimental message construction especially challenging. Showing the effect of such an individual-difference variable on the relative persuasiveness of gain- and loss-framed appeals requires ruling out variations in consequence desirability as an alternative explanation. 35. PMT is actually a bit more complex than this. Threat appraisal is said to depend not just on threat severity and threat vulnerability but also on the rewards of adopting a maladaptive response (e.g., the perceived rewards of not adopting the protective behavior); coping appraisal is said to depend not just on response efficacy and self-efficacy but also on response costs (perceived costs of taking the protective response, such as money, time, and so forth). But maladaptive rewards and response costs have received less research attention than the other four elements, and the simpler version presented here suffices to introduce the relevant general issues. (Moreover, the relation between self-efficacy and response costs is not lucid. After all, one reason why I might think that I can’t actually carry out a protective behavior such as an exercise program [low self-efficacy] is that it takes too much time [high response cost]. But PMT treats these separately.) 36. The mean correlations between PMT-related perceptions (perceived threat severity and so on) and persuasive outcomes (intentions and behaviors) were reported by Milne et al. (2000) to range roughly from .05 to .35, although relatively few cases were available for analysis. The paucity of cases reflects unhappy data-analytic decisions in primary research experiments in which messages are varied in an attempt to influence one or more of the four underlying perceptual states (e.g., an experiment in which participants receive either a message depicting the recommended action as either relatively easy or relative difficult to adopt, so as to influence perceived self-efficacy). When researchers have assessed the relevant perceptual states (e.g., perceived self-efficacy), those data have been used as a “manipulation check” to confirm that the messages produced appropriate perceptual differences—and then those data have been discarded, with the analysis focused on how the message conditions vary in their effects on the outcome variables (e.g., intention). The unfortunate consequence is that even when researchers have collected data relevant to the relationship of PMT’s perceptual states to persuasive 375



outcomes such as intentions and behaviors, research reports commonly have not reported such information. 37. For ethical reasons, when the message topic concerns a real (as opposed to fabricated) threat, the experimental conditions often involve contrasts between (for instance) a high-vulnerability message and a nomessage control condition (e.g., Yzer et al., 1998). 38. Expressed as correlations, the mean effects of message manipulations (of threat severity, threat vulnerability, response efficacy, and selfefficacy) on corresponding perceptions (perceived threat severity, perceived threat vulnerability, and so forth) range from roughly .30 to .45 (Witte & Allen, 2000); Milne et al. (2000) reported mean effects ranging from about .25 to .65 but analyzed a much smaller set of studies. The mean effects (expressed as correlations) of these message manipulations on persuasive outcomes (attitudes, intentions, and behaviors) were reported by Witte and Allen (2000) to be in the neighborhood of .10 to .20; Floyd et al. (2000), reviewing a different (and usually smaller) set of studies, reported mean effects corresponding to correlations roughly in a range from .20 to .40; de Hoog et al. (2007), reviewing effects of severity, susceptibility, and response efficacy manipulations, reported mean effects corresponding to correlations generally ranging from about .10 to .20— though with a notable lack of effect of vulnerability manipulations on attitude. 39. Notice that this way of defining the message variation is based on the properties of the communication, not the reactions of an audience. By contrast, sometimes threat appeal variations have been defined on the basis of evoked reactions (so that a strong threat appeal is one that arouses more fear than does a weak one). But this latter way of defining message variations should be dispreferred (for discussion, see O’Keefe, 2003; Tao & Bucy, 2007). 40. Even these estimates may be misleading. For example, Witte and Allen (2000) reported a mean correlation of .30 between threat appeal manipulations and aroused fear (across 51 cases). But this figure was inflated by (a) the exclusion of studies with a failed “manipulation check” (studies in which there was not a dependable difference in aroused fear between message conditions) and (b) the adjustment of individual effect sizes, before being analyzed, for factors such as range restriction (thereby increasing the size of the individual effects). An analysis that included all 376



studies and used unadjusted correlations would presumably yield a smaller mean effect. (This treatment passes over complexities such as questions about how to interpret postmessage fear reports and about the potential role of evoked emotions other than fear. For helpful discussion, see Shen & Dillard, 2014.) 41. There has been regrettably little attention to describing the particulars of threat appeal variations. The meta-analytic treatments of this literature commonly simply rely on the primary research categories (e.g., strong and weak appeals) and do not consider what specific message features might have been experimentally varied. The consequence is that we know rather less than we might about what particular message variations might produce the observed effects. 42. Unfortunately, threat appeal research results are often reported in ways that do not permit full examination of the relationships of interest (see note 36 above). For example, it is common that a researcher will create two message variations (with strong and weak threat appeals), check that they aroused different levels of fear (in a manipulation check), and then report the contrast between the two message conditions for the persuasionoutcome dependent variables (such as attitude and intention)—leaving unreported the direct relationship between the presumed mediating state (fear) and the persuasion outcome variables. This has been a rather widespread problem in persuasion research. A brief way of putting the problem is to say that assessments of intervening states, rather than properly being understood (and analyzed) as assessing mediating states, have instead unfortunately been seen (and analyzed) as providing manipulation checks for independent variables (for discussion, see O’Keefe, 2003). 43. Carey, McDermott, and Sarma’s (2013) meta-analytic results are not obviously an exception to this generalization. That meta-analysis examined studies that compared messages that addressed road safety using threat appeals against various control messages. Across four studies, threat messages aroused greater fear than control messages (mean effect size of r = .64); across 15 studies, these two kinds of messages were not associated with differences in driving practices (mean effect size of r = .03). But of those 15 studies, only two assessed both fear arousal and driving outcomes (and these two were not separately analyzed), so there is not much direct evidence in this data set concerning the question of whether messages that arouse greater fear are also generally more persuasive. Notice that in meta377



analyses of threat appeal research where the experimental contrast of interest is between high-intensity and low-intensity depictions of negative consequences in threat appeals (e.g., Witte & Allen, 2000), the metaanalytic results speak to the question of how message designers should implement threat appeals. In Carey et al.’s (2013) review, where the experimental contrast was between threat appeals and nonthreat appeals, the meta-analytic results speak to the question of whether road safety message designers should have any reason to prefer threat appeals over nonthreat (control) appeals. 44. The relevant relationships are almost certainly only roughly linear, not rectilinear. For instance, in a given persuasive circumstance, as message intensity increases, there might come a point at which aroused fear plateaus. That is, at some point further increases in message intensity might not yield any greater fear (or any greater persuasion). 45. Although this is a commonly mentioned limiting condition on the persuasive effects of variations in threat-related aspects of threat appeals, the evidence is surprisingly slim. Consider, for example: In an effort to locate such evidence, Peters et al.’s (2012) meta-analytic review was limited to studies that varied both threat (depicted severity, depicted vulnerability, or both) and efficacy (depicted response efficacy, depicted self-efficacy, or both) and that had behavioral outcomes. Only six studies were eventually included. The results indicated that threat variations had a statistically significant effect on behavioral adoption when depicted efficacy was high (mean effect size reported as a standardized mean difference of .31, which corresponds to a correlation of .15) but not when depicted efficacy was low (mean effect size reported as a standardized mean difference of –.31, which corresponds to a correlation of –.15), with those two mean effect sizes indicated as being significantly different. These results look to be consistent with the claimed limiting condition. (It should not pass unremarked that this interpretation relies on treating experimental variations in depicted efficacy as a proxy for perceived efficacy.) But the fragility of these findings can be seen by noticing that these results were obtained only with the exclusion of another apparently relevant study (Study 43; Chu, 1966). A reanalysis of the set of seven studies (i.e., including Study 43), with effect sizes converted from standardized mean differences to correlations and analyzed using the random-effects procedures of Borenstein and Rothstein (2005), yields results that are not reassuring: Threat variations did not have a statistically significant effect on behavioral adoption when depicted efficacy was high 378



(mean r = .22, 95% CI limits of –.08 and .48) or when depicted efficacy was low (mean r = –.05, 95% CI limits of –.24 and .15), with those two mean effect sizes not significantly different from each other, Q(1) = 2.2, p = .14. Peters et al. (2012, pp. 11–12) did offer a rationale for the exclusion of Study 43, but the point here is the substantial effect that a single study can have on these meta-analytic conclusions. In sum, it may well be the case that more intense depictions of threat severity (and correspondingly greater fear arousal) will be associated with greater persuasion only when a workable, effective solution is perceived to be in hand—but the evidence to date is not as robust as one might like. 46. This description of the EPPM is necessarily only a brief gloss (the most detailed description is Witte, 1998). There is room for some uncertainty about the EPPM’s predictions, at least in part because presentations of the EPPM sometimes run together questions about influences on protective intentions and actions and questions about message effects. To clarify: Read in one way, the EPPM—and protection motivation theory (PMT)—offer an account of what influences protective intentions and actions. There is an obvious parallel here with reasoned action theory (RAT; Chapter 6). RAT, the EPPM, and PMT each identify a set of determinants of (influences on, precursors to) intentions and behavior. The EPPM and PMT are narrower than RAT (because the EPPM and PMT are focused on protective behaviors specifically) and so naturally have a distinct set of determinants, but the overall analytical approach is quite similar. The EPPM goes beyond PMT because it incorporates Leventhal’s (1970) concepts of fear control and danger control processes and because it offers an elaborated account of the interplay between these processes and among the various determinants. Even so, the EPPM can be seen to have the same central focus as PMT: understanding what factors affect protective intentions and actions. Notice that one can have such an account without ever considering questions about persuasive messages or interventions. Each distinct determinant is of course a potential target for persuasive messages, but how and why messages influence those perceptual determinants make for a separate set of research questions. It would be possible to know that perceptual state X (e.g., perceived selfefficacy) is strongly related to protective behaviors without knowing what sorts of messages or interventions influence X. As a parallel case from RAT: It is possible to know that perceived behavior control (PBC) is generally correlated with behavioral intentions without knowing how to influence PBC. But the EPPM wants to offer not only an account of what influences protective actions and how those influences are interrelated 379



(e.g., Propositions 6–10 in Witte, 1998) but also an account of what happens when threat-related persuasive messages are encountered (e.g., Propositions 1, 2, and 5). These enterprises are naturally related but conceptually distinct. Questions about what happens when this or that perceptual state has a given value (e.g., what happens when perceived threat severity is high) are different from questions about what happens when this or that message feature has a given value (e.g., what happens when depicted threat severity is high). In this research domain, unfortunately, such differences are not always fully grasped. For example, as Popova (2012, p. 457) pointed out, “the conceptual difference between threat as a message characteristic and perceived threat is often overlooked in practice.” The upshot is that sorting out predictions (and, for that matter, empirical findings) in this research area can be quite challenging. 47. Notice that where fear control processes are activated, recipients may end up experiencing relatively little fear and hence exhibit relatively little persuasion. That is, the EPPM’s picture here is consistent with the generally positive relationship between fear and persuasion: If people aren’t experiencing much fear (because they don’t find the message contents scary, don’t think they’re vulnerable, aren’t thinking about the threat, or any other reason), then they’re not likely to be especially motivated to adopt the protective action. 48. Popova’s (2012) review of EPPM research nicely draws attention to ways in which the EPPM’s concepts and claims have not been formulated as carefully as one might like. But that review’s lack of a quantitative treatment of the relevant research is felt acutely. For example, where the EPPM expects two conditions to differ, a study that fails to find a statistically significant difference between those conditions does not necessarily represent evidence that is inconsistent with the EPPM (the observed effect might have been in the predicted direction even though not statistically significant). For all one knows, a meta-analytic integration of significant and nonsignificant primary research effects would produce a pattern of mean effect sizes wholly consistent with the EPPM. 49. These various lines of research also commonly focus on a single emotional reaction (fear, guilt, and so on). Although research is only beginning to untangle the complexities here, there is good reason to think that messages influence multiple emotions and that the interplay of evoked emotions may be important in influencing message effects (e.g., Dillard & Peck, 2001; Morales, Wu, & Fitzsimons, 2012). 380



50. Other compliance techniques have also received some research attention. Notable among these are the “that’s-not-all” technique, in which before any response is given to the initial request, the communicator makes the offer more attractive (see, e.g., Banas & Turner, 2011; Burger, Reed, DeCesare, Rauner, & Rozolis, 1999); the “low-ball” technique, in which the communicator initially obtains the receiver’s commitment to an action and then increases the cost of performing the action (see, e.g., Cialdini, Cacioppo, Bassett, & Miller, 1978; Guéguen, Pascual, & Dagot, 2002); and the “legitimizing paltry contributions” technique, in which fundraisers explicitly legitimize small contributions (e.g., by saying “even a penny helps”; for a illustrative study, see Cialdini & Schroeder, 1976; for a review, see Andrews, Carpenter, Shaw, & Boster, 2008). Cialdini and Griskevicius (2010) provide a useful general discussion of compliance techniques. 51. The following representation of FITD research should be treated cautiously, for two reasons. First, the extant systematic reviews of this topic are now rather dated (the most recent was reported in 1999), which means that their conclusions do not reflect accumulated subsequent work. Second, the data-analytic procedures of these reviews can leave room for doubt about the security of their conclusions. For example, Burger’s (1999) analysis aggregated raw frequencies across studies (rather than treating each study’s data as providing a separate case). This had the unfortunate consequence of making the analysis’s confidence intervals insensitive both to the number of studies and to the amount of betweenstudies variation. Aggregating raw frequencies in this way can also potentially lead to a problem known as Simpson’s paradox, discussed by Borenstein et al. (2009, pp. 303–309), but Burger (1999, p. 306n1) suggested that that problem was unlikely to arise in this data set. 52. In computing the reported compliance rate in the FITD conditions (55%), all participants who heard the first request (regardless of whether they agreed to it) were included in the denominator in Freedman and Fraser’s (1966) data analysis. This is the appropriate way of figuring the FITD compliance rate (as opposed to figuring it by computing the compliance proportion among only those who received the second request). If those who declined the initial request had been excluded from the analysis, then a higher target-request compliance rate in the FITD condition might have been explained as an artifact of having excluded dispositionally uncooperative persons (those generally unwilling to accede to requests) from the FITD condition denominator but not the control 381



condition denominator (or, alternatively expressed, as an artifact of having dispositionally cooperative persons overrepresented in the FITD condition by virtue of having passed through what amounted to the screening procedure of the initial request). 53. The average effect size in FITD studies is roughly equivalent to a correlation of between .10 and .15, with larger effects under optimal conditions (Beaman et al., 1983; Dillard et al., 1984; Fern et al., 1986). 54. Chartrand, Pinckert, and Burger (1999) found that if the same person makes both requests with no delay between them, the FITD technique may backfire. But even this effect is apparently not general. Burger’s (1999) review reported an advantage for FITD conditions over control conditions when the same person made both requests without a delay (overall effect corresponding to a correlation of .05 across 24 studies); when the same person made the requests but with a delay between them (correlation of .07, seven studies); when different persons made the requests without a delay (correlation of .11, five studies); and when different persons made the requests with a delay (correlation of .12, 28 studies). Taken at face value (see note 51 above), these reported overall effects underwrite a conclusion that FITD effects are unaffected by delay and a suspicion that FITD effects perhaps might be larger when different persons make the requests than when the same person makes them (but in the absence of appropriate statistical analyses—comparing the differences between the relevant effects—this can be only a suspicion). 55. The lack of an effect for the time interval between the requests is sometimes seen as inconsistent with the self-perception explanation (e.g., Dillard et al., 1984). But it is not clear what predictions the selfperception explanation would make here. On the one hand, it might be expected that with increasing delay between the two requests, the FITD effect would weaken (because there would be many opportunities, during the interval, for other events to undermine the self-attributions of helpfulness and cooperativeness). On the other hand, it might be predicted that with increasing delay between the requests, the FITD effect would become stronger (because it takes time for receivers to reflect on the causes of their behavior and so to make the required self-attributions). Or (as Beaman et al., 1983, noted) it might be that both these processes are at work and cancel each other out. 56. A number of other explanations have also been proposed for FITD 382



effects (e.g., Ahluwalia & Burnkrant, 1993; Gorassini & Olson, 1995), but at present none seems entirely satisfactory. For many explanations, there is little direct relevant evidence. And it is not always obvious how the explanations can accommodate the existing evidence concerning moderating factors. For example, Fennis, Janssen, and Vohs’s (2009; see also Fennis & Janssen, 2010) account invokes self-regulatory resource depletion processes (the effortful character of responding to the first request makes yielding to the second request more likely); but resource depletion presumably dissipates relatively quickly whereas FITD effects have been observed at some temporal remove (e.g., two weeks: Freedman & Fraser, 1966). 57. The average effect size in DITF studies is equivalent to a correlation of between .05 and .15, with larger effects under optimal conditions (Dillard et al., 1984; Feeley, Anker, & Aloe, 2012; Fern et al., 1986; O’Keefe & Hale, 1998, 2001). 58. This representation of DITF moderator-variable effects is unsubtle. In the most recent meta-analytic review (Feeley et al., 2012, concerning verbal outcomes), mean DITF effects were numerically larger with the same requester, with the same beneficiary, with no delay between requests, and with prosocial requests (than in the respective comparison conditions). And in that review, whether statistically significant mean DITF effects were obtained did vary depending on each of those moderators (e.g., there was a statistically significant DITF effect when the beneficiaries of the two requests were the same but not when the beneficiaries differed). But the only statistically significant difference in mean DITF effects between conditions of those moderators was for requester variation (the mean effect was significantly larger when the same person made both requests than when different people made the two requests). For other moderator variables, differences between mean effect sizes for different moderator conditions (e.g., between the mean effect size for prosocial requests and the mean effect size for nonprosocial requests) were not statistically significant. So, on the one hand, one cannot say that (for example) DITF effects are dependably larger with prosocial requests than with nonprosocial requests; but on the other hand, one can confidently recommend the use of the DITF strategy only with prosocial requests (because with nonprosocial requests, the obtained mean effect is not significantly different from zero). 59. The explanation might be defended against this criticism, however, by 383



suggesting that any concession merely needs to be large enough to trigger the reciprocal concessions norm; so long as the concession surpasses this threshold, the reciprocal concessions mechanism will be engaged (and thus increasing the size of the concession beyond that threshold would not affect the strategy’s effectiveness). This defense is certainly adequate as far as it goes, but consider that if larger concessions had been found to be associated with greater DITF effectiveness, such a result surely would have been counted as evidence supporting the reciprocal concessions explanation. Thus the failure to find such effects requires at a minimum some revision in the account (such as represented by the articulation of a threshold-model version of the explanation).



384



Chapter 12 Receiver Factors Individual Differences Topic-Specific Differences General Influences on Persuasion Processes Summary Transient Receiver States Mood Reactance Other Transient States Influencing Susceptibility to Persuasion Reducing Susceptibility: Inoculation, Warning, Refusal Skills Training Increasing Susceptibility: Self-Affirmation Conclusion For Review Notes



This chapter reviews research concerning the effects that various recipient characteristics have on persuasive outcomes. The discussion is organized around three main topics: individual differences (such as personality traits), transient receiver states (such as moods), and means of influencing receivers’ susceptibility to persuasion.



Individual Differences Individual differences (ways in which people vary with respect to relatively stable characteristics, such as personality variations) can influence persuasion in two broad ways: by virtue of their association with topic-specific differences or by virtue of their general influence on persuasion processes.



Topic-Specific Differences Some individual-difference variables can be associated with topic-specific differences in attitudes, beliefs, values, or behaviors—and hence (where relevant to the topic) such individual differences may be related to 385



persuasive effects. A convenient example is provided by the personality variable of self-monitoring (the degree to which a person regulates selfpresentation). As discussed in Chapter 3 (concerning functional approaches to attitude), high and low self-monitors differ in what they tend to value in certain consumer products; for instance, high self-monitors especially favor the image projection attributes of automobiles, whereas low self-monitors are more likely to value characteristics such as reliability. Hence high and low self-monitors are differentially influenced by corresponding persuasive appeals; high self-monitors react more favorably to image-oriented advertisements than to ads focused on product quality, whereas the opposite tendency is found for low self-monitors (e.g., Snyder & DeBono, 1985). To put the relevant point generally, this personality difference serves as a marker of differences in receiver values and hence is related to the success of persuasive appeals that vary in the degree to which those values are engaged. Similarly, variations in receiver age can sometimes serve as a marker of topic-specific variations in persuasion-relevant beliefs and attitudes. For instance, the motivations underlying volunteering may vary as a function of age: There is some evidence that as people become older, the importance of career-oriented motivations declines and the importance of the interpersonal relationships that can come from volunteering increases (Okun & Schultz, 2003). Thus when seeking to encourage volunteering, one might want to use different persuasive appeals for receivers varying in age—not because age per se is important but because age serves as a proxy for the relevant value differences. (See, relatedly, Williams & Drolet, 2005; Zhang, Fung, & Ching, 2009.) As another example, differences in receivers’ cultural backgrounds might serve to index various topic-specific differences relevant to persuasion. For instance, cultural variation may be associated with value differences and hence with differential effectiveness of corresponding persuasive appeals (e.g., Chang, 2006; Gregory, Munch, & Peterson, 2002; Laufer, Silvera, McBride, & Schertzer, 2010; for a review, see Hornikx & O’Keefe, 2009).1 Or receivers of varying cultural backgrounds might typically have different salient beliefs on a given topic, suggesting correspondingly different persuasive approaches (see, e.g., Marin, Marin, Perez-Stable, Sabogal, & Otero-Sabogal, 1990). (It is also possible that within a given cultural group, individual variation in, for example, cultural identification might provide still further bases for message adaptation. See Davis & Resnicow, 2012; Kreuter, Lukwago, Bucholtz, Clark, & Sanders386



Thompson, 2003.)



General Influences on Persuasion Processes The second way in which individual differences might influence persuasion is through their general influence on persuasion processes. For example, as discussed in Chapter 8 (concerning the elaboration likelihood model), individual differences in need for cognition (NFC) are associated with differences in elaboration motivation: Those higher in NFC are generally likely to have greater elaboration motivation than those lower in NFC. Other factors also influence elaboration motivation, of course, but the point here is that (ceteris paribus) across advocacy topics generally, one expects NFC variations to produce corresponding variations in elaboration likelihood. Variations in receiver self-esteem and intelligence may also affect persuasion processes generally. There is some evidence that persuasibility may be maximized at intermediate levels of self-esteem and at lower levels of intelligence (for reviews, see Rhodes & Wood, 1992).2 Explanations for these differences are at present somewhat speculative, but the most plausible accounts look to the possible influence of these personality characteristics on various general aspects of persuasion processes (see Rhodes & Wood, 1992). Concerning self-esteem, the suggestion has been that persons low in self-esteem are unlikely to pay sufficient attention to the message (by virtue of being withdrawn or distracted), and those high in self-esteem are likely to be confident in the correctness of their current opinions (and so, perhaps, be more likely to counterargue), thus making each group less likely to be persuaded than those of moderate levels of self-esteem (for potential complexities, see Sanaktekin & Sunar, 2008). With regard to intelligence, it may be that the greater knowledgeability commonly associated with greater intelligence enables more critical scrutiny of messages. Another (more complex) example is provided by the individual-difference variable of sensation seeking, which concerns preferences for novel, complex, and ambiguous stimuli and situations. High sensation seekers appear to use drugs and alcohol more frequently, and to begin using them at an earlier age, than do low sensation seekers (e.g., Zuckerman, 1979, pp. 278–294). In designing programs aimed at reducing drug use, then, this personality variable provides a means of identifying people who are the most important and appropriate targets for persuasive messages 387



(Stephenson et al., 1999). But high and low sensation seekers also differ in the kinds of messages to which they are especially susceptible; for example, the use of rapid edits, intense imagery, and surprise endings can make for more effective antidrug public service announcements for high sensation seekers (e.g., Morgan, Palmgreen, Stephenson, Hoyle, & Lorch, 2003; Niederdeppe, Davis, Farrelly, & Yarsevich, 2007; for complexities, see Kang, Cappella, & Fishbein, 2006). The implication is that sensation seeking might provide not only a basis for identifying members of the target audience but also a means of adapting messages to that audience (for a useful review, see Morgan, 2012).



Summary A great many individual individual-difference receiver characteristics have been examined for their possible relationships to persuasibility. For most such characteristics, the research evidence is commonly not extensive, and dependable generalizations seem hard to come by. (For some illustrative studies, see Geers, Handley, & McLarney, 2003; Guadagno & Cialdini, 2010; Gunnell & Ceci, 2010; Hirsh, Kang, & Bodenhausen, 2012; Lee & Bichard, 2006; Magee & Kalyanaraman, 2009; Resnicow et al., 2008; Saucier & Webster, 2009; Stephenson, Quick, & Hirsch, 2010; van ’t Riet, Ruiter, & de Vries, 2012; Williams-Piehota et al., 2009.) But, as this brief sketch indicates, individual differences may affect persuasion in a number of different ways, so perhaps it is unsurprising that research has so often yielded complex results. A given individual difference such as receiver age might potentially be related to general persuasibility differences (Krosnick & Alwin, 1989), to dispositional differences in information-processing inclinations (e.g., Williams & Drolet, 2005), and to topic-specific differences in persuasion-relevant beliefs and attitudes (as in the observed age-related differences in evaluations of volunteering outcomes; Okun & Schultz, 2003). Similarly, cultural variations may be related not only to variations in underlying values but also to some information-processing differences (e.g., Hornikx & Hoeken, 2007; Larkey & Gonzalez, 2007). It is likely to take some time to sort out completely the different pathways by which various individual-difference variables exert their influence on persuasive effects. (For one attempt, see Briñol & Petty, 2005. For other general discussions of individual differences and persuasion, see Briñol, Rucker, Tormala, & Petty, 2004; Shakarchi & Haugtvedt, 2004; W. Wood & Stagner, 1994.)



388



Transient Receiver States Whereas the previous section discussed the effects of relatively stable (individual-difference) receiver characteristics, persuasion can also potentially be influenced by more transient receiver states. Two such states are discussed here: mood and reactance.3



Mood There seems to be a natural appeal to the hypothesis that a receiver’s preexisting mood (affective state) will influence persuasion quite straightforwardly, such that positive moods will enhance persuasion and negative moods will diminish persuasion. And although the research evidence to date does not yet yield an entirely clear picture, it is nevertheless plain that this simple hypothesis will not suffice. Rather, the research in hand appears to suggest that receivers in (at least some kinds of) negative moods are more likely to engage in close message processing than are receivers in (at least some kinds of) positive moods. Expressed in terms of the elaboration likelihood model (ELM; see Chapter 8), mood influences elaboration likelihood. For example, Bless, Bohner, Schwarz, and Strack (1990) found that sad participants were persuaded by a counterattitudinal message if the message’s arguments were strong but not if the arguments were weak (indicating relatively high elaboration); by contrast, happy participants were equally persuaded by strong and by weak arguments (suggesting relatively low elaboration). (For similar results, see, e.g., Mackie & Worth, 1989. For a review, see Hullett, 2005.)4 Research has only begun to explore possible moderating conditions for this effect (circumstances under which this effect weakens or even reverses), so conclusions are not yet secure (see, e.g., Banas, Turner, & Shulman, 2012; Das, Vonkeman, & Hartmann, 2012; Shen, 2013; Sinclair, Moore, Mark, Soldat, & Lavis, 2010; Ziegler, 2013, 2014). However, one notable theme has been the suggestion that instead of referring generally to positive and negative moods (affective states), a more differentiated treatment of affective states will be needed—because, for example, different positive affective states may have different message-processing consequences. (For some illustrative studies, see Agrawal, Menon, & Aaker, 2007; Griskevicius, Shiota, & Neufeld, 2010. For some useful general discussions of affect and persuasion, see Bless & Schwarz, 1999; 389



Dillard & Seo, 2013; Nabi, 2007, 2010.)



Reactance Reactance is a motivational state that is aroused when a person’s freedom is perceived to be threatened or eliminated (Brehm, 1966; Brehm & Brehm, 1981). When a person believes that his or her freedom may be diminished, the person will be motivated to restore (defend, exercise) that freedom—perhaps by acting counter to the impending pressure. (For some general treatments of reactance, see Miron & Brehm, 2006; Quick, Shen, & Dillard, 2013.) Although reactance can be aroused by any number of circumstances (e.g., people may resist granting favors that they think would obligate them in the future, so as to retain their freedom of action), counterattitudinal persuasive messages are an obvious potential cause of reactance. Such communications seek to reduce recipients’ freedom (of belief and action) by suggesting that receivers adopt the advocated view, so such messages might potentially be seen as trying to manipulate or pressure the recipient. When a message arouses reactance, its persuasiveness is naturally diminished. Receivers experiencing reactance are likely to reject the advocated view; they may cling to their existing attitudes more strongly and perhaps even change in the direction opposite that sought by the message (a boomerang effect). For many years, reactance was treated as an unmeasurable state (identifiable only by its antecedents and consequences); however, recent work has developed useful assessments of reactance, which in turn have permitted reactance to be further unpacked. Specifically, where persuasive messages are concerned, reactance seems best conceived of as a combination of anger (an affective reaction) and counterarguing (a cognitive reaction; see Dillard & Shen, 2005; Rains, 2013). That is, reactance is not merely an affective (emotional) state but also involves cognitive activity (specifically, counterarguing). Research is only beginning to explore what makes reactance more or less likely to be evoked.5 For many potential influences, the evidence is as yet rather slim (for some illustrative studies, see Edwards, Li, & Lee, 2002; Feiler, Tost, & Grant, 2012; Fitzsimons & Lehmann, 2004; Lee & Lee, 2009; Reinhart & Anker, 2012; Seibel & Dowd, 2001; Silvia, 2005).6 390



However, a number of studies have found that explicit freedomthreatening language can evoke reactance and lead to diminished persuasiveness (compared with parallel messages without such language).7 For example, Quick and Considine (2008) compared exercise messages with “forceful” language (e.g., “it is impossible to deny all the evidence” of exercise benefits, “no other conclusion makes sense,” and so forth) and ones with non-forceful language (e.g., “there is pretty good evidence” of exercise benefits, “it’s a sensible conclusion,” and so on), finding that forceful language evoked reactance and reduced perceived message persuasiveness. (For similar results, see, e.g., Bensley & Wu, 1991; Burgoon et al., 2002; Dillard & Shen, 2005; Rains & Turner, 2007.)8 Avoiding such directive language is thus one way in which to reduce the likelihood of the arousal of reactance. Another potential strategy for minimizing reactance or its effect might be to emphasize the receiver’s freedom of choice (e.g., Miller, Lane, Deatrick, Young, & Potts, 2007). (See also the “but you are free to refuse” request strategy, as in Guéguen & Pascual, 2005; for a review, see Carpenter, 2013.) Indeed, in work on influencing addictive and related problematic health behaviors, the approach known as motivational interviewing specifically recommends that counselors avoid confrontation and instead affirm the client’s autonomy and capacity for self-direction (for a general treatment of motivational interviewing, see Miller & Rollnick, 2002; for some reviews of relevant research, see Hettema & Hendricks, 2010; Jensen et al., 2011; Knight, McGowan, Dickens, & Bundy, 2006; Morton et al., 2014).



Other Transient States A number of other temporary states (beyond moods and reactance) have also been investigated for their potential roles in persuasion processes. For example, researchers have examined the effects of counterfactual mindsets (in which a person thinks about what might have been; e.g., Krishnamurthy & Sivaraman, 2002; Tal-Or et al., 2004); of “bolstering” and “counterarguing” mindsets (which orient the recipient to have thoughts supporting or opposing the advocated view; Xu & Wyer, 2012); of low-level, concrete mindsets as opposed to higher-level, abstract mindsets (e.g., K. White, MacDonnell, & Dahl, 2011); of the receiver’s expecting to have to communicate with others about the subject (e.g., Nienhuis, Manstead, & Spears, 2001; Tal-Or, Nemets, & Ziv, 2009); and so forth. But the evidence in hand on these states is as yet too slim to 391



support confident conclusions.



Influencing Susceptibility to Persuasion How can people’s susceptibility to persuasive messages be influenced? Sometimes an advocate will have an interest in making people more resistant (reducing their susceptibility to persuasion by opposing messages), and sometimes a persuader will hope to make people less resistant (increasing their susceptibility to the persuader’s efforts). This section discusses ways of influencing people’s susceptibility to persuasion. (For a collection of papers relevant to the general topic of influencing susceptibility, see Knowles & Linn, 2004.)



Reducing Susceptibility: Inoculation, Warning, Refusal Skills Training It’s all very well to persuade someone to one’s point of view—but once persuaded, the person may be exposed to counterpersuasion, that is, persuasive messages advocating some opposing viewpoint. The question that naturally arises is how receivers might be made resistant to such persuasive efforts (recognizing that making receivers resistant to counterpersuasion may involve something different from persuading them in the first place).9 In what follows, three approaches to creating such resistance to persuasion are discussed: inoculation, warning, and refusal skills training.



Inoculation The fundamental ideas of inoculation can be usefully displayed through a biomedical metaphor. Consider the question of how persons can be made resistant to a disease virus (such as smallpox). One possibility is what might be called supportive treatments—making sure that people get adequate rest, a good diet, sufficient exercise, necessary vitamin supplements, and so on. The hope, obviously, is that this treatment will make it less likely that the disease will be contracted. But another approach to inducing resistance is inoculation (as with smallpox vaccines). An inoculation treatment consists of exposing persons to small doses of the disease virus. The dose is small (to avoid bringing on the disease itself) but is sufficient to stimulate and build the body’s defenses so that any later 392



massive attack (e.g., a smallpox epidemic) can be defeated. As this analogy suggests, the parallel approach for inducing resistance to persuasion consists of giving people “small doses” of the opposing view. Specifically, an inoculation treatment consists of exposing people to a weak attack on their current attitudes and then refuting that attack. That is, the advocate explicitly discusses and refutes some opposing argument.10 In experimental designs in this research area, receivers are initially exposed to a treatment designed to induce resistance to persuasion on a given topic and then are exposed to an attack message (i.e., a counterattitudinal message) on that topic to see whether the treatment has made them resistant to the attack. (For some general discussions of inoculation theory and research, see Compton, 2013; Compton & Pfau, 2005; Ivanov, 2012; Pfau & Szabo, 2004. For a classic treatment, see McGuire, 1964.) Such inoculation treatments do commonly create resistance to persuasion when compared with no-treatment control conditions. (For some illustrative studies, see Bither, Dolich, & Nell, 1971; Nabi, 2003; Pfau, Holbert, Zubric, Pasha, & Lin, 2000. For a review, see Banas & Rains, 2010.)11 That is, showing receivers refutations of weak opposing arguments makes receivers more resistant to persuasion (by subsequent attack messages) than they otherwise would have been. Notably, this resistance-creating effect of inoculation is not limited to attack messages that use the same arguments that were refuted in the inoculation treatment. That is, the resistance conferred by inoculation generalizes beyond the refuted arguments; receivers who have been inoculated are also more resistant (than they would have been) to novel opposing arguments (Banas & Rains, 2010). The implication is that an advocate need not try to inoculate receivers against all possible opposing arguments; application of an inoculation treatment to only some opposing arguments will create generalized resistance. Following the biomedical metaphor, however, supportive treatments are a possible alternative to inoculation. In persuasion, a supportive treatment would consist of providing people with arguments and information supporting their current views.12 One might imagine that supportive treatments could also create resistance to persuasion, by bolstering the existing attitude (just as, in the biomedical realm, supportive treatments— good diet, adequate rest, and so forth—might create some resistance to 393



disease). Research to date suggests that supportive treatments may indeed confer some resistance to persuasion (compared to no-treatment control conditions), but the evidence is not yet quite as decisive as one might want. (For examples of relevant research, see Bernard, Maio, & Olson, 2003; Rosnow, 1968. For a review, see Banas & Rains, 2010.)13 The more illuminating comparison, however, is between inoculation treatments and supportive treatments, and the evidence in hand plainly indicates an advantage for inoculation: Inoculation treatments confer greater resistance to persuasion than do supportive treatments. (For illustrative research, see Adams & Beatty, 1977; Kamins & Assael, 1987b. For a review, see Banas & Rains, 2010.)14 That is, providing refutation of weak counterarguments is more effective in making people resistant to counterpersuasion than is bolstering their existing attitudes by providing supportive material. Thus, given a choice between administering a supportive treatment or an inoculation treatment, advocates should presumably prefer inoculation.15 Evidence concerning the resistancecreating effects of combining inoculation and supportive treatments is unhappily slim (and has not been systematically reviewed); there is, however, some evidence suggesting that the combination may be more effective in conferring resistance than are supportive treatments alone (e.g., Koehler, 1972; McCroskey, Young, & Scott, 1972; Szybillo & Heslin, 1973).16 Explaining how and why inoculation induces resistance to persuasion has proved rather challenging. Given that the resistance created by inoculation treatments generalizes to novel arguments, the underlying mechanism is unlikely to involve receivers simply acquiring the contents of the particular refutation to which they are exposed; obviously, something more general happens, something that makes receivers resistant even to new attacks on their views. The most commonly invoked mechanism for explaining inoculation effects has been the receiver’s perception that their views are vulnerable to attack. In McGuire’s (1964) classic formulation of inoculation theory, the supposition was that if receivers are not aware that their beliefs might be opposed, then they will be unmotivated to prepare their cognitive defenses. An inoculation treatment was thought to stimulate the receiver’s natural defenses and so make them resistant to subsequent attack messages.17 That is, one effect of the inoculation treatment is (supposedly) to make the 394



recipient aware of the possibility of opposing arguments or views (because the recipient sees an opposing argument).18 But the idea of perceived vulnerability has not yet been entirely carefully specified. For example, does the receiver need to think that an attack message is actually about to be encountered? Or only that it might plausibly occur at some time in the future? Or perhaps merely that in the abstract, somebody somewhere might believe differently (never mind whether an actual attack is expected)? Is it necessary that the receiver think that the attack message (or the imagined interlocutor) has good reasons for the opposing view (reasons that might form the basis of good arguments against the receiver’s opinion)? Or perhaps is the mere recognition of the possibility of opposition (whether or not well-founded) sufficient?19 Even given some specification of the idea of vulnerability, the issue will then become explaining exactly how and why perceived vulnerability leads to resistance. Perhaps it stimulates counterarguing, or possibly it simply inclines the receiver to reject the opposing view out of hand without thinking about it very much.20 (For some discussion of alternative means of resistance, see Ahluwalia, 2000; Blumberg, 2000; Burkley, 2008.) In short, much remains to be learned about how inoculation creates resistance to persuasion.



Warning If one’s awareness that a belief is vulnerable to attack might be sufficient to lead one to bolster one’s defense of that belief (and thereby reduce the effectiveness of attacks on it), then perhaps simply warning a person of an impending counterattitudinal message will decrease the effectiveness of the attack once it is presented. A fair amount of research has been conducted concerning the effects of such warning on resistance to persuasion. Two sorts of warnings have been studied. One type simply warns receivers that they will hear a message intended to persuade them, without providing any information about the topic of the message, the viewpoint to be advocated, and so on. The other type of warning tells receivers the topic and position of the message. Both sorts of warnings can confer resistance to persuasion and appear to do so by stimulating counterarguing in the audience (e.g., Petty & 395



Cacioppo, 1977, 1979a; for a review, see W. Wood & Quinn, 2003).21 Topic-position warnings make it possible for receivers to engage in anticipatory counterarguing because the audience knows the issue to be discussed and the view to be advocated. Thus as the time interval between the topic-position warning and the onset of the message increases (up to a point, anyway), there is more opportunity for the audience to engage in counterarguing. For example, in one study, high school students were shown messages arguing that teenagers should not be allowed to drive. Students received a warning of the topic and position of the impending message, but the interval between the warning and the message varied (no delay between warning and message, a 2-minute delay, or a 10-minute delay). With increasing delay, there was increasing resistance to persuasion (Freedman & Sears, 1965). Persuasive-intent warnings, of course, do not permit anticipatory counterarguing because the receivers do not know the subject of the message; consequently, variations in the time interval between a persuasive-intent warning and the communication have little effect on resistance (e.g., R. G. Hass & Grady, 1975). But persuasive-intent warnings do apparently stimulate greater counterarguing during the persuasive message, thereby reducing receivers’ susceptibility to persuasion. Because warning apparently creates resistance by encouraging counterarguing, the effectiveness of warnings is influenced by factors that affect receivers’ motivation and ability to counterargue. When receivers are not motivated to counterargue (e.g., because the issue is insufficiently important to them) or are unable to counterargue (e.g., because accompanying distraction prevents them from doing so), then the resistance-inducing effects of warning are reduced (see, e.g., H. C. Chen, Reardon, Rea, & Moore, 1992; Neimeyer et al., 1991; Petty & Cacioppo, 1979a; Romero, Agnew, & Insko, 1996; for a review, see W. Wood & Quinn, 2003).22



Refusal Skills Training Another, more specialized approach to creating resistance to persuasion focuses on training the receiver in skills for refusing unwanted offers. The central idea is that in some circumstances, the key to resistance is being able to refuse offers or requests made by an influence agent. In particular, it has often been supposed that children and adolescents are commonly 396



unable to resist offers of illegal drugs, alcohol, or tobacco and so end up using these substances—even if they have negative attitudes about such substances. Hence it has been thought that one avenue to preventing substance use (or abuse) might be to provide training in how to refuse such offers. Refusal skills training is a different approach to resistance induction from inoculation or warning. Inoculation and warning seek to provide the receiver with certain sorts of cognitive defenses (hardening the initial attitude, preparing the receiver’s attitudinal defenses, encouraging mental counterarguing). In contrast, refusal skills training aims at equipping the receiver with certain communicative abilities. A good deal of research has explored refusal skills induction in the context of preventing children and adolescents from using or misusing drugs (alcohol, tobacco, marijuana, and so on). Three broad conclusions may be drawn from this research. First, it is possible to teach such refusal skills. Studies have found that resistance skills training does improve the quality of role-played refusals, participants’ perceived self-efficacy for refusing offers, and the like (see, e.g., Brown, Birch, Thyagaraj, Teufel, & Phillips, 2007; Langlois, Petosa, & Hallam, 1999; Wynn, Schulenberg, Maggs, & Zucker, 2000). Second, the programs that are most effective at teaching refusal skills commonly involve rehearsal with directed feedback (i.e., opportunities for participants to practice their refusal skills and to receive systematic evaluation of their performance). Simply encouraging participants to refuse offers or providing information about refusal skills seems less effective in developing such skills than is providing guided practice (see, e.g., Corbin, Jones, & Schulman, 1993; Turner et al., 1993). Third, refusal skills programs are generally not very effective in preventing or reducing drug, alcohol, or tobacco use/misuse. Evaluations of such programs commonly find that refusal skills (or refusal skills selfefficacy or exposure to refusal skills training) are unrelated to substance use or abuse (for examples and reviews, see Donaldson, Graham, & Hansen, 1994; Elder, Sallis, Woodruff, & Wildey, 1993; Gorman, 1995; D. C. Smith, Tabb, Fisher, & Cleeland, 2014; Wynn et al., 2000). Some successes have been reported (e.g., Donaldson, Graham, Piccinin, & Hansen, 1995; Hecht, Graham, & Elek, 2006), but in a few circumstances, boomerang effects—where refusal skills training has outcomes that appear 397



to encourage substance use—have also been observed (Biglan et al., 1987; Donaldson et al., 1995; S. Kim, McLeod, & Shantzis, 1989). In the United States, one particularly prominent refusal skills training program has been Drug Abuse Resistance Education (DARE); the program is aimed at children and adolescents and has featured police officers as influence agents. The traditional DARE curriculum has many elements, but its core is focused on teaching students skills for recognizing and resisting pressures to use drugs and alcohol. Despite widespread implementation, there is strikingly little evidence that DARE dependably reduces substance use (for reviews, see Ennett, Tobler, Ringwalt, & Flowelling, 1994; Pan & Bai, 2009; S. K. West & O’Neal, 2004).23 The lack of effectiveness of these refusal skills training programs may suggest that the key to preventing substance use/misuse is not found in the ability to refuse offers but rather lies somewhere else. (One possibility is that descriptive norms—the perceived prevalence of substance use among one’s peers—are a more important determinant of use than are refusal skills; see, e.g., Donaldson et al., 1994; Wynn et al., 2000. See, relatedly, LaBrie, Grossbard, & Hummer, 2009; Lai, Ho, & Lam, 2004; Vitoria, Salgueiro, Silva, & De Vries, 2009.) That is, the apparent relative ineffectiveness of the refusal skills programs aimed at preventing (or reducing) substance misuse does not mean that the general refusal skills induction strategy is somehow intrinsically defective as a mechanism for creating resistance to persuasion, only that it may not be especially helpful for these particular applications. Alternatively, one might think that, because some refusal skill programs appear to have been more successful than others in addressing substance use, the key to future program development is the identification of the relevant program ingredients. For some discussion along these lines, see Krieger et al. (2013) and Miller-Day and Hecht (2013).



Increasing Susceptibility: Self-Affirmation Some persuasive messages seem to be able to evoke a particular sort of defensive reaction in recipients—an avoidance motivation, in which recipients do not want to engage the message, want to avoid thinking about its information, perhaps want to avoid the topic entirely. This (avoidant, defensive) reaction is different from reactance. Where reactance is activated, recipients undertake counterarguing; that is, reactance is an 398



active form of resistance that engages the message. Defensive avoidance, on the other hand, represents a withdrawal from the message, an unwillingness to engage with it. It’s as though the message is somehow so threatening that people want to close themselves off from it.24 Consider, for example, that smokers may want to avoid information suggesting that smoking harms health, those who consume alcohol may not want to attend closely to messages describing alcohol risks, and so on. Generally speaking, people may well avoid information if attending to it seems capable of causing distress. And, obviously, such reactions are likely to minimize the success of persuasive messages.25 These avoidance behaviors can be seen to arise from a broad desire to maintain a positive self-image (itself a widely recognized general motivation; see, e.g., Briñol & Petty, 2005; Chaiken, Liberman, & Eagly, 1989; Prislin & Wood, 2005). For example, smokers may find it hard to hold a positive view of the self if they have to confront information suggesting they are harming themselves. The question that arises is how, in such circumstances, people can be made more susceptible to influence, more open to persuasion. The apparent motivational foundation for avoidance—the desire to maintain a positive self-image—suggests a possible avenue to minimizing these avoidance tendencies: self-affirmation. Self-affirmation refers to treatments aimed at affirming (confirming, supporting) the recipient’s positive characteristics or important values. Self-affirmation can be accomplished in a variety of ways, but the most common methods in research studies have had participants reflect on a core value (e.g., by writing about a value that is important to them, by describing instances in which they performed positive actions such as kindness behaviors, and so on). The idea is that active affirmation of some positive aspect of one’s self-concept will permit people to be open to information that would otherwise be threatening. (For some discussion of self-affirmation manipulations, see Armitage, Harris, & Arden, 2011; McQueen & Klein, 2006; Napper, Harris, & Epton, 2009.) Substantial evidence has accumulated that self-affirmation treatments can increase acceptance of subsequent messages with otherwise threatening contents. For example, self-affirmation has been reported to make coffee drinkers more accepting of information about the health risks of caffeine (Van Koningsbruggen, Das, & Roskos-Ewoldsen, 2009), to make sunbathers less defensive about sun-exposure risk information (Jessop, Simmonds, & Sparks, 2009), to make smokers more accepting of 399



information about smoking risks (Harris, Mayle, Mabbott, & Napper, 2007), to make opponents and proponents of capital punishment more open to opposing viewpoints (Cohen, Aronson, & Steele, 2000), and so forth. (For other examples, see Howell & Shepperd, 2012; Reed & Aspinwall, 1998; Schüz, Schüz, & Eid, 2013; Sparks, Jessop, Chapman, & Holmes, 2010; Van Koningsbruggen, & Das, 2009. For reviews, see Epton, Harris, Kane, van Koningsbruggen, & Sheeran, in press; Harris & Epton, 2009; Sweeney & Moyer, 2015.) At present, however, little can be confidently said about factors that might moderate these self-affirmation effects—the circumstances under which self-affirmation effects are most likely to occur, what sorts of selfaffirmation treatments might be most effective, whether individual differences affect the success of self-affirmation treatments, and so forth (for some illustrative studies, see Klein et al., 2010; Nan & Zhao, 2012; Pietersma & Dijkstra, 2011; Sherman et al., 2009). Similarly, research is only beginning to explore the mechanisms by which self-affirmation treatments have their effects (e.g., Crocker, Niiya, & Mischkowski, 2008; Klein & Harris, 2009; Van Koningsbruggen et al., 2009).26 But the manifest usefulness of self-affirmation recommends its continued investigation. (For some general discussions of self-affirmation theory and research, see J. Aronson, Cohen, & Nail, 1999; Harris, 2011; Harris & Epton, 2009, 2010; Sherman & Cohen, 2006.)



Conclusion Researchers have investigated a large number of recipient characteristics as possible influences on persuasive effectiveness; in particular, a great many individual-difference variables have received some attention. The present treatment provides only an overview of several especially prominent lines of research.



For Review 1. What are individual differences? Explain how individual-difference variables can be associated with topic-specific differences in attitudes, beliefs, values, or behavior. Give examples. Explain how individual-difference variables might be related to general differences in persuasion processes. Give examples. 2. Are people in positive moods generally more easily persuaded than 400



3.



4.



5.



6.



people in negative moods? Describe the effect of variation in moods on the extensiveness of message processing. What is reactance? How does reactance influence message persuasiveness? Is reactance purely an affective (emotional) state? Explain. Identify a message feature that might arouse reactance. Describe how persuaders might minimize the arousal of reactance. Describe the general idea of resistance to counterpersuasion. Identify two general ways persons might be made resistant to a disease virus. Describe supportive medical treatments; describe how inoculation against disease works. Describe inoculation treatments for inducing resistance to persuasion; are these treatments effective in creating resistance? Do inoculation treatments create resistance only to the particular attack arguments that are refuted, or does the resistance generalize to other attack arguments? Describe supportive treatments for inducing resistance to persuasion. Which treatment, supportive or inoculation, is more effective in creating resistance to persuasion? Describe one possible explanation for the resistance-creating effects of inoculation treatments. Can warning a person of an impending counterattitudinal message create resistance to persuasion? Distinguish two kinds of warnings. Explain the mechanism by which warning confers resistance to persuasion. Identify factors that influence the effectiveness of warnings. How might the effectiveness of warnings be influenced by the presence of distraction or by the degree of personal relevance of the topic to the receiver? What is refusal skills training? How is refusal skills training meant to create resistance to persuasion? Explain how refusal skills training is different from inoculation and warning as means of creating resistance to persuasion. Is it possible to teach refusal skills effectively? What are the most important elements in programs aimed at teaching refusal skills? What effect do refusal skills training programs have on substance use/misuse? Describe how persuasive messages might evoke defensive avoidance reactions from recipients. How are such reactions different from reactance? Describe one way of minimizing the arousal of defensive avoidance. What is self-affirmation? Can self-affirmation treatments enhance acceptance of threatening messages?



Notes 401



1. As discussed in Chapter 11 (concerning message factors), a number of individual-difference variables (such as “consideration of future consequences”) appear to be proxies for value differences (for a general treatment, see O’Keefe, 2013a). 2. Rhodes and Wood (1992) reviewed both self-esteem and intelligence effects on persuasibility. Their mean reported persuasibility difference between low and medium levels of self-esteem corresponds to a correlation of .12 (across nine cases), indicating greater persuasibility at medium levels; the difference between medium and high levels of selfesteem corresponds to a correlation of –.06 (across nine cases), again indicating greater persuasibility at medium levels. The mean persuasibility difference between low and high levels of intelligence corresponds to a correlation of –.14 (across seven cases), indicating greater persuasibility at lower levels. 3. Several other such states have been mentioned in the context of discussing messages aimed at arousing such states as fear or guilt (see Chapter 11 concerning message factors). 4. Hullett’s (2005) analysis is not quite as decisive on this point as one might suppose, for two reasons. First, Hullett’s analysis was limited to one particular index of message processing, namely, differentiated effects of messages varying in argument strength (the greater the observed difference in persuasiveness of strong and weak arguments, the greater the degree of message processing is taken to be). One cannot say whether studies with other indices of elaboration (e.g., number of thoughts reported, or memory for message contents) would yield similar conclusions. Second, the results were taken to suggest “some interference of more positive moods on message processing” (Hullett, 2005, p. 435), because the mean correlation between argument quality and attitude was larger in the negative mood condition (mean r = .39 across 21 cases) than in the positive mood condition (mean r = .29 across 12 cases). But the reported confidence intervals for those means suggests that those two mean effect sizes were not significantly different; indeed, a random-effects reanalysis (using the methods of Borenstein & Rothstein, 2005) of Hullett’s (2005, Table 1) effect sizes yielded a nonsignificant (p = .23) difference between negative (mean r = .377) and positive (mean r = .292) mood conditions. However, a random-effects reanalysis restricted to counterattitudinal and neutral messages (i.e., excluding proattitudinal messages) produced a significant (p = .030) difference in processing between mood conditions: mean rs of 402



.418 (based on 11 cases; 95% confidence interval limits of .278 and .541) and .236 (based on 8 cases; 95% confidence interval limits of .148 and .321) for negative and positive moods, respectively. (But see also Chapter 8, note 4.) 5. As Quick, Shen, and Dillard (2013) pointed out, a good deal of “reactance” research has not measured reactance directly but rather only inferred its presence. This makes for some uncertainty about how to interpret past findings, especially when alternative (nonreactance) interpretations are possible. 6. As discussed elsewhere in this chapter, if receivers are warned of an impending message described as intended to persuade, persuasiveness is commonly diminished (for a review, see W. Wood & Quinn, 2003). One might think that this effect reflects the arousal of reactance (generated by the manifest intent to influence), but similar resistance-to-persuasion effects are associated with warnings that do not mention the intent to persuade (they mention only the topic and message position), which suggests that reactance cannot explain the observed effects of intent-topersuade warnings (W. Wood & Quinn, 2003, p. 132). 7. Unfortunately this research literature does not contain a careful conceptual treatment of the relevant linguistic variations, which are variously described as “forceful,” “controlling,” or “freedom-threatening” language. Because the message features have not been specified, drawing dependable generalizations is difficult (and offering guidance to message designers is correspondingly challenging). 8. A number of studies that have found such manipulations to reduce persuasiveness have invoked reactance-based explanations but without assessing reactance. One may hope that future research will illuminate just how these language variations produce their effects (for some work along such lines, see, e.g., Craig & Blankenship, 2011; Haugtvedt, Shakarchi, Samuelson, & Liu, 2004; Silvia, 2006a). 9. One matter of some delicacy that is not treated here concerns the definition of resistance to persuasion, which poses more difficulties than one might initially suppose; a useful (if incomplete) discussion of this topic has been provided by Pryor and Steinfatt (1978, pp. 220–221). 10. As discussed in Chapter 11 (concerning message factors), a refutational two-sided message is one that both presents supporting 403



arguments and refutes opposing arguments. An inoculation treatment thus functionally consists of the refutational portion of a refutational two-sided message. 11. Banas and Rains’s (2010) fixed-effect analysis of adjusted effect sizes yielded a statistically significant advantage (in resistance creation) for inoculation treatments over no-treatment controls, with a mean d (standardized mean difference) of .43 across 41 cases (an effect that corresponds to an r of .21). A random-effects analysis (using the methods of Borenstein & Rothstein, 2005) of the unadjusted effect sizes (converting the reported ds to rs for the analysis) also yields a significant difference: mean r = .200 (95% CI limits of .160 and .238). 12. A supportive treatment thus amounts to a one-sided persuasive message (presenting only supportive materials). 13. Banas and Rains’s (2010) fixed-effect analysis of adjusted effect sizes yielded a statistically significant difference (in resistance creation) between supportive treatments and no-treatment controls, with a mean d (standardized mean difference) of .34 across 10 cases (an effect that corresponds to an r of .17). But a random-effects analysis (using the methods of Borenstein & Rothstein, 2005) of the unadjusted effect sizes (converting the reported ds to rs for the analysis) yields a nonsignificant (p = .052) difference: mean r = .130 (95% CI limits of –.001 and .256). Given the 95% confidence interval around that random-effects mean, however, smart money will surely bet that the population effect is positive. 14. Banas and Rains’s (2010) fixed-effect analysis of adjusted effect sizes yielded a statistically significant difference (in resistance creation) between inoculation treatments and supportive treatments, with a mean d (standardized mean difference) of .22 across 19 cases (an effect that corresponds to an r of .11). A random-effects analysis (using the methods of Borenstein & Rothstein, 2005) of the unadjusted effect sizes (converting the reported ds to rs for the analysis) also yields a significant difference: mean r = .099 (95% CI limits of .045 and .153). 15. This discussion has focused on the effects of inoculation on creating resistance to persuasion among people for whom the subsequent attack message is indeed an attack (i.e., is counterattitudinal). When an “inoculation” treatment is applied to people in advance of a proattitudinal message (i.e., what would otherwise be the “attack” message), on the other hand, that treatment is probably better conceived in different terms (e.g., as 404



a refutation of beliefs that the audience might hold); for some discussion, see M. Wood (2007). 16. Combining supportive and inoculation treatments amounts to creating a refutational two-sided persuasive message (discussed in Chapter 11), that is, a message that both gives arguments supporting the communicator’s view and gives refutations of counterarguments. Thus this point can be expressed in terms of message sidedness: Refutational two-sided messages may be more effective in conferring resistance to persuasion than are onesided messages. 17. This reasoning led to the expectation that “cultural truisms” (beliefs that a person rarely, if ever, hears attacked, such as “it’s a good idea to brush after every meal if possible”) would be especially vulnerable to attack, precisely because people were unpracticed at (and had no motivate to rehearse) defending those beliefs (McGuire, 1964). This reasoning also suggests that resistance-to-persuasion processes might differ between cultural truisms and more controversial beliefs. Specifically, (a) supportive treatments should differentially induce resistance in these two circumstances (with supportive treatments inducing more resistance on more controversial topics than on truisms), (b) refutational treatments should differentially induce resistance in these two circumstances (with inoculation producing greater resistance for truisms than for more controversial topics), (c) for truisms, refutational treatments will confer more resistance than supportive treatments, and (d) the difference in resistance-induction between the two kinds of treatments will be smaller for controversial topics than for truisms (such that it might even turn out that for controversial topics, refutational treatments and supportive treatments might not differ in resistance induction). However, the reports of much early inoculation research concerning truisms (e.g., McGuire & Papageorgis, 1961) do not contain sufficient statistical information to permit such questions to be addressed meta-analytically (Banas & Rains, 2010, p. 304). This lack is especially acutely felt because the contrast between truisms and controversial topics affords a natural basis for examination of the role putatively played by awareness of opposing views as a stimulus for arousing defenses. 18. Approached with the contrast between cultural truisms and controversial beliefs in mind, it is easy to see how inoculation might function this way for truisms. However, for more ordinary beliefs, people may already be entirely prepared to think that others might hold different 405



views—which would imply that inoculation ought have no special powers compared with, say, supportive treatments. And yet inoculation is demonstrably more effective than supportive treatments for conferring resistance on ordinary (nontruism) beliefs (Banas & Rains, 2010). 19. In some formulations, the essential element is said to be specific awareness of an impending counterattitudinal (i.e., attack) message (e.g., Compton & Pfau, 2005, p. 100). But this is rather different from a global awareness that opposing views are possible. Indeed, this approach as much as suggests that one necessary component of inoculation treatments is a warning of an impending counterattitudinal communication. But (as discussed elsewhere in this chapter) given that extant research indicates that such warnings alone (i.e., without any accompanying refutational material) create resistance to persuasion, it would look to be necessary to sort out exactly what contribution refutational material makes to the creation of resistance via inoculation. The apparent resistance-creating effect of warnings alone (as compared with no-treatment controls) corresponds to a mean correlation of .21 (W. Wood & Quinn, 2003), whereas the parallel effect for inoculation treatments is .20 (see note 11 above). Taken at face value, these results might suggest that the active ingredient in inoculation treatments is not the refutational component but rather the element of warning. 20. One might think that refutational inoculation treatments would create resistance by encouraging counterarguing (in response to subsequent attack messages), but what little evidence exists on this matter appears not to be encouraging (see, e.g., Benoit, 1991; Pfau et al., 1997, 2000). This is especially puzzling given that (a) warnings of impending counterattitudinal messages do stimulate counterarguing (as discussed shortly) and (b) such warnings may be responsible for the resistance-creating effects of inoculation treatments (as discussed in note 19 above). 21. In W. Wood and Quinn’s (2003) random-effects analysis, across 17 cases, the mean persuasion-inhibiting effect of warnings corresponded to a correlation of .21. The mean effects associated with the different kinds of warnings (warnings of intent to persuade, warnings of topic and position, or a combination of these) were statistically indistinguishable (though there were relatively few cases of each). 22. A number of complexities in the research literature on warning are passed over here. For example, on topics that are not especially personally 406



relevant to receivers, warnings sometimes seem to initially produce opinion change toward the to-be-advocated position (e.g., J. Cooper & Jones, 1970), but this change is apparently an anticipatory strategic shift meant to minimize the threat to self of having to change later in response to the message (and hence this effect evaporates when the expectation of the impending message is canceled); for discussion, see W. Wood and Quinn (2003). As another example of complexity, at least some prosocial solicitations appear to be made more persuasive if preceded by a warning (Kennedy, 1982); this, however, might reflect processes engaged when warning of an impending proattitudinal communication (e.g., such a warning, instead of eliciting the counterarguing engendered by warnings of counterattitudinal messages, might encourage supportive argumentation). 23. A random-effects analysis (using the methods of Borenstein & Rothstein, 2005) of the correlations in S. K. West and O’Neal’s (2004) Table 1 yields a nonsignificant (p = .18) mean correlation of .03 across 11 cases. A similar reanalysis of the effect sizes concerning drug use in Pan and Bai’s (2009) Table 1, converted to correlations, yields a significant (p = .02) mean correlation of .02 across 18 cases. The DARE program has undergone some revisions, but it is not clear that the changes have improved the program’s effectiveness (e.g., Vincus, Ringwalt, Harris, & Shamblen, 2010). 24. This treatment (of avoidance motivations) offers a simplified treatment of a complex subject. A number of different defensive reactions and processes might be distinguished (including a variety of avoidant defensive reactions), and it is not yet clear exactly how things will get sorted out conceptually or empirically. For useful discussions, see Blumberg (2000), Good and Abraham (2007), and van ’t Riet and Ruiter (2013). And one might place defensive reactions in a larger framework, such as Chaiken, Liberman, and Eagly’s (1989) differentiation of accuracy motivation (when recipients want to align their views with the facts), defense motivation (when recipients want to defend particular views), and impression motivation (when recipients want to make a good impression on others); or such as Briñol and Petty’s (2005) differentiation of knowledge, consistency, self-worth, and social approval motivations. For a general discussion of such frameworks in the context of defensive processing, see Eagly (2007). 25. Avoidant defensive reactions reduce persuasiveness, but presumably not simply because the recipient is avoiding thinking much about the issue. 407



After all, as the elaboration likelihood model (Chapter 8) suggests, even when there is little issue-relevant thinking (little elaboration), persuasion can still come about through the receiver’s use of heuristics. But in the case of avoidance motivation, recipients don’t want to think about the issue at all (and so don’t even use the cognitive shortcut of a heuristic). 26. A word of caution: Stapel and van der Linde’s (2011) report on selfaffirmation mechanisms was based on falsified data; see https://www.commissielevelt.nl/.



408



References Aaker, J. L., & Schmitt, B. (2001). Culture-dependent assimilation and differentiation of the self: Preferences for consumption symbols in the United States and China. Journal of Cross-Cultural Psychology, 32, 561–576. Aarts, H., Paulussen, T., & Schaalma, H. (1997). Physical exercise habit: On the conceptualization and formation of habitual health behaviours. Health Education Research, 12, 363–374. Abdulla, R. A. (2004). Entertainment-education in the Middle East: Lessons from the Egyptian oral rehydration therapy campaign. In A. Singhal, M. J. Cody, E. M. Rogers, & M. Sabido (Eds.), Entertainmenteducation and social change: History, research, and practice (pp. 301–320). Mahwah, NJ: Lawrence Erlbaum. Abelson, R. P. (1968). A summary of hypotheses on modes of resolution. In R. P. Abelson, E. Aronson, W. J. McGuire, T. M. Newcomb, M. J. Rosenberg, & P. H. Tannenbaum (Eds.), Theories of cognitive consistency: A sourcebook (pp. 716–720). Chicago: Rand McNally. Abelson, R. P. (1986). Beliefs are like possessions. Journal for the Theory of Social Behavior, 16, 223–250. Abelson, R. P. (1995). Statistics as principled argument. Hillsdale, NJ: Lawrence Erlbaum. Abelson, R. P., Kinder, D. R, Peters, M. D., & Fiske, S. T. (1982). Affective and semantic components in political person perception. Journal of Personality and Social Psychology, 42, 619–630. Abelson, R. P., & Prentice, D. A. (1989). Beliefs as possessions: A 409



functional perspective. In A. R Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function (pp. 361–381). Hillsdale, NJ: Lawrence Erlbaum. Abraham, C. (2008). Beyond stages of change: Multi-determinant continuum models of action readiness and menu-based interventions. Applied Psychology: An International Review, 57, 30–41. Abraham, C. (2012). Developing evidence-based content for health promotion materials. In C. Abraham & M. Kools (Eds.), Writing health communication: An evidence-based guide (pp. 83–98). Los Angeles: Sage. Abraham, C., & Sheeran, P. (2004). Deciding to exercise: The role of anticipated regret. British Journal of Health Psychology, 9, 269–278. Adams, J., & White, M. (2005). Why don’t stage-based activity promotion interventions work? Health Education Research, 20, 237–243. Adams, W. C., & Beatty, M. J. (1977). Dogmatism, need for social approval, and the resistance to persuasion. Communication Monographs, 44, 321–325. Adamval, R., & Wyer, R. S., Jr. (1998). The role of narratives in consumer information processing. Journal of Consumer Psychology, 7, 207–245. Adriaanse, M. A., de Ridder, D. T. D., & de Wit, J. B. F. (2009). Finding the critical cue: Implementation intentions to change one’s diet work best when tailored to personally relevant reasons for unhealthy eating. Personality and Social Psychology Bulletin, 35, 60–71. Adriaanse, M. A., Gollwitzer, P. M., de Ridder, D. T. D., de Wit, J. B. F., & Kroese, F. M. (2011). Breaking habits with implementation intentions: A test of underlying processes. Personality and Social Psychology Bulletin, 37, 502–513. 410



Adriaanse, M. A., Vinkers, C. D. W., de Ridder, D. T. D., Hox, J. J., & De Wit, J. B. F. (2011). Do implementation intentions help to eat a healthy diet? A systematic review and meta-analysis of the empirical evidence. Appetite, 56, 183–193. Agarwal, J., & Malhotra, N. K. (2005). An integrated model of attitude and affect: Theoretical foundation and an empirical investigation. Journal of Business Research, 58, 483–493. Agarwal, N., Menon, G., & Aaker, J. L. (2007). Getting emotional about health. Journal of Marketing Research, 44, 100–113. Aggarwal, P., Jun, S. Y., & Huh, J. H. (2011). Scarcity messages: A consumer competition perspective. Journal of Advertising, 40(3), 19–30. Agnew, C. R. (1998). Modal versus individually-derived beliefs about condom use: Measuring the cognitive underpinnings of the theory of reasoned action. Psychology and Health, 13, 271–287. Ahluwalia, R. (2000). Examination of psychological processes underlying resistance to persuasion. Journal of Consumer Research, 27, 217–232. Ahluwalia, R., & Burnkrant, R. E. (1993). A framework for explaining multiple request effectiveness: The role of attitude toward the request. Advances in Consumer Research, 20, 620–624. Aitken, C. K., McMahon, T. A., Wearing, A. J., & Finlayson, B. L. (1994). Residential water use: Predicting and reducing consumption. Journal of Applied Social Psychology, 24, 136–158. Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior (pp. 11–39). Berlin: Springer-Verlag.



411



Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. Ajzen, I. (2002). Perceived behavioral control, self-efficacy, locus of control, and the theory of planned behavior. Journal of Applied Social Psychology, 32, 1–19. Ajzen, I. (2011). The theory of planned behaviour: Reactions and reflections. Psychology & Health, 26, 1113–1127. Ajzen, I., Albarracín, D., & Hornik, R. (Eds.). (2007). Prediction and change of health behavior: Applying the reasoned action approach. Mahwah, NJ: Lawrence Erlbaum. Ajzen, I., & Cote, N. G. (2008). Attitudes and the prediction of behavior. In W. D. Crano & R. Prislin (Eds.), Attitudes and attitude change (pp. 289–311). New York: Psychology Press. Ajzen, I., Czasch, C., & Flood, M.G. (2009). From intentions to behavior: Implementation intention, commitment, and conscientiousness. Journal of Applied Social Psychology, 39, 1356–1372. Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84, 888–918. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall. Ajzen, I., & Fishbein, M. (2008). Scaling and testing multiplicative combinations in the expectancy-value model of attitudes. Journal of Applied Social Psychology, 38, 2222–2247. Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: 412



Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology, 22, 453–474. Ajzen, I., & Manstead, A. S. R. (2007). Changing health-related behaviours: An approach based on the theory of planned behaviour. In M. Hewstone, H. A. W. Schut, J. B. F. de Wit, K. van den Bos, & M. S. Stroebe (Eds.), The scope of social psychology: Theory and applications (pp. 43–63). New York: Psychology Press. Ajzen, I., Nichols, A. J., III, & Driver, B. L. (1995). Identifying salient beliefs about leisure activities: Frequency of elicitation versus response latency. Journal of Applied Social Psychology, 25, 1391–1410. Ajzen, I., & Sexton, J. (1999). Depth of processing, belief congruence, and attitude-behavior correspondence. In S. Chaiken & Y. Trope (Eds.), Dual-process models in social psychology (pp. 117–138). New York: Guilford. Akhtar, O., Paunesku, D., & Tormala, Z. L. (2013). Weak > strong: The ironic effect of argument strength on supportive advocacy. Personality and Social Psychology Bulletin, 39, 1214–1226. Akl, E. A., Oxman, A. D., Herrin, J., Vist, G. E., Terrenato, I., Sperati, F., … Schünemann, H. (2011). Framing of health information messages. Cochrane Database of Systematic Reviews, 2001 (12), CD006777. Albarracín, D., Johnson, B. T., Fishbein, M., & Muellerleile, P. A. (2001). Theories of reasoned action and planned behavior as models of condom use: A meta-analysis. Psychological Bulletin, 127, 142–161. Albarracín, D., & Wyer, R. S., Jr. (2001). Elaborative and nonelaborative processing of a behavior-related communication. Personality and Social Psychology Bulletin, 27, 691–705. Alden, D. L., & Crowley, A. E. (1995). Improving the effectiveness of 413



condom advertising: A research note. Health Marketing Quarterly, 12(4), 25–38. Alemi, F., Alemagno, S. A., Goldhagen, J., Ash, L., Finkelstein, B., Lavin, A., … Ghadiri, A. (1996). Computer reminders improve on-time immunization rates. Medical Care, 34, OS45–OS51. Algie, J., & Rossiter, J. R. (2010). Fear patterns: A new approach to designing road safety advertisements. Journal of Prevention and Intervention in the Community, 38, 264–279. Allcott, H., & Rogers, T. (2012). The short-run and long-run effects of behavioral interventions: Experimental evidence from energy conservation. NBER Working Paper No. w18492. Retrieved from SSRN: http://ssrn.com/abstract=2167595. Allen, C. T., Machleit, K. A., Kleine, S. S., & Notani, A. S. (2005). A place for emotion in attitude models. Journal of Business Research, 58, 494–499. Allen, M. (1991). Meta-analysis comparing the persuasiveness of onesided and two-sided messages. Western Journal of Speech Communication, 55, 390–404. Allen, M. (1993). Determining the persuasiveness of message sidedness: A prudent note about utilizing research summaries. Western Journal of Communication, 57, 98–103. Allen, M. (1998). Comparing the persuasive effectiveness of one- and twosided messages. In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 87–98). Cresskill, NJ: Hampton. Allen, M., Adamski, L., Bates, M., Bernhagen, M., Callendar, A., Casey, M., … Zirbel, C. (2002). Effect of timing of communicator identification and level of source credibility on attitude. Communication 414



Research Reports, 19, 46–55. Allen, M., & Preiss, R. W. (1997). Comparing the persuasiveness of narrative and statistical evidence using meta-analysis. Communication Research Reports, 14, 125–131. Allen, M. W., & Ng, S. H. (2003). Human values, utilitarian benefits and identification: The case of meat. European Journal of Social Psychology, 33, 37–56. Allen, M. W., Ng, S. H., & Wilson, M. (2002). A functional approach to instrumental and terminal values and the value-attitude-behaviour system of consumer choice. European Journal of Marketing, 36, 111–135. Allport, G. W. (1935). Attitudes. In C. Murchison (Ed.), A handbook of social psychology (pp. 798–844). Worcester, MA: Clark University Press. Alós-Ferrer, C., Granić, Đ.-G., Shi, F., & Wagner, A. K. (2012). Choices and preferences: Evidence from implicit choices and response times. Journal of Experimental Social Psychology, 48, 1336–1342. Alós-Ferrer, C., & Shi, F. (2012). Choice-induced preference change: In defense of the free-choice paradigm. Social Science Research Network (SSRN) Working Paper 2062507. doi: 10.2139/ssrn.2062507. Al-Rafee, S., & Dashti, A. E. (2012). A cross cultural comparison of the extended TPB: The case of digital piracy. Journal of Global Information Technology Management, 15(1), 5–24. Alwin, D. F. (1997). Feeling thermometers versus 7-point scales: Which are better? Sociological Methods and Research, 25, 318–340.



415



Amass, L., Bickel, W. K., Higgins, S. T., Budney, A. J., & Foerg, F. E. (1993). The taking of free condoms in a drug abuse treatment clinic: The effects of location and posters. American Journal of Public Health, 83, 1466–1468. Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness: A quantitative synthesis of effect size. International Journal of Advertising, 27, 209–234. Andersen, K. E. (1961). An experimental study of the interaction of artistic and nonartistic ethos in persuasion (Doctoral dissertation, University of Wisconsin-Madison). ProQuest No. 6103079. Andersen, R. E., Franckowiak, S. C., Snyder, J., Bartlett, S. J., & Fontaine, K. R. (1998). Can inexpensive signs encourage the use of stairs? Results from a community intervention. Annals of Internal Medicine, 129, 363–369. Anderson, C. A. (1983). Imagination and expectation: The effect of imagining behavioral scripts on personal intentions. Journal of Personality and Social Psychology, 45, 293–305. Anderson, L. (1970). An experimental study of reluctant and biased authority-based assertions. Journal of the American Forensic Association, 7, 79–84. Anderson, L. R. (1970). Prediction of negative attitude from congruity, summation, and logarithm formulae for the evaluation of complex stimuli. Journal of Social Psychology, 81, 37–48. Anderson, N. H. (1965). Averaging versus adding as a stimuluscombination rule in impression formation. Journal of Experimental Psychology, 70, 394–400.



416



Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78, 171–206. Anderson, N. H. (1981a). Foundations of information integration theory. New York: Academic Press. Anderson, N. H. (1981b). Integration theory applied to cognitive responses and attitudes. In R. E. Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 361–397). Hillsdale, NJ: Lawrence Erlbaum. Anderson, N. H. (Ed.). (1991). Contributions to information integration theory (Vols. 1–3). Hillsdale, NJ: Lawrence Erlbaum. Anderson, R. (2009). Comparison of indirect sources of efficacy information in pretesting messages for campaigns to prevent drunken driving. Journal of Public Relations Research, 21, 428–454. Anderson, R. B. (1995). Cognitive appraisal of performance capability in the prevention of drunken driving: A test of self-efficacy theory. Journal of Public Relations Research, 7, 205–229. Anderson, R. B. (2000). Vicarious and persuasive influences on efficacy expectations and intentions to perform breast self-examination. Public Relations Review, 26, 97–114. Anderson, R. B., & McMillion, P. Y. (1995). Effects of similar and diversified modeling on African American women’s efficacy expectations and intentions to perform breast self-examination. Health Communication, 7, 327–343. Andersson, E. K., & Moss, T. P. (2011). Imagery and implementation intention: A randomised controlled trial of interventions to increase exercise behaviour in the general population. Psychology of Sport and Exercise, 12, 63–70. 417



Andreoli, V., & Worchel, S. (1978). Effects of media, communicator, and message position on attitude change. Public Opinion Quarterly, 42, 59–70. Andrews, K. R., Carpenter, C. J., Shaw, A. S., & Boster, F. J. (2008). The legitimization of paltry favors effect: A review and meta-analysis. Communication Reports, 21, 59–69. Appel, M., & Mara, M. (2013). The persuasive influence of a fictional character’s trustworthiness. Journal of Communication, 63, 912–932. Appel, M., & Richter, T. (2007). Persuasive effects of fictional narratives increase over time. Media Psychology, 10, 113–134. Appel, M., & Richter, T. (2010). Transportation and need for affect in narrative persuasion: A mediated moderation model. Media Psychology, 13, 101–135. Applbaum, R. L., & Anatol, K. W. E. (1972). The factor structure of source credibility as a function of the speaking situation. Speech Monographs, 39, 216–222. Applbaum, R. L., & Anatol, K. W. E. (1973). Dimensions of source credibility: A test for repro-ducibility. Speech Monographs, 40, 231–237. Apsler, R., & Sears, D. O. (1968). Warning, personal involvement, and attitude change. Journal of Personality and Social Psychology, 9, 162–166. Areni, C. S., & Lutz, R. J. (1988). The role of argument quality in the elaboration likelihood model. Advances in Consumer Research, 15, 197–203.



418



Armitage, C. J. (2009). Is there utility in the transtheoretical model? British Journal of Health Psychology, 14, 195–210. Armitage, C. J., & Conner, M. (1999a). Distinguishing perceptions of control from self-efficacy: Predicting consumption of a low-fat diet using the theory of planned behavior. Journal of Applied Social Psychology, 29, 72–90. Armitage, C. J., & Conner, M. (1999b). The theory of planned behaviour: Assessment of predictive validity and “perceived control.” British Journal of Social Psychology, 38, 35–54. Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40, 471–499. Armitage, C. J., Conner, M., & Norman, P. (1999). Differential effects of mood on information processing: Evidence from the theories of reasoned action and planned behaviour. European Journal of Social Psychology, 29, 419–433. Armitage, C. J., Harris, P. R., & Arden, M. A. (2011). Evidence that selfaffirmation reduces alcohol consumption: Randomized exploratory trial with a new, brief means of self-affirming. Health Psychology, 30, 633–641. Armitage, C. J., Reid, J. C., & Spencer, C. P. (2011). Evidence that implementation intentions reduce single-occupancy car use in a rural population: Moderating effects of compliance with instructions. Transportmetrica, 7, 455–466. Armitage, C. J., & Reidy, J. G. (2008). Use of mental simulations to change theory of planned behaviour variables. British Journal of Health Psychology, 13, 513–524.



419



Armitage, C. J., & Talibudeen, L. (2010). Test of a brief theory of planned behaviour-based intervention to promote adolescent safe sex intentions. British Journal of Psychology 101, 155–172. Armstrong, A. W., Watson, A. J., Makredes, M., Frangos, J. E., Kimball, A. B., & Kvedar, J. C. (2009). Text-message reminders to improve sunscreen use: A randomized, controlled trial using electronic monitoring. Archives of Dermatology, 145, 1230–1236. Armstrong, C. L., & McAdams, M. J. (2009). Blogs of information: How gender cues and individual motivations influence perceptions of credibility. Journal of Computer-Mediated Communication, 14, 435–456. Armstrong, J. S. (2010). Persuasive advertising: Evidence-based principles. New York: Palgrave Macmillan. Arnold, W. E., & McCroskey, J. C. (1967). The credibility of reluctant testimony. Central States Speech Journal, 18, 97–103. Aronson, E. (1968). Dissonance theory: Progress and problems. In R. P. Abelson, E. Aronson, W. J. McGuire, T. M. Newcomb, M. J. Rosenberg, & P. H. Tannenbaum (Eds.), Theories of cognitive consistency: A sourcebook (pp. 5–27). Chicago: Rand McNally. Aronson, E. (1992). The return of the repressed: Dissonance theory makes a comeback. Psychological Inquiry, 3, 303–311. Aronson, E. (1999). Dissonance, hypocrisy, and the self-concept. In E. Harmon-Jones & J. Mills (Eds.), Cognitive dissonance: Progress on a pivotal theory in social psychology (pp. 103–126). Washington, DC: American Psychological Association. Aronson, E., & Carlsmith, J. M. (1963). Effect of the severity of threat on the devaluation of forbidden behavior. Journal of Abnormal and Social 420



Psychology, 66, 584–588. Aronson, E., Fried, C., & Stone, J. (1991). Overcoming denial and increasing the intention to use condoms through the induction of hypocrisy. American Journal of Public Health, 81, 1636–1638. Aronson, E., Turner, J. A., & Carlsmith, J. M. (1963). Communicator credibility and communication discrepancy as determinants of opinion change. Journal of Abnormal and Social Psychology, 67, 31–36. Aronson, J., Cohen, G., & Nail, P. R. (1999). Self-affirmation theory: An update and appraisal. In E. Harmon-Jones & J. Mills (Eds.), Cognitive dissonance: Progress on a pivotal theory in social psychology (pp. 127–147). Washington, DC: American Psychological Association. Arriaga, X. B., & Longoria, Z. N. (2011). Implementation intentions increase parent-teacher communication among Latinos. Basic and Applied Social Psychology, 33, 365–373. Ashford, S., Edmunds, J., & French, D. P. (2010). What is the best way to change self-efficacy to promote lifestyle and recreational physical activity? A systematic review with meta-analysis. British Journal of Health Psychology, 15, 265–288. Astrom, A. N., & Rise, J. (2001). Young adults’ intention to eat healthy food: Extending the theory of planned behavior. Psychology and Health, 16, 223–237. Atkins, A. L., Deaux, K. K., & Bieri, J. (1967). Latitude of acceptance and attitude change: Empirical evidence for a reformulation. Journal of Personality and Social Psychology, 6, 47–54. Atkinson, D. R., Winzelberg, A., & Holland, A. (1985). Ethnicity, locus of control for family planning, and pregnancy counselor credibility. Journal of Counseling Psychology, 32, 417–321. 421



Audi, R. (1972). On the conception and measurement of attitudes in contemporary Anglo-American psychology. Journal for the Theory of Social Behavior, 2, 179–203. Austin, E. W., de Vord, R. V., Pinkleton, B. E., & Epstein, E. (2008). Celebrity endorsements and their potential to motivate young voters. Mass Communication & Society, 11, 420–436. Austin, J., Alvero, A. M., & Olson, R. (1998). Prompting patron safety belt use at a restaurant. Journal of Applied Behavior Analysis, 31, 655–657. Austin, J., Sigurdsson, S. O., & Rubin, Y. S. (2006). An examination of the effects of delayed versus immediate prompts on safety belt use. Environment and Behavior, 38, 140–149. Averbeck, J. M., Jones, A., & Robertson, K. (2011). Prior knowledge and health messages: An examination of affect as heuristics and information as systematic processing for fear appeals. Southern Communication Journal, 76, 35–54. Axsom, D., Yates, S., & Chaiken, S. (1987). Audience response as a heuristic cue in persuasion. Journal of Personality and Social Psychology, 53, 30–40. Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008). Mixed-effects modeling with crossed random effects for subjects and items. Journal of Memory and Language, 59, 390–412. Babrow, A. S., & O’Keefe, D. J. (1984). Construct differentiation as a moderator of attitude-behavior consistency: A failure to confirm. Central States Speech Journal, 35, 160–165. Bagozzi, R. P. (1984). Expectancy-value attitude models: An analysis of critical measurement issues. International Journal of Research in Marketing, 1, 295–310. 422



Bagozzi, R. P. (1985). Expectancy-value attitude models: An analysis of critical theoretical issues. International Journal of Research in Marketing, 2, 43–60. Bagozzi, R. P., Baumgartner, H., & Yi, Y. (1992). State versus action orientation and the theory of reasoned action: An application to coupon usage. Journal of Consumer Research, 18, 505–518. Bagozzi, R. P., Lee, K. H., & Van Loo, M. F. (2001). Decisions to donate bone marrow: The role of attitudes and subjective norms across cultures. Psychology and Health, 16, 29–56. Bailis, D. S., Fleming, J. A., & Segall, A. (2005). Self-determination and functional persuasion to encourage physical activity. Psychology and Health, 20, 691–708. Bakker, A. B. (1999). Persuasive communication about AIDS prevention: Need for cognition determines the impact of message format. AIDS Education and Prevention, 11, 150–162. Bakker, M., van Dijk, A., & Wicherts, J. M. (2012). The rules of the game called psychological science. Perspectives on Psychological Science, 7, 543–554. Balmford, J., Borland, R., & Burney, S. (2008). Is contemplation a separate stage of change to precontemplation? International Journal of Behavioral Medicine, 15, 141–148. Bamberg, S. (2003). How does environmental concern influence specific environmentally related behaviors? A new answer to an old question. Journal of Environmental Psychology, 23, 21–32. Bamberg, S. (2007). Is a stage model a useful approach to explain car drivers’ willingness to use public transportation? Journal of Applied Social Psychology, 37, 1757–1783. 423



Banaji, M. R., & Heiphetz, L. (2010). Attitudes. In S. T. Fiske, D. T. Gilbert, & G. Lindzey (Eds.), Handbook of social psychology (5th ed., Vol. 1, pp. 353–393). Hoboken, NJ: Wiley. Banas, J. A., & Rains, S. A. (2010). A meta-analysis of research on inoculation theory. Communication Monographs, 77, 281–311. Banas, J. A., & Turner, M. M. (2011). Exploring the “that’s-not-all” effect: A test of theoretical explanations. Southern Communication Journal, 76, 305–322. Banas, J. A., Turner, M. M., & Shulman, H. (2012). A test of competing hypotheses of the effects of mood on persuasion. Communication Quarterly, 60, 143–164. Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman. Banerjee, S. C., & Greene, K. (2012). Role of transportation in the persuasion process: Cognitive and affective responses to antidrug narratives. Journal of Health Communication, 17, 564–581. Bansal, H. S., & Taylor, S. F. (2002). Investigating interactive effects in the theory of planned behavior in a service-provider switching context. Psychology and Marketing, 19, 407–425. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Baron, R. S., Baron, P. H., & Miller, N. (1973). The relation between distraction and persuasion. Psychological Bulletin, 80, 310–323.



424



Basil, D. Z., & Herr, P. M. (2006). Attitudinal balance and cause-related marketing: An empirical application of balance theory. Journal of Consumer Psychology, 16, 391–403. Basil, M., & Witte, K. (2012). Health risk message design using the extended parallel process model. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 41–58). Los Angeles: Sage. Batra, R., & Homer, P. M. (2004). The situational impact of brand image beliefs. Journal of Consumer Psychology, 14, 318–330. Baudhuin, E. S., & Davis, M. K. (1972). Scales for the measurement of ethos: Another attempt. Speech Monographs, 39, 296–301. Baumeister, R. F., Reis, H. T., & Delespaul, P. A. E. G. (1995). Subjective and experiential correlates of guilt in daily life. Personality and Social Psychology Bulletin, 21, 1256–1268. Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F. (1995). Personal narratives about guilt: Role in action control and interpersonal relationships. Basic and Applied Social Psychology, 178, 173–198. Bazzini, D. G., & Shaffer, D. R. (1995). Investigating the social-adjustive and value-expressive functions of well-grounded attitudes: Implications for change and for subsequent behavior. Motivation and Emotion, 19, 279–305. Beaman, A. L., Cole, C. M., Preston, M., Klentz, B., & Steblay, N. M. (1983). Fifteen years of foot-in-the-door research: A meta-analysis. Personality and Social Psychology Bulletin, 9, 181–196. Bearden, W. O., Shuptrine, F. K., & Teel, J. E. (1989). Self-monitoring and reactions to image appeals and claims about product quality. Advances in Consumer Research, 16, 703–710. 425



Beatty, M. J., & Behnke, R. R. (1980). Teacher credibility as a function of verbal content and paralinguistic cues. Communication Quarterly, 28(1), 55–59. Beatty, M. J., & Kruger, M. W. (1978). The effects of heckling on speaker credibility and attitude change. Communication Quarterly, 26(2), 46–50. Beauvois, J.-L., & Joule, R. V. (1999). A radical point of view on dissonance theory. In E. Harmon-Jones & J. Mills (Eds.), Cognitive dissonance: Progress on a pivotal theory in social psychology (pp. 43–70). Washington, DC: American Psychological Association. Becker, C. B., Smith, L. M., & Ciao, A. C. (2006). Peer-facilitated eating disorder prevention: A randomized effectiveness trial of cognitive dissonance and media advocacy. Journal of Counseling Psychology, 53, 550–555. Beisecker, T. D., & Parson, D. W. (1972). Introduction. In T. D. Beisecker & D. W. Parson (Eds.), The process of social influence (pp. 1–6). Englewood Cliffs, NJ: Prentice Hall. Belch, G. E., & Belch, M. A. (1987). The application of an expectancy value operationalization of function theory to examine attitudes of boycotters and nonboycotters of a consumer product. Advances in Consumer Research, 14, 232–236. Beltramini, R. F., & Sirsi, A. K. (1992). Physician information acquisition and believability. Journal of Health Care Marketing, 12(4), 52–59. Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 6, pp. 1–62). New York: Academic Press. Bennett, P. D., & Harrell, G. D. (1975). The role of confidence in understanding and predicting buyers’ attitudes and purchase intentions. 426



Journal of Consumer Research, 2, 110–117. Benoit, W. L. (1991). Two tests of the mechanism of inoculation theory. Southern Communication Journal, 56, 219–229. Bensley, L. S., & Wu, R. (1991). The role of psychological reactance in drinking following alcohol prevention messages. Journal of Applied Social Psychology, 21, 1111–1124. Bergin, A. E. (1962). The effect of dissonant persuasive communications upon changes in a self-referring attitude. Journal of Personality, 30, 423–438. Berkowitz, A. D. (2005). An overview of the social norms approach. In L. C. Lederman & L. P. Stewart (Eds.), Changing the culture of college drinking: A socially situated health communication campaign (pp. 193–214). Cresskill, NJ: Hampton Press. Berlo, D. K., Lemert, J. B., & Mertz, R. J. (1969). Dimensions for evaluating the acceptability of message sources. Public Opinion Quarterly, 33, 563–576. Bernard, M. M., Maio, G. R., & Olson, J. M. (2003). The vulnerability of values to attack: Inoculation of values and value-relevant attitudes. Personality and Social Psychology Bulletin, 29, 63–75. Berry, T. R., & Howe, B. L. (2005). The effects of exercise advertising on self-efficacy and decisional balance. American Journal of Health Behavior, 29, 117–126. Berscheid, E. (1985). Interpersonal attraction. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed., Vol. 2, pp. 413–484). New York: Random House.



427



Berscheid, E., & Walster, E. (1974). Physical attractiveness. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 7, pp. 157–215). New York: Academic Press. Betsch, T., Kaufmann, M., Lindow, F., Plessner, H., & Hoffmann, K. (2006). Different principles of information integration in implicit and explicit attitude formation. European Journal of Social Psychology, 36, 887–905. Biek, M., Wood, W., & Chaiken, S. (1996). Working knowledge, cognitive processing, and attitudes: On the determinants of bias. Personality and Social Psychology Bulletin, 22, 547–556. Biglan, A., Glasgow, R., Ary, D., Thompson, R., Severson, H., Lichtenstein, E., … Gallison, C. (1987). How generalizable are the effects of smoking prevention programs? Refusal skills training and parent messages in a teacher-administered program. Journal of Behavioral Medicine, 10, 613–628. Bilandzic, H., & Busselle, R. (2013). Narrative persuasion. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 200–219). Thousand Oaks, CA: Sage. Birkimer, J. C., Johnston, P. L., & Berry, M. M. (1993). Guilt and help from friends: Variables related to healthy behavior. Journal of Social Psychology, 133, 683–692. Biswas, D., Biswas, A., & Das, N. (2006). The differential effects of celebrity and expert endorsements on consumer risk perceptions: The role of consumer knowledge, perceived congruency, and product technology orientation. Journal of Advertising, 35(2), 17–31. Bither, S. W., Dolich, I. J., & Nell, E. B. (1971). The application of attitude immunization techniques in marketing. Journal of Marketing Research, 8, 56-61.



428



Blake, H., Lee, S., Stanton, T., & Gorely, T. (2008). Workplace intervention to promote stair-use in an NHS setting. International Journal of Workplace Health Management, 1, 162–175. Bless, H., Bohner, G., Schwarz, N., & Strack, F. (1990). Mood and persuasion: A cognitive response analysis. Personality and Social Psychology Bulletin, 16, 331–345. Bless, H., Mackie, D. M., & Schwarz, N. (1992). Mood effects on attitude judgments: Independent effects of mood before and after message elaboration. Journal of Personality and Social Psychology, 63, 585–595. Bless, H., & Schwarz, N. (1999). Sufficient and necessary conditions in dual-process models: The case of mood and information processing. In S. Chaiken & Y. Trope (Eds.), Dual-process models in social psychology (pp. 423–440). New York: Guilford. Blumberg, S. J. (2000). Guarding against threatening HIV prevention messages: An information-processing model. Health Education and Behavior, 27, 780–795. Bochner, S., & Insko, C. A. (1966). Communicator discrepancy, source credibility, and opinion change. Journal of Personality and Social Psychology, 4, 614–621. Bock, D. G., & Saine, T. J. (1975). The impact of source credibility, attitude valence, and task sensitivity on trait errors in speech evaluation. Speech Monographs, 42, 229–236. Bodenhausen, G. V., & Gawronski, B. (2013). Attitude change. In D. Reisberg (Ed.), The Oxford handbook of cognitive psychology. New York: Oxford University Press. doi:10.1093/oxfordhb/9780195376746.013.0060. Bodur, H. O., Brinberg, D., & Coupey, E. (2000). Belief, affect, and 429



attitude: Alternative models of the determinants of attitude. Journal of Consumer Psychology, 9, 17–28. Boen, F., Maurissen, K., & Opdenacker, J. (2010). A simple health sign increases stair use in a shopping mall and two train stations in Flanders, Belgium. Health Promotion International, 25, 183–191. Bohner, G., & Dickel, N. (2011). Attitudes and attitude change. Annual Review of Psychology, 62, 391–417. Bohner, G., Erb, H.-P., & Siebler, F. (2008). Information processing approaches to persuasion: Integrating assumptions for the dual- and single-processing perspectives. In W. D. Crano & R. Prislin (Eds.), Attitudes and attitude change (pp. 161–188). New York: Psychology Press. Bohner, G., Ruder, M., & Erb, H.-P. (2002). When expertise backfires: Contrast and assimilation effects in persuasion. British Journal of Social Psychology, 41, 495–519. Bohner, G., & Weinerth, T. (2001). Negative affect can increase or decrease message scrutiny: The affect interpretation hypothesis. Personality and Social Psychology Bulletin, 27, 1417–1428. Bolsen, T., Druckman, J. N., & Cook, F. L. (2014). How frames can undermine support for scientific adaptations: Politicization and the status-quo bias. Public Opinion Quarterly, 78, 1–26. Bond, R. M., Fariss, C. J., Jones, J. J., Kramer, A. D. I., Marlow, C., Settle, J. E., & Fowler, J. H. (2012). A 61-million-person experiment in social influence and political mobilization. Nature, 489, 295–298. Booth, A. R., Norman, P., Harris, P. R., & Goyder, E. (2014). Using the theory of planned behaviour and self-identity to explain chlamydia testing intentions in young people living in deprived areas. British 430



Journal of Health Psychology, 19, 101–112. Booth-Butterfield, S., & Reger, B. (2004). The message changes belief and the rest is theory: The “1% or less” milk campaign and reasoned action. Preventive Medicine, 39, 581–588. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis. Chichester, UK: Wiley. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2010). A basic introduction to fixed-effect and random-effects models for meta-analysis. Research Synthesis Methods, 1, 97–111. Borenstein, M., & Rothstein, H. (2005). Comprehensive meta-analysis (Version 2.2.023) [Computer software]. Englewood, NJ: Biostat. Borgida, E., & Campbell, B. (1982). Belief relevance and attitudebehavior consistency: The moderating role of personal experience. Journal of Personality and Social Psychology, 42, 239–247. Boster, F. J., & Mongeau, P. (1984). Fear-arousing persuasive messages. Communication Yearbook, 8, 330–375. Botta, R. A., Dunker, K., Fenson-Hood, K., Maltarich, S., & McDonald, L. (2008). Using a relevant threat, EPPM and interpersonal communication to change hand-washing behaviours on campus. Journal of Communication in Healthcare, 1, 373–381. Bouman, M. (2004). Entertainment-education television drama in the Netherlands. In A. Singhal, M. J. Cody, E. M. Rogers, & M. Sabido (Eds.), Entertainment-education and social change: History, research, and practice (pp. 225–242). Mahwah, NJ: Lawrence Erlbaum. Bowers, J. W., & Phillips, W. A. (1967). A note on the generality of 431



source-credibility scales Speech Monographs, 34, 185–186. Bradac, J. J. (1986). Threats to generalization in the use of elicited, purloined, and contrived messages in human communication research. Communication Quarterly, 34, 55–65. Bradac, J. J., Bowers, J. W., & Courtright, J. A. (1980). Lexical variations in intensity, immediacy, and diversity: An axiomatic theory and causal model. In R. N. St. Clair & H. Giles (Eds.), The social and psychological contexts of language (pp. 193–223). Hillsdale, NJ: Lawrence Erlbaum. Bradley, P. H. (1981). The folk-linguistics of women’s speech: An empirical examination. Communication Monographs, 48, 73–90. Brannon, L. A., & Brock, T. C. (1994). Test of schema correspondence theory of persuasion: Effects of matching an appeal to actual, ideal, and product “selves.” In E. M. Clark, T. C. Brock, & D. W. Stewart (Eds.), Attention, attitude, and affect in response to advertising (pp. 169–188). Hillsdale, NJ: Lawrence Erlbaum. Brannon, L. A., & Brock, T. C. (2001). Scarcity claims elicit extreme responding to persuasive messages: Role of cognitive elaboration. Personality and Social Psychology Bulletin, 27, 365–375. [Erratum notice: Personality and Social Psychology Bulletin, 27 (2001), 639.] Branstrom, R., Ullen, H., & Brandberg, Y. (2004). Attitudes, subjective norms and perception of behavioural control as predictors of sun-related behaviour in Swedish adults. Preventive Medicine, 39, 992–999. Brashers, D. E., & Jackson, S. (1999). Changing conceptions of message effects: A 24-year overview. Human Communication Research, 25, 457–477. Braverman, J. (2008). Testimonials versus informative persuasive 432



messages: The moderating effect of delivery mode and personal involvement. Communication Research, 35, 666–694. Brehm, J. W. (1956). Postdecision changes in the desirability of alternatives. Journal of Abnormal and Social Psychology, 52, 384–389. Brehm, J. W. (1966). A theory of psychological reactance. New York: Academic Press. Brehm, J. W. (2007). A brief history of dissonance theory. Social and Personality Psychology Compass, 1, 381–391. Brehm, S. S., & Brehm, J. W. (1981). Psychological reactance: A theory of freedom and control. New York: Academic Press. Breivik, E., & Supphellen, M. (2003). Elicitation of product attributes in an evaluation context: A comparison of three elicitation techniques. Journal of Economic Psychology, 24, 77–98. Brickell, T. A., Chatzisarantis, N. L. D., & Pretty, G. M. (2006). Autonomy and control: Augmenting the validity of the theory of planned behaviour in predicting exercise. Journal of Health Psychology, 11, 51–63. Bridle, C., Riemsma, R. P., Pattenden, J., Sowden, A. J., Mather, L., Watt, I. S., & Walker, A. (2005). Systematic review of the effectiveness of health behavior interventions based on the transtheoretical model. Psychology and Health, 20, 283–301. Brinberg, D., & Durand, J. (1983). Eating at fast-food restaurants: An analysis using two behavioral intention models. Journal of Applied Social Psychology, 13, 459–472. Briñol, P., & Petty, R. E. (2005). Individual differences in attitude change. 433



In D. Albarracín, B. T. Johnson, & M. P. Zanna (Eds.), Handbook of attitudes (pp. 575–615). Mahwah, NJ: Lawrence Erlbaum. Briñol, P., & Petty, R. E. (2009a). Persuasion: Insights from the selfvalidation hypothesis. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 41, pp. 69–118). New York: Academic Press. Briñol, P., & Petty, R. E. (2009b). Source factors in persuasion: A selfvalidation approach. European Review of Social Psychology, 20, 49–96. Briñol, P., Petty, R. E., & McCaslin, M. J. (2009). Changing attitudes on implicit versus explicit measures: What is the difference? In R. E. Petty, R. H. Fazio, & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 285–326). New York: Psychology Press. Briñol, P., Rucker, D. D., Tormala, Z. L., & Petty, R. E. (2004). Individual differences in resistance to persuasion: The role of beliefs and metabeliefs. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 83–104). Mahwah, NJ: Lawrence Erlbaum. Brock, T. C. (1965). Communicator–recipient similarity and decision change. Journal of Personality and Social Psychology, 1, 650–654. Brock, T. C. (1968). Implications of commodity theory for value change. In A. G. Greenwald, T. C. Brock, & T. M. Ostrom (Eds.), Psychological foundations of attitudes (pp. 243–275). New York: Academic Press. Brodsky, S. L., Griffin, M. P., & Cramer, R. J. (2010). The witness credibility scale: An outcome measure for expert witness research. Behavioral Sciences & the Law, 28, 892–907. Broemer, P. (2002). Relative effectiveness of differently framed health messages: The influence of ambivalence. European Journal of Social Psychology, 32, 685–703.



434



Brouwers, M. C., & Sorrentino, R. M. (1993). Uncertainty orientation and protection motivation theory: The role of individual differences in health compliance. Journal of Personality and Social Psychology, 65, 102–112. Brown, S., Birch, D., Thyagaraj, S., Teufel, J., & Phillips, C. (2007). Effects of a single-lesson tobacco prevention curriculum on knowledge, skill identification and smoking intention. Journal of Drug Education, 37, 55–69. Brown, S. P., Cron, W. L., & Slocum, J. W., Jr. (1997). Effects of goaldirected emotions on salesperson volitions, behavior, and performance: A longitudinal study. Journal of Marketing, 61 (1), 39–50. Brown, S. P., & Stayman, D. M. (1992). Antecedents and consequences of attitude toward the ad: A meta-analysis. Journal of Consumer Research, 19, 34–51. Brown, T. J., Ham, S. H., & Hughes, M. (2010). Picking up litter: An application of theory-based communication to influence tourist behaviour in protected areas. Journal of Sustainable Tourism, 18, 879–900. Browne, B. A., & Kaldenberg, D. O. (1997). Self–monitoring and image appeals in advertising. Psychological Reports, 81, 1267–1275. Brownstein, A. L. (2003). Biased predecision processing. Psychological Bulletin, 129, 545–568. Bryant, J., Brown, D., Silberberg, A. R., & Elliott, S. M. (1981). Effects of humorous illustrations in college textbooks. Human Communication Research, 8, 43–57. Budd, R. J. (1986). Predicting cigarette use: The need to incorporate measures of salience in the “theory of reasoned action.” Journal of 435



Applied Social Psychology, 16, 663–685. Budd, R. J., North, D., & Spencer, C. (1984). Understanding seat-belt use: A test of Bentler and Speckart’s extension of the theory of reasoned action. European Journal of Social Psychology, 14, 69–78. Budd, R. J., & Spencer, C. (1984). Latitude of rejection, centrality, and certainty: Variables affecting the relationship between attitudes, norms, and behavioural intentions. British Journal of Social Psychology, 23, 1–8. Buller, D. B., & Hall, J. R. (1998). The effects of distraction during persuasion. In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 155–173). Cresskill, NJ: Hampton. Burack, R. C., & Gimotty, P. A. (1997). Promoting screening mammography in inner-city settings: The sustained effectiveness of computerized reminders in a randomized controlled trial. Medical Care, 35, 921–932. Burger, J. M. (1999). The foot-in-the-door compliance procedure: A multiple-process analysis and review. Personality and Social Psychology Review, 3, 303–325. Burger, J. M., Bell, H., Harvey, K., Johnson, J., Stewart, C., Dorian, K., & Swedroe, M. (2010). Nutritious or delicious? The effect of descriptive norm information on food choice. Journal of Social and Clinical Psychology, 29, 228–242. Burger, J. M., & Caldwell, D. F. (2003). The effects of monetary incentives and labeling on the foot-in-the-door effect: Evidence for a self-perception process. Basic and Applied Social Psychology, 25, 235–241. Burger, J. M., & Guadagno, R. E. (2003). Self-concept clarity and the foot436



in-the-door procedure. Basic and Applied Social Psychology, 25, 79–86. Burger, J. M., LaSalvia, C. T., Hendricks, L. A., Mehdipour, T., & Neudeck, E. M. (2011). Partying before the party gets started: The effects of descriptive norms on pregaming behavior. Basic and Applied Social Psychology, 33, 220–227. Burger, J. M., Messian, N., Patel, S., del Prado, A., & Anderson, C. (2004). What a coincidence! The effects of incidental similarity on compliance. Personality and Social Psychology Bulletin, 30, 35–43. Burger, J. M., Reed, M., DeCesare, K., Rauner, S., & Rozolis, J. (1999). The effects of initial request size on compliance: More about the that’snot-all technique. Basic and Applied Social Psychology, 21, 243–249. Burger, J. M., & Shelton, M. (2011). Changing everyday health behaviors through descriptive norm manipulations. Social Influence, 6, 69–77. Burgoon, J. K. (1976). The ideal source: A reexamination of source credibility measurement. Central States Speech Journal, 27, 200–206. Burgoon, M., Alvaro, E. M., Broneck, K., Miller, C., Grandpre, J. R., Hall, J. R., & Frank, C. A. (2002). Using interactive media tools to test substance abuse prevention messages. In W. D. Crano & M. Burgoon (Eds.), Mass media and drug prevention: Classic and contemporary theories and research (pp. 67–87). Mahwah, NJ: Lawrence Erlbaum. Burgoon, M., Hall, J., & Pfau, M. (1991). A test of the “messages-asfixed-effect fallacy” argument: Empirical and theoretical implications of design choices. Communication Quarterly, 39, 18–34. Burke, C. J. (1953). A brief note on one-tailed tests. Psychological Bulletin, 50, 384–387.



437



Burkley, E. (2008). The role of self-control in resistance to persuasion. Personality and Social Psychology Bulletin, 34, 419–432. Burnell, P., & Reeve, A. (1984). Persuasion as a political concept. British Journal of Political Science, 14, 393–410. Burnkrant, R. E., & Howard, D. J. (1984). Effects of the use of introductory rhetorical questions versus statements on information processing. Journal of Personality and Social Psychology, 47, 1218–1230. Busselle, R., & Bilandzic, H. (2009). Measuring narrative engagement. Media Psychology, 12, 321–347. Buunk, A. P., & Dijkstra, P. (2011). Does attractiveness sell? Women’s attitude toward a product as a function of model attractiveness, gender priming, and social comparison orientation. Psychology and Marketing, 28, 958–973. Byrne, D. (1969). Attitudes and attraction. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 4, pp. 35–89). New York: Academic Press. Byrne, S., Guillory, J. E., Mathios, A. D., Avery, R. J., & Hart, P. S. (2012). The unintended consequences of disclosure: Effect of manipulating sponsor identification on the perceived credibility and effectiveness of smoking cessation advertisements. Journal of Health Communication, 17, 1119–1137. Cacioppo, J. T., Gardner, W. L., & Berntson, G. G. (1997). Beyond bipolar conceptualizations and measures: The case of attitudes and evaluative space. Personality and Social Psychology Review, 1, 3–25. Cacioppo, J. T., Harkins, S. G., & Petty, R. E. (1981). The nature of attitudes and cognitive responses and their relationships to behavior. In 438



R. E. Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 31–54). Hillsdale, NJ: Lawrence Erlbaum. Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42, 116–131. Cacioppo, J. T., & Petty, R. E. (1984). The elaboration likelihood model of persuasion. Advances in Consumer Research, 11, 673–675. Cacioppo, J. T., Petty, R. E., Feinstein, J. A., & Jarvis, W. B. G. (1996). Dispositional differences in cognitive motivation: The life and times of individuals varying in need for cognition. Psychological Bulletin, 119, 197–253. Cacioppo, J. T., Petty, R. E., Kao, C. F., & Rodriguez, R. (1986). Central and peripheral routes to persuasion: An individual difference perspective. Journal of Personality and Social Psychology, 51, 1032–1043. Cacioppo, J. T., Petty, R. E., & Stoltenberg, C. D. (1985). Processes of social influence: The elaboration likelihood model of persuasion. In P. C. Kendall (Ed.), Advances in cognitive-behavioral research and therapy (Vol. 4, pp. 215–274). New York: Academic Press. Cacioppo, J. T., von Hippel, W., & Ernst, J. M. (1997). Mapping cognitive structures and processes through verbal content: The thought-listing technique. Journal of Consulting and Clinical Psychology, 65, 928–940. Cafri, G., Kromrey, J. D., & Brannick, M. T. (2010). A meta-metaanalysis: Empirical review of statistical power, type I error rates, effect sizes, and model selection of meta-analyses published in psychology. Multivariate Behavioral Research, 45, 239–270. Cahill, K., Lancaster, T., & Green, N. (2010). Stage-based interventions for smoking cessation. Cochrane Database of Systematic Reviews, 439



2010(11), CD004492. Callaghan, R. C., & Taylor, L. (2007). Is stage-matched process-use associated with the transition from the preparation to action stage among smokers? Longitudinal test for the transtheoretical model. Addition Research and Theory, 15, 493–501. Calsyn, D. A., Hatch-Maillette, M. A., Doyle, S. R., Cousins, S., Chen, T., & Godinez, M. (2010). Teaching condom use skills: Practice is superior to observation. Substance Abuse, 31, 231–239. Cameron, K. A., & Campo, S. (2006). Stepping back from social norms campaigns: Comparing normative influences to other predictors of health behaviors. Health Communication, 20, 277–288. Campbell, G. (1776). The philosophy of rhetoric. Retrieved from http://people.cohums.ohio-state.edu/Ulman1/Campbell/. Campbell, M. C., & Kirmani, A. (2000). Consumers’ use of persuasion knowledge: The effects of accessibility and cognitive capacity on perceptions of an influence agent. Journal of Consumer Research, 27, 69–83. Campo, S., Askelson, N. M., Carter, K. D., & Losch, M. (2012). Segmenting audiences and tailoring messages using the extended parallel process model and cluster analysis to improve health campaigns. Social Marketing Quarterly, 18, 98–111. Campo, S., & Cameron, K. A. (2006). Differential effects of exposure to social norms campaigns: A cause for concern. Health Communication, 19, 209–219. Cann, A., Sherman, S. J., & Elkes, R. (1975). Effects of initial request size and timing of a second request on compliance: The foot in the door and the door in the face. Journal of Personality and Social Psychology, 32, 440



774–782. Cappella, J. N. (2006). Integrating message effects and behavior change theories: Organizing comments and unanswered questions. Journal of Communication, 56, S265–S279. Cappella, J. N. (2007). The role of discrete emotions in the theory of reasoned action and its successors: Quitting smoking in young adults. In I. Ajzen, D. Albarracín, & R. Hornik (Eds.), Prediction and change of health behavior: Applying the reasoned action approach (pp. 43–51). Mahwah, NJ: Lawrence Erlbaum. Cappella, J. N., Yzer, M., & Fishbein, M. (2003). Using beliefs about positive and negative consequences as the basis for designing message interventions for lowering risky behavior. In D. Romer (Ed.), Reducing adolescent risk (pp. 210–219). Thousand Oaks, CA: Sage. Carcioppolo, N., Jensen, J. D., Wilson, S. R., Collins, W. B., Carrion, M., & Linnemeier, G. (2013). Examining HPV threat-to-efficacy ratios in the extended parallel process model. Health Communication, 28, 20–28. Card, N. A. (2012). Applied meta-analysis for social science research. New York: Guilford. Cardenas, M. P., & Simons-Morton, B. G. (1993). The effect of anticipatory guidance on mothers’ self-efficacy and behavioral intentions to prevent burns caused by hot tap water. Patient Education and Counseling, 21, 117–123. Carey, R. N., McDermott, D. T., & Sarma, K. M. (2013). The impact of threat appeals on fear arousal and driver behavior: A meta-analysis of experimental research 1990–2011. PLOS ONE, 8(5), e62821. Carpenter, C. J. (2012a). A meta-analysis and an experiment investigating the effects of speaker disfluency on persuasion. Western Journal of 441



Communication, 76, 552–569. Carpenter, C. J. (2012b). A meta-analysis of the functional matching effect based on functional attitude theory. Southern Communication Journal, 77, 438–451. Carpenter, C. J. (2013). A meta-analysis of the effectiveness of the “but you are free” compliance-gaining technique. Communication Studies, 64, 6–17. Carpenter, C., Boster, F. J., & Andrews, K. R. (2013). Functional attitude theory. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 104–119). Los Angeles: Sage. Carpenter, J. M., & Green, M. C. (2012). Flying with Icarus: Narrative transportation and the persuasiveness of entertainment. In L. J. Shrum (Ed.), The psychology of entertainment media: Blurring the lines between entertainment and persuasion (2nd ed., pp. 169–194). New York: Routledge. Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319–333. Celuch, K., & Slama, M. (1995). “Getting along” and “getting ahead” as motives for self-presentation: Their impact on advertising effectiveness. Journal of Applied Social Psychology, 25, 1700–1713. Cerully, J. L., & Klein, W. M. P. (2010). Effects of emotional state on behavioral responsiveness to personal risk feedback. Journal of Risk Research, 13, 591–598. Cesario, J., Corker, K. S., & Jelinek, S. (2013). A self-regulatory 442



framework for message framing. Journal of Experimental Social Psychology, 49, 238–249. Cesario, J., Grant, H., & Higgins, E. T. (2004). Regulatory fit and persuasion: Transfer from “feeling right.” Journal of Personality and Social Psychology, 86, 388–404. Chaiken, S. (1979). Communicator physical attractiveness and persuasion. Journal of Personality and Social Psychology, 37, 1387–1397. Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39, 752–766. Chaiken, S. (1986). Physical appearance and social influence. In C. P. Herman, M. P. Zanna, & E. T. Higgins (Eds.), Physical appearance, stigma, and social behavior: The Ontario Symposium, vol. 3 (pp. 143–177). Hillsdale, NJ: Lawrence Erlbaum. Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna, J. M. Olson, & C. P. Herman (Eds.), Social influence: The Ontario Symposium, vol. 5 (pp. 3–39). Hillsdale, NJ: Lawrence Erlbaum. Chaiken, S., Duckworth, K. L., & Darke, P. (1999). When parsimony fails … Psychological Inquiry, 10, 118–123. Chaiken, S., & Eagly, A. H. (1983). Communication modality as a determinant of persuasion: The role of communicator salience. Journal of Personality and Social Psychology, 45, 241–256. Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic information processing within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 212–252). New York: Guilford.



443



Chaiken, S., & Stangor, C. (1987). Attitudes and attitude change. Annual Review of Psychology, 38, 575–630. Chaiken, S., & Trope, Y. (Eds.). (1999). Dual-process theories in social psychology. New York: Guilford. Chaiken, S., Wood, W., & Eagly, A. H. (1996). Principles of persuasion. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 702–742). New York: Guilford. Chandran, S., & Menon, G. (2004). When a day means more than a year: Effects of temporal framing on judgments of health risk. Journal of Consumer Research, 31, 375–389. Chang, C. (2004). Country of origin as a heuristic cue: The effects of message ambiguity and product involvement. Media Psychology, 6, 169–192. Chang, C. (2006). Cultural masculinity/femininity influences on advertising appeals. Journal of Advertising Research, 46, 315–323. Chang, C. (2010). Message framing and interpersonal orientation at cultural and individual levels: Involvement as a moderator. International Journal of Advertising, 29, 765–794. Chang, C.-T. (2007). Interactive effects of message framing, product perceived risk, and mood: The case of travel healthcare product advertising. Journal of Advertising Research, 47, 51–65. Chang, M.-J., & Gruner, C. R. (1981). Audience reaction to self– disparaging humor. Southern Speech Communication Journal, 46(4), 19–26. Chapman, J., Armitage, C. J., & Norman, P. (2009). Comparing 444



implementation intention interventions in relation to young adults’ intake of fruit and vegetables. Psychology and Health, 24, 317–332. Chartrand, T., Pinckert, S., & Burger, J. M. (1999). When manipulation backfires: The effects of time delay and requester on the foot-in-thedoor technique. Journal of Applied Social Psychology, 29, 211–221. Chatzisarantis, N., & Hagger, M. (2005). Effects of a brief intervention based on the theory of planned behavior on leisure-time physical activity participation. Journal of Sport & Exercise Psychology, 27, 470–487. Chatzisarantis, N. L. D., & Hagger, M. S. (2007). Mindfulness and the intention-behavior relationship within the theory of planned behavior. Personality and Social Psychology Bulletin, 33, 663–676. Chebat, J.-C., Filiatrault, P., Laroche, M., & Watson, C. (1988). Compensatory effects of cognitive characteristics of the source, the message, and the receiver upon attitude change. Journal of Psychology, 122, 609–621. Chebat, J.-C., Laroche, M., Baddoura, D., & Filiatrault, P. (1992). Effects of source likability on attitude change through message repetition. Advances in Consumer Research, 19, 353–358. Chen, H. C., Reardon, R., Rea, C., & Moore, D. J. (1992). Forewarning of content and involvement: Consequences for persuasion and resistance to persuasion. Journal of Experimental Social Psychology, 28, 523–541. Chen, M. F., & Tung, P. T. (2010). The moderating effect of perceived lack of facilities on consumers’ recycling intentions. Environment and Behavior, 42, 824–844. Chen, M. K., & Risen, J. L. (2009). Is choice a reliable predictor of choice? A comment on Sagarin and Skowronski. Journal of 445



Experimental Social Psychology, 45, 425–427. Chen, M. K., & Risen, J. L. (2010). How choice affects and reflects preferences: Revisiting the free-choice paradigm. Journal of Personality and Social Psychology, 99, 573–594. Chen, S., & Chaiken, S. (1999). The heuristic-systematic model in its broader context. In S. Chaiken & Y. Trope (Eds.), Dual-process theories in social psychology (pp. 73–96). New York: Guilford. Cheung, C. M.-Y., Sia, C.-L., & Kuan, K. K. Y. (2012). Is this review believable? A study of factors affecting the credibility of online consumer reviews from an ELM perspective. Journal of the Association for Information Systems, 13, 618–635. Cheung, S. F., Chan, D. K.-S., & Wong, Z. S.-Y. (1999). Reexamining the theory of planned behavior in understanding wastepaper recycling. Environment and Behavior, 31, 587–612. Chien, Y. H. (2011). Use of message framing and color in vaccine information to increase willingness to be vaccinated. Social Behavior and Personality, 39, 1063–1071. Cho, H., & Witte, K. (2004). A review of fear-appeal effects. In J. S. Seiter & R. H. Gass (Eds.), Perspectives on persuasion, social influence, and compliance gaining (pp. 223–238). Boston: Pearson Allyn and Bacon. Cho, H., & Witte, K. (2005). Managing fear in public health campaigns: A theory-based formative evaluation process. Health Promotion Practice, 6, 482–490. Chong, D., & Druckman, J. N. (2007). Framing theory. Annual Review of Political Science, 10, 103–126.



446



Chu, G. C. (1966). Fear arousal, efficacy, and imminency. Journal of Personality and Social Psychology, 4, 517–524. Chu, G. C. (1967). Prior familiarity, perceived bias, and one-sided versus two-sided communications. Journal of Experimental Social Psychology, 3, 243–254. Chung, S., Fink, E. L., & Kaplowitz, S. A. (2008). The comparative statics and dynamics of beliefs: The effect of message discrepancy and source credibility. Communication Monographs, 75, 158–189. Chung, S., Fink, E. L., Waks, L., Meffert, M. F., & Xie, X. (2012). Sequential information integration and belief trajectories: An experimental study using candidate evaluations. Communication Monographs, 79, 160–180. Churchill, S., & Jessop, D. C. (2011). Too impulsive for implementation intentions? Evidence that impulsivity moderates the effectiveness of an implementation intention intervention. Psychology & Health, 26, 517–530. Cialdini, R. B. (1984). Influence: How and why people agree to things. New York: William Morrow. Cialdini, R. B. (1987). Compliance principles of compliance professionals: Psychologists of necessity. In M. P. Zanna, J. M. Olson, & C. P. Herman (Eds.), Social influence: The Ontario Symposium, vol. 5 (pp. 165–184). Hillsdale, NJ: Lawrence Erlbaum. Cialdini, R. B. (2009). Influence: Science and practice (5th ed.). Boston: Pearson. Cialdini, R. B., Cacioppo, J. T., Bassett, R., & Miller, J. A. (1978). Lowball procedure for producing compliance: Commitment then cost. Journal of Personality and Social Psychology, 36, 463–476. 447



Cialdini, R. B., Demaine, L. J., Sagarin, B. J., Barrett, D. W., Rhoads, K., & Winter, P. L. (2006). Managing social norms for persuasive impact. Social Influence, 1, 3–15. Cialdini, R. B., & Goldstein, N. J. (2004). Social influence: Compliance and conformity. Annual Review of Psychology, 55, 591–621. Cialdini, R. B., & Griskevicius, V. (2010). Social influence. In R. F. Baumeister & E. J. Finkel (Eds.), Advanced social psychology: The state of the science (pp. 385–417). Oxford, UK: Oxford University Press. Cialdini, R. B., Griskevicius, V., Sundie, J. M., & Kenrick, D. T. (2007). Persuasion paralysis: When unrelated motives immobilize influence. Social Influence, 2, 4–17. Cialdini, R. B., & Schroeder, D. A. (1976). Increasing compliance by legitimizing paltry contributions: When even a penny helps. Journal of Personality and Social Psychology, 34, 599–604. Cialdini, R. B., & Trost, M. R. (1998). Social influence: Social norms, conformity, and compliance. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 2, pp. 151–192). Boston: McGraw-Hill. Cialdini, R. B., Vincent, J. E., Lewis, S. K., Catalan, J., Wheeler, D., & Darby, B. L. (1975). Reciprocal concessions procedure for inducing compliance: The door-in-the-face technique. Journal of Personality and Social Psychology, 31, 206–215. Clack, Z. A., Pitts, S. R., & Kellermann, A. L. (2000). Do reminder signs promote use of safety belts? Annals of Emergency Medicine, 36, 597–601. Clapp, J. D., Lange, J. E., Russell, C., Shillington, A., & Voas, R. B. 448



(2003). A failed norms social marketing campaign. Journal of Studies on Alcohol, 64, 409–414. Clark, H. H. (1973). The language-as-fixed-effect fallacy: A critique of language statistics in psychological research. Journal of Verbal Learning and Verbal Behavior, 12, 335–359. Clark, J. K., & Wegener, D. T. (2013). Message position, information processing, and persuasion: The discrepancy motives model. Advances in Experimental Social Psychology, 47, 189–232. Clark, J. K., Wegener, D. T., & Evans, A. T. (2011). Perceptions of source efficacy and persuasion: Multiple mechanisms for source effects on attitudes. European Journal of Social Psychology, 41, 596–607. Clark, J. K., Wegener, D. T., Habashi, M. M., & Evans, A. T. (2012). Source expertise and persuasion: The effects of perceived opposition or support on message scrutiny. Personality and Social Psychology Bulletin, 38, 90–100. Clark, R. A., & Stewart, R. (1971). Latitude of rejection as a measure of ego involvement. Speech Monographs, 38, 228–234. Clark, R. A., Stewart, R., & Marston, A. (1972). Scale values for highest and lowest levels of credibility. Central States Speech Journal, 23, 193–196. Clarkson, J. J., Tormala, Z. L., & Rucker, D. D. (2011). Cognitive and affective matching effects in persuasion: An amplification perspective. Personality and Social Psychology Bulletin, 37, 1415–1427. Clary, E. G., Snyder, M., Ridge, R. D., Copeland, J., Stukas, A. A., Haugen, J., & Miene, P. (1998). Understanding and assessing the motivations of volunteers: A functional approach. Journal of Personality and Social Psychology, 74, 1516–1530. 449



Clary, E. G., Snyder, M., Ridge, R. D., Miene, P. K., & Haugen, J. A. (1994). Matching messages to motives in persuasion: A functional approach to promoting volunteerism. Journal of Applied Social Psychology, 24, 1129–1149. Claypool, H. M., Mackie, D. M., Garcia-Marques, T., McIntosh, A., & Udall, A. (2004). The effects of personal relevance and repetition on persuasive processing. Social Cognition, 22, 310–335. Cohen, G. L., Aronson, J., & Steele, C. M. (2000). When beliefs yield to evidence: Reducing biased evaluation by affirming the self. Personality and Social Psychology Bulletin, 26, 1151–1164. Combs, D. J. Y., & Keller, P. S. (2010). Politicians and trustworthiness: Acting contrary to self-interest enhances trustworthiness. Basic and Applied Social Psychology, 32, 328–339. Compton, J. (2013). Inoculation theory. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 220–236). Thousand Oaks, CA: Sage. Compton, J. A., & Pfau, M. W. (2005). Inoculation theory of resistance to influence at maturity: Recent progress in theory development and application and suggestions for future research. Communication Yearbook, 29, 97–145. Conner, M. (2008). Initiation and maintenance of health behaviors. Applied Psychology: An International Review, 57, 42–50. Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior: A review and avenues for further research. Journal of Applied Social Psychology, 28, 1429–1464. Conner, M., & Armitage, C. J. (2008). Attitudinal ambivalence. In W. D. Crano & R. Prislin (Eds.), Attitudes and attitude change (pp. 261–286). 450



New York: Psychology Press. Conner, M., & Godin, G. (2007). Temporal stability of behavioural intention as a moderator of intention-health behaviour relationships. Psychology and Health, 22, 875–897. Conner, M., Godin, G., Norman, P., & Sheeran, P. (2011). Using the question-behavior effect to promote disease prevention behaviors: Two randomized controlled trials. Health Psychology, 30, 300–309. Conner, M., Graham, S., & Moore, B. (1999). Alcohol and intentions to use condoms: Applying the theory of planned behaviour. Psychology and Health, 14, 795–812. Conner, M., & Higgins, A. R. (2010). Long-term effects of implementation intentions on prevention of smoking uptake among adolescents: A cluster randomized controlled trial. Health Psychology, 29, 529–538. Conner, M., & McMillan, B. (1999). Interaction effects in the theory of planned behaviour: Studying cannabis use. British Journal of Social Psychology, 38, 195–222. Conner, M., Rhodes, R. E., Morris, B., McEachan, R., & Lawton, R. (2011). Changing exercise through targeting affective or cognitive attitudes. Psychology & Health, 26, 133–149. Conner, M., Sheeran, P., Norman, P., & Armitage, C. J. (2000). Temporal stability as a moderator of relationships in the theory of planned behaviour. British Journal of Social Psychology, 39, 469–493. Conner, M., Smith, N., & McMillan, B. (2003). Examining normative pressure in the theory of planned behaviour: Impact of gender and passengers on intentions to break the speed limit. Current Psychology, 22, 252–263. 451



Conner, M., & Sparks, P. (1996). The theory of planned behaviour and health behaviours. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and practice with social cognition models (pp. 121–162). Buckingham, UK: Open University Press. Conner, M., & Sparks, P. (2005). Theory of planned behaviour and health behaviour. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and practice with social cognition models (2nd ed., pp. 170–222). Maidenhead, UK: Open University Press. Conner, M., Sparks, P., Povey, R., James, R., Shepherd, R., & Armitage, C. J. (2002). Moderator effects of attitudinal ambivalence on attitudebehaviour relationships. European Journal of Social Psychology, 32, 705–718. Converse, J., Jr., & Cooper, J. (1979). The importance of decisions and free-choice attitude change: A curvilinear finding. Journal of Experimental Social Psychology, 15, 48–61. Cook, A. J., Kerr, G. N., & Moore, K. (2002). Attitudes and intentions towards purchasing GM food. Journal of Economic Psychology, 23, 557–572. Cooke, R., & French, D. P. (2008). How well do the theory of reasoned action and theory of planned behaviour predict intentions and attendance at screening programmes? A meta-analysis. Psychology & Health, 23, 745–765. Cooke, R., & Sheeran, P. (2004). Moderation of cognition-intention and cognition-behaviour relations: A meta-analysis of properties of variables from the theory of planned behaviour. British Journal of Social Psychology, 43, 159–186. Cooper, H., Hedges, L. V., & Valentine, J. C. (Eds.). (2009). The handbook of research synthesis and meta-analysis (2nd ed.). New York: Russell Sage Foundation. 452



Cooper, J. (1998). Unlearning cognitive dissonance: Toward an understanding of the development of dissonance. Journal of Experimental Social Psychology, 34, 562–575. Cooper, J. (2007). Cognitive dissonance: Fifty years of a classic theory. Los Angeles: Sage. Cooper, J., Darley, J. M., & Henderson, J. E. (1974). On the effectiveness of deviant- and conventional-appearing communicators: A field experiment. Journal of Personality and Social Psychology, 29, 752–757. Cooper, J., & Jones, R. A. (1970). Self-esteem and consistency as determinants of anticipatory opinion change. Journal of Personality and Social Psychology, 14, 312–320. Corbin, S. K. T., Jones, R. T., & Schulman, R. S. (1993). Drug refusal behavior: The relative efficacy of skills-based and information-based treatment. Journal of Pediatric Psychology, 18, 769–784. Corby, N. H., Enguidanos, S. M., & Kay, L. S. (1996). Development and use of role model stories in a community level HIV risk reduction intervention. Public Health Reports, 111 (Suppl. 1), 54–58. Côté, S. (2005). Reconciling the feelings-as-information and hedonic contingency models of how mood influences systematic information processing. Journal of Applied Social Psychology, 35, 1656–1679. Cotte, J., Coulter, R. A., & Moore, M. (2005). Enhancing or disrupting guilt: The role of ad credibility and perceived manipulative intent. Journal of Business Research, 58, 361–368. Cotton, J. L. (1985). Cognitive dissonance in selective exposure. In D. Zillmann & J. Bryant (Eds.), Selective exposure to communication (pp. 11–33). Hillsdale, NJ: Lawrence Erlbaum.



453



Courneya, K. S. (1994). Predicting repeated behavior from intention: The issue of scale correspondence. Journal of Applied Social Psychology, 24, 580–594. Courneya, K. S. (1995). Understanding readiness for regular physical activity in older individuals: An application of the theory of planned behavior. Health Psychology, 14, 80–87. Covey, J. (2014). The role of dispositional factors in moderating message framing effects. Health Psychology, 33, 52–65. Cox, B. S., Cox, A. B., & Cox, D. J. (2000). Motivating signage prompts safety belt use among drivers exiting senior communities. Journal of Applied Behavior Analysis, 33, 635–638. Craciun, C., Schüz, N., Lippke, S., & Schwarzer, R. (2012). A mediator model of sunscreen use: A longitudinal analysis of social-cognitive predictors and mediators. International Journal of Behavioral Medicine, 19, 65–72. Craig, T. Y., & Blankenship, K. L. (2011). Language and persuasion: Linguistic extremity influences message processing and behavioral intentions. Journal of Language and Social Psychology, 30, 290–310. Crandall, C. S., Glor, J., & Britt, T. W. (1997). AIDS-related stigmatization: Instrumental and symbolic attitudes. Journal of Applied Social Psychology, 27, 95–123. Crano, W. D., & Prislin, R. (1995). Components of vested interest and attitude-behavior consistency. Basic and Applied Social Psychology, 17, 1–21. Crawley, F. E., III (1990). Intentions of science teachers to use investigative teaching methods: A test of the theory of planned behavior. Journal of Research in Science Teaching, 27, 685–698. 454



Crites, S. L., Jr., Fabrigar, L. R., & Petty, R. E. (1994). Measuring the affective and cognitive properties of attitudes: Conceptual and methodological issues. Personality and Social Psychology Bulletin, 20, 619–634. Crocker, J., Niiya, Y., & Mischkowski, D. (2008). Why does writing about important values reduce defensiveness? Self-affirmation and the role of positive other-directed feelings. Psychological Science, 19, 740–747. Crockett, W. H. (1982). Balance, agreement, and positivity in the cognition of small social structures. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 15, pp. 1–57). New York: Academic Press. Cronen, V. E., & Conville, R. L. (1975). Fishbein’s conception of belief strength: A theoretical, methodological, and experimental critique. Speech Monographs, 42, 143–150. Crowley, A. E., & Hoyer, W. D. (1994). An integrative framework for understanding two-sided persuasion. Journal of Consumer Research, 20, 561–574. Croy, G., Gerrans, P., & Speelman, C. (2010). Injunctive social norms primacy over descriptive social norms in retirement savings decisions. International Journal of Aging and Human Development, 71, 259–282. Cruz, M. G. (1998). Explicit and implicit conclusions in persuasive messages. In M. Allen & R.W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 217–230). Cresskill, NJ: Hampton. Cuijpers, P. (2002). Peer-led and adult-led school drug prevention: A meta-analytic comparison. Journal of Drug Education, 32, 107–119. Cunningham, W. A., Packer, D. J., Kesek, A., & Van Bavel, J. J. (2009). Implicit measurement of attitudes: A physiological approach. In R. E. 455



Petty, R. H. Fazio, & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 485–512). New York: Psychology Press. Dahl, J., Enemo, I., Drevland, G. C. B., Wessel, E., Eilertsen, D. E., & Magnussen, S. (2007). Displayed emotions and witness credibility: A comparison of judgments by individuals and mock juries. Applied Cognitive Psychology, 21, 1145–1156. Dahlstrom, M. F. (2010). The role of causality in information acceptance in narratives: An example from science communication. Communication Research, 37, 857–875. Dal Cin, S., Zanna, M. P., & Fong, G. T. (2004). Narrative persuasion and overcoming resistance. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 175–191). Mahwah, NJ: Lawrence Erlbaum. Dale, A., & Strauss, A. (2009). Don’t forget to vote: Text message reminders as a mobilized tool. American Journal of Political Science, 53, 787–804. D’Alessio, D., & Allen, M. (2007). The selective exposure hypothesis and media choice processes. In R. W. Preiss, B. M. Gayle, N. Burrell, M. Allen, & J. Bryant (Eds.), Mass media effects research: Advances through meta-analysis (pp. 103–118). Mahwah, NJ: Lawrence Erlbaum. Dardis, F. E., & Shen, F. (2008). The influence of evidence type and product involvement on message-framing effects in advertising. Journal of Consumer Behaviour, 7, 222–238. Darke, P. R., Chaiken, S., Bohner, G., Einwiller, S., Erb, H.-P., & Hazlewood, J. D. (1998). Accuracy motivation, consensus information, and the law of large numbers: Effects on attitude judgment in the absence of argumentation. Personality and Social Psychology Bulletin, 24, 1205–1215.



456



Darker, C. D., French, D. P., Eves, F. F., & Sniehotta, F. F. (2010). An intervention to promote walking amongst the general population based on an ‘extended’ theory of planned behaviour: A waiting list randomised controlled trial. Psychology and Health, 25, 71–88. Darker, C. D., French, D. P., Longdon, S., Morris, K., & Eves, F. F. (2007). Are beliefs elicited biased by question order? A theory of planned behaviour belief elicitation study about walking in the UK general population. British Journal of Health Psychology, 12, 93–110. Darley, S. A., & Cooper, J. (1972). Cognitive consequences of forced noncompliance. Journal of Personality and Social Psychology, 24, 321–326. Das, E., & Fennis, B. M. (2008). In the mood to face the facts: When a positive mood promotes systematic processing of self-threatening information. Motivation and Emotion, 32, 221–230. Das, E., Vonkeman, C., & Hartmann, T. (2012). Mood as a resource in dealing with health recommendations: How mood affects information processing and acceptance of quitsmoking messages. Psychology and Health, 27, 116–127. Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35, 982–1003. Davis, M. H., & Runge, T. E. (1981). Beliefs and attitudes in a gubernatorial primary: Some limitations on the Fishbein model. Journal of Applied Social Psychology, 11, 93–113. Davis, R. E., & Resnicow, K. (2012). The cultural variance framework for tailoring health messages. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 115–135). Los Angeles: Sage.



457



Dean, M., Arvola, A., Vassallo, M., Lähteenmäki, L., Raats, M. M., Saba, A., & Shepherd, R. (2006). Comparison of elicitation methods for moral and affective beliefs in the theory of planned behaviour. Appetite, 47, 244–252. DeBono, K. G. (1987). Investigating the social-adjustive and valueexpressive functions of attitudes: Implications for persuasion processes. Journal of Personality and Social Psychology, 52, 279–287. DeBono, K. G. (2006). Self-monitoring and consumer psychology. Journal of Personality, 74, 715–737. DeBono, K. G., & Harnish, R. J. (1988). Source expertise, source attractiveness, and the processing of persuasive information: A functional approach. Journal of Personality and Social Psychology, 55, 541–546. DeBono, K. G., Leavitt, A., & Backus, J. (2003). Product packaging and product evaluation: An individual difference approach. Journal of Applied Social Psychology, 33, 513–521. DeBono, K. G., & Omoto, A. M. (1993). Individual differences in predicting behavioral intentions from attitude and subjective norm. Journal of Social Psychology, 133, 825–831. DeBono, K. G., & Packer, M. (1991). The effects of advertising appeal on perceptions of product quality. Personality and Social Psychology Bulletin, 17, 194–200. DeBono, K. G., & Snyder, M. (1989). Understanding consumer decisionmaking processes: The role of form and function in product evaluation. Journal of Applied Social Psychology, 19, 416–424. DeBono, K. G., & Telesca, C. (1990). The influence of source physical attractiveness on advertising effectiveness: A functional perspective. 458



Journal of Applied Social Psychology, 20, 1383–1395. de Bruijn, G. J. (2010). Understanding college students’ fruit consumption. Integrating habit strength in the theory of planned behaviour. Appetite, 54, 16–22. de Bruijn, G. J., & Gardner, B. (2011). Active commuting and habit strength: An interactive and discriminant analyses approach. American Journal of Health Promotion, 25, 27–36. de Bruijn, G. J., Kremers, S. P. J., Singh, A., van den Putte, B., & van Mechelen, W. (2009). Adult active transportation: Adding habit strength to the theory of planned behavior. American Journal of Preventive Medicine, 36, 189–194. de Bruijn, G. J., Kroeze, W., Oenema, A., & Brug, J. (2008). Saturated fat consumption and the theory of planned behaviour: Exploring additive and interactive effects of habit strength. Appetite, 51, 318–323. de Bruijn, G. J., & Rhodes, R. E. (2011). Exploring exercise behavior, intention and habit strength relationships. Scandinavian Journal of Medicine & Science in Sports, 21, 482–491. de Graaf, A., Hoeken, H., Sanders, J., & Beentjes, J. W. J. (2012). Identification as a mechanism of narrative persuasion. Communication Research, 39, 802–823. de Graaf, A., & Hustinx, L. (2011). The effect of story structure on emotion, transportation, and persuasion. Information Design Journal, 19, 142–154. de Hoog, N., Stroebe, W., & de Wit, J. (2007). The impact of vulnerability to and severity of a health risk on processing and acceptance of feararousing communications: A meta-analysis. Review of General Psychology, 11, 258–285. 459



de Hoog, N., Stroebe, W., & de Wit, J. B. F. (2008). The processing of fear-arousing communications: How biased processing leads to persuasion. Social Influence, 3, 84–113. De Houwer, J. (2009). Comparing measures of attitudes at the functional and procedural level: Analysis and implications. In R. E. Petty, R. H. Fazio, & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 362–390). New York: Psychology Press. De Houwer, J., Teige-Mocigemba, S., Spruyt, A., & Moors, A. (2009). Implicit measures: A normative analysis and review. Psychological Bulletin, 135, 347–368. DeJong, W. (1979). An examination of self-perception mediation of the foot-in-the-door effect. Journal of Personality and Social Psychology, 37, 2221–2239. DeJong, W., & Smith, S. W. (2013). Truth in advertising: Social norms marketing campaigns to reduce college drinking. In R. E. Rice & C. K. Atkin (Eds.), Public communication campaigns (4th ed., pp. 177–187). Los Angeles: Sage. Delia, J. G. (1975). Regional dialect, message acceptance, and perceptions of the speaker. Central States Speech Journal, 26, 188–194. Delia, J. G. (1976). A constructivist analysis of the concept of credibility. Quarterly Journal of Speech, 62, 361–375. Delia, J. G., Crockett, W. H., Press, A. N., & O’Keefe, D. J. (1975). The dependency of interpersonal evaluations on context-relevant beliefs about the other. Speech Monographs, 42, 10–19. DelVecchio, D., Henard, D. H., & Freling, T. H. (2006). The effect of sales promotion on post-promotion brand preference: A meta-analysis. Journal of Retailing, 82, 203–213. 460



Denizeau, M., Golsing, P., & Oberle, D. (2009). L’effet de l’ordre et du délai sur l’usage de trois modes de réduction de la dissonance cognitive: le changement d’attitude, la trivialisation et le déni de responsabilité [The effects of order and delay on the use of three modes of dissonance reduction: Attitude change, trivialization and denial of responsibility]. Année Psychologique, 109, 629–654. De Nooijer, J., van Assema, P., de Vet, E., & Brug, J. (2005). How stable are stages of change for nutrition behaviors in the Netherlands? Health Promotion International, 20, 27–32. Derose, S. F., Nakahiro, R. K., & Ziel, F. H. (2009). Automated messaging to improve compliance with diabetes test monitoring. American Journal of Managed Care, 15, 425–436. Detweiler, J. B., Bedell, B. T., Salovey, P., Pronin, E., & Rothman, A. J. (1999). Message framing and sunscreen use: Gain-framed messages motivate beach-goers. Health Psychology, 18, 189–196. Devine, D. J. (2012). Jury decision making: The state of the science. New York: New York University Press. de Vries, P., Aarts, H., & Midden, C. J. H. (2011). Changing simple energy-related consumer behaviors: How the enactment of intentions is thwarted by acting and non-acting habits. Environment and Behavior, 43, 612–633. de Wit, J. B. F., Stroebe, W., De Vroome, E. M. M., Sandfort, T. G. M., & van Griensven, G. J. P. (2000). Understanding AIDS preventive behavior with casual and primary partners in homosexual men: The theory of planned behavior and the information-motivation-behavioralskills model. Psychology and Health, 15, 325–340. Dexheimer, J. W., Talbot, T. R., Sanders, D. L., Rosenbloom, S. T., & Aronsky, D. (2008). Prompting clinicians about preventive care measures: A systematic review of randomized controlled trials. Journal 461



of the American Medical Informatics Association, 15, 311–318. De Young, R. (1989). Exploring the difference between recyclers and nonrecyclers: The role of information. Journal of Environmental Systems, 18, 341–351. De Young, R. (1990). Recycling as appropriate behavior: A review of survey data from selected recycling education programs in Michigan. Resources, Conservation, and Recycling, 3, 253–266. Dholakia, R. R. (1987). Source credibility effects: A test of behavioral persistence. Advances in Consumer Research, 14, 426–430. Diamond, G. A., & Cobb, M. D. (1996). The candidate as catastrophe: Latitude theory and the problems of political persuasion. In D. C. Mutz, P. M. Sniderman, & R. A. Brody (Eds.), Political persuasion and attitude change (pp. 225–247). Ann Arbor: University of Michigan Press. Dibonaventura, M. D., & Chapman, G. B. (2005). Moderators of the intention–behavior relationship in influenza vaccinations: Intention stability and unforeseen barriers. Psychology & Health, 20, 761–774. Dickerson, C. A., Thibodeau, R., Aronson, E., & Miller, D. (1992). Using cognitive dissonance to encourage water conservation. Journal of Applied Social Psychology, 22, 841–854. Dillard, A. J., Fagerlin, A., Dal Cin, S., Zikmund-Fisher, B. J., & Ubel, P. A. (2010). Narratives that address affective forecasting errors reduce perceived barriers to colorectal cancer screening. Social Science & Medicine, 71, 45–52. Dillard, J. P. (2011). An application of the integrative model to women’s intention to be vaccinated against HPV: Implications for message design. Health Communication, 26, 479–486. 462



Dillard, J. P., & Anderson, J. W. (2004). The role of fear in persuasion. Psychology and Marketing, 21, 909–926. Dillard, J. P., Hunter, J. E., & Burgoon, M. (1984). Sequential-request persuasive strategies: Meta-analysis of foot-in-the-door and door-in-theface. Human Communication Research, 10, 461–488. Dillard, J. P., & Nabi, R. L. (2006). The persuasive influence of emotion in cancer prevention and detection messages. Journal of Communication, 56, S123–S139. Dillard, J. P., & Peck, E. (2001). Persuasion and the structure of affect: Dual systems and discrete emotions as complementary models. Human Communication Research, 27, 38–68. Dillard, J. P., & Seo, K. (2013). Affect and persuasion. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 150–166). Thousand Oaks, CA: Sage. Dillard, J. P., & Shen, L. (2005). On the nature of reactance and its role in persuasive health communication. Communication Monographs, 72, 144–168. Dilliplane, S. (2010). Raising the specter of death: What terror management theory brings to the study of fear appeals. Communication Yearbook, 34, 93–131. Dingus, T. A., Hunn, B. P., & Wreggit, S. S. (1991). Two reasons for providing protective equipment as part of hazardous consumer product packaging. In Proceedings of the Human Factors Society 35th annual meeting (pp. 1039–1042). Santa Monica, CA: Human Factors Society. Di Noia, J., & Prochaska, J. O. (2010). Dietary stages of change and decisional balance: A meta-analytic review. American Journal of Health Behavior, 34, 618–632. 463



Ditto, P. H., Druley, J. A., Moore, K. A., Danks, J. H., & Smucker, W. D. (1996). Fates worse than death: The role of valued life activities in health–state evaluations. Health Psychology, 15, 332–343. DiVesta, F. J., & Merwin, J. C. (1960). The effects of need-oriented communications on attitude change. Journal of Abnormal and Social Psychology, 60, 80–85. Doll, J., & Ajzen, I. (1992). Accessibility and stability of predictors in the theory of planned behavior. Journal of Personality and Social Psychology, 63, 754–765. Doll, J., Ajzen, I., & Madden, T. J. (1991). Optimale skalierung und urteilsbildung in unter-schiedlichen einstellungsbereichen: Eine reanalyse [Optimal scaling and judgment in different attitude domains: A reanalysis]. Zeitschriftfur Sozialpsychologie, 22, 102–111. Doll, J., & Mallu, R. (1990). Individuierte einstellungsformation, einstellungsstruktur und einstellungs-verhalten-konsistenz [Individuated attitude formation, attitude structure and attitude-behavior consistency]. Zeitschrift fur Sozialpsychologie, 21, 2–14. Doll, J., & Orth, B. (1993). The Fishbein and Ajzen theory of reasoned action applied to contraceptive behavior: Model variants and meaningfulness. Journal of Applied Social Psychology, 23, 395–415. Donaldson, S. I., Graham, J. W., & Hansen, W. B. (1994). Testing the generalizability of intervening mechanism theories: Understanding the effects of adolescent drug use prevention interventions. Journal of Behavioral Medicine, 17, 195–216. Donaldson, S. I., Graham, J. W., Piccinin, A. M., & Hansen, W. B. (1995). Resistance-skills training and onset of alcohol use: Evidence for beneficial and potentially harmful effects in public schools and in private Catholic schools. Health Psychology, 14, 291–300.



464



Donnelly, J. H., Jr., & Ivancevich, J. M. (1970). Post-purchase reinforcement and back-out behavior. Journal of Marketing Research, 7, 399–400. Doob, A. N., Carlsmith, J. M., Freedman, J. L., Landauer, T. K., & Tom, S., Jr. (1969). Effect of initial selling price on subsequent sales. Journal of Personality and Social Psychology, 11, 345–350. Druckman, J. N. (2011). What’s it all about? Framing in political science. In G. Keren (Ed.), Perspectives on framing (pp. 279–302). New York: Psychology Press. Druckman, J. N., & Bolsen, T. (2012). How scientific evidence links attitudes to behaviors. In D. A. Dana (Ed.), The nanotechnology challenge: Creating legal institutions for uncertain risks (pp. 84–102). Cambridge, UK: Cambridge University Press. Drummond, A. J. (2011). Assimilation, contrast and voter projections of parties in left-right space: Does the electoral system matter? Party Politics, 17, 711–743. Duncan, T. E., Duncan, S. C., Beauchamp, N., Wells, J., & Ary, D. V. (2000). Development and evaluation of an interactive CD-ROM refusal skills program to prevent youth substance use: “Refuse to Use.” Journal of Behavioral Medicine, 23, 59–72. Dunlop, S. M., Wakefield, M., & Kashima, Y. (2010). Pathways to persuasion: Cognitive and experiential responses to health-promoting mass media messages. Communication Research, 37, 133–164. Durantini, M. R., Albarracín, D., Mitchell, A. L., Earl, A. N., & Gillette, J. C. (2006). Conceptualizing the influence of social agents of behavior change: A meta-analysis of the effectiveness of HIV-prevention interventions for different groups. Psychological Bulletin, 132, 212–248.



465



Dutta-Bergman, M. J. (2003). The linear interaction model of personality effects in health communication. Health Communication, 15, 101–116. Dwan, K., Gamble, C., Williamson, P. R., Kirkham, J. J., & the Reporting Bias Group. (2013). Systematic review of the empirical evidence of study publication bias and outcome reporting bias: An updated review. PLO SONE, 8(7), e66844. Eagly, A. H. (2007). In defence of ourselves: The effects of defensive processing on attitudinal phenomena. In M. Hewstone, H. A. W. Schut, J. B. F. de Wit, K. van den Bos, & M. S. Stroebe (Eds.), The scope of social psychology: Theory and applications (pp. 65–83). New York: Psychology Press. Eagly, A. H., & Chaiken, S. (1975). An attribution analysis of the effect of communicator characteristics on opinion change: The case of communicator attractiveness. Journal of Personality and Social Psychology, 32, 136–144. Eagly, A. H., & Chaiken, S. (1984). Cognitive theories of persuasion. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 17, pp. 267–359). New York: Academic Press. Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace Jovanovich. Eagly, A. H., & Chaiken, S. (1998). Attitude structure and function. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 1, pp. 269–322). Boston: McGraw-Hill. Eagly, A. H., Mladinic, A., & Otto, S. (1994). Cognitive and affective bases of attitudes toward social groups and social policies. Journal of Experimental Social Psychology, 30, 113–137. Eagly A. H., & Telaak, K. (1972). Width of the latitude of acceptance as a 466



determinant of attitude change. Journal of Personality and Social Psychology, 23, 388–397. Eagly, A. H., Wood, W., & Chaiken, S. (1978). Causal inferences about communicators and their effect on opinion change. Journal of Personality and Social Psychology, 36, 424–435. Eagly, A. H., Wood, W., & Chaiken, S. (1981). An attribution analysis of persuasion. In J. H. Harvey, W. Ickes, & R. F. Kidd (Eds.), New directions in attribution research (Vol. 3, pp. 37–62). Hillsdale, NJ: Lawrence Erlbaum. Earl, A., & Albarracín, D. (2007). Nature, decay, and spiraling of the effects of fear-inducing arguments and HIV counseling and testing: A meta-analysis of the short and long-term outcomes of HIV-prevention interventions. Health Psychology, 26, 496–506. Eckes, T., & Six, B. (1994). Fakten und fiktionen in der einstellungsverhaltens-forschung: Eine meta-analyse [Fact and fiction in attitudebehavior research: A meta-analysis]. Zeitschriftfür Sozialpsychologie, 25, 253–271. Edwards, K. (1990). The interplay of affect and cognition in attitude formation and change. Journal of Personality and Social Psychology, 59, 202–216. Edwards, S. M., Li, H. R., & Lee, J. H. (2002). Forced exposure and psychological reactance: Antecedents and consequences of the perceived intrusiveness of pop-up ads. Journal of Advertising, 31 (3), 83–95. Ehrlich, D., Guttman, I., Schönbach, P., & Mills, J. (1957). Postdecision exposure to relevant information. Journal of Abnormal and Social Psychology, 54, 98–102.



467



Eisend, M. (2002). Dimensions of credibility in marketing communication. In R. Zwick & T. Ping (Eds.), Asia Pacific advances in consumer research (Vol. 5, pp. 366–373). Valdosta, GA: Association for Consumer Research. Eisend, M. (2006). Two-sided advertising: A meta-analysis. International Journal of Research in Marketing, 23, 187–198. Eisend, M. (2007). Understanding two-sided persuasion: An empirical assessment of theoretical approaches. Psychology and Marketing, 24, 615–640. Eisend, M. (2008). Explaining the impact of scarcity appeals in advertising: Affect integration in a simultaneous presentation context. Journal of Advertising, 37(3), 33–41. Eisend, M. (2010). Explaining the joint effect of source credibility and negativity of information in two-sided messages. Psychology & Marketing, 27, 1032–1049. Eisend, M., & Langner, T. (2010). Immediate and delayed advertising effects of celebrity endorsers’ attractiveness and expertise. International Journal of Advertising, 29, 527–546. Eisenstadt, D., Leippe, M. R., Rivers, J. A., & Stambush, M. A. (2003). Counterattitudinal advocacy on a matter of prejudice: Effects of distraction, commitment, and personal importance. Journal of Applied Social Psychology, 33, 2123–2152. Elder, J. P., Sallis, J. F., Woodruff, S. I., & Wildey, M. B. (1993). Tobacco-refusal skills and tobacco use among high-risk adolescents. Journal of Behavioral Medicine, 16, 629–642. Elliott, M. A., & Ainsworth, K. (2012). Predicting university undergraduates’ binge-drinking behavior: A comparative test of the one468



and two-component theories of planned behavior. Addictive Behaviors, 37, 92–101. Elliott, M. A., & Armitage, C. J. (2006). Effects of implementation intentions on the self-reported frequency of drivers’ compliance with speed limits. Journal of Experimental Psychology: Applied, 12, 108–117. Elliott, M. A., Armitage, C. J., & Baughan, C. J. (2005). Exploring the beliefs underpinning drivers’ intentions to comply with speed limits. Transportation Research Part F: Traffic Psychology and Behaviour, 8, 459–480. Elliott, R., Jobber, D., & Sharp, J. (1995). Using the theory of reasoned action to understand organizational behaviour: The role of belief salience. British Journal of Social Psychology, 34, 161–172. Elms, A. C. (Ed.). (1969). Role playing, reward, and attitude change. New York: Van Nostrand Reinhold. Engstrom, E. (1994). Effects of nonfluencies on speaker’s credibility in newscast settings. Perceptual and Motor Skills, 78, 739–743. Ennett, S. T., Tobler, N. S., Ringwalt, C. L., & Flowelling, R. L. (1994). How effective is drug abuse resistance education? A meta-analysis of Project DARE outcome evaluations. American Journal of Public Health, 84, 1394–1401. Ennis, R., & Zanna, M. P. (1993). Attitudes, advertising, and automobiles: A functional approach. Advances in Consumer Research, 20, 662–666. Ennis, R., & Zanna, M. P. (2000). Attitude function and the automobile. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 395–415). Mahwah, NJ: Lawrence Erlbaum.



469



Epton, T., Harris, P. R., Kane, R., van Koningsbruggen, G. M., & Sheeran, P. (in press). The impact of self-affirmation on health-behavior change: A meta-analysis. Health Psychology. doi:10.1037/hea0000116. Erb, H.-P., Pierro, A., Mannetti, L., Spiegel, S., & Kruglanski, A. W. (2007). Biased processing of persuasive information: On the functional equivalence of cues and message arguments. European Journal of Social Psychology, 37, 1057–1075. Escalas, J. E. (2004). Imagine yourself in the product: Mental simulation, narrative transportation, and persuasion. Journal of Advertising, 33(2), 37–48. Escalas, J. E. (2007). Self referencing and persuasion: Narrative transportation versus analytic elaboration. Journal of Consumer Research, 33, 421–429. Escalas, J. E., & Luce, M. F. (2004). Understanding the effect of processfocused versus outcome-focused thought in response to advertising. Journal of Consumer Research, 31, 274–285. Esses, V. M., Haddock, G., & Zanna, M. P. (1993). Values, stereotypes, and emotions as determinants of intergroup attitudes. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 137–166). San Diego: Academic Press. Estabrooks, P., & Carron, A. V. (1998). The conceptualization and effect of control beliefs on exercise attendance in the elderly. Journal of Aging and Health, 10, 441–457. Evans, R. I., Rozelle, R. M., Lasater, T. M., Dembroski, T. M., & Allen, B. P. (1970). Fear arousal, persuasion, and actual versus implied behavioral change: New perspective utilizing a real-life dental hygiene program. Journal of Personality and Social Psychology, 16, 220–227.



470



Everson, E. S., Daley, A. J., & Ussher, M. (2007). Brief report: The theory of planned behaviour applied to physical activity in young people who smoke. Journal of Adolescence, 30, 347–351. Fabrigar, L. R., MacDonald, T. K., & Wegener, D. T. (2007). The structure of attitudes. In D. Albarracín, B. T. Johnson, & M. P. Zanna (Eds.), Handbook of attitudes and attitude change (pp. 79–124). Mahwah, NJ: Lawrence Erlbaum. Fabrigar, L. R., & Petty, R. E. (1999). The role of the affective and cognitive bases of attitudes in susceptibility to affectively and cognitively based persuasion. Personality and Social Psychology Bulletin, 25, 363–381. Fabrigar, L. R., Petty, R. E., Smith, S. M., & Crites, S. L., Jr. (2006). Understanding knowledge effects on attitude-behavior consistency: The role of relevance, complexity, and amount of knowledge. Journal of Personality and Social Psychology, 90, 556–577. Fairchild, A. J., & MacKinnon, D. P. (2009). A general model for testing mediation and moderation effects. Prevention Science, 10, 87–99. Falcione, R. L. (1974). The factor structure of source credibility scales for immediate superiors in the organizational context. Central States Speech Journal, 25, 63–66. Falomir-Pichastor, J. M., Butera, F., & Mugny, G. (2002). Persuasive constraint and expert versus non-expert influence in intention to quit smoking. European Journal of Social Psychology, 32, 209–222. Fazio, R. H., & Roskos-Ewoldsen, D. R. (2005). Acting as we feel: When and how attitudes guide behavior. In T. C. Brock & M. C. Green (Eds.). Persuasion: Psychological insights and perspectives (2nd ed., pp. 41–79). Thousand Oaks, CA: Sage.



471



Fazio, R. H., & Towles-Schwen, T. (1999). The MODE model of attitudebehavior processes. In S. Chaiken & Y. Trope (Eds.), Dual-process models in social psychology (pp. 97–116). New York: Guilford. Feeley, T. H., Anker, A. E., & Aloe, A. M. (2012). The door-in-the-face persuasive message strategy: A meta-analysis of the first 35 years. Communication Monographs, 79, 316–343. Feiler, D. C., Tost, L. P., & Grant, A. M. (2012). Mixed reasons, missed givings: The costs of blending egoistic and altruistic reasons in donation requests. Journal of Experimental Social Psychology, 48, 1322–1328. Fein, S., & Spencer, S. J. (1997). Prejudice as self-image maintenance: Affirming the self through derogating others. Journal of Personality and Social Psychology, 73, 31–44. Feldman, J. M. (1974). Note on the utility of certainty weights in expectancy theory. Journal of Applied Psychology, 59, 727–730. Feldman, L., Stroud, N. J., Bimber, B., & Wojcieszak, M. (2013). Assessing selective exposure in experiments: The implications of different methodological choices. Communication Methods and Measures, 7, 172–194. Fennis, B. M., & Janssen, L. (2010). Mindlessness revisited: Sequential request techniques foster compliance by draining self-control resources. Current Psychology, 29, 235–246. Fennis, B. M., Janssen, L., & Vohs, K. D. (2009). Acts of benevolence: A limited-resource account of compliance with charitable requests. Journal of Consumer Research, 35, 906–924. Fennis, B. M., & Stroebe, W. (2010). The psychology of advertising. New York: Psychology Press.



472



Ferguson, C. J., & Heene, M. (2012). A vast graveyard of undead theories: Publication bias and psychological science’s aversion to the null. Perspectives on Psychological Science, 7, 555–561. Fern, E. F., Monroe, K. B., & Avila, R. A. (1986). Effectiveness of multiple request strategies: A synthesis of research results. Journal of Marketing Research, 23, 144–152. Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Festinger, L. (Ed.). (1964). Conflict, decision, and dissonance. Stanford, CA: Stanford University Press. Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of Abnormal and Social Psychology, 58, 203–210. Festinger, L., & Walster, E. (1964). Post-decision regret and decision reversal. In L. Festinger (Ed.), Conflict, decision, and dissonance (pp. 100–110). Stanford, CA: Stanford University Press. Feufel, M. A., Schneider, T. R., & Berkel, H. J. (2010). A field test of the effects of instruction design on colorectal cancer self-screening accuracy. Health Education Research, 25, 709–723. Fiedler, K., Messner, C., & Bluemke, M. (2006). Unresolved problems with the “I,” the “A,” and the “T”: A logical and psychometric critique of the implicit association test (IAT). European Review of Social Psychology, 17, 74–147. Field, A. P., & Gillett, R. (2010). How to do a meta-analysis. British Journal of Mathematical & Statistical Psychology, 63, 665–694. Fine, B. J. (1957). Conclusion-drawing, communicator credibility, and 473



anxiety as factors in opinion change. Journal of Abnormal and Social Psychology, 54, 369–374. Fink, E. L., & Cai, D. A. (2013). Discrepancy models of belief change. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 84–103). Thousand Oaks, CA: Sage. Fischer, P. (2011). Selective exposure, decision uncertainty, and cognitive economy: A new theoretical perspective on confirmatory information search. Social and Personality Psychology Compass, 5, 751–762. Fischer, P., & Greitemeyer, T. (2010). A new look at selective-exposure effects: An integrative model. Current Directions in Psychological Science, 19, 384–389. Fischer, P., Lea, S., Kastenmüller, A., Greitemeyer, T., Fischer, J., & Frey, D. (2011). The process of selective exposure: Why confirmatory information search weakens over time. Organizational Behavior and Human Decision Processes, 114, 37–48. Fischer, P., Schulz-Hardt, S. & Frey, D. (2008). Selective exposure and information quantity: How different information quantities moderate decision makers’ preference for consistent and inconsistent information. Journal of Personality and Social Psychology, 94, 231–244. Fishbein, M. (1967a). A behavior theory approach to the relations between beliefs about an object and the attitude toward the object. In M. Fishbein (Ed.), Readings in attitude theory and measurement (pp. 389–400). New York: Wiley. Fishbein, M. (1967b). A consideration of beliefs, and their role in attitude measurement. In M. Fishbein (Ed.), Readings in attitude theory and measurement (pp. 257–266). New York: Wiley.



474



Fishbein, M. (2008). A reasoned action approach to health promotion. Medical Decision Making, 28, 834–844. Fishbein, M., & Ajzen, I. (1974). Attitudes towards objects as predictors of single and multiple behavioral criteria. Psychological Review, 81, 59–74. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior. Reading, MA: Addison-Wesley. Fishbein, M., & Ajzen, I. (1981). Attitudes and voting behaviour: An application of the theory of reasoned action. In G. M. Stephenson & J. M. Davis (Eds.), Progress in applied social psychology (Vol. 1, pp. 253–313). New York: Wiley. Fishbein, M., & Ajzen, I. (2010). Predicting and changing behavior: The reasoned action approach. New York: Psychology Press. Fishbein, M., Cappella, J., Hornik, R., Sayeed, S., Yzer, M., & Ahern, R. K. (2002). The role of theory in developing effective anti-drug public service announcements. In W. D. Crano & M. Burgoon (Eds.), Mass media and drug prevention: Classic and contemporary theories and research (pp. 89–117). Mahwah, NJ: Lawrence Erlbaum. Fishbein, M., & Hunter, R. (1964). Summation versus balance in attitude organization and change. Journal of Abnormal and Social Psychology, 69, 505–510. Fishbein, M., & Lange, R. (1990). The effects of crossing the midpoint on belief change: A replication and extension. Personality and Social Psychology Bulletin, 16, 189–199. Fishbein, M., & Middlestadt, S. (1995). Noncognitive effects on attitude formation and change: Fact or artifact? Journal of Consumer Psychology, 4, 181–202. 475



Fishbein, M., & Middlestadt, S. E. (1997). A striking lack of evidence for nonbelief-based attitude formation and change: A response to five commentaries. Journal of Consumer Psychology, 6, 107–115. Fishbein, M., & Yzer, M. C. (2003). Using theory to design effective health behavior interventions. Communication Theory, 13, 164–183. Fisher, J. D., Fisher, W. A., Misovich, S. J., Kimble, D. L., & Malloy, T. E. (1996). Changing AIDS-risk behavior: Effects of an intervention emphasizing AIDS risk reduction information, motivation, and behavioral skills in a college student population. Health Psychology, 15, 114–123. Fitzsimons, G. J., & Lehmann, D. R. (2004). Reactance to recommendations: When unsolicited advice yields contrary responses. Marketing Science, 23, 82–94. Flanagin, A. J., & Metzger, M. J. (2007). The role of site features, user attributes, and information verification behaviors on the perceived credibility of web-based information. New Media and Society, 9, 319–142. Flay, B. R., McFall, S., Burton, D., Cook, T. D., & Warnecke, R. B. (1993). Health behavior changes through television: The roles of de facto and motivated selection processes. Journal of Health and Social Behavior, 34, 322–335. Fleming, D. (1967). Attitude: The history of a concept. Perspectives in American History, 1, 287–365. Fleming, J. K., & Ginis, K. A. M. (2004). The effects of commercial exercise video models on women’s self-presentational efficacy and exercise task self-efficacy. Journal of Applied Sport Psychology, 16, 92–102.



476



Fleming, M. A., & Petty, R. E. (2000). Identity and persuasion: An elaboration likelihood approach. In D. J. Terry & M. A. Hogg (Eds.), Attitudes, behavior, and social context: The role of norms and group membership (pp. 171–199). Mahwah, NJ: Lawrence Erlbaum. Floyd, D. L., Prentice-Dunn, S., & Rogers, R. W. (2000). A meta-analysis of research on protection motivation theory. Journal of Applied Social Psychology, 30, 407–429. Fointiat, V. (2004). “I know what I have to do, but …”: When hypocrisy leads to behavioral change. Social Behavior and Personality, 32, 741–746. Fontenelle, G. A., Phillips, A. P., & Lane, D. M. (1985). Generalizing across stimuli as well as subjects: A neglected aspect of external validity. Journal of Applied Psychology, 70, 101–107. Forehand, M., Gastil, J., & Smith, M. A. (2004). Endorsements as voting cues: Heuristic and systematic processing in initiative elections. Journal of Applied Social Psychology, 34, 2215–2231. Fornara, F., Carrus, G., Passafaro, P., & Bonnes, M. (2011). Distinguishing the sources of normative influence on proenvironmental behaviors: The role of local norms in household waste recycling. Group Processes & Intergroup Relations, 14, 623–635. Freedman, J. L. (1964). Involvement, discrepancy, and change. Journal of Abnormal and Social Psychology, 69, 290–295. Freedman, J. L., & Fraser, S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology, 4, 195–202. Freedman, J. L., & Sears, D. O. (1965). Warning, distraction, and resistance to influence. Journal of Personality and Social Psychology, 1, 477



262–266. Freijy, T., & Kothe, E. J. (2013). Dissonance-based interventions for health behaviour change: A systematic review. British Journal of Health Psychology, 18, 310–337. French, D. P., & Cooke, R. (2012). Using the theory of planned behaviour to understand binge drinking: The importance of beliefs for developing interventions. British Journal of Health Psychology, 17, 1–17. French, D. P., Sutton, S., Hennings, S. J., Mitchell, J., Wareham, N. J., Griffin, S., … Kinmonth, A. L. (2005). The importance of affective beliefs and attitudes in the theory of planned behavior: Predicting intention to increase physical activity. Journal of Applied Social Psychology, 35, 1824–1848. Frewer, L. J., Howard, C., Hedderley, D., & Shepherd, R. (1996). What determines trust in information about food-related risks? Underlying psychological constructs. Risk Analysis, 16, 473–486. Frey, D. (1986). Recent research on selective exposure to information. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 41–80). New York: Academic Press. Fried, C. B. (1998). Hypocrisy and identification with transgressions: A case of undetected dissonance. Basic and Applied Social Psychology, 20, 145–154. Fried, C. B., & Aronson, E. (1995). Hypocrisy, misattribution, and dissonance reduction. Personality and Social Psychology Bulletin, 21, 925–933. Friedrich, J., Fetherstonhaugh, D., Casey, S., & Gallagher, D. (1996). Argument integration and attitude change: Suppression effects in the integration of one-sided arguments that vary in persuasiveness. 478



Personality and Social Psychology Bulletin, 22, 179–191. Fry, J. P., & Neff, R. A. (2009). Periodic prompts and reminders in health promotion and health behavior interventions: Systematic review. Journal of Medical Internet Research, 11(2), e16. Gagné, C., & Godin, G. (2000). The theory of planned behavior: Some measurement issues concerning belief-based variables. Journal of Applied Social Psychology, 30, 2173–2193. Gagné, C., & Godin, G. (2007). Does the easy-difficult item measure attitude or perceived behavioural control? British Journal of Health Psychology, 12, 543–557. Gallagher, K. M., & Updegraff, J. A. (2012). Health message framing effects on attitudes, intentions, and behavior: A meta-analytic review. Annals of Behavioral Medicine, 43, 101–116. Gallagher, K. M., Updegraff, J. A., Rothman, A. J., & Sims, L. (2011). Perceived susceptibility to breast cancer moderates the effect of gainand loss-framed messages on use of screening. Health Psychology, 30, 145–152. Gallagher, S., & Povey, R. (2006). Determinants of older adults’ intentions to vaccinate against influenza: A theoretical application. Journal of Public Health, 28, 139–144. Gangestad, S. W., & Snyder, M. (2000). Self-monitoring: Appraisal and re-appraisal. Psychological Bulletin, 126, 530–555. Garner, R. (2005). What’s in a name? Persuasion perhaps. Journal of Consumer Psychology, 15, 108–116. Garrett, R. K. (2009). Politically motivated reinforcement seeking: 479



Reframing the selective exposure debate. Journal of Communication, 59, 676–699. Garrett, R. K., & Stroud, N. J. (2014). Partisan paths to exposure diversity: Differences in proand counterattitudinal news consumption. Journal of Communication, 64, 680–701. Gasco, M., Briñol, P., & Horcajo, J. (2010). Cambio de actitudes hacia la imagen corporal: El efecto de la elaboración sobre la fuerza de las actitudes [Attitude change toward body image: The role of elaboration on attitude strength]. Psicothema, 22, 71–76. Gass, R. H., & Seiter, J. S. (2004). Embracing divergence: A definitional analysis of pure and borderline cases of persuasion. In J. S. Seiter & R. H. Gass (Eds.), Perspectives on persuasion, social influence, and compliance gaining (pp. 13–29). Boston: Pearson Allyn and Bacon. Gastil, J. (1992). Why we believe in democracy: Testing theories of attitude functions and democracy. Journal of Applied Social Psychology, 22, 423–450. Gaston, A., Cramp, A., & Prapavessis, H. (2012). Enhancing self-efficacy and exercise readiness in pregnant women. Psychology of Sport and Exercise, 13, 550–557. Gawronski, B., & Bodenhausen, G. V. (2007). What do we know about implicit attitude measures and what do we have to learn? In B. Wittenbrink & N. Schwarz (Eds.), Implicit measures of attitudes (pp. 265–286). New York: Guilford. Gawronski, B., & Strack, F. (Eds.). (2012). Cognitive consistency: A fundamental principle in social cognition. New York: Guilford. Gaziano, C., & McGrath, K. (1986). Measuring the concept of credibility. Journalism Quarterly, 63, 451–462. 480



Geers, A. W., Handley, I. M., & McLarney, A. R. (2003). Discerning the role of optimism in persuasion: The valence-enhancement hypothesis. Journal of Personality and Social Psychology, 85, 554–565. Gerend, M. A., & Cullen, M. (2008). Effects of message framing and temporal context on college student drinking behavior. Journal of Experimental Social Psychology, 44, 1167–1173. Gerend, M. A., & Shepherd, J. E. (2012). Predicting human papillomavirus vaccine uptake in young adult women: Comparing the health belief model and theory of planned behavior. Annals of Behavioral Medicine, 44, 171–180. Gierl, H., & Huettl, V. (2010). Are scarce products always more attractive? The interaction of different types of scarcity signals with products’ suitability for conspicuous consumption. International Journal of Research in Marketing, 27, 225–235. Giles, M., & Cairns, E. (1995). Blood donation and Ajzen’s theory of planned behaviour: An examination of perceived behavioural control. British Journal of Social Psychology, 34, 173–188. Giles, M., McClenahan, C., Armour, C., Millar, S., Rae, G., Mallett, J., & Stewart-Knox, B. (2014). Evaluation of a theory of planned behaviour– based breastfeeding intervention in Northern Irish Schools using a randomized cluster design. British Journal of Health Psychology, 19, 16–35. Glasman, L. R., & Albarracín, D. (2006). Forming attitudes that predict future behavior: A meta-analysis of the attitude–behavior relation. Psychological Bulletin, 132, 778–822. Glik, D., Berkanovic, E., Stone, K., Ibarra, L., Jones, M. C., Rosen, B., … Richardes, D. (1998). Health education goes Hollywood: Working with prime-time and daytime entertainment television for immunization promotion. Journal of Health Communication, 3, 263–284. 481



Glynn, C. J., Huge, M. E., & Lunney, C. A. (2009). The influence of perceived social norms on college students’ intention to vote. Political Communication, 26, 48–64. Göckeritz, S., Schultz, P. W., Rendon, T., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2010). Descriptive normative beliefs and conservation behavior: The moderating roles of personal involvement and injunctive normative beliefs. European Journal of Social Psychology, 40, 514–523. Godin, G., Gagné, C., & Sheeran, P. (2004). Does perceived behavioural control mediate the relationship between power beliefs and intention? British Journal of Health Psychology, 9, 557–568. Godin, G., & Kok, G. (1996). The theory of planned behavior: A review of its applications to health-related behaviors. American Journal of Health Promotion, 11, 87–98. Godin, G., Maticka-Tyndale, E., Adrien, A., Manson-Singer, S., Willms, D., & Cappon, P. (1996). Cross-cultural testing of three social cognitive theories: An application to condom use. Journal of Applied Social Psychology, 26, 1556–1586. Godin, G., Sheeran, P., Conner, M., Delage, G., Germain, M., BélangerGravel, A., & Naccache, H. (2010). Which survey questions change behavior? Randomized controlled trial of mere measurement interventions. Health Psychology, 29, 636–644. Godin, G., Valois, P., & Lepage, L. (1993). The pattern of influence of perceived behavioral control upon exercising behavior: An application of Ajzen’s theory of planned behavior. Journal of Behavioral Medicine, 16, 81–102. Goei, R., Boyson, A. R., Lyon-Callo, S. K., Schott, C., Wasilevich, E., & Cannarile, S. (2010). An examination of EPPM predictions when threat is perceived externally: An asthma intervention with school workers. 482



Health Communication, 25, 333–344. Goei, R., Lindsey, L. L. M., Boster, F. J., Skalski, P. D., & Bowman, J. M. (2003). The mediating roles of liking and obligation on the relationship between favors and compliance. Communication Research, 30, 178–197. Goethals, G. R., & Nelson, R. E. (1973). Similarity in the influence process: The belief-value distinction. Journal of Personality and Social Psychology, 25, 117–122. Goldman, M., McVeigh, J. F., & Richterkessing, J. L. (1984). Door-in-theface procedure: Reciprocal concession, perceptual contrast, or worthy person. Journal of Social Psychology, 123, 245–251. Goldstein, N. J., Cialdini, R. B., & Griskevicius, V. (2008). A room with a viewpoint: Using social norms to motivate environmental conservation in hotels. Journal of Consumer Research, 35, 472–482. Goldstein, N. J., & Mortensen, C. R. (2012). Social norms: A how-to (and how-not-to) guide. In D. T. Kendrick, N. J. Goldstein, & S. L. Braver (Eds.), Six degrees of social influence: Science, application, and the psychology of Robert Cialdini (pp. 68–78). New York: Oxford University Press. Gollust, S. E., Niederdeppe, J., & Barry, C. L. (2013). Framing the consequences of childhood obesity to increase public support for obesity prevention policy. American Journal of Public Health, 103, e96–e102. Gollwitzer, P. M., & Oettingen, G. (2008). The question-behavior effect from an action control perspective. Journal of Consumer Psychology, 18, 107–110. Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. In M. P. 483



Zanna (Ed.), Advances in experimental social psychology (Vol. 38, pp. 69–120). San Diego: Elsevier Academic Press. Good, A., & Abraham, C. (2007). Measuring defensive responses to threatening messages: A meta-analysis of measures. Health Psychology Review, 1, 208–229. Goodall, C. E. (2011). An overview of implicit measures of attitudes: Methods, mechanisms, strengths, and limitations. Communication Methods and Measures, 5, 203–222. Gorassini, D. R., & Olson, J. M. (1995). Does self-perception change explain the foot-in-the-door effect? Journal of Personality and Social Psychology, 69, 91–105. Gorman, D. R. (1995). Are school-based resistance skills training programs effective in preventing alcohol misuse? Journal of Alcohol and Drug Education, 41, 74–98. Gould, S. J. (1991). Bully for brontosaurus: Reflections in natural history. New York: Norton. Gould, S. J. (1993). Eight little piggies: Reflections in natural history. New York: Norton. Gourville, J. T. (1999). The effect of implicit versus explicit comparisons on temporal pricing claims. Marketing Letters, 10, 113–124. Granberg, D. (1982). Social judgment theory. Communication Yearbook, 6, 304–329. Granberg, D., & Campbell, K. E. (1977). Effect of communication discrepancy and ambiguity on placement and opinion shift. European Journal of Social Psychology, 7, 137–150. 484



Granberg, D., Kasmer, J., & Nanneman, T. (1988). An empirical examination of two theories of political perception. Western Political Quarterly, 41, 29–46. Granberg, D., & Steele, L. (1974). Procedural considerations in measuring latitudes of acceptance, rejection, and noncommitment. Social Forces, 52, 538–542. Grant, N. K., Fabrigar, L. R., & Lim, H. (2010). Exploring the efficacy of compliments as a tactic for securing compliance. Basic and Applied Social Psychology, 32, 226–233. Grasmick, H. G., Bursik, R. J., Jr., & Kinsey, K. A. (1991). Shame and embarrassment as deterrents to noncompliance with the law: The case of an antilittering campaign. Environment and Behavior, 23, 233–251. Green, B. F. (1954). Attitude measurement. In G. Lindzey (Ed.), Handbook of social psychology (Vol. 1, pp. 335–369). Reading, MA: Addison-Wesley. Green, D. P., Ha, S. E., & Bullock, J. G. (2010). Enough already about “black box” experiments: Studying mediation is more difficult than most scholars suppose. Annals of the American Academy of Political and Social Science, 628, 200–208. Green, M. C. (2008). Research challenges in narrative persuasion. Information Design Journal, 16, 47–52. Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79, 701–721. Green, M. C., & Clark, J. L. (2013). Transportation into narrative worlds: Implications for entertainment media influences on tobacco use. Addiction, 108, 477–484. 485



Green, M. C., Garst, J., Brock, T. C., & Chung, S. (2006). Fact versus fiction labeling: Persuasion parity despite heightened scrutiny of fact. Media Psychology, 8, 267–285. Greenberg, B. S., & Miller, G. R. (1966). The effects of low-credible sources on message acceptance. Speech Monographs, 33, 127–136. Greene, K., Krcmar, M., Rubin, D. L., Walters, L.H., & Hale, J. L. (2002). Elaboration in processing adolescent health messages: The impact of egocentrism and sensation seeking on message processing. Journal of Communication, 52, 812–831. Greenwald, A. G., Poehlman, T. A., Uhlmann, E. L., & Banaji, M. R. (2009). Understanding and using the implicit association test: III. Metaanalysis of predictive validity. Journal of Personality and Social Psychology, 97, 17–41. Gregory, G. D., Munch, J. M., & Peterson, M. (2002). Attitude functions in consumer research: Comparing value-attitude relations in individualist and collectivist cultures. Journal of Business Research, 55, 933–942. Gregory, W. L., Cialdini, R. B., & Carpenter, K. M. (1982). Self-relevant scenarios as mediators of likelihood estimates and compliance: Does imagining make it so? Journal of Personality and Social Psychology, 43, 89–99. Griskevicius, V., Shiota, M. N., & Neufeld, S. L. (2010). Influence of different positive emotions on persuasion processing: A functional evolutionary approach. Emotion, 10, 190–206. Gruner, C. R. (1967). Effect of humor on speaker ethos and audience information gain. Journal of Communication, 17, 228–233. Gruner, C. R. (1970). The effect of humor in dull and interesting 486



informative speeches. Central States Speech Journal, 21, 160–166. Gruner, C. R., & Lampton, W. E. (1972). Effects of including humorous material in a persuasive sermon. Southern Speech Communication Journal, 38, 188–196. Guadagno, R. E., & Cialdini, R. B. (2010). Preference for consistency and social influence: A review of current research findings. Social Influence, 5, 152–163. Guéguen, N., & Pascual, A. (2005). Improving the response rate to a street survey: An evaluation of the “but you are free to accept or to refuse” technique. Psychological Record, 55, 297–303. Guéguen, N., Pascual, A., & Dagot, L. (2002). Low-ball and compliance to a request: An application in a field setting. Psychological Reports, 91, 81–84. Guéguen, N., Pichot, N., & Le Dreff, G. (2005). Similarity and helping behavior on the web: The impact of the convergence of surnames between a solicitor and a subject in a request made by e-mail. Journal of Applied Social Psychology, 35, 423–429. Gunnell, J. J., & Ceci, S. J. (2010). When emotionality trumps reason: A study of individual processing style and juror bias. Behavioral Sciences & the Law, 28, 850–877. Guo, B. L., Aveyard, P., Fielding, A., & Sutton, S. (2009). Do the transtheoretical model processes of change, decisional balance and temptation predict stage movement? Evidence from smoking cessation in adolescents. Addiction, 104, 828–838. Hackman, C. L., & Knowlden, A. P. (2014). Theory of reasoned action and theory of planned behavior-based dietary interventions in adolescents and young adults: A systematic review. Adolescent Health, 487



Medicine, and Therapeutics, 5, 101–114. Hackman, J. R., & Anderson, L. R. (1968). The strength, relevance, and source of beliefs about an object in Fishbein’s attitude theory. Journal of Social Psychology, 76, 55–67. Haddock, G., Mio, G. R., Arnold, K., & Huskinson, T. (2008). Should persuasion be affective or cognitive? The moderating effects of need for affect and need for cognition. Personality and Social Psychology Bulletin, 34, 769–779. Haddock, G., & Zanna, M. P. (1998). Assessing the impact of affective and cognitive information in predicting attitudes toward capital punishment. Law and Human Behavior, 22, 325–339. Hagen, K. M., Gutkin, T. B., Wilson, C. P., & Oats, R. G. (1998). Using vicarious experience and verbal persuasion to enhance self-efficacy in pre-service teachers: “Priming the pump” for consultation. School Psychology Quarterly, 13, 169–178. Hagger, M. S., & Chatzisarantis, N. L. D. (2009). Integrating the theory of planned behaviour and self-determination theory in health behaviour: A meta-analysis. British Journal of Health Psychology, 14, 275–302. Hagger, M. S., Chatzisarantis, N. L. D., & Biddle, S. J. H. (2002). A metaanalytic review of the theories of reasoned action and planned behavior in physical activity: Predictive validity and the contribution of additional variables. Journal of Sport and Exercise Psychology, 24, 3–32. Hale, J. L., Householder, B. J., & Greene, K. L. (2002). The theory of reasoned action. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 259–286). Thousand Oaks, CA: Sage.



488



Hall, K. L., & Rossi, J. S. (2008). Meta-analytic examination of the strong and weak principles across 48 health behaviors. Preventive Medicine, 46, 266–274. Hall, P. A., Fong, G. T., Epp, L. J., & Elias, L. J. (2008). Executive function moderates the intention-behavior link for physical activity and dietary behavior. Psychology and Health, 23, 309–326. Hall, P. A., Zehr, C. E., Ng, M., & Zanna, M. P. (2012). Implementation intentions for physical activity in supportive and unsupportive environmental conditions: An experimental examination of intention– behavior consistency. Journal of Experimental Social Psychology, 48, 432–436. Hamilton, D. L., & Zanna, M. P. (1972). Differential weighting of favorable and unfavorable attributes in impressions of personality. Journal of Experimental Research in Personality, 6, 204–212. Hänze, M. (2001). Ambivalence, conflict, and decision making: Attitudes and feelings in Germany towards NATO’s military intervention in the Kosovo war. European Journal of Social Psychology, 31, 693–706. Hardeman, W., Johnson, M., Johnston, D. W., Bonetti, D., Wareham, N. J., & Kinmonth, A. L. (2002). Application of the theory of planned behaviour in behaviour change interventions: A systematic review. Psychology and Health, 17, 123–158. Harland, P., Staats, H., & Wilke, H. A. M. (1999). Explaining proenvironmental intention and behavior by personal norms and the theory of planned behavior. Journal of Applied Social Psychology, 29, 2505–2528. Harmon, R. R., & Coney, K. A. (1982). The persuasive effects of source credibility in buy and lease situations. Journal of Marketing Research, 19, 255–260.



489



Harmon-Jones, E. (1999). Toward an understanding of the motivation underlying dissonance effects: Is the production of aversive consequences necessary? In E. Harmon-Jones & J. Mills (Eds.), Cognitive dissonance: Progress on a pivotal theory in social psychology (pp. 71–99). Washington, DC: American Psychological Association. Harmon-Jones, E. (2002). A cognitive dissonance theory perspective on persuasion. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 99–116). Thousand Oaks, CA: Sage. Harmon-Jones, E., Amodio, D. M., & Harmon-Jones, C. (2010). Actionbased model of dissonance: On cognitive conflict and attitude change. In J. P. Forgas, J. Cooper, & W. D. Crano (Eds.), The psychology of attitudes and attitude change (pp. 163–181). New York: Psychology Press. Harmon-Jones, E., Brehm, J. W., Greenberg, J., Simon, L., & Nelson, D. E. (1996). Evidence that the production of aversive consequences is not necessary to create cognitive dissonance. Journal of Personality and Social Psychology, 70, 5–16. Harmon-Jones, E., & Mills, J. (Eds.). (1999). Cognitive dissonance: Progress on a pivotal theory in social psychology. Washington, DC: American Psychological Association. Harris, A. J. L., & Hahn, U. (2009). Bayesian rationality in evaluating multiple testimonies: Incorporating the role of coherence. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, 1366–1373. Harris, P. R. (2011). Self-affirmation and the self-regulation of health behavior change. Self and Identity, 10, 304–314. Harris, P. R., & Epton, T. (2009). The impact of self-affirmation on health cognition, health behaviour, and other health-related responses: A 490



narrative review. Social and Personality Psychology Compass, 3, 962–978. Harris, P. R., & Epton, T. (2010). The impact of self-affirmation on healthrelated cognition and health behaviour: Issues and prospects. Social and Personality Psychology Compass, 4, 439–454. Harris, P. R., Mayle, K., Mabbott, L., & Napper, L. (2007). Selfaffirmation reduces smokers’ defensiveness to graphic on-pack cigarette warning labels. Health Psychology, 26, 437–446. Hart, W., Albarracín, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009). Feeling validated versus being correct: A metaanalysis of selective exposure to information. Psychological Bulletin, 135, 555–588. Harte, T. (1976). The effects of evidence in persuasive communication. Central States Speech Journal, 27, 42–46. Harvey, O., & Rutherford, J. (1958). Gradual and absolute approaches to attitude change. Sociometry, 21, 61–68. Hass, J. W., Bagley, G. S., & Rogers, R. W. (1975). Coping with the energy crisis: Effects of fear appeals upon attitudes toward energy consumption. Journal of Applied Psychology, 60, 754–756. Hass, R. G. (1981). Effects of source characteristics on cognitive responses and persuasion. In R. E. Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 141–172). Hillsdale, NJ: Lawrence Erlbaum. Hass, R. G., & Grady, K. (1975). Temporal delay, type of forewarning, and resistance to influence. Journal of Experimental Social Psychology, 11, 459–469.



491



Hass, R. G., & Linder, D. E. (1972). Counterargument availability and the effects of message structure on persuasion. Journal of Personality and Social Psychology, 23, 219–233. Hassandra, M., Viachopoulos, S. P., Kosmidou, E., Hatzigeorgiadis, A., Goudas, M., & Theodorakis, Y. (2011). Predicting students’ intention to smoke by theory of planned behaviour variables and parental influences across school grade levels. Psychology & Health, 26, 1241–1258. Haugtvedt, C. P., Schumann, D. W., Schneier, W. L., & Warren, W. L. (1994). Advertising repetition and variation strategies: Implications for understanding attitude strength. Journal of Consumer Research, 21, 176–189. Haugtvedt, C. P., Shakarchi, R. J., Samuelson, B. M., & Liu, K. (2004). Consumer psychology and attitude change. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 283–297). Mahwah, NJ: Lawrence Erlbaum. Hausenblas, H. A., Carron, A. V., & Mack, D. E. (1997). Application of the theories of reasoned action and planned behavior to exercise behavior: A meta-analysis. Journal of Sport and Exercise Psychology, 19, 36–51. Head, K. J., Noar, S. M., Iannarino, N. T., & Harrington, N. G. (2013). Efficacy of text messaging-based interventions for health promotion: A meta-analysis. Social Science and Medicine, 97, 41–48. Hecht, M. L., Graham, J. W., & Elek, E. (2006). The drug resistance strategies intervention: Program effects on substance use. Health Communication, 20, 267–276. Hedeker, D., Flay, B. R., & Petraitis, J. (1996). Estimating individual influences of behavioral intentions: An application of random-effects modeling to the theory of reasoned action. Journal of Consulting and Clinical Psychology, 64, 109–120. 492



Hefner, D., Rothmund, T., Klimmt, C., & Gollwitzer, M. (2011). Implicit measures and media effects research: Challenges and opportunities. Communication Methods and Measures, 5, 181–202. Heider, F. (1946). Attitudes and cognitive organization. Journal of Psychology, 21, 107–112. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Heitland, K., & Bohner, G. (2010). Reducing prejudice via cognitive dissonance: Individual differences in preference for consistency moderate the effects of counter-attitudinal advocacy. Social Influence, 5, 164–181. Henley, N., & Donovan, R. J. (2003). Young people’s response to death threat appeals: Do they really feel immortal? Health Education Research, 18, 1–14. Herek, G. M. (1987). Can functions be measured? A new perspective on the functional approach to attitudes. Social Psychology Quarterly, 50, 285–303. Herek, G. M. (2000). The social construction of attitudes: Functional consensus and divergence in the U.S. public’s reactions to AIDS. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 325–364). Mahwah, NJ: Lawrence Erlbaum. Herek, G. M., & Capitanio, J. P. (1998). Symbolic prejudice or fear of infection? A functional analysis of AIDS-related stigma among heterosexual adults. Basic and Applied Social Psychology, 20, 230–241. Herr, P. M. (1995). Whither fact, artifact, and attitude: Reflections on the theory of reasoned action. Journal of Consumer Psychology, 4, 371–380. 493



Herzog, T. A. (2008). Analyzing the transtheoretical model using the framework of Weinstein, Rothman, and Sutton (1998): The example of smoking cessation. Health Psychology, 27, 548–556. Hether, H. J., Huang, G. C., Beck, V., Murphy, S. T., & Valente, T. W. (2008). Entertainment-education in a media-saturated environment: Examining the impact of single and multiple exposures to breast cancer storylines on two popular medical dramas. Journal of Health Communication, 13, 808–823. Hettema, J. E., & Hendricks, P. S. (2010). Motivational interviewing for smoking cessation: A meta-analytic review. Journal of Consulting and Clinical Psychology, 78, 868–884. Hetts, J. J., Boninger, D. S., Armor, D. A., Gleicher, F., & Nathanson, A. (2000). The influence of anticipated counterfactual regret on behavior. Psychology and Marketing, 17, 345–368. Hewes, D. E. (1983). Confessions of a methodological puritan: A response to Jackson and Jacobs. Human Communication Research, 9, 187–191. Hewgill, M. A., & Miller, G. R. (1965). Source credibility and response to feararousing communications. Speech Monographs, 32, 95–101. Hibbert, S., Smith, A., Davies, A., & Ireland, F. (2007). Guilt appeals: Persuasion knowledge and charitable giving. Psychology and Marketing, 24, 723–742. Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. Advances in Experimental Social Psychology, 30, 1–46. Highhouse, S. (2009). Designing experiments that generalize. Organizational Research Methods, 12, 554–566.



494



Hilligoss, B., & Rieh, S. Y. (2008). Developing a unifying framework of credibility assessment: Construct, heuristics, and interaction in context. Information Processing & Management, 44, 1467–1484. Hilmert, C. J., Kulik, J. A., & Christenfeld, N. J. S. (2006). Positive and negative opinion modeling: The influence of another’s similarity and dissimilarity. Journal of Personality & Social Psychology, 90, 440–452. Himmelfarb, S., & Arazi, D. (1974). Choice and source attractiveness in exposure to discrepant messages. Journal of Experimental Social Psychology, 10, 516–527. Hing, L. S. S., Li, W., & Zanna, M. P. (2002). Inducing hypocrisy to reduce prejudicial responses among aversive racists. Journal of Experimental Social Psychology, 38, 71–78. Hinyard, L. J., & Kreuter, M. W. (2007). Using narrative communication as a tool for health behavior change: A conceptual, theoretical, and empirical overview. Health Education and Behavior, 34, 777–792. Hirsh, J. B., Kang, S. K., & Bodenhausen, G. V. (2012). Personalized persuasion: Tailoring persuasive appeals to recipient personality traits. Psychological Science, 23, 578–581. Hodson, G., Maio, G. R., & Esses, V. M. (2001). The role of attitudinal ambivalence in susceptibility to consensus information. Basic and Applied Social Psychology, 23, 197–205. Hoeken, H., & Geurts, D. (2005). The influence of exemplars in fear appeals on the perception of self-efficacy and message acceptance. Information Design Journal + Document Design, 13, 238–248. Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA: Sage. 495



Høie, M., Moan, I. S., Rise, J., & Larsen, E. (2012). Using an extended version of the theory of planned behaviour to predict smoking cessation in two age groups. Addiction Research and Theory, 20, 42–54. Holbrook, M. B. (1977). Comparing multi-attribute attitude models by optimal scaling. Journal of Consumer Research, 4, 165–171. Holbrook, M. B. (1978). Beyond attitude structure: Toward the informational determinants of attitude. Journal of Marketing Research, 15, 545–556. Holbrook, M. B., & Hulbert, J. M. (1975). Multi-attribute attitude models: A comparative analysis. Advances in Consumer Research, 2, 375–388. Hong, S., & Park, H. S. (2012). Computer-mediated persuasion in online reviews: Statistical versus narrative evidence. Computers in Human Behavior, 28, 906–919. Hong, T. (2006). The influence of structural and message features on Web site credibility. Journal of the American Society for Information Science and Technology, 57, 114–127. Hopfer, S. (2012). Effects of narrative HPV vaccination intervention aimed at reaching college women: A randomized controlled trial. Prevention Science, 13, 173–182. Horai, J., Naccari, N., & Fatoullah, E. (1974). The effects of expertise and physical attractiveness upon opinion agreement and liking. Sociometry, 37, 601–606. Hornikx, J., & Hoeken, H. (2007). Cultural differences in the persuasiveness of evidence types and evidence quality. Communication Monographs, 74, 443–463.



496



Hornikx, J., & O’Keefe, D. J. (2009). Adapting consumer advertising appeals to cultural values: A meta-analytic review of effects on persuasiveness and ad liking. Communication Yearbook, 33, 39–71. Hosseinzadeh, H., & Hossain, S. Z. (2011). Functional analysis of HIV/AIDS stigma: Consensus or divergence? Health Education & Behavior, 38, 584–595. Hovland, C. I., Harvey, O. J., & Sherif, M. (1957). Assimilation and contrast effects in reactions to communication and attitude change. Journal of Abnormal and Social Psychology, 55, 244–252. Howard, D. J., & Kerin, R. A. (2011). The effects of name similarity on message processing and persuasion. Journal of Experimental Social Psychology 47, 63–71. Howell, J. L., & Shepperd, J. A. (2012). Reducing information avoidance through affirmation. Psychological Science, 23, 141–145. Hoyt, W. T. (1996). Antecedents and effects of perceived therapist credibility: A meta-analysis. Journal of Counseling Psychology, 43, 430–447. Hsieh, G., Hudson, S. E., & Kraut, R. E. (2011). Donate for credibility: How contribution incentives can improve credibility. In Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI) (pp. 3435–3438). New York: Association for Computing Machinery. doi:10.1145/1978942.1979454. Hu, Y. F., & Sundar, S. S. (2010). Effects of online health sources on credibility and behavioral intentions. Communication Research, 37, 105–132. Hübner, G., & Kaiser, F. G. (2006). The moderating role of the attitudesubjective norms conflict on the link between moral norms and 497



intention. European Psychologist, 11, 99–109. Huhmann, B. A., & Brotherton, T. P. (1997). A content analysis of guilt appeals in popular magazine advertisements. Journal of Advertising, 26(2), 35–45. Hukkelberg, S. S., Hagtvet, K. A., & Kovac, V. B. (2014). Latent interaction effects in the theory of planned behaviour applied to quitting smoking. British Journal of Health Psychology, 19, 83–100. Hullett, C. R. (2002). Charting the process underlying the change of valueexpressive attitudes: The importance of value-relevance in predicting the matching effect. Communication Monographs, 69, 158–178. Hullett, C. R. (2004). Using functional theory to promote sexually transmitted disease (STD) testing: The impact of value-expressive messages and guilt. Communication Research, 31, 363–396. Hullett, C. R. (2005). The impact of mood on persuasion: A meta-analysis. Communication Research, 32, 423–442. Hullett, C. R. (2006). Using functional theory to promote HIV testing: The impact of value-expressive messages, uncertainty, and fear. Health Communication, 20, 57–67. Hullett, C. R., & Boster, F. J. (2001). Matching messages to the values underlying value-expressive and social-adjustive attitudes: Reconciling an old theory with a contemporary measurement approach. Communication Monographs, 68, 133–153. Hunt, S. D. (1970). Post-transaction communications and dissonance reduction. Journal of Marketing, 34(3), 46–51. Hurwitz, J. (1986). Issue perception and legislative decision making: An 498



application of social judgment theory. American Politics Quarterly, 14, 150–185. Hurwitz, S. D., Miron, M. S., & Johnson, B. T. (1992). Source credibility and the language of expert testimony. Journal of Applied Social Psychology, 24, 1909–1939. Huskinson, T. L. H., & Haddock, G. (2004). Individual differences in attitude structure: Variance in the chronic reliance on affective and cognitive information. Journal of Experimental Social Psychology, 40, 82–90. Hustinx, L., van Enschot, R., & Hoeken, H. (2007). Argument quality in the elaboration likelihood model: An empirical study of strong and weak arguments in a persuasive message. In F. H. van Eemeren, J. A. Blair, C. A. Willard, & B. Garssen (Eds.), Proceedings of the Sixth Conference of the International Society for the Study of Argumentation (pp. 651–657). Amsterdam: Sic Sat. Huston, T. L., & Levinger, G. (1978). Interpersonal attraction and relationships. Annual Review of Psychology, 29, 115–156. Hutchison, A. J., Breckon, J. D., & Johnston, L. H. (2009). Physical activity behavior change interventions based on the transtheoretical model: A systematic review. Health Education and Behavior, 36, 829–845. Hyde, J., Hankins, M., Deale, A., & Marteau, T. M. (2008). Interventions to increase self-efficacy in the context of addiction behaviours: A systematic literature review. Journal of Health Psychology, 13, 607–623. Igartua, J.-J. (2010). Identification with characters and narrative persuasion through fictional feature films. Communications, 35, 347–373.



499



Igartua, J.-J., & Barrios, I. (2012). Changing real-world beliefs with controversial movies: Processes and mechanisms of narrative persuasion. Journal of Communication, 62, 514–531. Infante, D. A. (1978). Similarity between advocate and receiver: The role of instrumentality. Central States Speech Journal, 29, 187–193. Ioannidis, J. P. A. (2005). Why most published research findings are false. PLOS Medicine, 2, 696–701. Ioannidis, J. P. A. (2008). Why most discovered true associations are inflated. Epidemiology, 19, 640–648. Ito, T. A., & Cacioppo, J. T. (2007). Attitudes as mental and neural states of readiness: Using physiological measures to study implicit attitudes. In B. Wittenbrink & N. Schwarz (Eds.), Implicit measures of attitudes (pp. 125–158). New York: Guilford. Ivanov, B. (2012). Designing inoculation messages for health communication campaigns. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 73–93). Los Angeles: Sage. Iyengar, S., & Hahn, K. S. (2009). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59, 19–39. Izuma, K., & Murayama, K. (2013). Choice-induced preference change in the free-choice paradigm: A critical methodological review. Frontiers in Psychology, 4, 41. Jaccard, J., Radecki, C., Wilson, T., & Dittus, P. (1995). Methods for identifying consequential beliefs: Implications for understanding attitude strength. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 337–359). Mahwah, NJ: Lawrence Erlbaum. 500



Jackson, S. (1992). Message effects research: Principles of design and analysis. New York: Guilford. Jackson, S. (1993). How to do things to words: The experimental manipulation of message variables. Southern Communication Journal, 58, 103–114. Jackson, S., & Brashers, D. (1994). M > 1: Analysis of treatment × replication designs. Human Communication Research, 20, 356–389. Jackson, S., & Jacobs, S. (1983). Generalizing about messages: Suggestions for design and analysis of experiments. Human Communication Research, 9, 169–181. Jackson, S., O’Keefe, D. J., & Brashers, D. (1994). The messages replication factor: Methods tailored to messages as objects of study. Journalism Quarterly, 71, 984–996. Jacobson, R. P., Mortensen, C. R., & Cialdini, R. B. (2011). Bodies obliged and unbound: Differentiated response tendencies for injunctive and descriptive social norms. Journal of Personality and Social Psychology, 100, 433–448. Janis, I. L., & Mann, L. (1968). A conflict-theory approach to attitude change and decision making. In A. G. Greenwald, T. C. Brock, & T. M. Ostrom (Eds.), Psychological foundations of attitudes (pp. 327–360). New York: Academic Press. Janssen, L., Fennis, B. M., Pruyn, A. T. H., & Vohs, K. D. (2008). The path of least resistance: Regulatory resource depletion and the effectiveness of social influence techniques. Journal of Business Research, 61, 1041–1046. Jemmott, J. B. (2012). The reasoned action approach in HIV risk-reduction strategies for adolescents. Annals of the American Academy of Political 501



and Social Science, 640, 150–172. Jensen, C. D., Cushing, C. C., Aylward, B. S., Craig, J. T., Sorell, D. M., & Steele, R. G. (2011). Effectiveness of motivational interviewing interventions for adolescent substance use behavior change: A metaanalytic review. Journal of Consulting and Clinical Psychology, 79, 433–440. Jensen, J. D. (2008). Scientific uncertainty in news coverage of cancer research: Effects of hedging on scientists’ and journalists’ credibility. Human Communication Research, 34, 347–369. Jeong, E. S., Shi, Y., Baazova, A., Chiu, C., Nahai, A., Moons, W. G., & Taylor, S. E. (2011). The relation of approach/avoidance motivation and message framing to the effectiveness of charitable appeals. Social Influence, 6, 15–21. Jeong, S.-H., & Hwang, Y. (2012), Does multitasking increase or decrease persuasion? Effects of multitasking on comprehension and counterarguing. Journal of Communication, 62, 571–587. Jerit, J. (2009). How predictive appeals affect policy opinions. American Journal of Political Science, 53, 411–426. Jessop, D. C., Simmonds, L.V., & Sparks, P. (2009). Motivational and behavioral consequences of self-affirmation interventions: A study of sunscreen use among women. Psychology & Health, 24, 529–544. Jiang, L., Hoegg, J., Dahl, D.W., & Chattopadhyay, A. (2010). The persuasive role of incidental similarity on attitudes and purchase intentions in a sales context. Journal of Consumer Research, 36, 778–791. John, L. K., Loewenstein, G., & Prelec, D. (2012). Measuring the prevalence of questionable research practices with incentives for truth 502



telling. Psychological Science, 23, 524–532. Johnson, B. B. (2005). Testing and expanding a model of cognitive processing of risk information. Risk Analysis, 25, 631–650. Johnson, B. T. (1994). Effects of outcome-relevant involvement and prior information on persuasion. Journal of Experimental Social Psychology, 30, 556–579. Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-analysis. Psychological Bulletin, 106, 290–314. Johnson, B. T., & Eagly, A. H. (1990). Involvement and persuasion: Types, traditions, and the evidence. Psychological Bulletin, 107, 375–384. Johnson, B. T., Lin, H.-Y., Symons, C. S., Campbell, L. A., & Ekstein, G. (1995). Initial beliefs and attitudinal latitudes as factors in persuasion. Personality and Social Psychology Bulletin, 21, 502–511. Johnson, B. T., Smith-McLallen, A., Killeya, L. A., & Levin, K. D. (2004). Truth or consequences: Overcoming resistance to persuasion with positive thinking. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 215–233). Mahwah, NJ: Lawrence Erlbaum. Johnson, H. H., & Scileppi, J. A. (1969). Effects of ego-involvement conditions on attitude change to high and low credibility communicators. Journal of Personality and Social Psychology, 13, 31–36. Johnson, R. W., Kelly, R. J., & LeBlanc, B. A. (1995). Motivational basis of dissonance: Aversive consequences or inconsistency. Personality and Social Psychology Bulletin, 21, 850–855. [Erratum notice: Personality and Social Psychology Bulletin, 22 (1996), 222.]



503



Jonas, K., Broemer, P., & Diehl, M. (2000). Experienced ambivalence as a moderator of the consistency between attitudes and behaviors. Zeitschrift fur Sozialpsychologie, 31, 153–165. Jonas, K., Diehl, M., & Brömer, P. (1997). Effects of attitudinal ambivalence on information processing and attitude-intention consistency. Journal of Experimental Social Psychology, 33, 190–210. Jones, J. L., & Leary, M. R. (1994). Effects of appearance-based admonitions against sun exposure on tanning intentions in young adults. Health Psychology, 13, 86–90. Jones, R. A., & Brehm, J. W. (1967). Attitudinal effects of communicator attractiveness when one chooses to listen. Journal of Personality and Social Psychology, 6, 64–70. Jordan, A., Piotrowski, J. T., Bleakley, A., & Mallya, G. (2012). Developing media interventions to reduce household sugar-sweetened beverage consumption. Annals of the American Academy of Political and Social Science, 640, 118–135. Joule, R. V., & Martinie, M. A. (2008). Forced compliance, misattribution and trivialization. Social Behavior and Personality, 36, 1205–1212. Judah, G., Gardner, B., & Aunger, R. (2013). Forming a flossing habit: An exploratory study of the psychological determinants of habit formation. British Journal of Health Psychology, 18, 338–353. Judd, C. M., Kenny, D. A., & Krosnick, J. A. (1983). Judging the positions of political candidates: Models of assimilation and contrast. Journal of Personality and Social Psychology, 44, 952–963. Judd, C. M., Westfall, J., & Kenny, D. A. (2012). Treating stimuli as a random factor in social psychology: A new and comprehensive solution to a pervasive but largely ignored problem. Journal of Personality and 504



Social Psychology, 103, 54–69. Julka, D. L., & Marsh, K. L. (2005). An attitude functions approach to increasing organdonation participation. Journal of Applied Social Psychology, 35, 821–849. Jung, J. M., & Kellaris, J. M. (2004). Cross-national differences in proneness to scarcity effects: The moderating roles of familiarity, uncertainty avoidance, and need for cognitive closure. Psychology and Marketing, 21, 739–753. Jung, T. J., & Heald, G. R. (2009). The effects of discriminate message interventions on behavioral intentions to engage in physical activities. Journal of American College Health, 57, 527–535. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–291. Kaiser, F. G., Hübner, G., & Bogner, F. X. (2005). Contrasting the theory of planned behavior with the value-belief-norm model in explaining conservation behavior. Journal of Applied Social Psychology, 35, 2150–2170. Kaiser, F. G., & Schultz, P. W. (2009). The attitude-behavior relationship: A test of three models of the moderating role of behavioral difficulty. Journal of Applied Social Psychology, 39, 186–207. Kaiser, H. F. (1960). Directional statistical decisions. Psychological Review, 67, 160–167. Kamins, M. A. (1990). An investigation into the “match-up” hypothesis in celebrity advertising: When beauty may be only skin deep. Journal of Advertising, 19(1), 4–13.



505



Kamins, M. A., & Assael, H. (1987a). Moderating disconfirmation of expectations through the use of two-sided appeals: A longitudinal approach. Journal of Economic Psychology, 8, 237–254. Kamins, M. A., & Assael, H. (1987b). Two-sided versus one-sided appeals: A cognitive perspective on argumentation, source derogation, and the effect of disconfirming trial on belief change. Journal of Marketing Research, 24, 29–39. Kamins, M. A., & Marks, L. J. (1987). Advertising puffery: The impact of using two-sided claims on product attitude and purchase intention. Journal of Advertising, 16(4), 6–15. Kang, Y., Cappella, J., & Fishbein, M. (2006). The attentional mechanism of message sensation value: Interaction between message sensation value and argument quality on message effectiveness. Communication Monographs, 73, 351–378. Kang, Y.-S., & Kerr, P. M. (2006). Beauty and the beholder: Toward an integrative model of communication source effects. Journal of Consumer Research, 33, 123–130. Kantola, S. J., Syme, G. J., & Campbell, N. A. (1982). The role of individual differences and external variables in a test of the sufficiency of Fishbein’s model to explain behavioral intentions to conserve water. Journal of Applied Social Psychology, 12, 70–83. Kantola, S. J., Syme, G. J., & Campbell, N. A. (1984). Cognitive dissonance and energy conservation. Journal of Applied Psychology, 69, 416–421. Kao, D. T. (2007). Conclusion explicitness in message communication: The roles of NFC and knowledge in attitude formation. Social Behavior and Personality, 35, 819–826.



506



Kaplowitz, S. A., & Fink, E. L. (1997). Message discrepancy and persuasion. In G. A. Barnett & F. J. Boster (Eds.), Progress in communication sciences: Vol. 13. Advances in persuasion (pp. 75–106). Greenwich, CT: Ablex. Karlan, D., McConnell, M., Mullainathan, S., & Zinman, J. (2010). Getting to the top of mind: How reminders increase saving. National Bureau of Economic Research Working Paper No. 16205. Retrieved from http://www.nber.org/papers/w16205.pdf. Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24, 163–204. Katz, D., McClintock, C., & Sarnoff, I. (1957). The measurement of ego defense as related to attitude change. Journal of Personality, 25, 465–474. Keaveney, S. M., Huber, F., & Hemnann, A. (2007). A model of buyer regret: Selected prepurchase and postpurchase antecedents with consequences for the brand and the channel. Journal of Business Research, 12, 1207–1215. Keer, M., van den Putte, B., & Neijens, P. (2010). The role of affect and cognition in health decision making. British Journal of Social Psychology, 49, 143–153. Kellar, I., & Abraham, C. (2005). Randomized controlled trial of a brief research-based intervention promoting fruit and vegetable consumption. British Journal of Health Psychology, 10, 543–558. Keller, P. A., & Lehmann, D. R. (2008). Designing effective health communications: A meta-analysis. Journal of Public Policy and Marketing, 27, 117–130. Kelly, J. A., St. Lawrence, J. S., Stevenson, Y., Hauth, A. C., Kalichman, 507



S. C., Diaz, Y. E., … Morgan, M. G. (1992). Community AIDS/HIV risk reduction: The effects of endorsements by popular people in three cities. American Journal of Public Health, 82, 1483–1489. Kelly, K. J., Slater, M. D., & Karan, D. (2002). Image advertisements’ influence on adolescents’ perceptions of the desirability of beer and cigarettes. Journal of Public Policy and Marketing, 21, 295–304. Keltner, D., & Buswell, B. N. (1996). Evidence for the distinctness of embarrassment, shame, and guilt: A study of recalled antecedents and facial expressions of emotion. Cognition and Emotion, 10, 155–171. Kendzierski, D., & Whitaker, D. J. (1997). The role of self-schema in linking intentions with behavior. Personality and Social Psychology Bulletin, 23, 139–147. Keng, C.-J., & Liao, T.-H. (2009). Consequences of postpurchase dissonance: The mediating role of an external information search. Social Behavior and Personality, 37, 1327–1340. Kennedy, C., O’Reilly, K., & Sweat, M. (2009). Effectiveness of peer education interventions for HIV prevention in developing countries: A systematic review and meta-analysis. AIDS Education and Prevention, 21, 181–206. Kennedy, M. G., O’Leary, A., Beck, V., Pollard, K., & Simpson, P. (2004). Increases in calls to the CDC National STD and AIDS Hotline following AIDS-related episodes in a soap opera. Journal of Communication, 54, 287–301. Kennedy, N. B. (1982). Contact donors before making solicitations. Fund Raising Management, 13(9), 16–17. Kenski, K., & Fishbein, M. (2005). The predictive benefits of importance: Do importance ratings improve the prediction of political attitudes? 508



Journal of Applied Social Psychology, 35, 487–507. Kenworthy, J. B., Miller, N., Collins, B. E., Read, S. J., & Earleywine, M. (2011). A trans-paradigm theoretical synthesis of cognitive dissonance theory: Illuminating the nature of discomfort. European Review of Social Psychology, 22, 36–113. Kesselheim, A. S., Robertson, C. T., Myers, J. A., Rose, S. L., Gillet, V., Ross, K. M., … Avorn, J. (2012). A randomized study of how physicians interpret research funding disclosures. New England Journal of Medicine, 367, 1119–1127. Kessels, L. T. E., & Ruiter, R. A. C. (2012). Eye movement responses to health messages on cigarette packages. BMC Public Health, 12, 352–360. Kidder, L. H., & Campbell, D. T. (1970). The indirect testing of social attitudes. In G. F. Summers (Ed.), Attitude measurement (pp. 333–385). Chicago: Rand McNally. Kidwell, B., & Jewell, R. D. (2003). The moderated influence of internal control: An examination across health-related behaviors. Journal of Consumer Psychology, 13, 377–386. Kiesler, C. A., Collins, B. E., & Miller, N. (1969). Attitude change: A critical analysis of theoretical approaches. New York: Wiley. Kim, A., Stark, E., & Borgida, E. (2011). Symbolic politics and the prediction of attitudes toward federal regulation of reduced-exposure tobacco products. Journal of Applied Social Psychology, 41, 381–400. Kim, H. S., Bigman, C. A., Leader, A. E., Lerman, C., & Cappella, J. N. (2012). Narrative health communication and behavior change: The influence of exemplars in the news on intention to quit smoking. Journal of Communication, 62, 473–492. 509



Kim, M.-S., & Hunter, J. E. (1993a). Attitude-behavior relations: A metaanalysis of attitudinal relevance and topic. Journal of Communication, 43(1), 101–142. Kim, M.-S., & Hunter, J. E. (1993b). Relationships among attitudes, behavioral intentions, and behavior: A meta-analysis of past research, part 2. Communication Research, 20, 331–364. Kim, S., McLeod, J. H., & Shantzis, C. (1989). An outcome evaluation of refusal skills program as a drug abuse prevention strategy. Journal of Drug Education, 19, 363–371. King, A. J., Williams, E. A., Harrison, T. R., Morgan, S. E., & Havermahl, T. (2012). The “Tell Us Now” campaign for organ donation: Using message immediacy to increase donor registration rates. Journal of Applied Communication Research, 40, 229–246. King, M. (1978). Assimilation and contrast of presidential candidates’ issue positions, 1972. Public Opinion Quarterly, 41, 515–522. King, S. W., & Sereno, K. K. (1973). Attitude change as a function of degree and type of interpersonal similarity and message type. Western Speech, 37, 218–232. King, W. R., & He, J. (2006). A meta-analysis of the technology acceptance model. Information & Management, 43, 740–755. Kiviniemi, M. T., Voss-Humke, A. M., & Seifert, A. L. (2007). How do I feel about the behavior? The interplay of affective associations with behaviors and cognitive beliefs as influences on physical activity behavior. Health Psychology, 26, 152–158. Klass, E. T. (1978). Psychological effects of immoral actions: The experimental evidence. Psychological Bulletin, 85, 756–771.



510



Klein, W. M. P., & Harris, P. R. (2009). Self-affirmation enhances attentional bias toward threatening components of a persuasive message. Psychological Science, 20, 1463–1467. Klein, W. M. P., Lipkus, I. M., Scholl, S. M., McQueen, A., Cerully, J. L., & Harris, P. R. (2010). Self-affirmation moderates effects of unrealistic optimism and pessimism on reactions to tailored risk feedback. Psychology & Health, 25, 1195–1208. Klock, S. J., & Traylor, M. B. (1983). Older and younger models in advertising to older consumers: An advertising effectiveness experiment. Akron Business and Economic Review, 14(4), 48–52. Knäuper, B., McCollam, A., Rosen-Brown, A., Lacaille, J., Kelso, E., & Roseman, M. (2011). Fruitful plans: Adding targeted mental imagery to implementation intentions increases fruit consumption. Psychology and Health, 26, 601–617. Knäuper, B., Roseman, M., Johnson, P.J., & Krantz, L.H. (2009). Using mental imagery to enhance the effectiveness of implementation intentions. Current Psychology, 28, 181–186. Knight, K. M., McGowan, L., Dickens, C., & Bundy, C. (2006). A systematic review of motivational interviewing in physical health care settings. British Journal of Health Psychology, 11, 319–332. Knowles, E. S., & Linn, J. A. (Eds.). (2004). Resistance and persuasion. Mahwah, NJ: Lawrence Erlbaum. Koehler, J. W. (1972). Effects on audience opinion of one-sided and twosided speeches supporting and opposing a proposition. In T. D. Beisecker & D. W. Parson (Eds.), The process of social influence (pp. 351–369). Englewood Cliffs, NJ: Prentice Hall. Koestner, R., Horberg, E. J., Gaudreau, P., Powers, T., Di Dio, P., Bryan, 511



C., … Salter, N. (2006). Bolstering implementation plans for the long haul: The benefits of simultaneously boosting self-concordance or selfefficacy. Personality and Social Psychology Bulletin, 32, 1547–1558. Kok, G., Hospers, H. J., Harterink, P., & De Zwart, O. (2007). Socialcognitive determinants of HIV risk-taking intentions among men who date men through the Internet. AIDS Care, 19, 410–417. Koring, M., Richert, J., Lippke, S., Parschau, L., Reuter, T., & Schwarzer, R. (2012). Synergistic effects of planning and self-efficacy on physical activity. Health Education & Behavior, 39, 152–158. Kothe, E. J., Mullan, B. A., & Amaratunga, R. (2011). Randomised controlled trial of a brief theory-based intervention promoting breakfast consumption. Appetite, 56, 148–155. Kotowski, M. R., Smith, S. W., Johnstone, P. M., & Pritt, E. (2011). Using the EPPM to create and evaluate the effectiveness of brochures to reduce the risk for noise-induced hearing loss in college students. Noise and Health, 13, 261–271. Kraemer, H. C., Kiernan, M., Essex, M., & Kupfer, D. J. (2008). How and why criteria defining moderators and mediators differ between the Baron & Kenny and MacArthur approaches. Health Psychology, 27, S101–S109. Kraus, S. J. (1995). Attitudes and the prediction of behavior: A metaanalysis of the empirical literature. Personality and Social Psychology Bulletin, 21, 58–75. Kreuter, M. W., Green, M. C., Cappella, J. N., Slater, M. D., Wise, M. E., Storey, D., … Wooley, S. (2007). Narrative communication in cancer prevention and control: A framework to guide research and application. Annals of Behavioral Medicine, 33, 221–235.



512



Kreuter, M. W., Lukwago, S. N., Bucholtz, D. C., Clark, E. M., & Sanders-Thompson, V. (2003). Achieving cultural appropriateness in health promotion programs: Targeted and tailored approaches. Health Education and Behavior, 30, 133–146. Krieger, J. L., Coveleski, S., Hecht, M. L., Miller-Day, M., Graham, J. W., Pettigrew, J., & Kootsikas, A. (2013). From kids, through kids, to kids: Examining the social influence strategies used by adolescents to promote prevention among peers. Health Communication, 28, 683–695. Krieger, J. L., & Sarge, M. A. (2013). A serial multiple mediation model of message framing on intentions to receive the human papillomavirus (HPV) vaccine: Revisiting the role of threat and efficacy perceptions. Health Communication, 28, 5–19. Krishnamurthy, P., & Sivaraman, A. (2002). Counterfactual thinking and advertising responses. Journal of Consumer Research, 28, 650–658. Krosnick, J. A., & Alwin, D. F. (1989). Aging and susceptibility to attitude change. Journal of Personality and Social Psychology, 57, 416–425. Krosnick, J. A., Boninger, D. S., Chuang, Y. C., Berent, M. K., & Carnot, C. G. (1993). Attitude strength: One construct or many related constructs? Journal of Personality and Social Psychology, 65, 1132–1151. Krosnick, J. A., Judd, C. M., & Wittenbrink, B. (2005). The measurement of attitudes. In D. Albarracín, B. T. Johnson, & M. P. Zanna (Eds.), Handbook of attitudes and attitude change (pp. 21–76). Mahwah, NJ: Lawrence Erlbaum. Krosnick, J. A., & Petty, R. E. (1995). Attitude strength: An overview. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 1–24). Mahwah, NJ: Lawrence Erlbaum.



513



Kruglanski, A. W., Chen, X., Pierro, A., Mannetti, L., Erb, H.-P., & Spiegel, S. (2006). Persuasion according to the unimodel: Implications for cancer communication. Journal of Communication, 56, S105–S122. Kruglanski, A. W., & Stroebe, W. (2005). The influence of beliefs and goals on attitudes: Issues of structure, function, and dynamics. In D. Albarracín, B. T. Johnson, & M. P. Zanna (Eds.), The handbook of attitudes (pp. 323–368). Mahwah, NJ: Lawrence Erlbaum. Kruglanski, A. W., & Thompson, E. P. (1999a). The illusory second mode or, the cue is the message. Psychological Inquiry, 10, 182–193. Kruglanski, A. W., & Thompson, E. P. (1999b). Persuasion by a single route: A view from the unimodel. Psychological Inquiry, 10, 83–109. Kuhlmann, A. K. S., Kraft, J. M., Galavotti, C., Creek, T. L., Mooki, M., & Ntumy, R. (2008). Radio role models for the prevention of mother-tochild transmission of HIV and HIV testing among pregnant women in Botswana. Health Promotion International, 23, 260–268. Kumkale, G. T., Albarracín, D., & Seignourel, P. J. (2010). The effects of source credibility in the presence or absence of prior attitudes: Implications for the design of persuasive communication campaigns. Journal of Applied Social Psychology, 40, 1325–1356. Kwak, L., Kremers, S. P. J., van Baak, M. A., & Brug, J. (2007). A posterbased intervention to promote stair use in blue- and white-collar worksites. Preventive Medicine, 45, 177–181. LaBrie, J. W., Grossbard, J. R., & Hummer, J. F. (2009). Normative misperceptions and marijuana use among male and female college athletes. Journal of Applied Sport Psychology, 21, S77–S85. Laczniak, R. N., Muehling, D. D., & Carlson, L. (1991). Effects of motivation and ability on ad-induced cognitive processing. In R. 514



Holman (Ed.), Proceedings of the 1991 Conference of the American Academy of Advertising (pp. 81–87). New York: D’Arcy Masius Benton & Bowles. Lafferty, B. A., Goldsmith, R. E., & Newell, S. J. (2002). The dual credibility model: The influence of corporate and endorser credibility on attitudes and purchase intentions. Journal of Marketing Theory and Practice, 10(3), 1–12. Lai, M. K., Ho, S. K., & Lam, T. H. (2004). Perceived peer smoking prevalence and its association with smoking behaviours and intentions in Hong Kong Chinese adolescents. Addiction, 99, 1195–1205. Lajunen, T., & Rasanen, M. (2004). Can social psychological models be used to promote bicycle helmet use among teenagers? A comparison of the health belief model, theory of planned behavior, and the locus of control. Journal of Safety Research, 35, 115–123. Lally, P., & Gardner, B. (2013). Promoting habit formation. Health Psychology Review, 7(S1), S137–S158. Lalor, K. M., & Hailey, B. J. (1990). The effects of message framing and feelings of susceptibility to breast cancer on reported frequency of breast self-examination. International Quarterly of Community Health Education, 10, 183–192. Lam, S.-P. (2006). Predicting intention to save water: Theory of planned behavior, response efficacy, vulnerability, and perceived efficiency of alternative solutions. Journal of Applied Social Psychology, 36, 2803–2824. Landman, J., & Petty, R. (2000). “It could have been you”: How states exploit counterfactual thought to market lotteries. Psychology and Marketing, 17, 299–321.



515



Landy, D. (1972). The effects of an overheard audience’s reaction and attractiveness on opinion change. Journal of Experimental Social Psychology, 8, 276–288. Lane, K. A., Banaji, M. R., Nosek, B. A., & Greenwald, A. G. (2007). Understanding and using the Implicit Association Test: What we know (so far) about the method. In B. Wittenbrink & N. Schwarz (Eds.), Implicit measures of attitudes (pp. 59–102). New York: Guilford. Langlois, M. A., Petosa, R., & Hallam, J. S. (1999). Why do effective smoking prevention programs work? Student changes in social cognitive theory constructs. Journal of School Health, 69, 326–331. LaPiere, R. T. (1934). Attitudes vs. actions. Social Forces, 13, 230–237. Larimer, M. E., Kaysen, D. L., Lee, C. M., Kilmer, J. R., Lewis, M. A., Dillworth, T., … Neighbors, C. (2009). Evaluating level of specificity of normative referents in relation to personal drinking behavior. Journal of Studies on Alcohol and Drugs, S16, 115–121. Larkey, L. K., & Gonzalez, J. (2007). Storytelling for promoting colorectal cancer prevention and early detection among Latinos. Patient Education and Counseling, 67, 272–278. Larkey, L. K. & Hecht, M. (2010). A model of effects of narrative as culture-centric health promotion. Journal of Health Communication, 15, 114–135. Larkey, L. K., & Hill, A. L. (2012). Using narratives to promote health: A culture-centric approach. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 95–112). Los Angeles: Sage. Latimer, A. E., & Ginis, K. A. M. (2005a). Change in self-efficacy following a single strength training session predicts sedentary older adults’ subsequent motivation to join a strength training program. 516



American Journal of Health Promotion, 20, 135–138. Latimer, A. E., & Ginis, K. A. M. (2005b). The importance of subjective norms for people who care what others think of them. Psychology and Health, 20, 53–62. Latimer, A. E., Salovey, P., & Rothman, A. J. (2007). The effectiveness of gain-framed messages for encouraging disease prevention behavior: Is all hope lost? Journal of Health Communication, 12, 645–649. Laufer, D., Silvera, D. H., McBride, J. B., & Schertzer, S. M. B. (2010). Communicating charity successes across cultures: Highlighting individual or collective achievement? European Journal of Marketing, 44, 1322–1333. Lauver, D., & Knapp, T. R. (1993). Sum-of-products variables: A methodological critique. Research in Nursing and Health, 16, 385–391. Lavidge, R. J., & Steiner, G. A. (1961). A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25(6), 59–62. Lavine, H., & Snyder, M. (1996). Cognitive processing and the functional matching effect in persuasion: The mediating role of subjective perceptions of message quality. Journal of Experimental Social Psychology, 32, 580–604. Lavine, H., & Snyder, M. (2000). Cognitive processes and the functional matching effect in persuasion: Studies of personality and political behavior. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 97–131). Mahwah, NJ: Lawrence Erlbaum. Lawton, R., Conner, M., & McEachan, R. (2009). Desire or reason: Predicting health behaviors from affective and cognitive attitudes. Health Psychology, 28, 56–65. 517



Leader, A. E., Weiner, J. L., Kelly, B. J., Hornik, R. C., & Cappella, J. N. (2009). Effects of information framing on human papillomavirus vaccination. Journal of Women’s Health, 18, 225–233. Leavitt, C., & Kaigler-Evans, K. (1975). Mere similarity versus information processing: An exploration of source and message interaction. Communication Research, 2, 300–306. Lechner, L., de Vries, H., & Offermans, N. (1997). Participation in a breast cancer screening program: Influence of past behavior and determinants on future screening participation. Preventive Medicine, 26, 473–482. Lee, E.-J. (2008). When are strong arguments stronger than weak arguments? Deindividuation effects on message elaboration in computer-mediated communication. Communication Research, 35, 646–665. Lee, G., & Lee, W. J. (2009). Psychological reactance to online recommendation services. Information & Management, 46, 448–452. Lee, M. J., & Bichard, S. L. (2006). Effective message design targeting college students for the prevention of binge-drinking: Basing design on rebellious risk-taking tendency. Health Communication, 20, 299–308. Lennon, S. J., Davis, L. L., & Fairhurst, A. (1988). Evaluations of apparel advertising as a function of self-monitoring. Perceptual and Motor Skills, 66, 987–996. Leone, L., Perugini, M., & Bagozzi, R. P. (2005). Emotions and decision making: Regulatory focus moderates the influence of anticipated emotions on action evaluation. Cognition and Emotion, 19, 1175–1198. Leshner, G., Bolls, P., & Thomas, E. (2009). Scare’em or disgust’em: The effects of graphic health promotion messages. Health Communication, 518



24, 447–451. Lester, R.T., Ritvo, P., Mills, E.J., Kariri, A., Karanja, S., Chung, M.H., … Plummer, F.A. (2010). Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): A randomised trial. The Lancet, 376, 1838–1845. Leventhal, H. (1970). Findings and theory in the study of fear communications. Advances in Experimental Social Psychology, 5, 119–186. Leventhal, H., Jones, S., & Trembly, G. (1966). Sex differences in attitude and behavior change under conditions of fear and specific instructions. Journal of Experimental Social Psychology, 2, 387–399. Leventhal, H., & Mora, P. A. (2008). Predicting outcomes or modeling process? Commentary on the health action process approach. Applied Psychology: International Review, 57, 51–65. Levin, I. P., & Gaeth, G. J. (1988). How consumers are affected by the framing of attribute information before and after consuming the product. Journal of Consumer Research, 15, 374–378. Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are not created equal: A typology and critical analysis of framing effects. Organizational Behavior and Human Decision Processes, 76, 149–188. Levin, K. D., Nichols, D. R., & Johnson, B. T. (2000). Involvement and persuasion: Attitude functions for the motivated processor. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 163–194). Mahwah, NJ: Lawrence Erlbaum. Levine, T., Asada, K. J., & Carpenter, C. (2009). Sample sizes and effect sizes are negatively correlated in meta-analyses: Evidence and implications of a publication bias against non-significant findings. 519



Communication Monographs, 76, 286–302. Levitan, L. C., & Visser, P. S. (2008). The impact of the social context on resistance to persuasion: Effortful versus effortless responses to counterattitudinal information. Journal of Experimental Social Psychology, 44, 640–649. Levy, D. A., Collins, B. E., & Nail, P. R. (1998). A new model of interpersonal influence characteristics. Journal of Social Behavior and Personality, 13, 715–733. Lewis, M. A., & Neighbors, C. (2006). Social norms approaches using descriptive drinking norms education: A review of the research on personalized normative feedback. Journal of American College Health, 54, 213–218. Libby, L. K., Shaeffer, E. M., Eibach, R. P., & Slemmer, J. A. (2007). Picture yourself at the polls: Visual perspective in mental imagery affects self-perception and behavior. Psychological Science, 18, 199–203. Lieberman, D. A. (2012). Digital games for health behavior change: Research, design, and future directions. In S. M. Noar & N. G. Harrington (Eds.), eHealth applications: Promising strategies for behavior change (pp. 110–127). New York: Routledge. Lieberman, D. A. (2013). Designing digital games, social media, and mobile technologies to motivate and support health behavior change. In R. E. Rice & C. K. Atkin (Eds.), Public communication campaigns (4th ed., pp. 273–287). Los Angeles: Sage. Likert, R. (1932). A technique for the measurement of attitudes. Archives of Psychology, 22(140), 1–55. Lim, S. (2013). College students’ credibility judgments and heuristics 520



concerning Wikipedia. Information Processing and Management, 49, 405–419. Linder, D. E., Cooper, J., & Jones, E. E. (1967). Decision freedom as a determinant of the role of incentive magnitude in attitude change. Journal of Personality and Social Psychology, 6, 245–254. Linder, D. E., & Worchel, S. (1970). Opinion change as a result of effortfully drawing a counterattitudinal conclusion. Journal of Experimental Social Psychology, 6, 432–448. Lindsey, L. L. M., & Ah Yun, K. (2003). Examining the persuasive effect of statistical messages: A test of mediating relationships. Communication Studies, 54, 306–321. Lipkus, I. M., Green, L. G., & Marcus, A. (2003). Manipulating perceptions of colorectal cancer threat: Implications for screening intentions and behaviors. Journal of Health Communication, 8, 213–228. Lippke, S., Wiedemann, A.U., Ziegelmann, J. P., Reuter, T., & Schwarzer, R. (2009). Self-efficacy moderates the mediation of intentions into behavior via plans. American Journal of Health Behavior, 33, 521–529. Lippke, S., Ziegelmann, J. P., Schwarzer, R., & Velicer, W. F. (2009). Validity of stage assessment in the adoption and maintenance of physical activity and fruit and vegetable consumption. Health Psychology, 28, 183–193. Littell, J. H., & Girvin, H. (2002). Stages of change: A critique. Behavior Modification, 26, 223–273. Lord, K. R., Lee, M. S., & Sauer, P. L. (1995). The combined influence hypothesis: Central and peripheral antecedents of attitude toward the ad. Journal of Advertising, 24(1), 73–85. 521



Love, G. D., Mouttapa, M., & Tanjasiri, S. P. (2009). Everybody’s talking: Using entertainment-education video to reduce barriers to discussion of cervical cancer screening among Thai women. Health Education Research, 24, 829–838. Lowe, R., Eves, F., & Carroll, D. (2002). The influence of affective and instrumental beliefs on exercise intentions and behavior: A longitudinal analysis. Journal of Applied Social Psychology, 32, 1241–1252. Lu, A. S., Baranowski, T., Thompson, D., & Buday, R. (2012). Story immersion of video games for youth health promotion: A review of literature. Games for Health Journal, 1, 199–204. Luchok, J. A., & McCroskey, J. C. (1978). The effect of quality of evidence on attitude change and source credibility. Southern Speech Communication Journal, 43, 371–383. Lupia, A., & McCubbins, M. D. (1998). The democratic dilemma: Can citizens learn what they need to know? Cambridge, UK: Cambridge University Press. Luszczynska, A. (2004). Change in breast self-examination behavior: Effects of intervention on enhancing self-efficacy. International Journal of Behavioral Medicine, 11, 95–104. Luszczynska, A., & Tryburcy, M. (2008). Effects of a self-efficacy intervention on exercise: The moderating role of diabetes and cardiovascular diseases. Applied Psychology: An International Review, 57, 644–659. Lutz, R. J. (1975). Changing brand attitudes through modification of cognitive structure. Journal of Consumer Research, 1(4), 49–59. Luzzo, D. A., Hasper, P., Albert, K. A., Bibby, M. A., & Martinelli, E. A., Jr. (1999). Effects of self-efficacy-enhancing interventions on the 522



math/science self-efficacy and career interests, goals, and actions of career undecided college students. Journal of Consulting Psychology, 46, 233–243. Lynn, M. (1991). Scarcity effects on value: A quantitative review of the commodity theory literature. Psychology and Marketing, 8, 43–57. MacKenzie, S. B., Lutz, R. J., & Belch, G. E. (1986). The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations. Journal of Marketing Research, 23, 130–143. MacKenzie, S. B., & Spreng, R. A. (1992). How does motivation moderate the impact of central and peripheral processing on brand attitudes and intentions? Journal of Consumer Research, 18, 519–529. Mackie, D. M., & Queller, S. (2000). The impact of group membership on persuasion: Revisiting “who says what to whom with what effect?” In D. J. Terry & M. A. Hogg (Eds.), Attitudes, behavior, and social context: The role of norms and group-membership (pp. 135–155). Mahwah, NJ: Lawrence Erlbaum. Mackie, D. M., & Worth, L. T. (1989). Processing deficits and the mediation of positive affect in persuasion. Journal of Personality and Social Psychology, 57, 27–40. Maddux, J. E., & Rogers, R. W. (1980). Effects of source expertness, physical attractiveness, and supporting arguments on persuasion: A case of brains over beauty. Journal of Personality and Social Psychology, 39, 235–244. Magee, R. G., & Kalyanaraman, S. (2009). Effects of worldview and mortality salience in persuasion processes. Media Psychology, 12, 171–194. Magnini, V. P., Garcia, C., & Honeycutt, E. D. (2010). Identifying the 523



attributes of an effective restaurant chain endorser. Cornell Hospitality Quarterly, 51, 238–250. Mahler, H. I. M., Kulik, J. A., Butler, H. A., Gerrard, M., & Gibbons, F. X. (2008). Social norms information enhances the efficacy of an appearance-based sun protection intervention. Social Science & Medicine, 67, 321–329. Maio, G. R., Esses, V. M., & Bell, D. W. (2000). Examining conflict between components of attitudes: Ambivalence and inconsistency are distinct constructs. Canadian Journal of Behavioural Science, 32, 71–83. Maio, G. R., Haddock, G., Watt, S. E., & Hewstone, M. (2009). Implicit measures in applied contexts: An illustrative examination of antiracism advertising. In R. E. Petty, R. H. Fazio, & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 327–357). New York: Psychology Press. Maio, G. R., & Olson, J. M. (1994). Value-attitude-behaviour relations: The moderating role of attitude functions. British Journal of Social Psychology, 33, 301–312. Maio, G. R., & Olson, J. M. (1995). Relations between values, attitudes, and behavioral intentions: The moderating role of attitude function. Journal of Experimental Social Psychology, 31, 266–285. Maio, G. R., & Olson, J. M. (2000a). What is a “value-expressive” attitude? In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitude (pp. 249–269). Mahwah, NJ: Lawrence Erlbaum. Maio, G. R., & Olson, J. M. (Eds.). (2000b). Why we evaluate: Functions of attitude. Mahwah, NJ: Lawrence Erlbaum. Maloney, E. K., Lapinski, M. K., & Witte, K. (2011). Fear appeals and persuasion: A review and update of the extended parallel process model. 524



Social and Personality Psychology Compass, 5, 206–219. Malotte, C. K., Jarvis, B., Fishbein, M., Kamb, K., Iatesta, M., Hoxworth, T., … the Project RESPECT Study Group. (2000). Stage of change versus an integrated psychosocial theory as a basis for developing effective behaviour change interventions. AIDS Care, 12, 357–364. Mangleburg, T. F., Sirgy, M. J., Grewal, D., Axsom, D., Hatzios, M., Claiborne, C. B., & Bogle, T. (1998). The moderating effect of prior experience in consumers’ use of user–image based versus utilitarian cues in brand attitude. Journal of Business and Psychology, 13, 101–113. Manis, M. (1960). The interpretation of opinion statements as a function of recipient attitude. Journal of Abnormal and Social Psychology, 60, 340–344. Mann, T., Sherman, D., & Updegraff, J. (2004). Dispositional motivations and message framing: A test of the congruency hypothesis in college students. Health Psychology, 23, 330–334. Mannetti, L., Pierro, A., & Kruglanski, A. (2007). Who regrets more after choosing a non-status-quo option? Post decisional regret under need for cognitive closure. Journal of Economic Psychology, 28, 186–196. Mannetti, L., Pierro, A., & Livi, S. (2004). Recycling: Planned and selfexpressive behaviour. Journal of Environmental Psychology, 24, 227–236. Manning, M. (2009). The effects of subjective norms on behavior in the theory of planned behaviour: A meta-analysis. British Journal of Social Psychology, 48, 649–705. Manstead, A. S. R. (2000). The role of moral norms in the attitudebehavior relation. In D. J. Terry & M. A. Hogg (Eds.), Attitudes, 525



behavior, and social context: The role of norms and group membership (pp. 11–30). Mahwah, NJ: Lawrence Erlbaum. Marcus, A. C., Crane, L. A., Kaplan, C. R, Reading, A. E., Savage, E., Gunning, J., … Berek, J. S. (1992). Improving adherence to screening follow-up among women with abnormal pap smears: Results from a large clinic-based trial of three intervention strategies. Medical Care, 30, 216–230. Marin, G., Marin, B. V., Perez-Stable, E. J., Sabogal, F., & Otero-Sabogal, R. (1990). Cultural differences in attitudes and expectancies between Hispanic and non–Hispanic white smokers. Hispanic Journal of Behavioral Sciences, 12, 422–436. Marquart, J., O’Keefe, G. J., & Gunther, A. C. (1995). Believing in biotech: Farmers’ perceptions of the credibility of BGH information sources. Science Communication, 16, 388–402. Martin, B. A. S., Lang, B., & Wong, S. (2003). Conclusion explicitness in advertising: The moderating role of need for cognition (NFC) and argument quality (AQ) on persuasion. Journal of Advertising, 32(4), 57–65. Martin, J., Slade, P., Sheeran, P., Wright, A., & Dibble, T. (2011). “Ifthen” planning in one-to-one behaviour change counselling is effective in promoting contraceptive adherence in teenagers. Journal of Family Planning and Reproductive Health Care, 37, 85–88. Marttila, J., & Nupponen, R. (2003). Assessing stage of change for physical activity: How congruent are parallel methods? Health Education Research, 18, 419–428. Mason, T. E., & White, K. M. (2008). Applying an extended model of the theory of planned behaviour to breast self-examination. Journal of Health Psychology, 13, 946–956.



526



Masser, B., & France, C. R. (2010). An evaluation of a donation coping brochure with Australian non-donors. Transfusion and Apheresis Science, 43, 291–297. Maticka-Tyndale, E., & Barnett, J. P. (2010). Peer-led interventions to reduce HIV risk of youth: A review. Evaluation and Program Planning, 33, 98–112. Mattern, J. L., & Neighbors, C. (2004). Social norms campaigns: Examining the relationship between changes in perceived norms and changes in drinking levels. Journal of Studies on Alcohol, 65, 489–493. Mazor, K. M., Baril, J., Dugan, E., Spencer, F., Burgwinkle, P., & Gurwitz, J. H. (2007). Patient education about anticoagulant medication: Is narrative evidence or statistical evidence more effective? Patient Education and Counseling, 69, 145–157. McConnell, A. R., Niedermeier, K. E., Leibold, J. M., El-Alayli, A. G., Chin, P. P., & Kuiper, N. M. (2000). What if I find it cheaper somewhere else?: Role of prefactual thinking and anticipated regret in consumer behavior. Psychology and Marketing, 17, 281–298. McCroskey, J. C. (1966). Scales for the measurement of ethos. Speech Monographs, 33, 65–72. McCroskey, J. C., & Young, T. J. (1981). Ethos and credibility: The construct and its measurement after three decades. Central States Speech Journal, 32, 24–34. McCroskey, J. C., Young, T. J., & Scott, M. D. (1972). The effects of message sidedness and evidence on inoculation against counterpersuasion in small group communication. Speech Monographs, 39, 205–212. McEachan, R. R. C., Conner, M., Taylor, N. J., & Lawton, R. J. (2011). 527



Prospective prediction of health-related behaviours with the theory of planned behaviour: A meta-analysis. Health Psychology Review, 5, 97–144. McGilligan, C., McClenahan, C., & Adamson, G. (2009). Attitudes and intentions to performing testicular self-examination: Utilizing an extended theory of planned behavior. Journal of Adolescent Health, 44, 404–406. McGinnies, E. (1973). Initial attitude, source credibility, and involvement as factors in persuasion. Journal of Experimental Social Psychology, 9, 285–296. McGuire, W. J. (1964). Inducing resistance to persuasion: Some contemporary approaches. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 1, pp. 191–229). New York: Academic Press. McGuire, W. J. (1985). Attitudes and attitude change. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed., Vol. 2, pp. 233–346). New York: Random House. McGuire, W. J., & Papageorgis, D. (1961). The relative efficacy of various types of prior belief-defense in producing immunity against persuasion. Journal of Abnormal and Social Psychology, 62, 327–337. McIntyre, P., Barnett, M. A., Harris, R. J., Shanteau, J., Skowronski, J. J., & Klassen, M. (1987). Psychological factors influencing decisions to donate organs. Advances in Consumer Research, 14, 331–334. McKay-Nesbitt, J., Manchanda, R. V., Smith, M. C., & Huhmann, B. A. (2011). Effects of age, need for cognition, and affective intensity on advertising effectiveness. Journal of Business Research, 64, 12–17. McNeil, B. J., Pauker, S. G., Sox, H. C., Jr., & Tversky, A. (1982). On the 528



elicitation of preferences for alternative therapies. New England Journal of Medicine, 306, 1259–1262. McQueen, A., & Klein, W. M. P. (2006). Experimental manipulations of self-affirmation: A systematic review. Self and Identity, 5, 289–354. McQueen, A., Kreuter, M. W., Kalesan, B., & Alcaraz, K. I. (2011). Understanding narrative effects: The impact of breast cancer survivor stories on message processing, attitudes, and beliefs among African American women. Health Psychology, 30, 674–682. McRee, A. L., Reiter, P. L., Chantala, K., & Brewer, N. T. (2010). Does framing human papillomavirus vaccine as preventing cancer in men increase vaccine acceptability? Cancer Epidemiology, Biomarkers & Prevention, 19, 1937–1944. Meijinders, A., Midden, C., Olofsson, A., Ohman, S., Matthes, J., Bondarenko, O., … Rusanen, M. (2009). The role of similarity cues in the development of trust in sources of information about GM food. Risk Analysis, 29, 1116–1128. Mellor, S., Barclay, L. A., Bulger, C. A., & Kath, L. M. (2006). Augmenting the effect of verbal persuasion on self-efficacy to serve as a steward: Gender similarity in a union environment. Journal of Occupational and Organizational Psychology, 79, 121–129. Melnyk, V., van Herpen, E., Fischer, A. R. H., & van Trijp, H. C. M. (2011). To think or not to think: The effect of cognitive deliberation on the influence of injunctive versus descriptive social norms. Psychology & Marketing, 28, 709–729. Mercier, H., & Strickland, B. (2012). Evaluating arguments from the reaction of the audience. Thinking and Reasoning, 18, 365–378. Merrill, S., Grofman, B., & Adams, J. (2001). Assimilation and contrast 529



effects in voter projections of party locations: Evidence from Norway, France, and the USA. European Journal of Political Research, 40, 199–221. Metzger, M. J., Flanagin, A. J., Eyal, K., Lemus, D. R., & McCann, R. M. (2003). Credibility for the 21st century: Integrating perspectives on source, message, and media credibility in the contemporary media environment. Communication Yearbook, 27, 293–335. Metzger, M. J., Flanagin, A. J., & Medders, R. B. (2010). Social and heuristic approaches to credibility evaluation online. Journal of Communication, 60, 413–439. Mevissen, F. E. F., Meertens, R. M., Ruiter, R. A. C., & Schaalma, H. P. (2010). Testing implicit assumptions and explicit recommendations: The effects of probability information on risk perception. Journal of Health Communication, 15, 578–589. Meyerowitz, B. E., & Chaiken, S. (1987). The effect of message framing on breast self-examination attitudes, intentions, and behavior. Journal of Personality and Social Psychology, 52, 500–510. Miarmi, L., & DeBono, K. G. (2007). The impact of distractions on heuristic processing: Internet advertisements and stereotype use. Journal of Applied Social Psychology, 37, 539–548. Michie, S., Dormandy, E., French, D. P., & Marteau, T. M. (2004). Using the theory of planned behaviour to predict screening uptake in two contexts. Psychology and Health, 19, 705–718. [Erratum notice: Psychology and Health, 20 (2005), 275.] Michie, S., Dormandy, E., & Marteau, T. M. (2004). Increasing screening uptake amongst those intending to be screened: The use of action plans. Patient Education and Counseling, 55, 218–222.



530



Micu, C. C., Coulter, R. A., & Price, L. L. (2009). How product trial alters the effects of model attractiveness. Journal of Advertising, 38(2), 69–81. Middlestadt, S. E. (2007). What is the behavior? Strategies for selecting the behavior to be addressed by health promotion interventions. In I. Ajzen, D. Albarracín, & R. Hornik (Eds.), Prediction and change of health behavior: Applying the reasoned action approach (pp. 129–147). Mahwah, NJ: Lawrence Erlbaum. Middlestadt, S. E. (2012). Beliefs underlying eating better and moving more: Lessons learned from comparative salient belief elicitations with adults and youths. Annals of the American Academy of Political and Social Science, 640, 81–100. Millar, M. G., & Millar, K. U. (1998). Effects of prior experience and thought on the attitude-behavior relation. Social Behavior and Personality, 26, 105–114. Miller, C. H., Lane, L. T., Deatrick, L. M., Young, A. M., & Potts, K. A. (2007). Psychological reactance and promotional health messages: The effects of controlling language, lexical concreteness, and the restoration of freedom. Human Communication Research, 33, 219–240. Miller, G. R., & Baseheart, J. (1969). Source trustworthiness, opinionated statements, and response to persuasive communication. Speech Monographs, 36, 1–7. Miller, G. R., & Burgoon, J. K. (1982). Factors affecting assessments of witness credibility. In N. L. Kerr & R. M. Bray (Eds.), Psychology of the courtroom (pp. 169–194). New York: Academic Press. Miller, R. L. (1977). The effects of postdecisional regret on selective exposure. European Journal of Social Psychology, 7, 121–127. Miller, W. R., & Rollnick, S. (2002). Motivational interviewing: Preparing 531



people for change (2nd ed.). New York: Guilford. Miller-Day, M., & Hecht, M. L. (2013). Narrative means to preventative ends: A narrative engagement framework for designing prevention interventions. Health Communication, 28, 657–670. Mills, J., & Kimble, C. E. (1973). Opinion change as a function of perceived similarity of the communicator and subjectivity of the issue. Bulletin of the Psychonomic Society, 2, 35–36. Milne, S., Orbell, S., & Sheeran, P. (2002). Combining motivational and volitional interventions to promote exercise participation: Protection motivation theory and implementation intentions. British Journal of Health Psychology, 7, 163–184. Milne, S., Sheeran, P., & Orbell, S. (2000). Prediction and intervention in health-related behavior: A meta-analytic review of protection motivation theory. Journal of Applied Social Psychology, 30, 106–143. Milton, A. C., & Mullan, B. A. (2012). An application of the theory of planned behavior: A randomized controlled food safety pilot intervention for young adults. Health Psychology, 31, 250–259. Miniard, P. W., & Barone, M. J. (1997). The case for noncognitive determinants of attitude: A critique of Fishbein and Middlestadt. Journal of Consumer Psychology, 6, 77–91. Miron, A. M., & Brehm, J. W. (2006). Reaktanztheorie: 40 Jahre später [Reactance theory: 40 years later]. Zeitschrift fur Sozialpsychologie, 37, 9–18. Mishra, S. I., Chavez, L. R., Magana, J. R., Nava, P., Valdez, R. B., & Hubbell, F. A. (1998). Improving breast cancer control among Latinas: Evaluation of a theory-based educational program. Health Education and Behavior, 25, 653–670. 532



Misra, S., & Beatty, S. E. (1990). Celebrity spokesperson and brand congruence: An assessment of recall and affect. Journal of Business Research, 21, 159–173. Mitchell, A. A. (1986). The effect of verbal and visual components of advertisements on brand attitudes and attitude toward the advertisement. Journal of Consumer Research, 13, 12–24. Mitchell, M. M., Brown, K. M., Morris-Villagran, M., & Villagran, P. D. (2001). The effects of anger, sadness and happiness on persuasive message processing: A test of the negative state relief model. Communication Monographs, 68, 347–359. Mittal, B., Ratchford, B., & Prabhakar, P. (1990). Functional and expressive attributes as determinants of brand-attitude. In J. N. Sheth (Ed.), Research in marketing (Vol. 10, pp. 135–155). Greenwich, CT: JAI. Mittelstaedt, J. D., Riesz, P. C., & Burns, W. J. (2000). Why are endorsements effective? Sorting among theories of product and endorser effects. Journal of Current Issues and Research in Advertising, 22(1), 55–65. Moan, I. S., & Rise, J. (2005). Quitting smoking: Applying an extended version of the theory of planned behavior to predict intention and behavior. Journal of Applied Biobehavioral Research, 10, 39–68. Moan, I. S., & Rise, J. (2011). Predicting intentions not to “drink and drive” using an extended version of the theory of planned behaviour. Accident Analysis & Prevention, 43, 1378–1384. Mondak, J. J. (1990). Perceived legitimacy of Supreme Court decisions: Three functions of source credibility. Political Behavior, 12, 363–384. Mongeau, P. A. (1998). Another look at fear-arousing persuasive appeals. 533



In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances through metaanalysis (pp. 53–68). Cresskill, NJ: Hampton. Mongeau, P. A. (2013). Fear appeals. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 184–199). Thousand Oaks, CA: Sage. Montaño, D. E., Thompson, B., Taylor, V. M., & Mahloch, J. (1997). Understanding mammography intention and utilization among women in an inner city public hospital clinic. Preventive Medicine, 26, 817–824. Moons, W. G., & Mackie, D. M. (2007), Thinking straight while seeing red: The influence of anger on information processing. Personality and Social Psychology Bulletin, 33, 706–720. Morales, A. C., Wu, E. C., & Fitzsimons, G. J. (2012). How disgust enhances the effectiveness of fear appeals. Journal of Marketing Research, 49, 383–393. Morgan, S. E. (2012). Designing high sensation value messages for the sensation seeking audience. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 231–247). Los Angeles: Sage. Morgan, S. E., Cole, H. P., Struttmann, T., & Piercy, L. (2002). Stories or statistics? Farmers’ attitudes toward messages in an agricultural safety campaign. Journal of Agricultural Safety and Health, 8, 225–239. Morgan, S. E., Palmgreen, P., Stephenson, M. T., Hoyle, R. H., & Lorch, E. P. (2003). Associations between message features and subjective evaluations of the sensation value of antidrug public service announcements. Journal of Communication, 53, 512–526. Morman, M. T. (2000). The influence of fear appeals, message design, and masculinity on men’s motivation to perform the testicular self-exam. 534



Journal of Applied Communication Research, 28, 91–116. Morton, K., Beauchamp, M., Prothero, A., Joyce, L., Saunders, L., Spencer-Bowdage, S., … Pedlar, C. (2014). The effectiveness of motivational interviewing for health behaviour change in primary care settings: A systematic review. Health Psychology Review. doi: 10.1080/17437199.2014.882006. Mowen, J. C., Wiener, J. L., & Joag, S. (1987). An information integration analysis of how trust and expertise combine to influence source credibility and persuasion. Advances in Consumer Research, 14, 564. Moyer-Gusé, E. (2008). Toward a theory of entertainment persuasion: Explaining the persuasive effects of entertainment-education messages. Communication Theory, 18, 407–425. Moyer-Gusé, E., Chung, A. H., & Jain, P. (2011). Identification with characters and discussion of taboo topics after exposure to an entertainment narrative about sexual health. Journal of Communication, 61, 387–406. Moyer-Gusé, E., Jain, P., & Chung, A. H. (2012). Reinforcement or reactance? Examining the effect of an explicit persuasive appeal following an entertainment-education narrative. Journal of Communication, 62, 1010–1027. Moyer-Gusé, E., & Nabi, R. L. (2009). Explaining the effects of narrative in an entertainment television program: Overcoming resistance to persuasion. Human Communication Research, 36, 26–52. Munn, W. C., & Gruner, C R. (1981). “Sick” jokes, speaker sex, and informative speech. Southern Speech Communication Journal, 46, 411–418. Murphy, S. T., Frank, L. B., Chatterjee, J. S., & Baezconde-Garbanati, L. 535



(2013). Narrative versus nonnarrative: The role of identification, transportation, and emotion in reducing health disparities. Journal of Communication, 63, 116–137. Murphy, S. T., Frank, L. B., Moran, M. B., & Patnoe-Woodley, P. (2011). Involved, transported, or emotional? Exploring the determinants of change in knowledge, attitudes, and behavior in entertainmenteducation. Journal of Communication, 61, 407–431. Murray-Johnson, L., Witte, K., Liu, W.-Y., Hubbell, A. P., Sampson, J., & Morrison, K. (2001). Addressing cultural orientations in fear appeals: Promoting AIDS-protective behaviors among Mexican immigrant and African American adolescents and American and Taiwanese college students. Journal of Health Communication, 6, 335–358. Murray-Johnson, L., Witte, K., Patel, D., Orrego, V., Zuckerman, C., Maxfield, A. M., & Thimons, E. D. (2004). Using the extended parallel process model to prevent noise-induced hearing loss among coal miners in Appalachia. Health Education and Behavior, 31, 741–755. Muthusamy, N., Levine, T. R., & Weber, R. (2009). Scaring the already scared: Some problems with HIV/AIDS fear appeals in Namibia. Journal of Communication, 59, 317–344. Myers, L. B., & Horswill, M. S. (2006). Social cognitive predictors of sun protection intention and behavior. Behavioral Medicine, 32, 57–63. Nabi, R. L. (1998). The effect of disgust-eliciting visuals on attitudes toward animal experimentation. Communication Quarterly, 46, 472–484. Nabi, R. L. (2002). Discrete emotions and persuasion. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 289–308). Thousand Oaks, CA: Sage.



536



Nabi, R. L. (2003). “Feeling” resistance: Exploring the role of emotionally evocative visuals in inducing inoculation. Media Psychology, 5, 199–223. Nabi, R. L. (2007). Emotion and persuasion: A social cognitive perspective. In D. R. Roskos-Ewoldsen & J. L. Monahan (Eds.), Communication and social cognition: Theories and methods (pp. 377–398). Mahwah, NJ: Lawrence Erlbaum. Nabi, R. L. (2010). The case for emphasizing discrete emotions in communication research. Communication Monographs, 77, 153–159. Nail, P. R., Misak, J. E., & Davis, R. M. (2004). Self-affirmation versus self-consistency: A comparison of two competing self-theories of dissonance phenomena. Personality and Individual Differences, 36, 1893–1905. Nakanishi, M., & Bettman, J. R. (1974). Attitude models revisited: An individual level analysis. Journal of Consumer Research, 1(3), 16–21. Nan, X. (2008). The pursuit of self-regulatory goals. Journal of Advertising, 37(1), 17–27. Nan, X. (2009). The influence of source credibility on attitude certainty: Exploring the moderating effects of timing of source identification and individual need for cognition. Psychology and Marketing, 26, 321–332. Nan, X., & Zhao, X. (2010). The influence of liking for antismoking PSAs on adolescents’ smoking-related behavioral intentions. Health Communication, 25, 459–469. Nan, X., & Zhao, X. (2012). When does self-affirmation reduce negative responses to antismoking messages? Communication Studies, 63, 482–497.



537



Napper, L., Harris, P. R., & Epton, T. (2009). Developing and testing a self-affirmation manipulation. Self and Identity, 8, 45–62. Napper, L. E., Wood, M. M., Jaffe, A., Fisher, D. G., Reynolds, G. L., & Klahn, J. A. (2008). Convergent and discriminant validity of three measures of stage of change. Psychology of Addictive Behaviors, 22, 362–271. Neimeyer, G. J., MacNair, R., Metzler, A. E., & Courchaine, K. (1991). Changing personal beliefs: Effects of forewarning, argument quality, prior bias, and personal exploration. Journal of Social and Clinical Psychology, 10, 1–20. Nelson, T. E., Clawson, R. A., & Oxley, Z. M. (1997). Media framing of a civil liberties conflict and its effect on tolerance. American Political Science Review, 91, 567–583. Nelson, T. E., Oxley, Z. M., & Clawson, R. A. (1997). Toward a psychology of framing effects. Political Behavior, 19, 221–246. Nestler, S., & Egloff, B. (2012). Interactive effect of dispositional cognitive avoidance, magnitude of threat, and response efficacy on the persuasive impact of threat communications. Journal of Individual Differences, 33, 94–100. Newell, S. J., & Goldsmith, R. E. (2001). The development of a scale to measure perceived corporate credibility. Journal of Business Research, 52, 235–247. Ng, J. Y. Y., Tam, S. F., Yew, W. W., & Lam, W. K. (1999). Effects of video modeling on self-efficacy and exercise performance of COPD patients. Social Behavior and Personality, 27, 475–486. Nickerson, D. W., & Rogers, T. (2010). Do you have a voting plan? Implementation intentions, voter turnout, and organic plan making. 538



Psychological Science, 21, 194–199. Niederdeppe, J., Davis, K. C., Farrelly, M. C., & Yarsevich, J. (2007). Stylistic features, need for sensation, and confirmed recall of national smoking prevention advertisements. Journal of Communication, 57, 272–292. Niederdeppe, J., Kim, H. K., Lundell, H., Fazili, F., & Frazier, B. (2012). Beyond counterarguing: Simple elaboration, complex integration, and counterelaboration in response to variations in narrative focus and sidedness. Journal of Communication, 62, 758–777. Niederdeppe, J., Porticella, N., & Shapiro, M. A. (2012). Using theory to identify beliefs associated with support for policies to raise the price of high-fat and high-sugar foods. Journal of Health Communication, 17, 90–104. Niederdeppe, J., Shapiro, M. A., & Porticella, N. (2011). Attributions of responsibility for obesity: Narrative communication reduces reactive counterarguing among liberals. Human Communication Research, 37, 295–323. Nienhuis, A. E., Manstead, A. S. R., & Spears, R. (2001). Multiple motives and persuasive communication: Creative elaboration as a result of impression motivation and accuracy motivation. Personality & Social Psychology Bulletin, 27, 118–132. Nigbur, D., Lyons, E., & Uzzell, D. (2010). Attitudes, norms, identity and environmental behaviour: Using an expanded theory of planned behaviour to predict participation in a kerbside recycling programme. British Journal of Social Psychology, 49, 259–284. Noar, S. M., Benac, C. N., & Harris, M. S. (2007). Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin, 133, 673–693.



539



Noar, S. M., & Mehrotra, P. (2011). Toward a new methodological paradigm for testing theories of health behavior and health behavior change. Patient Education and Counseling, 82, 468–474. Nocon, M., Müller-Riemenschneider, F., Nitzschke, K., & Willich, S.N. (2010). Review article: Increasing physical activity with point-of-choice prompts—a systematic review. Scandinavian Journal of Public Health, 38, 633–638. Nolan, J. M., Schultz, P. W., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2008). Normative social influence is underdetected. Personality and Social Psychology Bulletin, 34, 913–923. Norman, P., Bennett, P., & Lewis, H. (1998). Understanding binge drinking among young people: An application of the theory of planned behavior. Health Education Research, 13, 163–169. Norman, P., & Conner, M. (2006). The theory of planned behaviour and binge drinking: Assessing the moderating role of past behaviour within the theory of planned behaviour. British Journal of Health Psychology, 11, 55–70. Norman, P., & Cooper, Y. (2011). The theory of planned behaviour and breast self-examination: Assessing the impact of past behaviour, context stability and habit strength. Psychology & Health, 26, 1156–1172. Norman, P., & Hoyle, S. (2004). The theory of planned behavior and breast self-examination: Distinguishing between perceived control and self-efficacy. Journal of Applied Social Psychology, 34, 694–708. Norman, P., & Smith, L. (1995). The theory of planned behaviour and exercise: An investigation into the role of prior behaviour, behavioural intentions and attitude variability. European Journal of Social Psychology, 25, 403–415.



540



Norman, R. (1976). When what is said is important: A comparison of expert and attractive sources. Journal of Experimental Social Psychology, 12, 294–300. Notani, A. S. (1998). Moderators of perceived behavioral control’s predictiveness in the theory of planned behavior: A meta-analysis. Journal of Consumer Psychology, 7, 247–271. O’Carroll, R. E., Dryden, J., Hamilton-Barclay, T., & Ferguson, E. (2011). Anticipated regret and organ donor registration: A pilot study. Health Psychology, 30, 661–664. Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39–52. O’Hara, B. S., Netemeyer, R. G., & Burton, S. (1991). An examination of the relative effects of source expertise, trustworthiness, and likability. Social Behavior and Personality, 19, 305–314. Ojala, M. (2008). Recycling and ambivalence: Quantitative and qualitative analyses of household recycling among young adults. Environment and Behavior, 40, 777–797. O’Keefe, D. J. (1987). The persuasive effects of delaying identification of high- and low-credibility communicators: A meta-analytic review. Central States Speech Journal, 38, 63–72. O’Keefe, D. J. (1990). Persuasion: Theory and research. Newbury Park, CA: Sage. O’Keefe, D. J. (1993). The persuasive effects of message sidedness variations: A cautionary note concerning Allen’s (1991) meta-analysis. Western Journal of Communication, 57, 87–97.



541



O’Keefe, D. J. (1997). Standpoint explicitness and persuasive effect: A meta-analytic review of the effects of varying conclusion articulation in persuasive messages. Argumentation and Advocacy, 34, 1–12. O’Keefe, D. J. (1998). Justification explicitness and persuasive effect: A meta-analytic review of the effects of varying support articulation in persuasive messages. Argumentation and Advocacy, 35, 61–75. O’Keefe, D. J. (1999a). How to handle opposing arguments in persuasive messages: A meta-analytic review of the effects of one-sided and twosided messages. Communication Yearbook, 22, 209–249. O’Keefe, D. J. (1999b). Three reasons for doubting the adequacy of the reciprocal-concessions explanation of door-in-the-face effects. Communication Studies, 50, 211–220. O’Keefe, D. J. (1999c). Variability of persuasive message effects: Metaanalytic evidence and implications. Document Design, 1, 87–97. O’Keefe, D. J. (2000). Guilt and social influence. Communication Yearbook, 23, 67–101. O’Keefe, D. J. (2002a). Guilt as a mechanism of persuasion. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 329–344). Thousand Oaks, CA: Sage. O’Keefe, D. J. (2002b). The persuasive effects of variation in standpoint articulation. In F. H. van Eemeren (Ed.), Advances in pragma-dialectics (pp. 65–82). Amsterdam: Sic Sat. O’Keefe, D. J. (2003). Message properties, mediating states, and manipulation checks: Claims, evidence, and data analysis in experimental persuasive message effects research. Communication Theory, 13, 251–274.



542



O’Keefe, D. J. (2011a). The asymmetry of predictive and descriptive capabilities in quantitative communication research: Implications for hypothesis development and testing. Communication Methods and Measures, 5, 113–125. O’Keefe, D. J. (2011b). Generalizing about the persuasive effects of message variations: The case of gain-framed and loss-framed appeals. In T. van Haaften, H. Jansen, J. de Jong, & W. Koetsenruijter (Eds.), Bending opinion: Essays on persuasion in the public domain (pp. 117–131). Leiden, Netherlands: Leiden University Press. O’Keefe, D. J. (2012a). Conviction, persuasion, and argumentation: Untangling the ends and means of influence. Argumentation, 26, 19–32. O’Keefe, D. J. (2012b). From psychological theory to message design: Lessons from the story of gain-framed and loss-framed persuasive appeals. In H. Cho (Ed.), Health communication message design: Theory, research, and practice (pp. 3–20). Thousand Oaks, CA: Sage. O’Keefe, D. J. (2013a). The relative persuasiveness of different forms of arguments-from-consequences: A review and integration. In C. T. Salmon (Ed.), Communication Yearbook 36 (pp. 109–135). New York: Routledge. O’Keefe, D. J. (2013b). The relative persuasiveness of different message types does not vary as a function of the persuasive outcome assessed: Evidence from 29 meta-analyses of 2,062 effect sizes for 13 message variations. In E. L. Cohen (Ed.), Communication Yearbook 37 (pp. 221–249). New York: Routledge. O’Keefe, D. J., & Figgé, M. (1999). Guilt and expected guilt in the doorin-the face technique. Communication Monographs, 66, 312–324. O’Keefe, D. J., & Hale, S. L. (1998). The door-in-the-face influence strategy: A random-effects meta-analytic review. Communication Yearbook, 21, 1–33. 543



O’Keefe, D. J., & Hale, S. L. (2001). An odds-ratio-based meta-analysis of research on the door-in-the-face influence strategy. Communication Reports, 14, 31–38. O’Keefe, D. J., & Jensen, J. D. (2006). The advantages of compliance or the disadvantages of noncompliance? A meta-analytic review of the relative persuasive effectiveness of gain-framed and loss-framed messages. Communication Yearbook, 30, 1–43. O’Keefe, D. J., & Jensen, J. D. (2007). The relative persuasiveness of gain-framed and loss-framed messages for encouraging disease prevention behaviors: A meta-analytic review. Journal of Health Communication, 12, 623–644. O’Keefe, D. J., & Jensen, J. D. (2009). The relative persuasiveness of gain-framed and loss-framed messages for encouraging disease detection behaviors: A meta-analytic review. Journal of Communication, 59, 296–316. O’Keefe, D. J., & Jensen, J. D. (2011). The relative effectiveness of gainframed and loss-framed persuasive appeals concerning obesity-related behaviors: Meta-analytic evidence and implications. In R. Batra, P. A. Keller, & V. J. Strecher (Eds.), Leveraging consumer psychology for effective health communications: The obesity challenge (pp. 171–185). Armonk, NY: M. E. Sharpe. O’Keefe, D. J., & Shepherd, G. J. (1982). Interpersonal construct differentiation, attitudinal confidence, and the attitude-behavior relationship. Central States Speech Journal, 33, 416–423. Okun, M. A., & Schultz, A. (2003). Age and motives for volunteering: Testing hypotheses derived from socioemotional selectivity theory. Psychology and Aging, 18, 231–239. Orbell, S., & Hagger, M. (2006). Temporal framing and the decision to take part in Type 2 diabetes screening: Effects of individual differences 544



in consideration of future consequences on persuasion. Health Psychology, 25, 537–548. Orbell, S., & Kyriakaki, M. (2008). Temporal framing and persuasion to adopt preventive health behavior: Moderating effects of individual differences in consideration of future consequences on sunscreen use. Health Psychology, 27, 770–779. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. Urbana: University of Illinois Press. Osgood, C. E., & Tannenbaum, P. H. (1955). The principle of congruity in the prediction of attitude change. Psychological Review, 62, 42–55. O’Sullivan, B., McGee, H., & Keegan, O. (2008). Comparing solutions to the “expectancy-value muddle” in the theory of planned behaviour. British Journal of Health Psychology, 13, 789–802. Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J., & Tetlock, P. E. (2013). Predicting ethnic and racial discrimination: A meta-analysis of IAT criterion studies. Journal of Personality and Social Psychology, 105, 171–192. Ouellette, J. A., & Wood, W. (1998). Habit and intention in everyday life: The multiple processes by which past behavior predicts future behavior. Psychological Bulletin, 124, 54–74. Paek, H.-J., Oh, H. J., & Hove, T. (2012). How media campaigns influence children’s physical activity: Expanding the normative mechanisms of the theory of planned behavior. Journal of Health Communication, 17, 869–885. Pan, L. Y., & Chiou, J. S. (2011). How much can you trust online information? Cues for perceived trustworthiness of consumer-generated online information. Journal of Interactive Marketing, 25, 67–74. 545



Pan, W., & Bai, H. (2009). A multivariate approach to a meta-analytic review of the effectiveness of the D.A.R.E. program. International Journal of Environmental Research and Public Health, 6, 267–277. Pappas-DeLuca, K. A., Kraft, J. M., Galavotti, C., Warner, L., Mooki, M., Hastings, P., … Kilmarx, P. H. (2008). Entertainment-education radio serial drama and outcomes related to HIV testing in Botswana. AIDS Education and Prevention, 20, 486–504. Park, H. S., Klein, K. A., Smith, S., & Martell, D. (2009). Separating subjective norms, university descriptive and injunctive norms, and U. S. descriptive and injunctive norms for drinking behavior intentions. Health Communication, 24, 746–751. Park, H. S., Levine, T. R., Westermann, C. Y. K., Orfgen, T., & Foregger, S. (2007). The effects of argument quality and involvement type on attitude formation and attitude change: A test of dual-process and social judgment predictions. Human Communication Research, 33, 81–102. Park, H. S., & Smith, S. W. (2007). Distinctiveness and influence of subjective norms, personal descriptive and injunctive norms, and societal descriptive and injunctive norms on behavioral intent: A case of two behaviors critical to organ donation. Human Communication Research, 33, 194–218. Parker, D., Manstead, A. S. R., & Stradling, S. G. (1995). Extending the theory of planned behaviour: The role of personal norm. British Journal of Social Psychology, 34, 127–137. Parker, D., Stradling, S. G., & Manstead, A. S. R. (1996). Modifying beliefs and attitudes to exceeding the speed limit: An intervention study based on the theory of planned behavior. Journal of Applied Social Psychology, 26, 1–19. Parschau, L., Richert, J., Koring, M., Ernsting, A., Lippke, S., & Schwarzer, R. (2012). Changes in social-cognitive variables are 546



associated with stage transitions in physical activity. Health Education Research, 27, 129–140. Parsons, A., Lycett, D., & Aveyard, P. (2011). Response to Spring et al.: What is the best method to assess the effect of combined interventions for smoking cessation and post-cessation weight gain? Addiction, 106, 675–676. Parvanta, S., Gibson, L., Forquer, H., Shapiro-Luft, D., Dean, L., Freres, D., … Hornik, R. (2013). Applying quantitative approaches to the formative evaluation of antismoking campaign messages. Social Marketing Quarterly, 19, 242–264. Patzer, G. L. (1983). Source credibility as a function of communicator physical attractiveness. Journal of Business Research, 11, 229–241. Pearce, W. B., & Brommel, B. J. (1972). Vocalic communication in persuasion. Quarterly Journal of Speech, 58, 298–306. Peay, M. Y. (1980). Changes in attitudes and beliefs in two-person interaction situations. European Journal of Social Psychology, 10, 367–377. Pechmann, C. (1990). How do consumer inferences moderate the effectiveness of two-sided messages? Advances in Consumer Research, 17, 337–341. Pechmann, C. (1992). Predicting when two-sided ads will be more effective than one-sided ads: The role of correlational and correspondent inferences. Journal of Marketing Research, 29, 441–453. Peng, W., Crouse, J. C., & Lin, J.-H. (2013). Using active video games for physical activity promotion: A systematic review of the current state of research. Health Education & Behavior, 40, 171–192.



547



Perez, M., Becker, C. B., & Ramirez, A. (2010). Transportability of an empirically supported dissonance-based prevention program for eating disorders. Body Image, 7, 179–186. Perloff, R. M. (2014). The dynamics of persuasion: Communication and attitudes in the 21st century (5th ed.). New York: Routledge. Pertl, M., Hevey, D., Thomas, K., Craig, A., Chuinneagáin, S. N., & Maher, L. (2010). Differential effects of self-efficacy and perceived control on intention to perform skin cancer-related health behaviours. Health Education Research, 25, 769–779. Peters, G.-J. Y., Ruiter, R. A. C., & Kok, G. (2013). Threatening communication: A critical re-analysis and a revised meta-analytic test of fear appeal theory. Health Psychology Review, 7(suppl. 1), S8–S31. doi: 10.1080/17437199.2012.703527. Peters, R. G., Covello, V. T., & McCallum, D. B. (1997). The determinants of trust and credibility in environmental risk communication: An empirical study. Risk Analysis, 17, 43–54. Petkova, K. G., Ajzen, I., & Driver, B. L. (1995). Salience of anti-abortion beliefs and commitment to an attitudinal position: On the strength, structure, and predictive validity of anti-abortion attitudes. Journal of Applied Social Psychology, 25, 463–483. Petrova, P. K., & Cialdini, R. B. (2005). Fluency of consumption imagery and the backfire effects of imagery appeals. Journal of Consumer Research, 32, 442–452. Petty, R. E., & Briñol, P. (2006). Understanding social judgment: Multiple systems and processes. Psychological Inquiry, 17, 217–223. Petty, R. E., & Briñol, P. (2010). Attitude change. In R. F. Baumeister & E. J. Finkel (Eds.), Advanced social psychology: The state of the 548



science (pp. 217–259). Oxford, UK: Oxford University Press. Petty, R. E., & Briñol, P. (2012a). The elaboration likelihood model. In P. A. M. Van Lange, A. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (Vol. 1, pp. 224–245). London: Sage. Petty, R. E., & Briñol, P. (2012b). A multiprocess approach to social influence. In D. T. Kendrick, N. J. Goldstein, & S. L. Braver (Eds.), Six degrees of social influence: Science, application, and the psychology of Robert Cialdini (pp. 49–58). New York: Oxford University Press. Petty, R. E., Briñol, P., Loersch, C., & McCaslin, M. J. (2009). The need for cognition. In M. R. Leary & R. H. Hoyle (Eds.), Handbook of individual differences in social behavior (pp. 318–329). New York: Guilford. Petty, R. E., Briñol, P., Tormala, Z. L., & Wegener, D. T. (2007). The role of metacognition in social psychology. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (2nd ed., pp. 254–284). New York: Guilford. Petty, R. E., & Brock, T. C. (1981). Thought disruption and persuasion: Assessing the validity of attitude change experiments. In R. E. Petty, T. M. Ostrom, & T. C. Brock (Eds.), Cognitive responses in persuasion (pp. 55–79). Hillsdale, NJ: Lawrence Erlbaum. Petty, R. E., & Cacioppo, J. T. (1977). Forewarning, cognitive responding, and resistance to persuasion. Journal of Personality and Social Psychology, 35, 645–655. Petty, R. E., & Cacioppo, J. T. (1979a). Effects of forewarning of persuasive intent and involvement on cognitive responses and persuasion. Personality and Social Psychology Bulletin, 5, 173–176. Petty, R. E., & Cacioppo, J. T. (1979b). Issue involvement can increase or 549



decrease persuasion by enhancing message-relevant cognitive responses. Journal of Personality and Social Psychology, 37, 1915–1926. Petty, R. E., & Cacioppo, J. T. (1986a). Communication and persuasion: Central and peripheral routes to attitude change. New York: SpringerVerlag. Petty, R. E., & Cacioppo, J. T. (1986b). The elaboration likelihood model of persuasion. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 123–205). New York: Academic Press. Petty, R. E., & Cacioppo, J. T. (1990). Involvement and persuasion: Tradition versus integration. Psychological Bulletin, 107, 367–374. Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a determinant of argument-based persuasion. Journal of Personality and Social Psychology, 41, 847–855. Petty, R. E., Cacioppo, J. T., & Haugtvedt, C. P. (1992). Ego-involvement and persuasion: An appreciative look at the Sherifs’ contribution to the study of self-relevance and attitude change. In D. Granberg & G. Sarup (Eds.), Social judgment and intergroup relations: Essays in honor of Muzafer Sherif (pp. 147–174). New York: Springer-Verlag. Petty, R. E., Cacioppo, J. T., & Heesacker, M. (1981). Effects of rhetorical questions on persuasion: A cognitive response analysis. Journal of Personality and Social Psychology, 40, 432–440. Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research, 10, 135–146. Petty, R. E., Cacioppo, J. T., Strathman, A. J., & Priester, J. R. (2005). To think or not to think: Exploring two routes to persuasion. In T. C. Brock 550



& M. C. Green (Eds.). Persuasion: Psychological insights and perspectives (2nd ed., pp. 81–116). Thousand Oaks, CA: Sage. Petty, R. E., Fazio, R. H., & Briñol, P. (Eds.). (2009a). Attitudes: Insights from the new implicit measures. New York: Psychology Press. Petty, R. E., Fazio, R. H., & Briñol, P. (2009b). The new implicit measures: An overview. In R. E. Petty, R. H. Fazio, & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 3–18). New York: Psychology Press. Petty, R. E., Haugtvedt, C. P., & Smith, S. M. (1995). Elaboration as a determinant of attitude strength: Creating attitudes that are persistent, resistant, and predictive of behavior. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 93–130). Mahwah, NJ: Lawrence Erlbaum. Petty, R. E., & Krosnick, J. A. (Eds.). (1995). Attitude strength: Antecedents and consequences. Mahwah, NJ: Lawrence Erlbaum. Petty, R. E., & Wegener, D. T. (1998a). Attitude change: Multiple roles for persuasion variables. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 1, pp. 323–390). Boston: McGraw-Hill. Petty, R. E., & Wegener, D. T. (1998b). Matching versus mismatching attitude functions: Implications for scrutiny of persuasive messages. Personality and Social Psychology Bulletin, 24, 227–240. Petty, R. E., & Wegener, D. T. (1999). The elaboration likelihood model: Current status and controversies. In S. Chaiken & Y. Trope (Eds.), Dual-process theories in social psychology (pp. 41–72). New York: Guilford. Petty, R. E., Wegener, D. T., Fabrigar, L. R., Priester, J. R., & Cacioppo, J. 551



T. (1993). Conceptual and methodological issues in the elaboration likelihood model: A reply to the Michigan State critics. Communication Theory, 3, 336–362. Petty, R. E., Wells, G. L., & Brock, T. C. (1976). Distraction can enhance or reduce yielding to propaganda: Thought disruption versus effort justification. Journal of Personality and Social Psychology, 34, 874–884. Petty, R. E., Wells, G. L., Heesacker, M., Brock, T. C., & Cacioppo, J. T. (1983). The effects of recipient posture on persuasion: A cognitive response analysis. Personality and Social Psychology Bulletin, 9, 209–222. Petty, R. E., Wheeler, S. C., & Bizer, G. Y. (1999). Is there one persuasion process or more? Lumping versus splitting in attitude change theories. Psychological Inquiry, 10, 156–163. Petty, R. E., Wheeler, S. C., & Bizer, G. Y. (2000). Attitude functions and persuasion: An elaboration likelihood approach to matched versus mismatched messages. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 133–162). Mahwah, NJ: Lawrence Erlbaum. Pfau, M., Holbert, R. L., Zubric, S. J., Pasha, N. H., & Lin, W.-K. (2000). Role and influence of communication modality in the process of resistance to persuasion. Media Psychology, 2, 1–33. Pfau, M., & Szabo, E. A. (2004). Inoculation and resistance to persuasion. In J. S. Seiter & R. H. Gass (Eds.), Perspectives on persuasion, social influence, and compliance gaining (pp. 265–286). Boston: Pearson Allyn and Bacon. Pfau, M., Tusing, K. J., Koerner, A. F., Lee, W., Godbold, L. C., Penaloza, L. C., … Hong, Y.-H. (1997). Enriching the inoculation construct: The role of critical components in the process of resistance. Human 552



Communication Research, 24, 187–215. Pierro, A., Mannetti, L., Erb, H.-P., Spiegel, S. & Kruglanski, A. W. (2005). Informational length and order of presentation as determinants of persuasion. Journal of Experimental Social Psychology, 41, 458–469. Pietersma, S., & Dijkstra, A. (2011). Do behavioural health intentions engender health behaviour change? A study on the moderating role of self-affirmation on actual fruit intake versus vegetable intake. British Journal of Health Psychology, 16, 815–827. Polonec, L. D., Major, A. M., & Atwood, L. E. (2006). Evaluating the believability and effectiveness of the social norms message “most students drink 0 to 4 drinks when they party.” Health Communication, 20, 23–34. Polyorat, K., Alden, D. L., & Kim, E. S. (2007). Impact of narrative versus factual print ad copy on product evaluation: The mediating role of ad message involvement. Psychology and Marketing, 24, 539–554. Popova, L. (2012). The extended parallel process model: Illuminating the gaps in research. Health Education and Behavior, 39, 455–473. Pornpitakpan, C. (2004). The persuasiveness of source credibility: A critical review of five decades’ evidence. Journal of Applied Social Psychology, 34, 243–281. Porzig-Drummond, R., Stevenson, R., Case, T., & Oaten, M. (2009). Can the emotion of disgust be harnessed to promote hand hygiene? Experimental and field-based tests. Social Science & Medicine, 68, 1006–1012. Posavac, E. J., Kattapong, K. R., & Dew, D. E., Jr. (1999). Peer-based interventions to influence health-related behaviors and attitudes: A meta-analysis. Psychological Reports, 85, 1179–1194. 553



Povey, R., Conner, M., Sparks, P., James, R., & Shepherd, R. (2000). Application of the theory of planned behaviour to two dietary behaviours: Roles of perceived control and self-efficacy. British Journal of Health Psychology, 5, 121–139. Powers, P. (2007). Persuasion and coercion: A critical review of philosophical and empirical approaches. HEC Forum, 19, 125–143. Prati, G., Pietrantoni, L., & Zani, B. (2011). Influenza vaccination: The persuasiveness of messages among people aged 65 years and older. Health Communication, 27, 413–420. Pratkanis, A. R. (2007). Social influence analysis: An index of tactics. In A. R. Pratkanis (Ed.), The science of social influence: Advances and future progress (pp. 17–82). New York: Psychology Press. Pratkanis, A. R., & Greenwald, A. G. (1989). A sociocognitive model of attitude structure and function. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 22, pp. 245–285). New York: Academic Press. Pratkanis, A. R., Greenwald, A. G., Ronis, D. L., Leippe, M. R., & Baumgardner, M. H. (1986). Consumer-product and sociopolitical messages for use in studies of persuasion. Personality and Social Psychology Bulletin, 12, 536–538. Praxmarer, S. (2011). How a presenter’s perceived attractiveness affects persuasion for attractiveness-unrelated products. International Journal of Advertising, 30, 839–865. Preacher, K. J., & Hayes, A. F. (2008). Contemporary approaches to assessing mediation in communication research. In A. F. Hayes, M. D. Slater, & L. B. Snyder (Eds.), The Sage sourcebook of advanced data analysis methods for communication research (pp. 13–54). Thousand Oaks, CA: Sage.



554



Preiss, R. W., & Allen, M. (1998). Performing counterattitudinal advocacy: The persuasive impact of incentives. In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 231–242). Cresskill, NJ: Hampton Press. Prentice, D. A., & Carlsmith, K. M. (2000). Opinions and personality: On the psychological functions of attitudes and other valued possessions. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 223–248). Mahwah, NJ: Lawrence Erlbaum. Prentice-Dunn, S., McMath, B. F., & Cramer, R. J. (2009). Protection motivation theory and stages of change in sun protective behavior. Journal of Health Psychology, 14, 297–305. Prestwich, A., Conner, M., Lawton, R., Bailey, W., Litman, J., & Molyneaux, V. (2005). Individual and collaborative implementation intentions and the promotion of breast self-examination. Psychology and Health, 20, 743–760. [Erratum notice: Psychology & Health, 21 (2006), 143.] Prestwich, A., Kellar, I., Parker, R., MacRae, S., Learmonth, M., Sykes, B., … Castle, H. (2014). How can self-efficacy be increased? Metaanalysis of dietary interventions. Health Psychology Review, 8, 270–285. Prestwich, A., Perugini, M., & Hurling, R. (2008). Goal desires moderate intention-behaviour relations. British Journal of Social Psychology, 47, 49–73. Priester, J. R., & Fleming, M. A. (1997). Artifact or meaningful theoretical constructs? Examining evidence for nonbelief- and belief-based attitude change processes. Journal of Consumer Psychology, 6, 67–76. Priester, J. R., & Petty, R. E. (1995). Source attributions and persuasion: Perceived honesty as a determinant of message scrutiny. Personality and Social Psychology Bulletin, 21, 637–654. 555



Primack, B. A., Carroll, M. V., McNamara, M., Klem, M. L., King, B., Rich, M., … Nayak, S. (2012). Role of video games in improving health-related outcomes: A systematic review. American Journal of Preventive Medicine, 42, 630–638. Prince, M. A., & Carey, K. B. (2010). The malleability of injunctive norms among college students. Addictive Behaviors, 35, 940–947. Prislin, R. (1987). Attitude-behaviour relationship: Attitude relevance and behaviour relevance. European Journal of Social Psychology, 17, 483–485. Prislin, R., & Wood, W. (2005). Social influence in attitudes and attitude change. In D. Albarracín, B. T. Johnson, & M. P. Zanna (Eds.), The handbook of attitudes (pp. 671–706). Mahwah, NJ: Lawrence Erlbaum. Prochaska, J. O. (1994). Strong and weak principles for progressing from precontemplation to action on the basis of twelve problem behaviors. Health Psychology, 13, 47–51. Prochaska, J. O., & DiClemente, C. C. (1984). The transtheoretical approach: Crossing traditional boundaries of therapy. Homewood, IL: Dow Jones-Irwin. Prochaska, J. O., DiClemente, C. C., Velicer, W. F., & Rossi, J. S. (1992). Criticisms and concerns of the transtheoretical model in light of recent research. British Journal of Addiction, 87, 825–828. Prochaska, J. O., Redding, C. A., & Evers, K. E. (2002). The transtheoretical model and stages of change. In K. Glanz, B. K. Rimer, & F. M. Lewis (Eds.), Health behavior and health education: Theory, research, and practice (3rd ed., pp. 99–120). San Francisco: JosseyBass. Prochaska, J. O., Velicer, W. F., Rossi, J. S., Goldstein, M. G., Marcus, B. 556



H., Rakowski, W., … Rossi, S. R. (1994). Stages of change and decisional balance for 12 problem behaviors. Health Psychology, 13, 39–46. Pryor, B., & Steinfatt, T. M. (1978). The effects of initial belief level on inoculation theory and its proposed mechanisms. Human Communication Research, 4, 217–230. Puckett, J. M., Petty, R. E., Cacioppo, J. T., & Fischer, D. L. (1983). The relative impact of age and attractiveness stereotypes on persuasion. Journal of Gerontology, 38, 340–343. Quick, B. L., Bates, B. R., & Quinlan, M. R. (2009). The utility of anger in promoting clean indoor air policies. Health Communication, 24, 548–561. Quick, B. L., & Considine, J. R. (2008). Examining the use of forceful language when designing exercise persuasive messages for adults: A test of conceptualizing reactance arousal as a two-step process. Health Communication, 23, 483–491. Quick, B. L., Shen, L., & Dillard, J. P. (2013). Reactance theory and persuasion. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 167–183). Thousand Oaks, CA: Sage. Quinlan, K. B., & McCaul, K. D. (2000). Matched and mismatched interventions with young adult smokers: Testing a stage theory. Health Psychology, 19, 165–171. Raaijmakers, J. G. W., Schrijnemakers, J. M. C., & Gremmen, F. (1999). How to deal with “the language-as-fixed-effect fallacy”: Common misconceptions and alternative solutions. Journal of Memory and Language, 41, 416–426.



557



Radecki, C. M., & Jaccard, J. (1999). Signing an organ donation letter: The prediction of behavior from behavioral intentions. Journal of Applied Social Psychology, 29, 1833–1853. Raghubir, P., & Corfman, K. (1999). When do price promotions affect pretrial brand evaluations? Journal of Marketing Research, 36, 211–222. Rains, S. A. (2013). The nature of psychological reactance revisited: A meta-analytic review. Human Communication Research, 39, 47–73. Rains, S. A., & Turner, M. M. (2007). Psychological reactance and persuasive health communication: A test and extension of the intertwined model. Human Communication Research, 33, 241–269. Randall, D. M., & Wolff, J. A. (1994). The time interval in the intentionbehaviour relationship: Meta-analysis. British Journal of Social Psychology, 33, 405–418. Ratcliff, C. D., Czuchry, M., Scarberry, N. C., Thomas, J. C., Dansereau, D. F., & Lord, C. G. (1999). Effects of directed thinking on intentions to engage in beneficial activities: Actions versus reasons. Journal of Applied Social Psychology, 29, 994–1009. Reed, M. B., & Aspinwall, L. G. (1998). Self-affirmation reduces biased processing of health-risk information. Motivation and Emotion, 22, 99–132. Regan, D. T., & Fazio, R. (1977). On the consistency between attitudes and behavior: Look to the method of attitude formation. Journal of Experimental Social Psychology, 13, 28–45. Reichert, T., Heckler, S. E., & Jackson, S. (2001). The effects of sexual social marketing appeals on cognitive processing and persuasion. Journal of Advertising, 30(1), 13–27.



558



Reid, A. E., & Aiken, L. S. (2013). Correcting injunctive norm misperceptions motivates behavior change: A randomized controlled sun protection intervention. Health Psychology, 32, 551–560. Reimer, T. (2003). Direkte und indirekte effekte der argumentqualität: Der einfluss der argumentstärke auf die wahrgenommene expertise eines kommunikators [Direct and indirect effects of argument quality: The impact of argument strength on the perceived expertise of a communicator]. Zeitschrift für Sozialpsychologie, 34, 243–255. Reinard, J. C. (1998). The persuasive effects of testimonial assertion evidence. In M. Allen & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 69–86). Cresskill, NJ: Hampton. Reinhart, A. M., & Anker, A. E. (2012). An exploration of transportation and psychological reactance in organ donation PSAs. Communication Research Reports, 29, 274–284. Renes, R. J., Mutsaers, K., & van Woerkum, C. (2012). The difficult balance between entertainment and education: A qualitative evaluation of a Dutch health-promoting documentary series. Health Promotion Practice, 13, 259–264. Resnicow, K., Davis, R. E., Zhang, G., Konkel, J., Strecher, V. J., Shaikh, A. R., … Weise, C. (2008). Tailoring a fruit and vegetable intervention on novel motivational constructs: Results of a randomized study. Annals of Behavioral Medicine, 35, 159–170. Rhine, R. J., & Severance, L. J. (1970). Ego-involvement, discrepancy, source credibility, and attitude change. Journal of Personality and Social Psychology, 16, 175–190. Rhodes, N., & Ewoldsen, D. R. (2013). Outcomes of persuasion: Behavioral, cognitive, and social. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 53–69). Thousand Oaks, CA: Sage. 559



Rhodes, N., & Wood, W. (1992). Self-esteem and intelligence affect influenceability: The mediating role of message reception. Psychological Bulletin, 111, 156–171. Rhodes, R. E., Blanchard, C. M., Courneya, K. S., & Plotnikoff, R. C. (2009). Identifying belief-based targets for the promotion of leisure-time walking. Health Education & Behavior, 36, 381–393. Rhodes, R. E., Blanchard, C. M., & Matheson, D. H. (2006). A multicomponent model of the theory of planned behaviour. British Journal of Health Psychology, 11, 119–137. Rhodes, R. E., & Dickau, L. (2013). Moderators of the intention-behaviour relationship in the physical activity domain: A systematic review. British Journal of Sports Medicine, 37, 215–225. Richard, R., de Vries, N. K., & van der Pligt, J. (1998). Anticipated regret and precautionary sexual behavior. Journal of Applied Social Psychology, 28, 1411–1428. Richard, R., van der Pligt, J., & de Vries, N. (1996a). Anticipated affect and behavioral choice. Basic and Applied Social Psychology, 18, 111–129. Richard, R., van der Pligt, J., & de Vries, N. (1996b). Anticipated regret and time perspective: Changing sexual risk-taking behavior. Journal of Behavioral Decision Making, 9, 185–199. Richert, J., Schüz, N., & Schüz, B. (2013). Stages of health behavior change and mindsets. Health Psychology, 32, 273–282. Ricketts, M., Shanteau, J., McSpadden, B., & Fernandez-Medina, K. (2010). Using stories to battle unintentional injuries: Narratives in safety and health communication. Social Science & Medicine, 70, 1441–1449.



560



Riemsma, R. P., Pattenden, J., Bridle, C., Sowden, A. J., Mather, L., Watt, I. S. & Walker, A. (2003). Systematic review of the effectiveness of stage based interventions to promote smoking cessation. BMJ, 326, 1175–1181. Rietveld, T., & van Hout, R. (2007). Analysis of variance for repeated measures designs with word materials as a nested random or fixed factor. Behavior Research Methods, 39, 735–747. Rimal, R. N., Bose, K., Brown, J., Mkandawire, G., & Folda, L. (2009). Extending the purview of the risk perception attitude framework: Findings from HIV/AIDS prevention research in Malawi. Health Communication, 24, 210–218. Rimal, R. N., & Juon, H.S. (2010). Use of the risk perception attitude framework for promoting breast cancer prevention. Journal of Applied Social Psychology, 40, 287–310. Rimal, R. N., & Real, K. (2003). Perceived risk and efficacy beliefs as motivators of change: Use of the risk perception attitude (RPA) framework to understand health behaviors. Human Communication Research, 29, 370–399. Rise, J., Sheeran, P., & Hukkelberg, S. (2010). The role of self-identity in the theory of planned behavior: A meta-analysis. Journal of Applied Social Psychology, 40, 1085–1105. Risen, J. L., & Chen, M. K. (2010). How to study choice-induced attitude change: Strategies for fixing the free-choice paradigm. Social and Personality Psychology Compass, 4, 1151–1164. Rittle, R. H. (1981). Changes in helping behavior: Self- versus situational perceptions as mediators of the foot-in-the-door effect. Personality and Social Psychology Bulletin, 7, 431–437.



561



Rivis, A., & Sheeran, P. (2003). Descriptive norms as an additional predictor in the theory of planned behaviour: A meta-analysis. Current Psychology, 22, 218–233. Robins, D., Holmes, J., & Stansbury, M. (2010). Consumer health information on the web: The relationship of visual design and perceptions of credibility. Journal of the American Society for Information Science and Technology, 61, 13–29. Robinson, J. K., Turrisi, R., & Stapleton, J. (2007). Efficacy of a partner assistance intervention designed to increase skin self-examination performance. Archives of Dermatology, 143, 37–41. Rodgers, W. M., Conner, M., & Murray, T. C. (2008). Distinguishing among perceived control, perceived difficulty, and self-efficacy as determinants of intentions and behaviour. British Journal of Social Psychology, 47, 607–630. Roehrig, M., Thompson, J. K., Brannick, M., & van den Berg, P. (2006). Dissonance-based eating disorder prevention program: A preliminary dismantling investigation. International Journal of Eating Disorders, 39, 1–10. Rogers, R. W., & Prentice-Dunn, S. (1997). Protection motivation theory. In D. Gochman (Ed.), Handbook of health behavior research: Vol. 1. Personal and social determinants (pp. 113–132). New York: Plenum. Rokeach, M. (1973). The nature of human values. New York: Free Press. Romero, A. A., Agnew, C. R., & Insko, C. A. (1996). The cognitive mediation hypothesis revisited: An empirical response to methodological and theoretical criticism. Personality and Social Psychology Bulletin, 22, 651–665. Rosen, C. S. (2000). Is the sequencing of change processes by stage 562



consistent across health problems? A meta-analysis. Health Psychology, 19, 593–604. Rosen, J., & Haaga, D. A. F. (1998). Facilitating cooperation in a social dilemma: A persuasion approach. Journal of Psychology, 132, 143–153. Rosen, S. (1961). Postdecision affinity for incompatible information. Journal of Abnormal and Social Psychology, 63, 188–190. Rosenberg, M. J. (1956). Cognitive structure and attitudinal affect. Journal of Abnormal and Social Psychology, 53, 367–372. Rosenberg, M. J., & Hovland, C. I. (1960). Cognitive, affective, and behavioral components of attitudes. In C. I. Hovland & M. J. Rosenberg (Eds.), Attitude organization and change: An analysis of consistency among attitude components (pp. 1–14). New Haven, CT: Yale University Press. Rosenzweig, E., & Gilovich, T. (2012). Buyer’s remorse or missed opportunity? Differential regrets for material and experiential purchases. Journal of Personality and Social Psychology, 102, 215–223. Roskos-Ewoldsen, D. R., & Fazio, R. H. (1992). The accessibility of source likability as a determinant of persuasion. Personality and Social Psychology Bulletin, 18, 19–25. Roskos-Ewoldsen, D. R., & Fazio, R. H. (1997). The role of belief accessibility in attitude formation. Southern Communication Journal, 62, 107–116. Rosnow, R. L. (1968). One-sided vs. two-sided communication under indirect awareness of persuasive intent. Public Opinion Quarterly, 32, 95–101.



563



Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion. Personality and Social Psychology Review, 5, 296–320. Ruiter, R. A. C., Abraham, C., & Kok, G. (2001). Scary warnings and rational precautions: A review of the psychology of fear appeals. Psychology and Health, 16, 613–630. Ruiter, R. A. C., Kessels, L. T. E., Peters, G.-J. Y., & Kok, G. (2014). Sixty years of fear appeal research: Current state of the evidence. International Journal of Psychology, 49, 63–70. Ruiter, R. A. C., & Kok, G. (2012). Planning to frighten people? Think again! In C. Abraham & M. Kools (Eds.), Writing health communication: An evidence-based guide (pp. 117–133). Los Angeles: Sage. Ruiter, R. A. C., Verplanken, B., De Cremer, D., & Kok, G. (2004). Danger and fear control in response to fear appeals: The role of need for cognition. Basic and Applied Social Psychology, 26, 13–24. Ruiz, S., & Sicilia, M. (2004). The impact of cognitive and/or affective processing styles on consumer response to advertising appeals. Journal of Business Research, 57, 657–664. Ryerson, W. N., & Teffera, N. (2004). Organizing a comprehensive national plan for entertainment-education in Ethiopia. In A. Singhal, M. J. Cody, E. M. Rogers, & M. Sabido (Eds.), Entertainment-education and social change: History, research, and practice (pp. 177–190). Mahwah, NJ: Lawrence Erlbaum. Sagarin, B. J., & Skowronski, J. J. (2009a). The implications of imperfect measurement for free-choice carry-over effects: Reply to M. Keith Chen’s (2008) “Rationalization and cognitive dissonance: Do choices affect or reflect preferences?” Journal of Experimental Social Psychology, 45, 421–423. 564



Sagarin, B. J., & Skowronski, J. J. (2009b). In pursuit of the proper null: Reply to Chen and Risen (2009). Journal of Experimental Social Psychology, 45, 428–430. Sailors, J. J. (2011). Preventing childhood obesity by persuading mothers to breastfeed: Matching appeal type to personality. In R. Batra, P. A. Keller, & V. J. Strecher (Eds.), Leveraging consumer psychology for effective health communications: The obesity challenge (pp. 253–271). Armonk, NY: M. E. Sharpe. Sakaki, H. (1980). [Communication discrepancy and ego involvement as determinants of attitude change]. Journal of the Nihon University College of Industrial Technology, 13, 1–9. Salovey, P., Schneider, T. R., & Apanovitch, A. M. (2002). Message framing in the prevention and early detection of illness. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 391–406). Thousand Oaks, CA: Sage. Sampson, E. E., & Insko, C. A. (1964). Cognitive consistency and performance in the autokinetic situation. Journal of Abnormal and Social Psychology, 68, 184–192. Sanaktekin, O. H., & Sunar, D. (2008). Persuasion and relational versus personal bases of self-esteem: Does the message need to be one- or twosided? Social Behavior and Personality, 36, 1315–1332. Sandberg, T., & Conner, M. (2008). Anticipated regret as an additional predictor in the theory of planned behaviour: A meta-analysis. British Journal of Social Psychology, 47, 589–606. Sandberg, T., & Conner, M. (2009). A mere measurement effect for anticipated regret: Impacts on cervical screening attendance. British Journal of Social Psychology, 48, 221–236.



565



Sarup, G., Suchner, R. W., & Gaylord, G. (1991). Contrast effects and attitude change: A test of the two-stage hypothesis of social judgment theory. Social Psychology Quarterly, 54, 364–372. Saucier, D., & Webster, R. (2009). Social vigilantism: Measuring individual differences in belief superiority and resistance to persuasion. Personality and Social Psychology Bulletin, 36, 19–32. Sawyer, A. G., & Howard, D. J. (1991). Effects of omitting conclusions in advertisements to involved and uninvolved audiences. Journal of Marketing Research, 28, 467–474. Sayeed, S., Fishbein, M., Hornik, R., Cappella, J., & Ahern, R. K. (2005). Adolescent marijuana use intentions: Using theory to plan an intervention. Drugs: Education, Prevention, and Policy, 12, 19–34. Schepers, J., & Wetzels, M. (2007). A meta-analysis of the technology acceptance model: Investigating subjective norm and moderation effects. Information and Management, 44, 90–103. Scher, S. J., & Cooper, J. (1989). Motivational basis of dissonance: The singular role of behavioral consequences. Journal of Personality and Social Psychology, 56, 899–906. Schlehofer, M. M., & Thompson, S. C. (2011). Individual differences in mediators and reactions to a personal safety threat message. Basic and Applied Social Psychology, 33, 194–205. Schulz, P. J., & Meuffels, B. (2012). Justifying age thresholds for mammographic screening: An application of pragma-dialectical argumentation theory. Health Communication, 27, 167–178. Schüz,, B., Sniehotta, F. F., Mallach, N., Wiedemann, A. U., & Schwarzer, R. (2009). Predicting transitions from preintentional, intentional and actional stages of change. Health Education Research, 24, 64–75. 566



Schüz, B., Sniehotta, F. F., & Schwarzer, R. (2007). Stage-specific effects of an action control intervention on dental flossing. Health Education Research, 22, 332–341. Schüz, N., Schüz, B., & Eid, M. (2013). When risk communication backfires: Randomized controlled trial on self-affirmation and reactance to personalized risk feedback in highrisk individuals. Health Psychology, 32, 561–570. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 1–65). San Diego: Academic Press. Schwarz, N. (2008). Attitude measurement. In W. D. Crano & E. Prislin (Eds.), Attitudes and attitude change (pp. 41–60). New York: Psychology Press. Schwarzer, R. (2008a). Modeling health behavior change: How to predict and modify the adoption and maintenance of health behaviors. Applied Psychology: An International Review, 57, 1–29. Schwarzer, R. (2008b). Response: Some burning issues in research on health behavior change. Applied Psychology: An International Review, 57, 84–93. Schwarzer, R., Cao, D. S., & Lippke, S. (2010). Stage-matched minimal interventions to enhance physical activity in Chinese adolescents. Journal of Adolescent Health, 47, 533–539. Schwarzer, R., & Fuchs, R. (1996). Self-efficacy and health behaviours. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and practice with social cognition models (pp. 163–196). Buckingham, UK: Open University Press.



567



Schwarzer, R., Richert, J., Kreausukon, P., Remme, L., Wiedemann, A. U., & Reuter, T. (2010). Translating intentions into nutrition behaviors via planning requires self-efficacy: Evidence from Thailand and Germany. International Journal of Psychology, 45, 260–268. Schweitzer, D., & Ginsburg, G. P. (1966). Factors of communicator credibility. In C. W. Backman & P. F. Secord (Eds.), Problems in social psychology (pp. 94–102). New York: McGraw-Hill. Schwenk, G., & Moser, G. (2009). Intention and behavior: A Bayesian meta-analysis with focus on the Ajzen-Fishbein model in the field of environmental behavior. Quality & Quantity, 43, 743–756. Sears, D. O. (1965). Biased indoctrination and selectivity of exposure to new information. Sociometry, 28, 363–376. See, Y. H. M., Petty, R. E., & Evans, L. M. (2009). The impact of perceived message complexity and need for cognition on information processing and attitudes. Journal of Research in Personality, 43, 880–889. Segan, C. J., Borland, R., & Greenwood, K. M. (2004). What is the right thing at the right time? Interactions between stages and processes of change among smokers who make a quit attempt. Health Psychology, 23, 86–93. Segar, M. L., Updegraff, J. A., Zikmund-Fisher, B. J., & Richardson, C. R. (2012). Physical activity advertisements that feature daily well-being improve autonomy and body image in overweight women but not men. Journal of Obesity, 2012, article ID 354721. Seibel, C. A., & Dowd, E. T. (2001). Personality characteristics associated with psychological reactance. Journal of Clinical Psychology, 57, 963–969.



568



Sestir, M., & Green, M. C. (2010). You are who you watch: Identification and transportation effects on temporary self-concept. Social Influence, 5, 272–288. Shakarchi, R. J., & Haugtvedt, C. P. (2004). Differentiating individual differences in resistance to persuasion. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 105–113). Mahwah, NJ: Lawrence Erlbaum. Shani, Y., & Zeelenberg, M. (2007). When and why do we want to know? How experienced regret promotes post-decision information search. Journal of Behavioral Decision Making, 20, 207–222. Sharan, M., & Valente, T. W. (2002). Spousal communication and family planning adoption: Effects of a radio drama serial in Nepal. International Family Planning Perspectives, 28, 16–25. Sharot, T., Fleming, S. M., Yu, X., Koster, R., & Dolan, R. J. (2012). Is choice-induced preference change long lasting? Psychological Science, 23, 1123–1129. Shavitt, S. (1989). Operationalizing functional theories of attitude. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function (pp. 311–337). Hillsdale, NJ: Lawrence Erlbaum. Shavitt, S. (1990). The role of attitude objects in attitude functions. Journal of Experimental Social Psychology, 26, 124–148. Shavitt, S., & Fazio, R. H. (1990). Effects of attribute salience on the consistency of product evaluations and purchase predictions. Advances in Consumer Research, 17, 91–97. Shavitt, S., & Lowrey, T. M. (1992). Attitude functions in advertising effectiveness: The interactive role of product type and personality type. Advances in Consumer Research, 19, 323–328. 569



Shavitt, S., & Nelson, M. R. (2000). The social-identity function in person perception: Communicated meanings of product preferences. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 37–57). Mahwah, NJ: Lawrence Erlbaum. Shavitt, S., & Nelson, M. R. (2002). The role of attitude functions in persuasion and social judgment. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 137–153). Thousand Oaks, CA: Sage. Shea, S., DuMouchel, W., & Bahamonde, L. (1996). A meta-analysis of 16 randomized controlled trials to evaluate computer-based clinical reminder systems for preventive care in the ambulatory setting. Journal of the American Medical Informatics Association, 3, 399–409. Sheeran, P. (2002). Intention-behavior relations: A conceptual and empirical review. European Review of Social Psychology, 12, 1–36. Sheeran, P., & Abraham, C. (2003). Mediator of moderators: Temporal stability of intention and the intention-behavior relation. Personality and Social Psychology Bulletin, 29, 205–215. Sheeran, P., Milne, S., Webb, T. L., & Gollwitzer, P. M. (2005). Implementation intentions and health behaviour. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and practice with social cognition models (2nd ed., pp. 276–323). Maidenhead, UK: Open University Press. Sheeran, P., & Orbell, S. (1998). Do intentions predict condom use? Metaanalysis and examination of six moderator variables. British Journal of Social Psychology, 37, 231–250. Sheeran, P., & Orbell, S. (1999a). Augmenting the theory of planned behavior: Roles for anticipated regret and descriptive norms. Journal of Applied Social Psychology, 29, 2107–2142.



570



Sheeran, P., & Orbell, S. (1999b). Implementation intentions and repeated behaviour: Augmenting the predictive validity of the theory of planned behaviour. European Journal of Social Psychology, 29, 349–369. Sheeran, P., & Orbell, S. (2000a). Self-schemas and the theory of planned behaviour. European Journal of Social Psychology, 30, 533–550. Sheeran, P., & Orbell, S. (2000b). Using implementation intentions to increase attendance for cervical cancer screening. Health Psychology, 19, 283–289. Shen, L. (2013). Incidental affect and message processing: Revisiting the competing hypotheses. Communication Studies, 64, 337–352. Shen, L., & Bigsby, E. (2013). The effects of message features: Content, structure, and style. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 20–35). Thousand Oaks, CA: Sage. Shen, L., & Dillard, J. P. (2014). Threat, fear, and persuasion: Review and critique of questions about functional form. Review of Communication Research, 2, 94–114. Shepherd, G. J. (1985). Linking attitudes and behavioral criteria. Human Communication Research, 12, 275–284. [Erratum notice: Human Communication Research, 12 (1985), 358.] Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modifications and future research. Journal of Consumer Research, 15, 325–343. Sherif, C. W. (1980). Social values, attitudes, and involvement of the self. In M. M. Page (Ed.), Nebraska Symposium on Motivation 1979: Beliefs, attitudes, and values (pp. 1–64). Lincoln: University of 571



Nebraska Press. Sherif, C. W., Kelly, M., Rodgers, H. L., Jr., Sarup, G., & Tittler, B. I. (1973). Personal involvement, social judgment and action. Journal of Personality and Social Psychology, 27, 311–328. Sherif, C. W., Sherif, M., & Nebergall, R. E. (1965). Attitude and attitude change: The social judgment-involvement approach. Philadelphia: W. B. Saunders. Sherif, M., & Hovland, C. I. (1961). Social judgment: Assimilation and contrast effects in communication and attitude change. New Haven, CT: Yale University Press. Sherif, M., & Sherif, C. W. (1967). Attitude as the individual’s own categories: The social judgment-involvement approach to attitude and attitude change. In C. W. Sherif & M. Sherif (Eds.), Attitude, egoinvolvement, and change (pp. 105–139). New York: Wiley. Sherman, D. K., & Cohen, G. L. (2006). The psychology of self-defense: Self-affirmation theory. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 38, pp. 183–242). San Diego: Academic Press. Sherman, D. K., Cohen, G. L., Nelson, L. D., Nussbaum, A. D., Bunyan, D. P., & Garcia, J. (2009). Affirmed yet unaware: Exploring the role of awareness in the process of selfaffirmation. Journal of Personality and Social Psychology, 97, 745–764. Sherman, D. K., Mann, T., & Updegraff, J. A. (2006). Approach/avoidance orientation, message framing, and health behavior: Understanding the congruency effect. Motivation and Emotion, 30, 165–169. Sherman, R. T., & Anderson, C. A. (1987). Decreasing premature termination from psychotherapy. Journal of Social and Clinical 572



Psychology, 5, 298–312. Shiv, B., Edell, J. A., & Payne, J. W. (1997). Factors affecting the impact of negatively and positively framed ad messages. Journal of Consumer Research, 24, 285–294. Shiv, B., Edell, J. A., & Payne, J. W. (2004). Does elaboration increase or decrease the effectiveness of negatively versus positively framed messages? Journal of Consumer Research, 31, 199–208. Siegel, J. T., Alvaro, E. M., Crano, W. D., Lac, A., Ting, S., & Jones, S. P. (2008). A quasi-experimental investigation of message appeal variations on organ donor registration rates. Health Psychology, 27, 170–178. Siegrist, M., Earle, T. C., & Gutscher, H. (2003). Test of a trust and confidence model in the applied context of electromagnetic field (EMF) risks. Risk Analysis, 23, 705–716. Siemer, M., & Joormann, J. (2003). Power and measure of effect size in analysis of variance with fixed versus random nested factors. Psychological Methods, 8, 497–517. Siero, F. W., & Doosje, B. J. (1993). Attitude change following persuasive communication: Integrating social judgment theory and the elaboration likelihood model. European Journal of Social Psychology, 23, 541–554. Sieverding, M., Matterne, U., & Ciccarello, L. (2010). What role do social norms play in the context of men’s cancer screening intention and behavior? Application of an extended theory of planned behavior. Health Psychology, 29, 72–81. Silk, K. J., Weiner, J., & Parrott, R. L. (2005). Gene cuisine or frankenfood? The theory of reasoned action as an audience segmentation strategy for messages about genetically modified foods. Journal of Health Communication, 10, 751–767. 573



Silverthorne, C. P., & Mazmanian, L. (1975). The effects of heckling and media of presentation on the impact of a persuasive communication. Journal of Social Psychology, 96, 229–236. Silvia, P. J. (2005). Deflecting reactance: The role of similarity in increasing compliance and reducing resistance. Basic and Applied Social Psychology, 27, 277–284. Silvia, P. J. (2006a). Reactance and the dynamics of disagreement: Multiple paths from threatened freedom to resistance to persuasion. European Journal of Social Psychology, 36, 673–685. Silvia, P. J. (2006b). A skeptical look at dispositional reactance. Personality and Individual Differences, 40, 1291–1297. Simoni, J., Nelson, K., Franks, J., Yard, S., & Lehavot, K. (2011). Are peer interventions for HIV efficacious? A systematic review. AIDS and Behavior, 15, 1589–1595. Simons, H. W., Berkowitz, N. N., & Moyer, R. J. (1970). Similarity, credibility, and attitude change: A review and a theory. Psychological Bulletin, 73, 1–16. Simsekoglu, O., & Lajunen, T. (2008). Social psychology of seat belt use: A comparison of theory of planned behavior and health belief model. Transportation Research Part F: Traffic Psychology and Behaviour, 11, 181–191. Sinclair, R. C., Moore, S. E., Mark, M. M., Soldat, A. S., & Lavis, C. A. (2010). Incidental moods, source likeability, and persuasion: Liking motivates message elaboration in happy people. Cognition & Emotion, 24, 940–961. Sjoberg, L. (1982). Attitude-behavior correlation, social desirability, and perceived diagnostic value. British Journal of Social Psychology, 21, 574



283–292. Skalski, P., Tamborini, R., Glazer, E., & Smith, S. (2009). Effects of humor on presence and recall of persuasive messages. Communication Quarterly, 57, 136–153. Skowronski, J. J., & Carlston, D. E. (1989). Negativity and extremity biases in impression formation: A review of explanations. Psychological Bulletin, 105, 131–142. Slater, M. D. (1991). Use of message stimuli in mass communication experiments: A methodological assessment and discussion. Journalism Quarterly, 68, 412–421. Slater, M. D. (2002). Involvement as goal-directed strategic processing: Extending the elaboration likelihood model. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 175–194). Thousand Oaks, CA: Sage. Slater, M. D., & Rouner, D. (1996). How message evaluation and source attributes may influence credibility assessment and belief change. Journalism and Mass Communication Quarterly, 73, 974–991. Slater, M. D., & Rouner, D. (2002). Entertainment-education and elaboration likelihood: Understanding the processing of narrative persuasion. Communication Theory, 12, 173–191. Slaunwhite, J. M., Smith, S. M., Fleming, M. T., & Fabrigar, L. R. (2009). Using normative messages to increase healthy behaviours. International Journal of Workplace Health Management, 2, 231–244. Smerecnik, C. M. R., & Ruiter, R. A. C. (2010). Fear appeals in HIV prevention: The role of anticipated regret. Psychology, Health, & Medicine, 15, 550–559.



575



Smidt, K. E., & DeBono, K. G. (2011). On the effects of product name on product evaluation: An individual difference perspective. Social Influence, 6, 131–141. Smit, E. G., van Meurs, L., & Neijens, P. C. (2006). Effects of advertising likeability: A 10-year perspective. Journal of Advertising Research, 46, 73–83. Smith, A. J., & Clark, R. D., III. (1973). The relationship between attitudes and beliefs. Journal of Personality and Social Psychology, 26, 321–326. Smith, D. C., Tabb, K. M., Fisher, D., & Cleeland, L. (2014). Drug refusal skills training does not enhance outcomes of African American adolescents with substance use problems. Journal of Substance Abuse Treatment, 46, 274–279. Smith, J. K., Gerber, A. S., & Orlich, A. (2003). Self-prophecy effects and voter turnout: An experimental replication. Political Psychology, 24, 593–604. Smith, J. L. (1996). Expectancy, value, and attitudinal semantics. European Journal of Social Psychology, 26, 501–506. Smith, J. R., & McSweeney, A. (2007). Charitable giving: The effects of revised theory of planned behavior model in predicting donating intentions and behavior. Journal of Community & Applied Social Psychology, 17, 363–386. Smith, J. R., & Terry, D. J. (2003). Attitude-behaviour consistency: The role of group norms, attitude accessibility, and mode of behavioural decision-making. European Journal of Social Psychology, 33, 591–608. Smith, J. R., Terry, D. J., Manstead, A. S. R., Louis, W. R., Kotterman, D., & Wolfs, J. (2008). The attitude-behavior relationship in consumer conduct: The role of norms, past behaviors, and self-identity. Journal of 576



Social Psychology, 148, 311–334. Smith, M. B., Bruner, J. B., & White, R. W. (1956). Opinions and personality. New York: Wiley. Smith, M. J. (1978). Discrepancy and the importance of attitudinal freedom. Human Communication Research, 4, 308–314. Smith, R. A., & Boster, F. J. (2009). Understanding the influence of others on perceptions of a message’s advocacy: Testing a two-step model. Communication Monographs, 76, 333–350. Smith, R. A., Downs, E., & Witte, K. (2007). Drama theory and entertainment education: Exploring the effects of a radio drama on behavioral intentions to limit HIV transmissions in Ethiopia. Communication Monographs, 74, 133–153. Smith, R. E., & Hunt, S. D. (1978). Attributional processes and effects in promotional situations. Journal of Consumer Research, 5, 149–158. Smith, R. E., & Swinyard, W. R. (1983). Attitude-behavior consistency: The impact of product trial versus advertising. Journal of Marketing Research, 20, 257–267. Smith, S. M., Fabrigar, L. R., & Norris, M. E. (2008). Reflecting on six decades of selective exposure research: Progress, challenges, and opportunities. Social and Personality Psychology Compass, 2, 464–493. Smith, S. M., Haugtvedt, C. P., & Petty, R. E. (1994). Need for cognition and the effects of repeated expression on attitude accessibility and extremity. Advances in Consumer Research, 21, 234–237. Smith, S. W., Atkin, C. K., Martell, D. C., Allen, R., & Hembroff, L. (2006). A social judgment theory approach to conducting formative 577



research in a social norms campaign. Communication Theory, 16, 141–152. Smith-McLallen, A. (2005). Is it true? (When) does it matter? The roles of likelihood and desirability in argument judgments and attitudes (Doctoral dissertation). Retrieved from UMI (UMI No. AAT-3187759). Snyder, M. (1982). When believing means doing: Creating links between attitudes and behavior. In M. P. Zanna, E. T. Higgins, & C. P. Herman (Eds.), Consistency in social behavior: The Ontario Symposium, vol. 2 (pp. 105–130). Hillsdale, NJ: Lawrence Erlbaum. Snyder, M., & DeBono, K. G. (1985). Appeals to image and claims about quality: Understanding the psychology of advertising. Journal of Personality and Social Psychology, 49, 586–597. Snyder, M., & DeBono, K. G. (1987). A functional approach to attitudes and persuasion. In M. P. Zanna, J. M. Olson, & C. P. Herman (Eds.), Social influence: The Ontario Symposium, vol. 5 (pp. 107–125). Hillsdale, NJ: Lawrence Erlbaum. Snyder, M., & DeBono, K. G. (1989). Understanding the functions of attitudes: Lessons from personality and social behavior. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function (pp. 339–359). Hillsdale, NJ: Lawrence Erlbaum. Snyder, M., & Gangestad, S. (1986). On the nature of self-monitoring: Matters of assessment, matters of validity. Journal of Personality and Social Psychology, 51, 125–139. Snyder, M., & Kendzierski, D. (1982). Acting on one’s attitudes: Procedures for linking attitude and behavior. Journal of Experimental Social Psychology, 18, 165–183. Snyder, M., & Rothbart, M. (1971). Communicator attractiveness and 578



opinion change. Canadian Journal of Behavioural Science, 3, 377–387. Soley, L. C. (1986). Copy length and industrial advertising readership. Industrial Marketing Management, 15, 245–251. Solomon, S., Greenberg, J., Psyczynski, T., & Pryzbylinski, J. (1995). The effects of mortality salience on personally-relevant persuasive appeals. Social Behavior and Personality, 23, 177–190. Sorrentino, R. M., Bobocel, D. R., Gitta, M. Z., Olson, J. M., & Hewitt, E. C. (1988). Uncertainty orientation and persuasion: Individual differences in the effects of personal relevance on social judgments. Journal of Personality and Social Psychology, 55, 357–371. Spangenberg, E. R., & Greenwald, A. G. (1999). Social influence by requesting self-prophesy. Journal of Consumer Psychology, 8, 61–69. Spangenberg, E. R., Obermiller, C., & Greenwald, A. G. (1992). A field test of subliminal selfhelp audiotapes: The power of expectancies. Journal of Public Policy and Marketing, 11(2), 26–36. Spangenberg, E. R., Sprott, D. E., Grohmann, B., & Smith, R. J. (2003). Mass-communicated prediction requests: Practical application and a cognitive dissonance explanation for self-prophesy. Journal of Marketing, 67(3), 47–62. Sparks, P., Hedderley, D., & Shepherd, R. (1991). Expectancy-value models of attitude: A note on the relationship between theory and methodology. European Journal of Social Psychology, 21, 261–271. Sparks, P., Jessop, D. C., Chapman, J., & Holmes, K. (2010). Proenvironmental actions, climate change, and defensiveness: Do selfaffirmations make a difference to people’s motives and beliefs about making a difference? British Journal of Social Psychology 49, 553–568.



579



Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why experiments are often more effective than mediational analyses in examining psychological processes. Journal of Personality and Social Psychology, 89, 845–851. Spring, B., Howe, D., Berendsen, M., McFadden, H. G., Hitchcock, K., Rademaker, A. W., & Hitsman, B. (2009). Behavioral intervention to promote smoking cessation and prevent weight gain: A systematic review and meta-analysis. Addiction, 104, 1472–1486. Spring, B., Rademaker, A. W., McFadden, H. G., & Hitsman, B. (2011). Reducing bias in systematic reviews of behavioral interventions: A response to Parsons et al. Addiction, 106, 676–678. Sprott, D. E., Spangenberg, E. R., Block, L. G., Fitzsimons, G. J., Morwitz, V. G., & Williams, P. (2006). The question–behavior effect: What we know and where we go from here. Social Influence, 1, 128–137. Stapel, D. A., & van der Linde, L. A. J. G. (2011). What drives selfaffirmation effects? On the importance of differentiating value affirmation and attribute affirmation. Journal of Personality and Social Psychology, 101, 34–45. See also https://www.commissielevelt.nl/. Stead, M., Tagg, S., MacKintosh, A. M., & Eadie, D. (2005). Development and evaluation of a mass media theory of planned behaviour intervention to reduce speeding. Health Education Research, 20, 36–50. Steadman, L., & Quine, L. (2004). Encouraging young males to perform testicular self-examination: A simple, but effective, implementation intentions intervention. British Journal of Health Psychology, 9, 479–487. Steadman, L., & Rutter, D. R. (2004). Belief importance and the theory of planned behaviour: Comparing modal and ranked modal beliefs in predicting attendance at breast screening. British Journal of Health 580



Psychology, 9, 447–463. Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the self. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 261–302). San Diego: Academic Press. Steenhaut, S., & Van Kenhove, P. (2006). The mediating role of anticipated guilt in consumers’ ethical decision-making. Journal of Business Ethics, 69, 269–288. Steffen, V. J., & Gruber, V. A. (1991). Direct experience with a cancer self-exam: Effects on cognitions and behavior. Journal of Social Psychology, 13, 165–177. Steinfatt, T. M. (1977). Measurement, transformations, and the real world: Do the numbers represent the concept? Et Cetera, 34, 277–289. Stephenson, M. T., Palmgreen, P., Hoyle, R. H., Donohew, L., Lorch, E. P., & Colon, S. E. (1999). Short-term effects of an anti-marijuana media campaign targeting high sensation seeking adolescents. Journal of Applied Communication Research, 27, 175–195. Stephenson, M. T., Quick, B. L., & Hirsch, H. A. (2010). Evidence in support of a strategy to target authoritarian and permissive parents in antidrug media campaigns. Communication Research, 37, 73–104. Stephenson, M. T., Witte, K., Vaught, C., Quick, B. L., Booth-Butterfield, S., Patel, D., & Zuckerman, C. (2005). Using persuasive messages to encourage voluntary hearing protection among coal miners. Journal of Safety Research, 36, 9–17. Sternthal, B., Dholakia, R., & Leavitt, C. (1978). The persuasive effect of source credibility: Tests of cognitive response. Journal of Consumer Research, 4, 252–260.



581



Steward, W. T., Schneider, T. R., Pizarro, J., & Salovey, P. (2003). Need for cognition moderates responses to framed smoking-cessation messages. Journal of Applied Social Psychology, 33, 2439–2464. Stice, E. (1992). The similarities between cognitive dissonance and guilt: Confession as relief of dissonance. Current Psychology Research and Reviews, 11, 69–78. Stice, E., Chase, A., Stormer, S., & Appel, A. (2001). A randomized trial of a dissonance-based eating disorder prevention program. International Journal of Eating Disorders, 29, 247–262. Stice, E., Marti, C. N., Spoor, S., Presnell, K., & Shaw, H. (2008). Dissonance and healthy weight eating disorder prevention programs: Long-term effects from a randomized efficacy trial. Journal of Consulting and Clinical Psychology, 76, 329–340. Stice, E., Shaw, H., Becker, C. B., & Rohde, P. (2008). Dissonance-based interventions for the prevention of eating disorders: Using persuasion principles to promote health. Prevention Science, 9, 114–128. Stiff, J. B., & Mongeau, P. A. (2003). Persuasive communication (2nd ed.). New York: Guilford. Stone, J. (2012). Consistency as a basis for behavioral interventions: Using hypocrisy and cognitive dissonance to motivate behavior change. In B. Gawronski & F. Strack (Eds.), Cognitive consistency: A fundamental principle in social cognition (pp. 326–347). New York: Guilford. Stone, J., Aronson, E., Crain, A. L., Winslow, M. P., & Fried, C. B. (1994). Inducing hypocrisy as a means of encouraging young adults to use condoms. Personality and Social Psychology Bulletin, 20, 116–128. Stone, J., & Cooper, J. (2001). A self-standards model of cognitive dissonance. Journal of Experimental Social Psychology, 37, 228–243. 582



Stone, J., & Fernandez, N. C. (2008a). How behavior shapes attitudes: Cognitive dissonance processes. In W. D. Crano & R. Prislin (Eds.), Attitudes and attitude change (pp. 313–334). New York: Psychology Press. Stone, J., & Fernandez, N. C. (2008b). To practice what we preach: The use of hypocrisy and cognitive dissonance to motivate behavior change. Social and Personality Psychology Compass, 2, 1024–1051. Stone, J., & Fernandez, N. C. (2011). When thinking about less failure causes more dissonance: The effect of elaboration and recall on behavior change following hypocrisy. Social Influence, 6, 199–211. Stone, J., & Focella, E. (2011). Hypocrisy, dissonance and the selfregulation processes that improve health. Self and Identity, 10, 295–303. Stone, J., Wiegand, A. W., Cooper, J., & Aronson, E. (1997). When exemplification fails: Hypocrisy and the motive for self-integrity. Journal of Personality and Social Psychology, 72, 54–65. Strathman, A., Gleicher, F., Boninger, D. S., & Edwards, C. S. (1994). The consideration of future consequences: Weighing immediate and distant outcomes of behavior. Journal of Personality and Social Psychology, 66, 742–752. Straughan, R. D., & Lynn, M. (2002). The effects of salesperson compensation on perceptions of salesperson honesty. Journal of Applied Social Psychology, 32, 719–731. Stroud, N. J. (2008). Media use and political predispositions: Revisiting the concept of selective exposure. Political Behavior, 30, 341–366. Struckman-Johnson, D., & Struckman-Johnson, C. (1996). Can you say condom? It makes a difference in fear-arousing AIDS prevention public service announcements. Journal of Applied Social Psychology, 26, 583



1068–1083. Studts, J. L., Ruberg, J. L., McGuffin, S. A., & Roetzer, L. M. (2010). Decisions to register for the National Marrow Donor Program: Rational vs. emotional appeals. Bone Marrow Transplantation, 45, 422–428. Stukas, A. A., Snyder, M., & Clary, E. G. (2008). The social marketing of volunteerism: A functional approach. In C. P. Haugtvedt, P. M. Herr, & F. R. Kardes (Eds.), Handbook of consumer psychology (pp. 959–979). New York: Lawrence Erlbaum. Sunnafrank, M., (1991). Interpersonal attraction and attitude similarity: A communication-based assessment. Communication Yearbook, 14, 451–483. Sutton, S. (1992). Shock tactics and the myth of the inverted U. British Journal of Addiction, 87, 517–519. Sutton, S. (1998). Predicting and explaining intentions and behavior: How well are we doing? Journal of Applied Social Psychology, 28, 1317–1338. Sutton, S. (2000). A critical review of the transtheoretical model applied to smoking cessation. In P. Norman, C. Abraham, & M. Conner (Eds.), Understanding and changing health behaviour: From health beliefs to self-regulation (pp. 207–225). Amsterdam: Harwood. Sutton, S. (2002). Using social cognition models to develop health behaviour interventions: Problems and assumptions. In D. Rutter & L. Quine (Eds.), Changing health behaviour: Intervention and research with social cognition models (pp. 193–208). Buckingham, UK: Open University Press. Sutton, S. (2004). Determinants of health-related behaviours: Theoretical and methodological issues. In S. Sutton, A. Baum, & M. Johnston 584



(Eds.), The SAGE handbook of health psychology (pp. 94–126). London: Sage. Sutton, S. (2005a). Another nail in the coffin of the transtheoretical model? A comment on West (2005). Addiction, 100, 1043–1046. Sutton, S. (2005b). Stage theories of health behaviour. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and practice with social cognition models (2nd ed., pp. 223–275). Maidenhead, UK: Open University Press. Sutton, S. (2008). How does health action process approach (HAPA) bridge the intention-behavior gap? An examination of the mode’s causal structure. Applied Psychology: An International Review, 57, 66–74. Sutton, S., French, D. P., Hennings, S. J., Mitchell, J., Wareham, N. J., Griffin, S., … Kinmonth, A. L. (2003). Eliciting salient beliefs in research on the theory of planned behaviour: The effect of question wording. Current Psychology, 22, 234–251. Sutton, S., McVey, D., & Glanz, A. (1999). A comparative test of the theory of reasoned action and the theory of planned behavior in the prediction of condom use intentions in a national sample of English young people. Health Psychology, 18, 72–81. Sutton, S. R. (1982). Fear-arousing communications: A critical examination of theory and research. In J. R. Eiser (Ed.), Social psychology and behavioral medicine (pp. 303–337). New York: Wiley. Swartz, T. A. (1984). Relationship between source expertise and source similarity in an advertising context. Journal of Advertising, 13(2), 49–55. Swasy, J. L., & Munch, J. M. (1985). Examining the target of receiver elaborations: Rhetorical question effects on source processing and 585



persuasion. Journal of Consumer Research, 11, 877–886. Sweeney, A. M., & Moyer, A. (2015). Self-affirmation and responses to health messages: A meta-analysis on intentions and behavior. Health Psychology, 34, 149–159. Szilagyi, P., Vann, J., Bordley, C., Chelminski, A., Kraus, R., Margolis, P., & Rodewald, L. (2002). Interventions aimed at improving immunization rates. Cochrane Database Systematic Review, 2002(4), CD003941. Szybillo, G. J., & Heslin, R. (1973). Resistance to persuasion: Inoculation theory in a marketing context. Journal of Marketing Research, 10, 396–403. Tal-Or, N., Boninger, D. S., Poran, A., & Gleicher, F. (2004). Counterfactual thinking as a mechanism in narrative persuasion. Human Communication Research, 30, 301–328. Tal-Or, N., & Cohen, J. (2010). Understanding audience involvement: Conceptualizing and manipulating identification and transportation. Poetics, 38, 402–418. Tal-Or, N., Nemets, E., & Ziv, S. (2009). The transmitter-persistence effect: Resolving the dispute. Social Influence, 4, 274–281. Tao, C.-C., & Bucy, E. P. (2007). Conceptualizing media stimuli in experimental research: Psychological versus attribute-based definitions. Human Communication Research, 33, 397–426. Terry, D. J., & Hogg, M. A. (1996). Group norms and the attitudebehavior relationship: A role for group identification. Personality and Social Psychology Bulletin, 22, 776–793. Terwel, B. W., Harinck, F., Ellemers, N., & Daamen, D. D. L. (2009). 586



Competence-based and integrity-based trust as predictors of acceptance of carbon dioxide capture and storage (CCS). Risk Analysis, 29, 1129–1140. Tewksbury, D., & Scheufele, D. A. (2009). News framing theory and research. In J. Bryant & M. B. Oliver (Eds.), Media effects: Advances in theory and research (3rd ed., pp. 17–33). New York: Routledge. Theodorakis, Y. (1994). Planned behavior, attitude strength, role identity, and the prediction of exercise behavior. Sport Psychologist, 8, 149–165. Thomas, K., Hevey, D., Pertl, M., Chuinneagáin, S. N., Craig, A., & Maher, L. (2011). Appearance matters: The frame and focus of health messages influences beliefs about skin cancer. British Journal of Health Psychology, 16, 418–429. Thompson, E. P., Kruglanski, A. W., & Spiegel, S. (2000). Attitudes as knowledge structures and persuasion as a specific case of subjective knowledge acquisition. In G. R. Maio & J. M. Olson (Eds.), Why we evaluate: Functions of attitudes (pp. 59–95). Mahwah, NJ: Lawrence Erlbaum. Thompson, R., & Haddock, G. (2012). Sometimes stories sell: When are narrative appeals most likely to work? European Journal of Social Psychology, 42, 92–102. Thomsen, C. J., Borgida, E., & Lavine, H. (1995). The causes and consequences of personal involvement. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 191–214). Mahwah, NJ: Lawrence Erlbaum. Thuen, F., & Rise, J. (1994). Young adolescents’ intention to use seat belts: The role of attitudinal and normative beliefs. Health Education Research, 9, 215–223.



587



Thurstone, L. L. (1931). The measurement of social attitudes. Journal of Abnormal and Social Psychology, 26, 249–269. Till, B. D., & Busier, M. (2000). The match-up hypothesis: Physical attractiveness, expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of Advertising, 29(3), 1–14. Todorov, A., Chaiken, S., & Henderson, M. D. (2002). The heuristicsystematic model of social information processing. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 195–211). Thousand Oaks, CA: Sage. Tormala, Z. L., Briñol, P., & Petty, R. E. (2006). When credibility attacks: The reverse impact of source credibility on persuasion. Journal of Experimental Social Psychology, 42, 684–691. Tormala, Z. L., Briñol, P., & Petty, R. E. (2007) Multiple roles for source credibility under high elaboration: It’s all in the timing. Social Cognition, 25, 536–552. Tormala, Z. L., & Rucker, D. D. (2007). Attitude certainty: A review of past findings and emerging perspectives. Social and Personality Psychology Compass, 1, 469–492. Tormala, Z. L., Rucker, D. D., & Seger, C. R. (2008). When increased confidence yields increased thought: A confidence-matching hypothesis. Journal of Experimental Social Psychology, 44, 141–147. Törn, F. (2012). Revisiting the match-up hypothesis: Effects of brandincongruent celebrity endorsements. Journal of Current Issues & Research in Advertising, 33, 20–36. Trafimow, D. (2007). Distinctions pertaining to Fishbein and Ajzen’s theory of reasoned action. In I. Ajzen, D. Albarracín, & R. Hornik (Eds.), Prediction and change of health behavior: Applying the reasoned 588



action approach (pp. 23–42). Mahwah, NJ: Lawrence Erlbaum. Trafimow, D., & Duran, A. (1998). Some tests of the distinction between attitude and perceived behavioural control. British Journal of Social Psychology, 37, 1–14. Trafimow, D., & Sheeran, P. (1998). Some tests of the distinction between cognitive and affective beliefs. Journal of Experimental Social Psychology, 34, 378–397. Trafimow, D., Sheeran, P., Conner, M., & Finlay, K. A. (2002). Evidence that perceived behavioural control is a multidimensional construct: Perceived control and perceived difficulty. British Journal of Social Psychology, 41, 101–121. Trumbo, C. W. (1999). Heuristic-systematic information processing and risk judgment. Risk Analysis, 19, 391–400. Tseng, D. S., Cox, E., Plane, M. B., & Hia, K. (2001). Efficacy of patient letter reminders on cervical cancer screening: A meta-analysis. Journal of General Internal Medicine, 16, 563–568. Tuah, N. A. A., Amiel, C., Qureshi, S., Car, J., Kaur, B., & Majeed, A. (2011). Transtheoretical model for dietary and physical exercise modification in weight loss management for overweight and obese adults. Cochrane Database of Systematic Reviews 2011(10), CD008066. Tufte, E. R. (1997). Visual explanations: Images and quantities, evidence and narrative. Cheshire, CT: Graphics Press. Tukachinsky, R., & Tokunaga, R. S. (2013). The effects of engagement with entertainment. In E. L. Cohen (Ed.), Communication Yearbook 37 (pp. 287–321). New York: Routledge.



589



Tuppen, C. J. S. (1974). Dimensions of communicator credibility: An oblique solution. Speech Monographs, 41, 253–260. Turner, G. E., Burciaga, C., Sussman, S., Klein-Selski, E., Craig, S., Dent, C. W., … Flay, B. (1993). Which lesson components mediate refusal assertion skill improvement in school-based adolescent tobacco use prevention? International Journal of the Addictions, 28, 749–766. Turner, M. M. (2012). Using emotional appeals in health messages. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 59–71). Los Angeles: Sage. Turner, M. M., & Underhill, J. C. (2012). Motivating emergency preparedness behaviors: The differential effects of guilt appeals and actually anticipating guilty feelings. Communication Quarterly, 60, 545–559. Tusing, K. J., & Dillard, J. P. (2000). The psychological reality of the door-in-the-face: It’s helping, not bargaining. Journal of Language and Social Psychology, 19, 5–25. Twyman, M., Harvey, N., & Harries, C. (2008). Trust in motives, trust in competence: Separate factors determining the effectiveness of risk communication. Judgment and Decision Making Journal, 2, 111–120. Tykocinski, O. E., & Bareket-Bojmel, L. (2009). The lost e-mail technique: Use of an implicit measure to assess discriminatory attitudes toward two minority groups in Israel. Journal of Applied Social Psychology, 39, 62–81. Updegraff, J. A., & Rothman, A. J. (2013). Health message framing: Moderators, mediators, and mysteries. Social and Personality Psychology Compass, 7, 668–679. Updegraff, J. A., Sherman, D. K., Luyster, F. S., & Mann, T. L. (2007). 590



The effects of message quality and congruency on perceptions of tailored health communications. Journal of Experimental Social Psychology, 43, 249–257. Uribe, R., Manzur, E., & Hidalgo, P. (2013). Exemplars’ impacts in marketing communication campaigns. Journal of Business Research, 66, SI1787–SI1790. Usdin, S., Singhal, A., Shongwe, T., Goldstein, S., & Shabalala, A. (2004). No short cuts in entertainment-education: Designing Soul City step-bystep. In A. Singhal, M. J. Cody, E. M. Rogers, & M. Sabido (Eds.), Entertainment-education and social change: History, research, and practice (pp. 153–175). Mahwah, NJ: Lawrence Erlbaum. Vakratsas, D., & Ambler, T. (1999). How advertising works: What do we really know? Journal of Marketing, 63(1), 26–43. Valentino, N. A., Banks, A. J., Hutchings, V. L., & Davis, A. K. (2009). Selective exposure in the Internet age: The interaction between anxiety and information utility. Political Psychology, 30, 591–614. Valois, P., Desharnais, R., Godin, G., Perron, J., & LeComte, C. (1993). Psychometric properties of a perceived behavioral control multiplicative scale developed according to Ajzen’s theory of planned behavior. Psychological Reports, 72, 1079–1083. van den Hende, E. A., Dahl, D. W., Schoormans, J. P. L., & Snelders, D. (2012). Narrative transportation in concept tests for really new products: The moderating effect of reader-protagonist similarity. Journal of Product Innovation Management, 29, 157–170. van der Pligt, J., & de Vries, N. K. (1998a). Belief importance in expectancy-value models of attitudes. Journal of Applied Social Psychology, 28, 1339–1354.



591



van der Pligt, J., & de Vries, N. K. (1998b). Expectancy-value models of health behaviour: The role of salience and anticipated affect. Psychology and Health, 13, 289–305. van der Pligt, J., de Vries, N. K., Manstead, A. S. R., & van Harreveld, F. (2000). The importance of being selective: Weighing the role of attribute importance in attitudinal judgment. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 32, pp. 135–200). San Diego: Academic Press. van Enschot-van Dijk, R., Hustinx, L., & Hoeken, H. (2003). The concept of argument quality in the elaboration likelihood model: A normative and empirical approach to Petty and Cacioppo’s “strong” and “weak” arguments. In F. H. van Eemeren, J. A. Blair, C. A. Willard, & A. F. Snoeck Henkemans (Eds.), Anyone who has a view: Theoretical contributions to the study of argumentation (pp. 319–335). Amsterdam: Kluwer. van Harreveld, F., Schneider, I. K., Nohlen, H., & van der Pligt, J. (2012). The dynamics of ambivalence: Evaluative conflict in attitudes and decision making. In B. Gawronski & F. Strack (Eds.), Cognitive consistency: A fundamental principle in social cognition (pp. 267–284). New York: Guilford. van Harreveld, F., van der Pligt, J., & de Vries, N. K. (1999). Attitudes towards smoking and the subjective importance of attributes: Implications for changing risk-benefit ratios. Swiss Journal of Psychology, 58, 65–72. van Ittersum, K., Pennings, J. M. E., Wansink, B., & van Trijp, H. C. M. (2007). The validity of attribute-importance measurement: A review. Journal of Business Research, 60, 1177–1190. Van Koningsbruggen, G. M., & Das, E. (2009). Don’t derogate this message! Self-affirmation promotes online type 2 diabetes risk test taking. Psychology & Health, 24, 635–650. 592



Van Koningsbruggen, G. M., Das, E., & Roskos-Ewoldsen, D. R. (2009). How self-affirmation reduces defensive processing of threatening health information: Evidence at the implicit level. Health Psychology, 28, 563–568. van Laer, T., de Ruyter, K., Visconti, L. M., & Wetzels, M. (2014). The extended transportation-imagery model: A meta-analysis of the antecedents and consequences of consumers’ narrative transportation. Journal of Consumer Research, 40, 797–817. van Leeuwen, L., Renes, R. J., & Leeuwis, C. (2013). Televised entertainment-education to prevent adolescent alcohol use: Perceived realism, enjoyment, and impact. Health Education and Behavior, 40, 193–205. Van Osch, L., Lechner, L., Reubsaet, A., & de Vries, H. (2010). From theory to practice: an explorative study into the instrumentality and specificity of implementation intentions. Psychology and Health, 25, 351–364. Van Overwalle, F., & Jordens, K. (2002). An adaptive connectionist model of cognitive dissonance. Personality and Social Psychology Review, 6, 204–231. van ’t Riet, J., Cox, A. D., Cox, D., Zimet, G. D., De Bruijn, G.-J., van den Putte, B., … Ruiter, R. A. C. (2014). Does perceived risk influence the effects of message framing? A new investigation of a widely held notion. Psychology and Health, 29, 933–949. van ’t Riet, J., & Ruiter, R. A. C. (2013). Defensive reactions to healthpromoting information: An overview and implications for future research. Health Psychology Review, 7(Suppl. 1), S104–S136. van ’t Riet, J., Ruiter, R. A.C., & de Vries, H. (2012). Avoidance orientation moderates the effect of threatening messages. Journal of Health Psychology, 17, 14–25. 593



van ’t Riet, J., Ruiter, R. A. C., Werrij, M. Q., & de Vries, H. (2010). Selfefficacy moderates message-framing effects: The case of skin-cancer detection. Psychology and Health, 25, 339–349. Vaughn, L. A., Childs, K. E., Maschinski, C., Niño, N. P., & Ellsworth, R. (2010). Regulatory fit, processing fluency, and narrative persuasion. Social and Personality Psychology Compass, 4(12), 1181–1192. Vaughn, L. A., Hesse, S. J., Petkova, Z., & Trudeau, L. (2009). “This story is right on”: The impact of regulatory fit on narrative engagement and persuasion. European Journal of Social Psychology, 39, 447–456. Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda on interventions. Decision Sciences, 39, 273–315. Venkatraman, M. P., Marlino, D., Kardes, F. R., & Sklar, K. B. (1990). The interactive effects of message appeals and individual differences on information processing and persuasion. Psychology and Marketing, 7, 85–96. Verplanken, B. (1991). Persuasive communication of risk information: A test of cue versus message processing effects in a field experiment. Personality and Social Psychology Bulletin, 17, 188–193. Vervloet, M., Linn, A. J., van Weert, J. C. M., de Bakker, D. H., Bouvy, M. L., & van Dijk, L. (2012). The effectiveness of interventions using electronic reminders to improve adherence to chronic medication: A systematic review of the literature. Journal of the American Medical Informatics Association, 19, 696–704. Vet, R., de Wit, J. B. F., & Das, E. (2011). The efficacy of social role models to increase motivation to obtain vaccination against hepatitis B among men who have sex with men. Health Education Research, 26, 192–200.



594



Vincus, A. A., Ringwalt, C., Harris, M. S., & Shamblen, S. R. (2010). A short-term, quasi-experimental evaluation of D.A.R.E’s revised elementary school curriculum. Journal of Drug Education, 40, 37–49. Visser, P. S., Bizer, G. Y., & Krosnick, J. A. (2006). Exploring the latent structure of strength-related attitude attributes. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 38, pp. 1–68). San Diego: Elsevier Academic Press. Vitoria, P. D., Salgueiro, M. F., Silva, S. A., & de Vries, H. (2009). The impact of social influence on adolescent intention to smoke: Combining types and referents of influence. British Journal of Health Psychology, 14, 681–699. Waalen, J., Bruning, A. L., Peters, M. J., & Blau, E. M. (2009). A telephone-based intervention for increasing the use of osteoporosis medication: A randomized controlled trial. American Journal of Managed Care, 15, E60–E70. Wachtler, J., & Counselman, E. (1981). When increased liking for a communicator decreases opinion change: An attribution analysis of attractiveness. Journal of Experimental Social Psychology, 17, 386–395. Wagner, W. (1984). Social comparison of opinions: Similarity, ability, and the value-fact distinction. Journal of Psychology, 117, 197–202. Wall, A.-M., Hinson, R. E., & McKee, S. A. (1998). Alcohol outcome expectancies, attitudes toward drinking and the theory of planned behavior. Journal of Studies on Alcohol, 59, 409–419. Wallace, D. S., Paulson, R. M., Lord, C. G., & Bond, C. F., Jr. (2005). Which behaviors do attitudes predict? Meta-analyzing the effects of social pressure and perceived difficulty. Review of General Psychology, 9, 214–227.



595



Walster, E., Aronson, E., & Abrahams, D. (1966). On increasing the persuasiveness of a low prestige communicator. Journal of Experimental Social Psychology, 2, 325–342. Walther, E., & Weil, R. (2012). Balance principles in attitude formation and change: The desire to maintain consistent cognitions about people. In B. Gawronski & F. Strack (Eds.), Cognitive consistency: A fundamental principle in social cognition (pp. 351–368). New York: Guilford. Walther, J. B., Liang, Y. H., Ganster, T., Wohn, D. Y., & Emington, J. (2012). Online reviews, helpfulness ratings, and consumer attitudes: An extension of congruity theory to multiple sources in web 2.0. Journal of Computer-Mediated Communication, 18, 97–112. Walther, J. B., Wang, Z., & Loh, T. (2004). The effect of top-level domains and advertisements on health web site credibility. Journal of Medical Internet Research, 6, e24. Wan, C. S., & Chiou, W. B. (2010). Inducing attitude change toward online gaming among adolescent players based on dissonance theory: The role of threats and justification of effort. Computers & Education, 54, 162–168. Wang, X. (2009). Integrating the theory of planned behavior and attitude functions: Implications for health campaign design. Health Communication, 24, 426–434. Wang, X. (2012). The role of attitude functions and self-monitoring in predicting intentions to register as organ donors and to discuss organ donation with family. Communication Research, 39, 26–47. Warburton, J., & Terry, D. J. (2000). Volunteer decision making by older people: A test of a revised theory of planned behavior. Basic and Applied Social Psychology, 22, 245–258.



596



Ward, C. D., & McGinnies, E. (1973). Perception of communicator’s credibility as a function of when he is identified. Psychological Record, 23, 561–562. Wathen, C. N., & Burkell, J. (2002). Believe it or not: Factors influencing credibility on the web. Journal of the American Society for Information Science and Technology, 53, 134–144. Watt, S. E., Maio, G. R., Haddock, G., & Johnson, B. T. (2008). Attitude functions in persuasion: Matching, involvement, self-affirmation, and hierarchy. In W. D. Crano & R. Prislin (Eds.), Attitudes and attitude change (pp. 189–211). New York: Psychology Press. Webb, T. L., & Sheeran, P. (2007). How do implementation intentions promote goal attainment? A test of component processes. Journal of Experimental Social Psychology, 43, 295–302. Webel, A. R., Okonsky, J., Trompeta, J., & Holzemer, W. L. (2010). A systematic review of the effectiveness of peer-based interventions on health related behaviors in adults. American Journal of Public Health, 100, 247–263. Wechsler, H., Nelson, T. F., Lee, J. E., Seibring, M., Lewis, C., & Keeling, R. P. (2003). Perception and reality: A national evaluation of social norms marketing interventions to reduce college students’ heavy alcohol use. Journal of Studies on Alcohol, 64, 484–494. Wegener, D. T., & Claypool, H. M. (1999). The elaboration continuum by any other name does not smell as sweet. Psychological Inquiry, 10, 176–181. Wegener, D. T., & Petty, R. E. (2001). Understanding effects of mood through the elaboration likelihood and flexible correction models. In L. L. Martin & G. L. Clore (Eds.), Theories of mood and cognition: A user’s guidebook (pp. 177–201). Mahwah, NJ: Lawrence Erlbaum.



597



Weigel, R. H., & Newman, L. S. (1976). Increasing attitude-behavior correspondence by broadening the scope of the behavioral measure. Journal of Personality and Social Psychology, 33, 793–802. Weilbacher, W. M. (2001). Point of view: Does advertising cause a “hierarchy of effects”? Journal of Advertising Research, 41(6), 19–26. Weinberger, M. G., & Dillon, W. R. (1980). The effects of unfavorable product rating information. Advances in Consumer Research, 7, 528–532. Weinstein, N. D. (2000). Perceived probability, perceived severity, and health-protective behavior. Health Psychology, 19, 65–74. Weinstein, N. D. (2007). Misleading tests of health behavior theories. Annals of Behavioral Medicine, 33, 1–10. Weinstein, N. D., Lyon, J. E., Sandman, P. M., & Cuite, C. L. (1998). Experimental evidence for stages of health behavior change: The precaution adoption process model applied to home radon testing. Health Psychology, 17, 445–453. Weinstein, N. D., Rothman, A. J., & Sutton, S. R. (1998). Stage theories of health behavior: Conceptual and methodological issues. Health Psychology, 17, 290–299. Weinstein, N. D., & Sandman, P. M. (1992). A model of the precaution adoption process: Evidence from home radon testing. Health Psychology, 11, 170–180. Weinstein, N. D., & Sandman, P. M. (2002). Reducing the risks of exposure to radon gas: An application of the precaution adoption process model. In D. Rutter & L. Quine (Eds.), Changing health behaviour: Intervention and research with social cognition models (pp. 66–86). Buckingham, UK: Open University Press. 598



Wells, G. L., & Windschitl, P. D. (1999). Stimulus sampling and social psychological experimentation. Personality and Social Psychology Bulletin, 25, 1115–1125. Wenzel, M. (2005). Misperceptions of social norms about tax compliance: From theory to intervention. Journal of Economic Psychology, 26, 862–883. West, M. D. (1994). Validating a scale for the measurement of credibility: A covariance structure modeling approach. Journalism Quarterly, 71, 159–168. West, R. (2005). Time for a change: Putting the transtheoretical (stages of change) model to rest. Addiction, 100, 1036–1039. West, S. K., & O’Neal, K. K. (2004). Project D.A.R.E. outcome effectiveness revisited. American Journal of Public Health, 94, 1027–1029. Whately, R. (1828). The elements of rhetoric (2nd ed.). Oxford, UK: Murray and Parker. White, G. L., & Gerard, H. B. (1981). Postdecision evaluation of choice alternatives as a function of valence of alternatives, choice, and expected delay of choice consequences. Journal of Research in Personality, 15, 371–382. White, K., MacDonnell, R., & Dahl, D. W. (2011). It’s the mind-set that matters: The role of construal level and message framing in influencing consumer efficacy and conservation behaviors. Journal of Marketing Research, 48, 472–485. White, K. M., Robinson, N. G., Young, R. M., Anderson, P. J., Hyde, M. K., Greenbank, S., … Baskerville, D. (2008). Testing an extended theory of planned behaviour to predict young people’s sun safety in a 599



high risk area. British Journal of Health Psychology, 13, 435–448. White, K. M., Smith, J. R., Terry, D. J., Greenslade, J. H., & McKimmie, B. M. (2009). Social influence in the theory of planned behaviour: The role of descriptive, injunctive, and in-group norms. British Journal of Social Psychology, 48, 135–158. Whitehead, J. L., Jr. (1968). Factors of source credibility. Quarterly Journal of Speech, 54, 59–63. Whitehead, J. L., Jr. (1971). Effects of authority-based assertion on attitude and credibility. Speech Monographs, 38, 311–315. Whitelaw, S., Baldwin, S., Bunton, R., & Flynn, D. (2000). The status of evidence and outcomes in stages of change research. Health Education Research, 15, 707–718. Whittaker, J. O. (1963). Opinion change as a function of communicationattitude discrepancy. Psychological Reports, 13, 763–772. Whittaker, J. O. (1965). Attitude change and communication-attitude discrepancy. Journal of Social Psychology, 65, 141–147. Whittaker, J. O. (1967). Resolution of the communication discrepancy issue in attitude change. In C. W. Sherif & M. Sherif (Eds.), Attitude, ego-involvement, and change (pp. 159–177). New York: Wiley. Whittier, D. K., Kennedy, M. G., St. Lawrence, J. S., Seeley, S., & Beck, V. (2005). Embedding health messages into entertainment television: Effect on gay men’s response to a syphilis outbreak. Journal of Health Communication, 10, 251–259. Wicker, A. W. (1969). Attitudes versus actions: The relationship of verbal and overt behavioral responses to attitude objects. Journal of Social 600



Issues, 25(4), 41–78. Wicklund, R. A., & Brehm, J. W. (1976). Perspectives on cognitive dissonance. Hillsdale, NJ: Lawrence Erlbaum. Widgery, R. N., & Ruch, R. S. (1981). Beauty and the Machiavellian. Communication Quarterly, 29, 297–301. Wieber, F., Odenthal, G., & Gollwitzer, P. (2010). Self-efficacy feelings moderate implementation intention effects. Self and Identity, 9, 177–194. Wilkin, H. A., Valente, T. W., Murphy, S., Cody, M. J., Huang, G., & Beck, V. (2007). Does entertainment-education work with Latinos in the United States? Identification and the effects of a telenovela breast cancer storyline. Journal of Health Communication, 12, 455–470. Williams, P., & Drolet, A. (2005). Age-related differences in responses to emotional advertisements. Journal of Consumer Research, 32, 343–354. Williams-Piehota, P., Latimer, A. E., Katulak, N. A., Cox, A., Silvera, S. A. N., Mowad, L., & Salovey, P. (2009). Tailoring messages to individual differences in monitoring-blunting styles to increase fruit and vegetable intake. Journal of Nutrition Education and Behavior, 41, 398–405. Williams-Piehota, P., Pizarro, J., Silvera, S. A. N., Mowad, L., & Salovey, P. (2006). Need for cognition and message complexity in motivating fruit and vegetable intake among callers to the Cancer Information Service. Health Communication, 19, 75–84. Wilmot, W. W. (1971a). Ego-involvement: A confusing variable in speech communication research. Quarterly Journal of Speech, 57, 429–436.



601



Wilmot, W. W. (1971b). A test of the construct and predictive validity of three measures of ego involvement. Speech Monographs, 38, 217–227. Wilson, E. J., & Sherrell, D. L. (1993). Source effects in communication and persuasion research: A meta-analysis of effect size. Journal of the Academy of Marketing Science, 21, 101–112. Winterbottom, A., Bekker, H. L., Conner, M., & Mooney, A. (2008). Does narrative information bias individual’s decision making? A systematic review. Social Science and Medicine, 67, 2079–2088. Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59, 329–349. Witte, K. (1998). Fear as motivator, fear as inhibitor: Using the extended parallel process model to explain fear appeal successes and failures. In P. A. Andersen & L. K. Guerrero (Eds.), Handbook of communication and emotion: Research, theory, applications, and contexts (pp. 423–450). San Diego: Academic Press. Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Education and Behavior, 27, 591–615. Wittenbrink, B. (2007). Measuring attitudes through priming. In B. Wittenbrink & N. Schwarz (Eds.), Implicit measures of attitudes (pp. 17–58). New York: Guilford. Wittenbrink, B., & Schwarz, N. (Eds.). (2007). Implicit measures of attitudes. New York: Guilford. Wogalter, M. S., & Barlow, T. (1990). Injury severity and likelihood in warnings. In Proceedings of the Human Factors Society 34th Annual Meeting (pp. 580–583). Santa Monica, CA: Human Factors Society.



602



Wojcieszak, M. E., & Mutz, D. C. (2009). Online groups and political discourse: Do online discussion spaces facilitate exposure to political disagreement? Journal of Communication, 59, 40–56. Wong, N. C. H., & Cappella, J. N. (2009). Antismoking threat and efficacy appeals: Effects on smoking cessation intentions for smokers with low and high readiness to quit. Journal of Applied Communication Research, 37, 1–20. Wood, M. L. M. (2007). Rethinking the inoculation analogy: Effects of subjects with differing preexisting beliefs. Human Communication Research, 33, 357–378. Wood, W. (1982). Retrieval of attitude-relevant information from memory: Effects on susceptibility to persuasion and on intrinsic motivation. Journal of Personality and Social Psychology, 42, 798–810. Wood, W., & Eagly, A. H. (1981). Stages in the analysis of persuasive messages: The role of causal attributions and message comprehension. Journal of Personality and Social Psychology, 40, 246–259. Wood, W., & Kallgren, C. A. (1988). Communicator attributes and persuasion: Recipients’ access to attitude-relevant information in memory. Personality and Social Psychology Bulletin, 14, 172–182. Wood, W., Kallgren, C. A., & Preisler, R. M. (1985). Access to attituderelevant information in memory as a determinant of persuasion: The role of message attributes. Journal of Experimental Social Psychology, 21, 73–85. Wood, W., & Quinn, J. M. (2003). Forewarned and forearmed? Two metaanalytic syntheses of forewarnings of influence appeals. Psychological Bulletin, 129, 119–138. Wood, W., Rhodes, N., & Biek, M. (1995). Working knowledge and 603



attitude strength: An information-processing analysis. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 283–313). Mahwah, NJ: Lawrence Erlbaum. Wood, W., & Stagner, B. (1994). Why are some people easier to influence than others? In S. Shavitt & T. C. Brock (Eds.), Persuasion: Psychological insights and perspectives (pp. 149–174). Boston: Allyn and Bacon. Woodside, A. G. (2004). Advancing means–end chains by incorporating Heider’s balance theory and Fournier’s consumer–brand relationship typology. Psychology and Marketing, 21, 279–294. Woodside, A. G., & Chebat, J.-C. (2001). Updating Heider’s balance theory in consumer behavior: A Jewish couple buys a German car and additional buying-consuming transformation stories. Psychology and Marketing, 18, 475–496. Woodside, A. G., & Davenport, J. W., Jr. (1974). The effect of salesman similarity and expertise on consumer purchasing behavior. Journal of Marketing Research, 11, 198–202. Wyer, N. A. (2010). Selective self-categorization: Meaningful categorization and the in-group persuasion effect. Journal of Social Psychology, 150, 452–470. Wyer, R. S., Jr. (1974). Cognitive organization and change: An information processing approach. Potomac, MD: Lawrence Erlbaum. Wynn, S. R., Schulenberg, J., Maggs, J. L., & Zucker, R. A. (2000). Preventing alcohol misuse: The impact of refusal skills and norms. Psychology of Addictive Behaviors, 14, 36–47. Xu, A. J., & Wyer, R. S., Jr. (2012). The role of bolstering and counterarguing mind-sets in persuasion. Journal of Consumer Research, 604



38, 920–932. Ybarra, O., & Trafimow, D. (1998). How priming the private self or collective self affects the relative weights of attitudes and subjective norms. Personality and Social Psychology Bulletin, 24, 362–370. Yi, M. Y., Yoon, J. J., Davis, J. M., & Lee, T. (2013). Untangling the antecedents of initial trust in Web-based health information: The roles of argument quality, source expertise, and user perceptions of information quality and risk. Decision Support Systems, 55, 284–295. Yi, Y., & Yoo, J. (2011). The long-term effects of sales promotions on brand attitude across monetary and non-monetary promotions. Psychology and Marketing, 28, 879–896. Yzer, M. (2007). Does perceived control moderate attitudinal and normative effects on intention? A review of conceptual and methodological issues. In I. Ajzen, D. Albarracín, & R. Hornik (Eds.), Prediction and change of health behavior: Applying the reasoned action approach (pp. 111–127). Mahwah, NJ: Lawrence Erlbaum. Yzer, M. (2012a). The integrative model of behavioral prediction as a tool for designing health messages. In H. Cho (Ed.), Health communication message design: Theory and practice (pp. 21–40). Los Angeles: Sage. Yzer, M. (2012b). Perceived behavioral control in reasoned action theory: A dual-aspect interpretation. Annals of the American Academy of Political and Social Science, 640, 101–117. Yzer, M. (2013). Reasoned action theory: Persuasion as belief-based behavior change. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (2nd ed., pp. 120–136). Thousand Oaks, CA: Sage. Yzer, M. C., Cappella, J. N., Fishbein, M., Hornik, R., Sayeed, S., & 605



Ahern, R. K. (2004). The role of distal variables in behavior change: Effects of adolescents’ risk for marijuana use on intention to use marijuana. Journal of Applied Social Psychology, 34, 1229–1250. Yzer, M. C., Fisher, J. D., Bakker, A. B., Siero, F. W., & Misovich, S. J. (1998). The effects of information about AIDS risk and self-efficacy on women’s intentions to engage in AIDS preventive behavior. Journal of Applied Social Psychology, 28, 1837–1852. Yzer, M. C., Southwell, B. G., & Stephenson, M. T. (2013). Inducing fear as a public communication campaign strategy. In R. E. Rice & C. K. Atkin (Eds.), Public communication campaigns (4th ed., pp. 163–176). Los Angeles: Sage. Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribution approach to studying the arousal properties of dissonance. Journal of Personality and Social Psychology, 29, 703–709. Zanna, M. P., & Rempel, J. K. (1988). Attitudes: A new look at an old concept. In D. Bar-Tal & A. W. Kruglanski (Eds.), The social psychology of knowledge (pp. 315–334). Cambridge, UK: Cambridge University Press. Zebregs, S., van den Putte, B., Neijens, P., & de Graaf, A. (2015). The differential impact of statistical and narrative evidence on beliefs, attitude, and intention: A meta-analysis. Health Communication, 30, 282–289. Zeelenberg, M., & Pieters, R. (2007). A theory of regret regulation 1.0. Journal of Consumer Psychology, 17, 3–18. Zhang, X., Fung, H., & Ching, B. H. (2009). Age differences in goals: Implications for health promotion. Aging and Mental Health, 13, 336–348.



606



Zhao, X. S., Lynch, J. G., & Chen, Q. M. (2010). Reconsidering Baron and Kenny: Myths and truths about meditation analysis. Journal of Consumer Research, 37, 197–206. Ziegler, R. (2010). Mood, source characteristics, and message processing: A mood-congruent expectancies approach. Journal of Experimental Social Psychology, 46, 743–752. Ziegler, R. (2013). Mood and processing of proattitudinal and counterattitudinal messages. Personality and Social Psychology Bulletin, 39, 482–495. Ziegler, R. (2014). Mood and processing effort: The mood-congruent expectancies approach. Advances in Experimental Social Psychology, 49, 287–355. Ziegler, R., & Diehl, M. (2001). The effect of multiple source information on message scrutiny: The case of source expertise and likability. Swiss Journal of Psychology, 60, 253–263. Ziegler, R., & Diehl, M. (2011). Mood and multiple source characteristics: Mood congruency of source consensus status and source trustworthiness as determinants of message scrutiny. Personality and Social Psychology Bulletin, 37, 1016–1030. Ziegler, R., Dobre, B., & Diehl, M. (2007). Does matching versus mismatching message content to attitude functions lead to biased processing? The role of message ambiguity. Basic and Applied Social Psychology, 29, 268–278. Zimbardo, P. G., Weisenberg, M., Firestone, I., & Levy, B. (1965). Communicator effectiveness in producing public conformity and private attitude change. Journal of Personality, 33, 233–255. Zuckerman, M. (1979). Sensation seeking: Beyond the optimal level of 607



arousal. Hillsdale, NJ: Lawrence Erlbaum. Zuckerman, M., Gioioso, C., & Tellini, S. (1988). Control orientation, selfmonitoring, and preference for image versus quality approach to advertising. Journal of Research in Personality, 22, 89–100. Zwarun, L., & Hall, A. (2012). Narrative persuasion, transportation, and the role of need for cognition in online viewing of fantastical films. Media Psychology, 15, 327–355.



608



Author Index Aaker, J. L., 54, 157, 222, 255 Aarts, H., 116 Abdulla, R. A., 219 Abelson, R. P., 38, 67, 96–97, 180 Abraham, C., 111, 113, 116, 119, 143, 233, 267 Abrahams, D., 192 Ackers, M., 221 Adams, J., 25, 139 Adams, W. C., 258 Adamski, L., 196 Adamson, G., 118 Adamval, R., 217 Adriaanse, M. A., 114–116 Adrien, A., 131 Agarwal, J., 67 Aggarwal, P., 75 Agnew, C. R., 61, 260 Agrawal, N., 157, 255 Ah Yun, K., 241 Ahern, R. K., 106, 116–117, 131 Ahluwalia, R., 250, 259 Aiken, L. S., 107 Ainsworth, K., 121 Aitken, C. K., 91 Ajzen, I., 4, 6, 10–12, 17, 57, 61, 64–65, 68, 73, 99, 103, 106–107, 109, 113, 115–116, 120, 122–123, 126–127, 130–131, 185 Akhar, O., 212 Akl, E. A., 244 Albarracín, D., 10–12, 72, 84, 96, 103–104, 106, 116–117, 155, 158, 196, 204, 231 Albert, K. A., 111 Alcaraz, K. I., 219 Alden, D. L., 193, 217 Alemagno, S. A., 221 Alemi, F., 221 Alexander, G., 254 Algie, J., 232 609



Allcott, H., 116 Allen, B. P., 216 Allen, C. T., 67 Allen, M., 81, 89, 96, 139, 145, 196, 223, 229–230, 241–242, 246– 247 Allen, M. W., 44, 50 Allen, R., 29 Allport, G. W., 4 Aloe, A. M., 236, 250–251 Alós-Ferrer, C., 81, 95 Al-Rafee, S., 130 Alvaro, E. M., 186, 256 Alvero, A. M., 221 Alwin, D. F., 6, 254 Amaratunga, R., 104 Amass, L., 221 Ambler, T., 142 Amiel, C., 139 Amodio, D. M., 93 Amos, C., 207 Anatol, K. W. E., 189, 193, 210 Andersen, K. E., 189 Andersen, R. E., 221 Anderson, C., 212 Anderson, C. A., 240 Anderson, J. P., 254 Anderson, J. W., 233 Anderson, L., 193 Anderson, L. R., 62 Anderson, N. H., 66, 73 Anderson, P. J., 103 Anderson, R. B., 111 Andersson, E. K., 114 Andreoli, V., 158 Andrews, K. R., 35, 55, 249 Anker, A. E., 218, 236, 250–251, 256 Apanovitch, A. M., 226 Appel, A., 88 Appel, M., 217, 219 Applbaum, R. L., 189, 193, 210 Apsler, R., 153 610



Arazi, D., 199 Arden, M. A., 262 Areni, C. S., 165 Armitage, C. J., 10, 15, 60, 72, 103–104, 106–107, 109, 113–116, 118, 120, 123, 127–128, 130–131, 140, 174, 240, 262 Armor, D. A., 119 Armour, C., 104 Armstrong, A. W., 221 Armstrong, C. L., 194 Armstrong, J. S., xv Arnold, K., 75 Arnold, W. E., 211 Aronsky, D., 221 Aronson, E., 13, 26, 87, 89–91, 93, 97, 192 Aronson, J., 93, 262–263 Arriaga, X. B., 114 Arvola, A., 121 Ary, D. V., 111, 261 Asada, K. J., 182 Ash, L., 221 Ashford, S., 111 Askelson, N. M., 232 Aspinwall, L. G., 262 Assael, H., 181, 258 Astrom, A. N., 131 Atkin, C. K., 29 Atkins, A. L., 25 Atkinson, D. R., 203 Atwood, L. E., 108 Audi, R., 4 Aunger, R., 116 Austin, E. W., 207 Austin, J., 221 Averbeck, J. M., 155 Avery, R. J., 158 Aveyard, P., 133, 243 Avila, R. A., 234, 236, 250 Avom, J., 192 Axsom, D., 38, 154, 159 Aylward, B. S., 256



611



Baayen, R. H., 186 Baazova, A., 227 Babrow, A. S., 10 Backus, J., 53 Baezconde-Garbanati, L., 217 Bagley, G. S., 69, 74, 166 Bagozzi, R. P., 64–65, 72, 118, 121, 130 Bahamonde, L., 221 Bai, H., 261, 267 Bailey, W., 115 Bailis, D. S., 50 Bakker, A. B., 111, 174, 246 Bakker, M., 182 Bala, H., 121 Baldwin, S., 140 Ball, T. B., 221 Balmford, J., 139 Bamberg, S., 11, 139 Banaji, M. R., 5, 9 Banas, J. A., 152, 249, 255, 258, 265–266 Bandura, A., 100, 128 Banerjee, S. C., 218 Banks, A. J., 85 Bansal, H. S., 123 Baranowski, T., 241 Barclay, L. A., 111 Bareket-Bojmel, L., 9 Baril, J., 241 Barlow, T., 69 Barnett, J. P., 212 Barnett, M. A., 145 Baron, P. H., 154 Baron, R. M., 185 Baron, R. S., 154 Barone, M. J., 67 Barrett, D. W., 108 Barrios, I., 218 Barry, C. L., 223 Bartlett, S. J., 221 Baseheart, J., 180 Basil, D. Z., 95 612



Basil, M., 232 Baskerville, D., 103 Bassett, R., 249 Bates, B. R., 233 Bates, D. M., 186 Bates, M., 196 Batra, R., 70 Baudhuin, E. S., 189 Baughan, C. J., 109, 128 Baumeister, R. F., 93, 97 Baumgardner, M. H., 181 Baumgartner, H., 130 Bazzini, D. G., 42 Beaman, A. L., 234, 250 Bearden, W. O., 42 Beatty, M. J., 159, 189, 258 Beatty, S. E., 213 Beauchamp, M., 256 Beauchamp, N., 111 Beauvois, J.-L., 93 Beck, V., 219, 220 Becker, C. B., 88 Bedell, B. T., 226 Beentjes, J. W. J., 218 Behnke, R. R., 189 Beisecker, T. D., 17 Bekker, H. L., 220 Bélanger-Gravel, A., 95 Belch, G. E., 63, 67 Belch, M. A., 63 Bell, D. W., 152 Bell, H., 108 Beltramini, R. F., 192 Bem, D. J., 74 Benac, C. N., 146 Bennett, P., 123 Benoit, W. L., 266 Bensley, L. S., 256 Berek, J. S., 110, 128 Berendsen, M., 224, 243 Berent, M. K., 31 613



Bergin, A. E., 197 Berkanovic, E., 220 Berkel, H. J., 110 Berkowitz, A. D., 108 Berkowitz, N. N., 198, 200–201 Berlo, D. K., 189 Bernard, M. M., 258 Bernhagen, M., 196 Bernstein, G., 110, 128 Berntson, G. G., 74, 225 Berry, M. M., 118, 233 Berry, T. R., 111 Berscheid, E., 201, 204, 211 Betsch, T., 66 Bibby, M. A., 111 Bichard, S. L., 254 Bickel, W. K., 221 Biddle, S. J. H., 112, 116–117 Biek, M., 172 Bieri, J., 25 Biglan, A., 261 Bigman, C. A., 217 Bigsby, E., 237 Bilandzic, H., 218, 220 Bimber, B., 84 Birch, D., 260 Birkimer, J. C., 118, 233 Biswas, A., 207 Biswas, D., 207 Bither, S. W., 258 Bizer, G. Y., 15, 43–44, 54, 168, 169, 171, 174–175 Blake, H., 221 Blanchard, C. M., 60, 116, 123 Blank, D., 244 Blankenship, K. L., 265 Blanton, H., 9 Blau, E. M., 221 Bleakley, A., 117 Bless, H., 150, 152, 158, 255 Block, L. G., 95 Bluemke, M., 9 614



Blumberg, S. J., 259, 267 Bobocel, D. R., 181 Bochner, S., 197, 211 Bock, D. G., 197 Bodenhausen, G. V., 17, 254 Bodur, H. O., 61, 67 Boen, F., 221 Bogle, T., 38 Bogner, F. X., 103 Bohner, G., 17, 88, 152, 159, 166, 197, 255 Bolan, G., 138 Bolls, P., 233 Bolsen, T., 70, 174 Bond, C. F., Jr., 10, 12 Bond, R. M., 108 Bondarenko, O., 203 Bonetti, D., 104 Boninger, D. S., 31, 119, 219, 222, 257 Bonnes, M., 117 Booth, A. R., 131 Booth-Butterfield, S., 97, 117, 187 Bordley, C., 221 Borenstein, M., 96, 130, 182–183, 242, 248, 250, 264–265, 267 Borgida, E., 30, 38, 171 Borland, R., 133, 139 Bose, K., 232 Boster, F. J., 34–35, 49–50, 55, 212, 230, 249 Botta, R. A., 230 Bouman, M., 220 Bouvy, M. L., 221 Bowers, J. W., 180, 189, 193, 210 Bowman, J. M., 212 Boyson, A. R., 232 Bradac, J. J., 180, 181 Bradley, P. H., 190 Brandberg, Y., 103 Brannick, M. T., 88, 183 Brannon, L. A., 55, 75 Branstrom, R., 103 Brasfield, T. L., 107 Brashers, D., 183, 186–187 615



Braverman, J., 219 Brechan, I., 84, 96 Breckon, J. D., 139 Brehm, J. W., 78, 80–81, 97, 199–200, 255–256 Brehm, S. S., 255 Breivik, E., 57, 121 Brewer, N. T., 244 Brickell, T. A., 102 Bridle, C., 139 Brinberg, D., 61, 67, 103, 115 Briñol, P., 9, 15, 17–18, 148, 151, 154, 161, 168–169, 173–175, 191, 196–197, 208, 211, 255, 262, 267 Britt, T. W., 38, 44 Brock, T. C., 55, 75, 154–155, 200, 218 Brodsky, S. L., 190 Broemer, P., 10, 226 Brömer, P., 152 Brommel, B. J., 189, 191, 193, 210 Broneck, K., 256 Brotherton, T. P., 119 Brouwers, M. C., 229 Brown, D., 194 Brown, J., 232 Brown, K. M., 157 Brown, S., 260 Brown, S. P., 67, 118, 207 Brown, T. J., 60 Browne, B. A., 42 Brownstein, A. L., 95 Brug, J., 116, 139, 221 Bruner, J. B., 37, 53, 55 Bruning, A. L., 221 Bryan, C., 114 Bryant, J., 194 Bucholtz, D. C., 253 Bucy, E. P., 184, 247 Buday, R., 241 Budd, R. J., 72, 106, 127 Budney, A. J., 221 Bulger, C. A., 111 Buller, D. B., 154 616



Bullock, J. G., 185 Bundy, C., 256 Bunton, R., 140 Bunyan, D. P., 262 Burack, R. C., 221 Burciaga, C., 261 Burger, J. M., 108, 212, 234–235, 249–250 Burgoon, J. K., 211–212 Burgoon, M., 186, 234, 236, 250, 256 Burgwinkle, P., 241 Burke, C. J., 96 Burkell, J., 211 Burkley, E., 259 Burnell, P., 2 Burney, S., 139 Burnkrant, R. E., 181, 250 Burns, W. J., 213 Bursik, R. J., Jr., 119 Burton, D., 85, 261 Burton, S., 195 Busier, M., 213 Busselle, R., 218, 220 Buswell, B. N., 97 Butera, F., 191, 197, 211 Butler, H. A., 108 Butts, J., 221 Buunk, A. P., 205 Byrne, D., 201 Byrne, S., 158 Cacioppo, J. T., 9, 74, 148–155, 157–161, 163–165, 171–173, 181, 195, 225, 249, 259–260 Cafri, G., 183 Cahill, K., 138–139 Cai, D. A., 26 Cairns, E., 123 Callaghan, R. C., 133 Callendar, A., 196 Calsyn, D. A., 111 Calvi, J., 254 Cameron, K. A., 108 617



Campbell, D. T., 9 Campbell, G., 147 Campbell, K. E., 25, 215 Campbell, L. A., 25 Campbell, M. C., 192 Campbell, N. A., 91, 106, 127 Campo, S., 108, 232 Cannarile, S., 232 Cao, D. S., 137 Capitanio, J. P., 38, 44 Cappella, J. N., 60, 104, 106, 116–118, 131, 217, 223, 232, 241, 254 Cappon, P., 131 Car, J., 139 Carcioppolo, N., 232 Card, N. A., xviii, 182–183 Cardenas, M. P., 110 Carey, K. B., 107 Carey, R. N., 247 Carlsmith, J. M., 26, 86–89 Carlsmith, K. M., 38 Carlson, L., 155 Carlston, D. E., 74 Carnot, C. G., 31 Carpenter, C. J., 35, 42–43, 54–55, 182, 191, 240, 249, 256 Carpenter, J. M., 219–220 Carpenter, K. M., 240 Carrion, M., 232 Carroll, D., 121 Carroll, M. V., 241 Carron, A. V., 103, 109, 112, 123 Carrus, G., 117 Carter, K. D., 232 Carver, C. S., 227 Case, T., 233 Casey, M., 196 Casey, S., 160 Castle, H., 111 Catalan, J., 235–236 Ceci, S. J., 254 Celuch, K., 42, 50 Cerully, J. L., 217, 262 618



Cesario, J., 54, 227, 245 Chaiken, S., 4, 11–12, 17, 19, 34–35, 49–50, 55, 61, 66–67, 69, 72– 73, 76, 85, 93, 104, 106, 123, 149, 154, 158–159, 168, 172–174, 181, 190, 192–193, 198–199, 204–205, 226, 262, 267 Chan, C. W., 241 Chan, D. K.-S., 109, 123 Chandran, S., 244 Chang, C., 159, 227, 253 Chang, C.-T., 226 Chang, M. J., 194 Chantala, K., 244 Chapman, G. B., 113 Chapman, J., 115, 130, 262 Chartrand, T., 250 Chase, A., 88 Chatterjee, J. S., 217 Chattopadhyay, A., 212 Chatzisarantis, N. L. D., 102–103, 112–113, 116–117 Chavez, L. R., 111 Chebat, J.-C., 95, 197–198 Chelminski, A., 221 Chen, H. C., 260 Chen, M. F., 97, 120 Chen, M. K., 81, 95 Chen, Q. M., 185 Chen, S., 149, 158 Chen, T., 111 Chen, X., 166 Cheung, C. M.-Y., 208 Cheung, S. F., 109, 123 Chien, Y. H., 226 Childs, K. E., 219–220 Chin, P. P., 118–119 Ching, B. H., 253 Chiou, J. S., 194 Chiou, W. B., 88 Chiu, C., 227 Cho, H., 232–233 Chong, D., 70 Christenfeld, N. J. S., 159 Chu, G. C., 223, 248 619



Chuang, Y. C., 31 Chuinneagáin, S. N., 123, 223 Chung, A. H., 215, 218–219 Chung, M. H., 221 Chung, S., 26, 66, 154 Churchill, S., 115 Cialdini, R. B., 75, 108, 122, 126–127, 158–159, 173, 235–237, 240, 249, 254 Ciao, A. C., 88 Ciccarello, L., 103, 126 Clack, Z. A., 242 Claiborne, C. B., 38 Clapp, J. D., 108 Clark, E. M., 241, 253 Clark, H. H., 186 Clark, J. K., 26, 161, 197, 208 Clark, J. L., 220 Clark, R. A., 31, 212 Clarkson, J. J., 75 Clary, E. G., 39, 42, 46–47 Clawson, R. A., 70 Claypool, H. M., 154, 168, 175 Cleeland, L., 261 Cobb, M. D., 32 Cody, M. J., 219–220 Cohen, G. L., 93, 262–263 Cohen, J., 219 Cole, C. M., 234, 250 Cole, H. P., 217 Collins, B. E., 17, 19, 24, 30, 34, 37, 93 Collins, W. B., 232 Colon, S. E., 254 Combs, D. J. Y., 192 Compton, J. A., 258, 266 Coney, K. A., 197 Conner, M., 10, 15, 60, 72, 75, 95, 103–104, 106, 109, 113–121, 123, 127–128, 130–131, 143, 174, 220 Considine, J. R., 256 Conville, R. L., 64 Cook, A. J., 131 Cook, F. L., 70 620



Cook, T. D., 85 Cooke, R., 103–104, 112–113, 116 Cooper, H., 182 Cooper, J., 78, 87, 89, 91, 93, 97, 199–200, 205, 267 Cooper, Y., 115 Copeland, J., 39, 42, 47 Corbin, S. K. T., 261 Corby, N. H., 111 Corfman, K., 89 Corker, K. S., 227, 245 Costiniuk, C., 244 Cote, N. G., 11 Côté, S., 152 Cotte, J., 233 Cotton, J. L., 84–85 Coulter, R. A., 204, 233 Counselman, E., 199 Coupey, E., 61, 67 Courchaine, K., 260 Courneya, K. S., 60, 109, 113, 116 Courtright, J. A., 180 Cousins, S., 111 Coveleski, S., 261 Covello, V. T., 193 Covey, J., 226 Cox, A., 254 Cox, A. B., 221 Cox, A. D., 244 Cox, B. S., 221 Cox, D., 244 Cox, D. J., 221 Cox, E., 221 Craciun, C., 143 Craig, A., 123, 223 Craig, J. T., 256 Craig, S., 261 Craig, T. Y., 265 Crain, A. L., 13, 90–91 Cramer, R. J., 137, 190 Cramp, A., 111 Crandall, C. S., 38, 44 621



Crane, L. A., 110, 128 Crano, W. D., 10, 186 Crawley, F. E., III, 123 Creek, T. L., 219–220 Crites, S. L., Jr., 10, 67 Crocker, J., 263 Crockett, W. H., 64, 70, 76 Cron, W. L., 118 Cronen, V. E., 64 Crouse, J. C., 241 Crowley, A. E., 193, 223 Croy, G., 116 Cruz, M. G., 240 Cuijpers, P., 204 Cuite, C. L., 138 Cullen, M., 226 Cunningham, W. A., 9 Cushing, C. C., 256 Czasch, C., 115 Czerwinski, A., 196 Czuchry, M., 240 D’Alessio, D., 81 Daamen, D. D. L., 195, 211 Dagot, L., 249 Dahl, D. W., 212, 218, 257 Dahl, J., 194 Dahlstrom, M. F., 219 Dal Cin, S., 217, 219 Dale, A., 221 Daley, A. J., 103 Dancy, B., 256 Danks, J. H., 230 Dansereau, D. F., 240 Darby, B. L., 235–236 Dardis, F. E., 226, 241 Darke, P. R., 159, 168 Darker, C. D., 111, 121, 131 Darley, J. M., 199–200, 205 Darley, S. A., 87 Das, E., 107, 171, 255, 262–263 622



Das, N., 207 Dashti, A. E., 130 Davenport, J. W., Jr., 200–201 Davidson, D. J., 186 Davies, A., 233 Davis, A. K., 85 Davis, F. D., 121 Davis, J. M., 208 Davis, K. C., 254 Davis, L. L., 42 Davis, M. K., 189 Davis, R. E., 253–254 Davis, R. M., 93 de Bakker, D. H., 221 de Bruijn, G. J., 116 De Bruijn, G.-J., 244 de Cremer, D., 233 de Graaf, A., 218–219, 241 de Hoog, N., 229–230, 232, 242, 247 de Houwer, J., 8, 17 de Nooijer, J., 139 de Ridder, D. T. D., 114–116 de Ruyter, K., 218–219, 241 de Vet, E., 139 de Vord, R. V., 207 de Vries, H., 115, 118, 122, 226, 233, 244, 254, 261 de Vries, N. K., 14, 57, 62, 70, 72, 118–119 de Vries, P., 116 de Vroome, E. M. M., 115 de Wit, J. B. F., 107, 114–116, 229–230, 232, 242, 247 de Young, R., 97 de Zwart, O., 120 Deale, A., 111 Dean, L., 60 Dean, M., 121 Deatrick, L. M., 256 Deaux, K. K., 25 DeBono, K. G., 37, 39, 42–44, 46, 49, 53, 130, 155, 253 DeCesare, K., 249 Decker, L., 196 DeJong, W., 108, 234 623



del Prado, A., 212 Delage, G., 95 Delespaul, P. A. E. G., 97 Delia, J. G., 64, 70, 189, 202–203, 210 DelVecchio, D., 89 Demaine, L. J., 108 Dembroski, T. M., 216 Denizeau, M., 78 Dent, C. W., 261 Derose, S. F., 221 Desharnais, R., 109 Detweiler, J. B., 226 Devine, D. J., xv Dew, D. E., Jr., 204 Dexheimer, J. W., 221 Dholakia, R., 197 Dholakia, R. R., 197, 211–212 Di Dio, P., 114 Di Noia, J., 134–135, 145 Diamond, G. A., 32 Diaz, Y. E., 107 Dibble, T., 114 Dibonaventura, M. D., 113 Dickau, L., 113 Dickel, N., 17 Dickens, C., 256 Dickerson, C. A., 91 DiClemente, C. C., 132–133 Diehl, M., 10, 44, 152, 171, 208 Dijkstra, A., 262 Dijkstra, P., 205 Dillard, A. J., 217 Dillard, J. P., 103–104, 123, 131, 232–234, 236–237, 247, 249–250, 255–256, 264 Dilliplane, S., 233 Dillon, W. R., 211 Dillworth, T., 108 Dingus, T. A., 110 Ditto, P. H., 230 Dittus, P., 72 DiVesta, F. J., 74 624



Dobre, B., 44 Dolan, R. J., 81 Dolich, I. J., 258 Doll, J., 11, 73, 106 Donaldson, S. I., 261 Donnelly, J. H., Jr., 83 Donohew, L., 254 Donovan, R. J., 230 Doob, A. N., 88–89 Doosje, B. J., 25 Dorian, K., 108 Dormandy, E., 103, 115, 130 Dowd, E. T., 256 Downs, E., 219–220 Doyle, S. R., 111 Drevland, G. C. B., 194 Driver, B. L., 57, 61 Drolet, A., 253–254 Druckman, J. N., 70, 174 Druley, J. A., 230 Drummond, A. J., 28 Dryden, J., 119, 233 Duckworth, K. L., 168 Dugan, E., 241 DuMouchel, W., 221 Duncan, S. C., 111 Duncan, T. E., 111 Dunker, K., 230 Dunlop, S. M., 218–219 Duran, A., 109 Durand, J., 103, 115 Durantini, M. R., 204 Dutta-Bergman, M. J., 42 Dwan, K., xviii, 96 Eadie, D., 104, 117 Eagly, A. H., 4, 11–12, 17, 19, 25, 30, 34–35, 49–50, 55, 61, 64, 66– 67, 69, 72–74, 76, 84–85, 93, 96, 104, 106, 123, 159, 163, 171, 190, 192–193, 198–199, 262, 267 Earl, A. N., 204, 231 Earle, T. C., 190 625



Earleywine, M., 93 Eckes, T., 10–11, 112 Edell, J. A., 181 Edmunds, J., 111 Edwards, C. S., 222 Edwards, K., 75 Edwards, S. M., 256 Egloff, B., 233 Ehrlich, D., 81 Eibach, R. P., 240 Eid, M., 262 Eilertsen, D. E., 194 Einwiller, S., 159 Eisend, M., 193, 198, 207, 211, 223, 243–244 Eisenstadt, D., 88 Ekstein, G., 25 El-Alayli, A. G., 118–119 Elder, J. P., 261 Elek, E., 261 Elias, L. J., 113 Ellemers, N., 195, 211 Elliot, M. A., 104 Elliott, M. A., 109, 115, 121, 128 Elliott, R., 62, 72 Elliott, S. M., 194 Ellsworth, R., 219–220 Elms, A. C., 87 Emington, J., 95 Enemo, I., 194 Engstrom, E., 191 Enguidanos, S. M., 111 Ennett, S. T., 261 Ennis, R., 38–39, 46, 74 Epp, L. J., 113 Epstein, E., 207 Epton, T., 262–263 Erb, H.-P., 159, 166, 168, 197 Ernst, J. M., 149 Ernsting, A., 143 Erwin, D. O., 241 Escalas, J. E., 218, 240 626



Esses, V. M., 72, 152, 159 Essex, M., 185 Estabrooks, P., 109, 123 Estambale, B., 221 Evans, A. T., 161, 197, 208 Evans, L. M., 172 Evans, R. I., 216 Evers, K. E., 132, 133, 135, 144–145 Everson, E. S., 103 Eves, F. F., 111, 121, 131 Ewoldsen, D. R., 15 Eyal, K., 208 Fabrigar, L. R., 10, 15, 67, 75, 84, 108, 161, 174, 212 Fagerlin, A., 217 Fairchild, A. J., 185 Fairhurst, A., 42 Falcione, R. L., 189, 193, 210 Faller, C., 261 Falomir-Pichastor, J. M., 191, 197, 211 Fariss, C. J., 108 Farrelly, M. C., 254 Fatoullah, E., 204–205 Fazili, F., 219 Fazio, R. H., 9, 11–12, 17, 57, 70, 198 Feeley, T. H., 236, 250–251 Feiler, D. C., 256 Fein, S., 36 Feinstein, J. A., 154, 172 Feldman, L., 84 Fennis, B. M., 142, 158, 171, 250 Fenson-Hood, K., 230 Ferguson, C. J., xviii Ferguson, E., 119, 233 Fern, E. F., 234, 236, 250 Fernandez, N. C., 78, 91, 93 Fernandez-Medina, K., 218 Festinger, L., 76–77, 82, 86–87 Fetherstonhaugh, D., 160 Feufel, M. A., 110 Fiedler, K., 9 627



Field, A. P., 182 Fielding, A., 133 Figgé, M., 237 Filiatrault, P., 197–198 Fine, B. J., 215 Fink, E. L., 26, 34, 66 Finkelstein, B., 221 Finlay, K. A., 103, 123 Finlayson, B. L., 91 Fiore, C., 134 Firestone, I., 199 Fischer, A. R. H., 122 Fischer, D. L., 160 Fischer, J., 81 Fischer, P., 81 Fishbein, M., 4, 6, 10, 17, 26, 56–57, 60, 62, 64–68, 72, 74, 99, 103– 104, 106–107, 113, 116–117, 120, 122–123, 126–127, 130–131, 138, 166, 185, 254 Fisher, D., 261 Fisher, D. G., 139 Fisher, J. D., 111, 246 Fisher, W. A., 111 Fiske, S. T., 67 Fitzsimons, G. J., 95, 249, 256 Flanagin, A. J., 208 Flay, B. R., 85, 127, 261 Fleming, D., 4 Fleming, J. A., 50 Fleming, J. K., 111 Fleming, M. A., 67, 204 Fleming, M. T., 108 Fleming, S. M., 81 Flood, M. G., 115 Flowelling, R. L., 261 Floyd, D. L., 229, 242, 246 Flynn, D., 140 Foerg, F. E., 221 Folda, L., 232 Fong, G. T., 113, 185, 219 Fontaine, K. R., 221 Fontenelle, G. A., 186 628



Foregger, S., 157, 163 Forehand, M., 159 Fornara, F., 117 Forquer, H., 60 Fowler, J. H., 108 France, C. R., 217 Franckowiak, S. C., 221 Frangos, J. E., 221 Frank, C. A., 256 Frank, L. B., 217, 219 Franks, J., 212 Fraser, S. C., 234, 250 Frazier, B., 219 Freedman, J. L., 26, 88–89, 234, 250, 260 Freijy, T., 91 Freling, T. H., 89 French, D. P., 74, 103–104, 111–112, 116, 121, 131 Freres, D., 60 Frewer, L. J., 190 Frey, D., 81, 84–85 Fried, C. B., 13, 18, 89–91, 97 Friedrich, J., 160 Fry, J. P., 221 Fuchs, R., 142 Fung, H., 253 Gaeth, G. J., 244 Gagné, C., 65, 109, 123, 127–128 Galavotti, C., 219–220 Gallagher, D., 160 Gallagher, K. M., 226, 244–245 Gallagher, S., 118 Gallison, C., 261 Gamble, C., xviii, 96 Gangestad, S. W., 39 Ganster, T., 95 Garcia, C., 207 Garcia, J., 262 Garcia-Marques, T., 154 Gardner, B., 116 Gardner, W. L., 74, 225 629



Garner, R., 212 Garrett, R. K., 85 Garst, J., 154 Gasco, M., 151 Gass, R. H., 2, 17 Gastil, J., 37, 39, 44–45, 159 Gaston, A., 111 Gaudreau, P., 114 Gawronski, B., 17, 95 Gaylord, G., 25 Gaziano, C., 190 Geers, A. W., 254 Gelmon, L. J., 221 Gerard, H. B., 80 Gerber, A. S., 95 Gerend, M. A., 103, 226 Germain, M., 95 Gerrans, P., 116 Gerrard, M., 108 Geurts, D., 111 Ghadiri, A., 221 Gibbons, F. X., 108 Gibson, L., 60 Gierl, H., 75 Giles, M., 104, 123 Gillet, V., 192 Gillett, R., 182 Gillette, J. C., 204 Gilovich, T., 82 Gimotty, P. A., 221 Ginis, K. A. M., 111, 130 Ginsburg, G. P., 189 Gioioso, C., 40, 42 Girvin, H., 139–140 Gitta, M. Z., 181 Glanz, A., 109 Glasgow, R., 261 Glasman, L. R., 10–12 Glazer, E., 194 Gleicher, F., 119, 219, 222, 257 Glik, D., 220 630



Glor, J., 38, 44 Glynn, C. J., 108 Glynn, R. J., 192 Göckeritz, S., 122 Godbold, L. C., 266 Godin, G., 65, 95, 103, 109, 112–113, 123, 127–128, 131 Godinez, M., 111 Goei, R., 212, 232 Goethals, G. R., 203 Goldhagen, J., 221 Goldman, M., 237 Goldman, R., 150, 153, 158, 195 Goldsmith, R. E., 190, 207 Goldstein, M. G., 134 Goldstein, N. J., 108, 122, 127, 237 Goldstein, S., 219 Gollust, S. E., 223 Gollwitzer, M., 9 Gollwitzer, P. M., 95, 114, 116, 216 Golsing, P., 78 Gonzalez, J., 217, 255 Good, A., 267 Goodall, C. E., 8 Gorassini, D. R., 235, 250 Gordon, L., 220 Gorely, T., 221 Gorman, D. R., 261 Goudas, M., 103 Gould, S. J., xvii Gourville, J. T., 244 Goyder, E., 131 Grady, K., 260 Graham, J. W., 261 Graham, S., 118, 120 Granberg, D., 19, 21, 23, 25, 28, 215 Grandpre, J. R., 256 Granić, Ð.-G., 81 Grant, A. M., 256 Grant, H., 54 Grant, N. K., 212 Grasmick, H. G., 119 631



Green, B. F., 7 Green, D. P., 185 Green, L. G., 69 Green, M. C., 154, 218–220, 241 Green, N., 138–139 Greenbank, S., 103 Greenberg, B. S., 211–212 Greenberg, J., 97, 181 Greene, K. L., 103, 186, 218 Greenslade, J. H., 97 Greenwald, A. G., 9, 46, 55, 95, 130, 181 Greenwood, K. M., 133 Gregory, G. D., 253 Gregory, W. L., 240 Greitemeyer, T., 81 Gremmen, F., 186 Grewal, D., 38 Griffin, M. P., 190 Griffin, S., 74, 121 Griskevicius, V., 75, 108, 122, 127, 249, 255 Grofman, B., 25 Grohmann, B., 95 Grossbard, J. R., 261 Gruber, V. A., 11 Gruner, C. R., 194 Guadagno, R. E., 254 Guéguen, N., 212, 240, 249, 256 Guillory, J. E., 158 Gunnell, J. J., 254 Gunning, J., 110, 128 Gunther, A. C., 211 Guo, B. L., 133 Gurwitz, J. H., 241 Gutkin, T. B., 111 Gutscher, H., 190 Gutteling, J., 203 Guttman, I., 81 Ha, S. E., 185 Haaga, D. A. F., 174 Habashi, M. M., 197 632



Habyarimana, J., 221 Hackman, C. L., 104 Hackman, J. R., 62 Haddock, G., 9, 35, 67, 72, 74–75, 219 Hagen, K. M., 111 Hagger, M. S., 54, 103, 112–113, 116–117, 222 Hagtvet, K. A., 123 Hahn, K. S., 84 Hahn, U., 66 Hailey, B. J., 181 Hale, J. L., 103, 186 Hale, S. L., 236, 250 Hall, A., 218 Hall, J. R., 154, 186, 256 Hall, K. L., 134–135 Hall, P. A., 113, 115 Hallam, J. S., 260 Ham, S. H., 60 Hamilton, D. L., 74 Hamilton-Barclay, T., 119, 233 Handley, I. M., 254 Hankins, M., 111 Hansen, W. B., 261 Hänze, M., 152 Hardeman, W., 74, 104, 121 Harinck, F., 195, 211 Harkins, S. G., 149 Harland, P., 120, 131 Harlow, L. L., 134 Harmon, R. R., 197 Harmon-Jones, C., 93 Harmon-Jones, E., 78, 93, 97 Harnish, R. J., 44 Harries, C., 190 Harrington, N. G., 221, 242 Harris, A. J. L., 66 Harris, M. S., 146, 267 Harris, P. R., 131, 262–263 Harris, R. J., 145 Harrison, T. R., 221 Hart, P. S., 158 633



Hart, W., 84, 96 Harte, T., 241 Harterink, P., 120 Hartmann, T., 171, 255 Hartwick, J., 103, 112 Harvey, K., 108 Harvey, N., 190 Harvey, O. J., 20, 26 Hasper, P., 111 Hass, J. W., 69, 74, 166 Hass, R. G., 201, 223, 260 Hassandra, M., 103 Hastings, P., 220 Hatch-Maillette, M. A., 111 Hatzigeorgiadis, A., 103 Hatzios, M., 38 Haugen, J. A., 39, 42, 46–47 Haugtvedt, C. P., 151–152, 154, 163, 255, 265 Hausenblas, H. A., 103, 112 Hauth, A. C., 107 Havermahl, T., 221 Hayes, A. F., 185 Hazlewood, J. D., 159 He, J., 121 Head, K. J., 221, 242 Heald, G. R., 60, 117 Heatherton, T. F., 93 Hecht, M. L., 220, 261 Heckler, S. E., 186 Hedderley, D., 65, 190 Hedeker, D., 127 Hedges, L. V., 96, 130, 182–183, 250 Heene, M., xviii Heesacker, M., 154, 181 Hefner, D., 9 Heiphetz, L., 5 Heitland, K., 88 Hembroff, L., 29 Hemnann, A., 82 Henard, D. H., 89 Henderson, J. E., 199–200, 205 634



Henderson, M. D., 149 Hendricks, L. A., 108 Hendricks, P. S., 256 Henley, N., 230 Hennings, S. J., 74, 121 Herek, G. M., 38–40, 44 Herr, P. M., 67, 74, 95 Herrin, J., 244 Herzog, T. A., 140 Heslin, R., 259 Hesse, S. J., 218 Hether, H. J., 220 Hettema, J. E., 256 Hetts, J. J., 119 Hevey, D., 123, 223 Hewgill, M. A., 191 Hewitt, E. C., 181 Hewstone, M., 9 Hia, K., 221 Hibbert, S., 233 Hidalgo, P., 217, 241 Higgins, A. R., 114 Higgins, E. T., 54, 227 Higgins, J. P. T., 96, 130, 182–183, 250 Higgins, S. T., 221 Highhouse, S., 181 Hill, A. L., 220 Hilligoss, B., 190 Hilmert, C. J., 159 Himmelfarb, S., 199 Hinson, R. E., 123 Hinyard, L. J., 220, 241 Hirsch, H. A., 254 Hirsh, J. B., 254 Hitchcock, K., 224, 243 Hitsman, B., 224, 243 Ho, S. K., 261 Hodson, G., 159 Hoegg, J., 212 Hoeken, H., 111, 165, 218, 254 Hoffmann, K., 66 635



Hogg, M. A., 130 Høie, M., 120 Holbert, R. L., 258, 266 Holbrook, M. B., 61–62, 73, 181 Holland, A., 203 Holmes, G., 207 Holmes, J., 208 Holmes, K., 241, 262 Holzemer, W. L., 212 Homer, P. M., 70 Honeycutt, E. D., 207 Hong, S., 241 Hong, T., 208 Hong, Y.-H., 266 Hopfer, S., 219 Horai, J., 204–205 Horberg, E. J., 114 Horcajo, J., 151 Hornik, R. C., 60, 106, 116–117, 131, 223 Hornikx, J., 222, 253–254 Horswill, M. S., 123 Hospers, H. J., 120 Hossain, S. Z., 44 Hosseinzadeh, H., 44 Householder, B. J., 103 Houston, T., 241 Hove, T., 103, 104 Hovland, C. I., 19–20, 22, 25, 29, 34, 74, 215 Howard, C., 190 Howard, D. J., 181, 208, 212, 240 Howard, G., 196 Howe, B. L., 111 Howe, D., 224, 243 Howell, J. L., 262 Hox, J. J., 114 Hoxworth, T., 138 Hoyer, W. D., 223 Hoyle, R. H., 254 Hoyle, S., 123 Hoyt, W. T., 211 Hsieh, G., 192 636



Hu, Y. F., 208 Huang, G. C., 219–220 Hubbell, A. P., 230 Hubbell, F. A., 111 Huber, F., 82 Hübner, G., 103, 120 Hudson, S. E., 192 Huettl, V., 75 Huge, M. E., 108 Hughes, M., 60 Huh, J. H., 75 Huhmann, B. A., 119, 174 Hukkelberg, S., 131 Hukkelberg, S. S., 123 Hulbert, J. M., 62 Hullett, C. R., 43, 49–50, 171, 255, 264 Hummer, J. F., 261 Hunn, B. P., 110 Hunt, S. D., 83, 193 Hunter, J. E., 10–11, 112, 234, 236, 250 Hunter, R., 66 Hurling, R., 113 Hurwitz, J., 25 Hurwitz, S. D., 191 Huskinson, T. L. H., 74–75 Hustinx, L., 165, 219 Huston, T. L., 201 Hutchings, V. L., 85 Hutchison, A. J., 139 Hwang, Y., 155 Hyde, J., 111 Hyde, M. K., 103 Iannarino, N. T., 221, 242 Iatesta, M., 138 Ibarra, L., 220 Igartua, J. J., 218 Infante, D. A., 200 Insko, C. A., 197–198, 211, 260 Ioannidis, J. P. A., xviii, 127, 182 Ireland, F., 233 637



Ito, T. A., 9 Ivancevich, J. M., 83 Ivanov, B., 258 Iyengar, S., 84 Izuma, K., 81 Jaccard, J., 9, 72, 113 Jack, W., 221 Jackson, S., 179, 181, 183, 186–187 Jacobs, S., 179, 181 Jacobson, R. P., 122 Jaffe, A., 139 Jain, P., 215, 218–219 James, R., 10, 109 Janssen, L., 158, 250 Jarvis, B., 138 Jarvis, W. B. G., 154, 172 Jelinek, S., 227, 245 Jemmott, J. B., 104 Jensen, C. D., 256 Jensen, J. D., 186, 194, 225–226, 232, 244–245 Jeong, E. S., 227 Jeong, S.-H., 155 Jerit, J., 69 Jessop, D. C., 115, 262 Jewell, R. D., 123 Jiang, L., 212 Joag, S., 195 Jobber, D., 62, 72 Jochum, R., 114 Joffe, S., 192 John, L. K., 182 Johnson, B. B., 152 Johnson, B. T., 25, 30, 35, 69, 72, 103–104, 106, 116–117, 163, 165– 166, 171–172, 181, 191 Johnson, H. H., 158, 211 Johnson, J., 108 Johnson, M., 104 Johnson, P. J., 115 Johnson, R. W., 97 Johnston, D. W., 104 638



Johnston, L. H., 139 Johnston, P. L., 118, 233 Johnstone, P. M., 232 Jonas, K., 10, 152 Jones, A., 155 Jones, E. E., 89 Jones, J. J., 108 Jones, J. L., 223 Jones, M. C., 220 Jones, R. A., 199–200, 267 Jones, R. T., 261 Jones, S., 216 Jones, S. P., 186 Joormann, J., 186 Jordan, A., 117 Jordan, B., 196 Jordens, K., 93 Joule, R. V., 89, 93 Joyce, L., 256 Judah, G., 116 Judd, C. M., 5, 28, 186 Julka, D. L., 43 Jun, S. Y., 75 Jung, J. M., 75 Jung, T. J., 60, 117 Juon, H. S., 232 Kahneman, D., 244 Kaigler-Evans, K., 200 Kaiser, F. G., 10, 103, 120 Kaiser, H. F., 96 Kaldenberg, D. O., 42 Kalesan, B., 219 Kalichman, S. C., 107 Kallgren, C. A., 155, 159–160, 172, 181 Kalyanaraman, S., 254 Kamb, K., 138 Kamins, M. A., 181, 213, 258 Kane, R., 262 Kang, S. K., 254 Kang, Y., 254 639



Kang, Y.-S., 159, 199, 205 Kantola, S. J., 91, 106, 127 Kao, C. F., 151, 172 Kao, D. T., 240 Kaplan, C. R., 110, 128 Kaplowitz, S. A., 26, 34 Karan, D., 207 Karanja, S., 221 Kardes, F. R., 181 Kariri, A., 221 Karlan, D., 221 Kashima, Y., 218–219 Kasmer, J., 28 Kath, L. M., 111 Kattapong, K. R., 204 Katulak, N. A., 254 Katz, D., 35–37, 40, 46, 48–49, 51, 53 Kaufmann, M., 66 Kaur, B., 139 Kay, L. S., 111 Kaysen, D. L., 108 Keane, J., 103 Keaveney, S. M., 82 Keegan, O., 61, 127 Keeling, R. P., 108 Keer, M., 121 Kellar, I., 111 Kellaris, J. M., 75 Keller, P. A., 242 Keller, P. S., 192 Kellermann, A. L., 242 Kelly, B. J., 223 Kelly, J. A., 107 Kelly, K. J., 207 Kelly, M., 25 Kelly, R. J., 97 Kelso, E., 240 Keltner, D., 97 Kendzierski, D., 10, 12, 113 Keng, C. J., 81 Kennedy, C., 212 640



Kennedy, M. G., 220 Kennedy, N. B., 267 Kenny, D. A., 28, 185–186 Kenrick, D. T., 75 Kenworthy, J. B., 93 Kerin, R. A., 208, 212 Kerr, G. N., 131 Kerr, P. M., 159, 199, 205 Kesek, A., 9 Kesselheim, A. S., 192 Kessels, L. T. E., 231, 233 Kidder, L. H., 9 Kidwell, B., 123 Kiernan, M., 185 Kiesler, C. A., 19, 24, 30, 34, 37 Killeya, L. A., 69, 165–166 Kilmarx, P. H., 220 Kilmer, J. R., 108 Kim, A., 38 Kim, E. S., 217 Kim, H. K., 219 Kim, H. S., 217, 219 Kim, M.-S., 10, 11, 112 Kim, S., 261 Kimani, J., 221 Kimball, A. B., 221 Kimble, C. E., 203 Kimble, D. L., 111 Kinder, D. R., 67 King, A. J., 221 King, B., 241 King, M., 28 King, S. W., 200, 203 King, W. R., 121 Kinmonth, A. L., 74, 104, 121 Kinsey, K. A., 119 Kirkham, J. J., xviii, 96 Kirmani, A., 192 Kiviniemi, M. T., 121 Klahn, J. A., 139 Klass, E. T., 93 641



Klassen, M., 145 Klein, K. A., 122–123 Klein, W. M. P., 217, 262–263 Kleine, S. S., 67 Klein-Selski, E., 261 Klem, M. L., 241 Klentz, B., 234, 250 Klimmt, C., 9 Klock, S. J., 200 Knapp, T. R., 64 Knäuper, B., 115 Knäuper, B., 240 Knight, K. M., 256 Knowlden, A. P., 104 Knowles, E. S., 257 Koehler, J. W., 259 Koerner, A. F., 266 Koestner, R., 114 Kok, G., 103, 112, 120, 231–233, 248 Konkel, J., 254 Koob, J. J., 107 Kootsikas, A., 261 Kopperhaver, T., 220 Koring, M., 114–115, 143 Kosmidou, E., 103 Koster, R., 81 Kothe, E. J., 91, 104 Kotowski, M. R., 232 Kotterman, D., 103 Kovac, V. B., 123 Kraemer, H. C., 185 Kraft, J. M., 219–220 Kramer, A. D. I., 108 Krantz, L. H., 115 Kraus, R., 221 Kraus, S. J., 10–11 Kraut, R. E., 192 Krcmar, M., 186 Kreausukon, P., 114 Kremers, S. P. J., 116, 221 Kreuter, M. W., 219–220, 241, 253 642



Krieger, J. L., 223, 232, 261 Krishnamurthy, P., 257 Kroese, F. M., 116 Kroeze, W., 116 Kromrey, J. D., 183 Krosnick, J. A., 5, 15, 18, 28, 31, 171, 174, 254 Kruger, M. W., 159, 189 Kruglanski, A. W., 72, 82, 166–169, 173, 175 Kuan, K. K. Y., 208 Kuhlmann, A. K. S., 219–220 Kuiper, N. M., 118–119 Kujawski, E., 196 Kulik, J. A., 108, 159 Kumkale, G. T., 158, 196 Kupfer, D. J., 185 Kvedar, J. C., 221 Kwak, L., 221 Kyriakaki, M., 222 LaBrie, J. W., 261 Lac, A., 186 Lacaille, J., 240 Laczniak, R. N., 155 Lafferty, B. A., 207 Lähteenmäki, L., 121 Lai, M. K., 261 Lajunen, T., 103, 118 Lally, P., 116 Lalor, K. M., 181 Lam, S.-P., 103 Lam, T. H., 261 Lam, W. K., 111 Lampton, W. E., 194 Lancaster, T., 138–139 Landauer, T. K., 88–89 Landman, J., 119 Landy, D., 159 Lane, D. M., 186 Lane, K. A., 9 Lane, L. T., 256 Lang, B., 240 643



Lange, J. E., 108 Lange, R., 26 Langlois, M. A., 260 Langner, T., 198, 207 LaPiere, R. T., 17 Lapinski, M. K., 232 Larimer, M. E., 108 Larkey, L. K., 217, 220, 254 Laroche, M., 197–198 Larsen, E., 120 LaSalvia, C. T., 108 Lasater, T. M., 216 Latimer, A. E., 111, 130, 226–227, 254 Laufer, D., 253 Lauver, D., 64 Lavidge, R. J., 142 Lavin, A., 221 Lavine, H., 30, 42–43, 46, 171 Lavis, C. A., 171, 208, 255 Lawton, R. J., 75, 103, 106, 115–117, 121, 127–128 Le Dreff, G., 212 Leader, A. E., 217, 223 Learmonth, M., 111 Leary, M. R., 223 Leavitt, A., 53 Leavitt, C., 197, 200, 211–212 LeBlanc, B. A., 97 Lechner, L., 115, 118 LeComte, C., 109 Lee, C. M., 108 Lee, E.-J., 157 Lee, G., 256 Lee, J. E., 108 Lee, J. H., 256 Lee, K. H., 130 Lee, M. J., 254 Lee, M. S., 207 Lee, S., 221 Lee, T., 208 Lee, W., 266 Lee, W. J., 256 644



Leeuwis, C., 219 Lehavot, K., 212 Lehmann, D. R., 242, 256 Leibold, J. M., 118–119 Leippe, M. R., 88, 181 Lemert, J. B., 189 Lemus, D. R., 208 Lennon, S. J., 42 Leone, L., 118 Lepage, L., 109 Lerman, C., 60, 217 Leshner, G., 233 Lester, R. T., 221 Leventhal, H., 143, 216, 248 Levin, I. P., 244 Levin, K. D., 30, 69, 163, 165–166, 171 Levine, T. R., 157, 163, 182, 231 Levinger, G., 201 Levitan, L. C., 157 Levy, B., 199 Levy, D. A., 17 Lewis, C., 108 Lewis, H., 123 Lewis, M. A., 108 Lewis, S. K., 235–236 Li, H. R., 256 Liang, Y. H., 95 Liao, T.-H., 81 Libby, L. K., 240 Liberman, A., 262, 267 Lichtenstein, E., 261 Lieberman, D. A., 241 Likert, R., 7 Lim, H., 212 Lim, S., 208 Lin, H.-Y., 25 Lin, J.-H., 241 Lin, W.-K., 258, 266 Lindberg, M. J., 84, 96 Linder, D. E., 89, 223, 240 Lindow, F., 66 645



Lindsey, L. L. M., 212, 241 Linn, A. J., 221 Linn, J. A., 257 Linnemeier, G., 232 Lipkus, I. M., 69, 262 Lippke, S., 114–115, 137, 139, 143 Litman, J., 115 Littell, J. H., 139–140 Liu, K., 265 Liu, W.-Y., 230 Livi, S., 97, 131 Loersch, C., 154 Loewenstein, G., 182 Loh, T., 208 Longdon, S., 121, 131 Longoria, Z. N., 114 Lorch, E. P., 254 Lord, C. G., 10, 12, 240 Lord, K. R., 207 Losch, M., 232 Louis, W. R., 103 Love, G. D., 220 Lowe, R., 121 Lowrey, T. M., 43, 53 Lu, A. S., 241 Luce, M. F., 240 Luchok, J. A., 191 Lukwago, S. N., 253 Lundell, H., 219 Lunney, C. A., 108 Lupia, A., 198 Luszczynska, A., 111 Lutz, R. J., 67, 69, 74, 165 Luyster, F. S., 181 Luzzo, D. A., 111 Lycett, D., 243 Lynch, J. G., 185 Lynn, M., 75, 192 Lyon, J. E., 138 Lyon-Callo, S. K., 232 Lyons, E., 97, 103, 131 646



Mabbott, L., 262 MacDonald, T. K., 15, 174 MacDonnell, R., 257 Machleit, K. A., 67 Mack, D. E., 103, 112 MacKenzie, S. B., 67, 151 Mackie, D. M., 152, 154, 204, 233, 255 MacKinnon, D. P., 185 MacKintosh, A. M., 104, 117 MacNair, R., 260 MacRae, S., 111 Madden, T. J., 73, 109 Maddux, J. E., 205–206 Magana, J. R., 111 Magee, R. G., 254 Maggs, J. L., 260–261 Magnini, V. P., 207 Magnussen, S., 194 Maher, L., 123, 223 Mahler, H. I. M., 108 Mahloch, J., 106 Maio, G. R., 9, 35, 39, 48–51, 152, 159, 258 Majeed, A., 139 Major, A. M., 108 Makredes, M., 221 Malhotra, N. K., 67 Mallach, N., 143 Mallett, J., 104 Malloy, T. E., 111 Mallu, R., 11 Mallya, G., 60, 117 Maloney, E. K., 232 Malotte, C. K., 138 Maltarich, S., 230 Manchanda, R. V., 174 Mangleburg, T. F., 38 Manis, M., 25 Mann, T., 181, 227 Mannetti, L., 82, 97, 131, 166, 168 Manning, M., 103, 112, 116–117, 122–123, 127, 129, 131 Manson-Singer, S., 131 647



Manstead, A. S. R., 62, 70, 72, 103, 109, 116, 118–120, 257 Manzur, E., 217, 241 Mara, M., 219 Marcus, A. C., 69, 110, 128 Marcus, B. H., 134 Margolis, P., 221 Marin, B. V., 116, 253 Marin, G., 116, 253 Mark, M. M., 171, 208, 255 Marks, L. J., 181 Marlino, D., 181 Marlow, C., 108 Marquart, J., 211 Marra, C. A., 221 Marsh, K. L., 43 Marston, A., 212 Marteau, T. M., 103, 111, 115, 130 Martell, C., 29 Martell, D., 122–123 Marti, C. N., 88 Martin, B. A. S., 240 Martin, J., 114 Martinelli, E. A., Jr., 111 Martinie, M. A., 89 Marttila, J., 139 Maschinski, C., 219–220 Mason, H. R. C., 261 Mason, T. E., 103 Masser, B., 217 Mather, L., 139 Matheson, D. H., 123 Mathios, A. D., 158 Maticka-Tyndale, E., 131, 212 Mattern, J. L., 108 Matterne, U., 103, 126 Matthes, J., 203 Maurissen, K., 221 Maxfield, A. M., 232 May, K., 196 Mayle, K., 262 Mazmanian, L., 159 648



Mazor, K. M., 241 McAdams, M. J., 194 McBride, J. B., 253 McCallum, D. B., 193 McCann, R. M., 208 McCaslin, M. J., 9, 154 McCaul, K. D., 138 McClenahan, C., 104, 118 McClintock, C., 40 McCollam, A., 240 McConnell, A. R., 118–119 McConnell, M., 221 McCroskey, J. C., 189, 191, 193, 210–211, 259 McCubbins, M. D., 198 McDermott, D. T., 247 McDonald, L., 230 McEachan, R. R. C., 75, 103, 106, 116–117, 121, 127–128 McFadden, H. G., 224, 243 McFall, S., 85 McGee, H., 61, 127 McGilligan, C., 118 McGinnies, E., 197, 210 McGowan, L., 256 McGrath, K., 190 McGuffin, S. A., 241 McGuire, W. J., 180, 215, 258–259, 266 McIntosh, A., 154 McIntyre, P., 145 McKay-Nesbitt, J., 174 McKee, S. A., 123 McKimmie, B. M., 97 McLarney, A. R., 254 McLeod, J. H., 261 McMahon, T. A., 91 McMath, B. F., 137 McMillan, B., 120, 123, 131 McMillion, P. Y., 111 McNamara, M., 241 McNeil, B. J., 244 McQueen, A., 219, 262 McRee, A. L., 244 649



McSpadden, B., 218 McSweeney, A., 116, 120 McVeigh, J. F., 237 McVey, D., 109 Medders, R. B., 208 Meertens, R. M., 69 Meffert, M. F., 66 Mehdipour, T., 108 Mehrotra, P., 130 Meijinders, A., 203 Mellor, S., 111 Melnyk, V., 122 Menon, G., 157, 244, 255 Mercier, H., 159 Merrill, L., 84, 96 Merrill, S., 25 Mertz, R. J., 189 Merwin, J. C., 74 Messian, N., 212 Messner, C., 9 Metzger, M. J., 208 Metzler, A. E., 260 Meuffels, B., 241 Mevissen, F. E. F., 69 Meyerowitz, B. E., 181, 226 Miarmi, L., 155 Michie, S., 103, 115, 130 Micu, C. C., 204 Midden, C. J. H., 116, 203 Middlestadt, S. E., 57, 60, 67, 74, 121, 126 Miene, P. K., 39, 42, 46–47 Millar, K. U., 11 Millar, M. G., 11 Millar, S., 104 Miller, C., 256 Miller, C. H., 256 Miller, D., 91 Miller, G. R., 180, 191, 211–212 Miller, J. A., 249 Miller, N., 19, 24, 30, 34, 37, 93, 154 Miller, R. L., 95 650



Miller, W. R., 256 Miller-Day, M., 261 Mills, E. J., 221 Mills, J., 78, 81, 203 Milne, S., 114, 229, 246 Milton, A. C., 117 Minassian, L., 220 Miniard, P. W., 67 Mio, G. R., 75 Miron, A. M., 256 Miron, M. S., 191 Misak, J. E., 93 Mischkowski, D., 263 Mishra, S. I., 111 Misovich, S. J., 111, 246 Misra, S., 213 Mitchell, A. A., 67 Mitchell, A. L., 204 Mitchell, G., 9 Mitchell, J., 74, 121 Mitchell, M. M., 157 Mittal, B., 41 Mittelstaedt, J. D., 213 Mkandawire, G., 232 Mladinic, A., 64, 67, 74 Moan, I. S., 120, 131 Moldovan-Johnson, M., 60 Molyneaux, V., 115 Mondak, J. J., 161 Mongeau, P. A., 229–230, 232–233, 237 Monroe, K. B., 234, 236, 250 Montaño, D. E., 106 Montoya, H. D., 108 Mooki, M., 219–220 Mooney, A., 220 Moons, W. G., 227, 233 Moore, B., 118, 120 Moore, D. J., 260 Moore, K., 131 Moore, K. A., 230 Moore, M., 233 651



Moore, S. E., 171, 208, 255 Moors, A., 8 Mora, P. A., 143 Morales, A. C., 249 Moran, M. B., 219 Morgan, M. G., 107 Morgan, S. E., 217, 221, 254 Morman, M. T., 218 Morris, B., 75 Morris, K., 121, 131 Morrison, K., 230 Morris-Villagran, M., 157 Mortensen, C. R., 122 Morton, K., 256 Morwitz, V. G., 95 Moser, G., 112 Moss, T. P., 114 Mouttapa, M., 220 Mowad, L., 174, 254 Mowen, J. C., 195 Moyer, A., 262 Moyer, R. J., 198, 200–201 Moyer-Gusé, E., 215, 218–220, 241 Muehling, D. D., 155 Muellerleile, P. A., 72, 103–104, 106, 116–117 Mugny, G., 191, 197, 211 Mullainathan, S., 221 Mullan, B. A., 104, 117 Müller-Riemenschneider, F., 221 Munch, J. M., 181, 253 Munn, W. C., 194 Murayama, K., 81 Murphy, S. T., 217, 219–220 Murray, T. C., 123 Murray-Johnson, L., 230, 232 Muthusamy, N., 231 Mutsaers, K., 220 Mutz, D. C., 85 Myers, J. A., 192 Myers, L. B., 123



652



Nabi, R. L., 219, 233, 255, 258 Naccache, H., 95 Naccari, N., 204–205 Nahai, A., 227 Nail, P. R., 17, 93, 263 Najafzadeh, M., 221 Nakahiro, R. K., 221 Nan, X., 174, 196, 207, 211, 217, 262 Nanneman, T., 28 Napper, L. E., 139, 262 Nathanson, A., 119 Nava, P., 111 Nayak, S., 241 Nebergall, R. E., 19, 23–25, 30, 33–34, 215 Neff, R. A., 221 Neighbors, C., 108 Neijens, P., 241 Neijens, P. C., 121, 207 Neimeyer, G. J., 260 Nell, E. B., 258 Nelson, D. E., 97 Nelson, K., 212 Nelson, L. D., 262 Nelson, M. R., 35, 46 Nelson, R. E., 203 Nelson, T. E., 70 Nelson, T. F., 108 Nemets, E., 257 Nestler, S., 233 Netemeyer, R. G., 195 Neudeck, E. M., 108 Neufeld, S. L., 255 Newell, S. J., 190, 207 Newman, L. S., 11 Ng, J. Y. Y., 111 Ng, M., 115 Ng, S. H., 44, 50 Ngugi, E., 221 Nichols, A. J., III, 57 Nichols, D. R., 30, 69, 163, 171 Nickerson, D. W., 114 653



Niederdeppe, J., 60, 218–219, 223, 254 Niedermeier, K. E., 118–119 Nienhuis, A. E., 257 Nigbur, D., 97, 103, 131 Niiya, Y., 263 Niño, N. P., 219–220 Nitzschke, K., 221 Noar, S. M., 130, 146, 221, 242 Nocon, M., 221 Nohlen, H., 15, 174 Nolan, J. M., 127 Norman, P., 95, 109, 113, 115–116, 123, 130–131 Norman, R., 205 Norris, M. E., 84 North, D., 106 Nosek, B. A., 9 Notani, A. S., 67, 103 Ntumy, R., 219–220 Nupponen, R., 139 Nussbaum, A. D., 262 Oaten, M., 233 Oats, R. G., 111 Oberle, D., 78 Obermiller, C., 130 O’Carroll, R. E., 119, 233 Odenthal, G., 114 Oenema, A., 116 Oettingen, G., 95 Offermans, N., 118 Oh, H. J., 103, 104 Ohanian, R., 190, 207 O’Hara, B. S., 195 Ohman, S., 203 Ojala, M., 97 O’Keefe, D. J., xviii, 10, 18, 34, 54, 64, 69–70, 79, 93, 147, 160, 165–166, 184, 186–187, 191, 193, 196, 211, 215–216, 222–227, 233, 236–237, 240–245, 247, 250, 253, 264 O’Keefe, G. J., 211 Okonsky, J., 212 Okun, M. A., 253–254 654



O’Leary, A., 220 Olofsson, A., 203 Olson, J. M., 35, 39, 48–51, 181, 235, 250, 258 Olson, P., 196 Olson, R., 221 Omoto, A. M., 130 O’Neal, K. K., 261, 267 Opdenacker, J., 221 Orbell, S., 54, 112, 114, 118–119, 130–131, 222, 229, 246 O’Reilly, K., 212 Orfgen, T., 157, 163 Orlich, A., 95 Orrego, V., 232 Orth, B., 106 Osgood, C. E., 5, 76 O’Sullivan, B., 61, 127 Oswald, F. L., 9 Otero-Sabogal, R., 116, 253 Otto, S., 64, 67, 74 Ouellette, J. A., 112, 115 Oxley, Z. M., 70 Oxman, A. D., 244 Packer, D. J., 9 Packer, M., 42, 44, 53 Paek, H.-J., 103–104 Palmgreen, P., 254 Pan, L. Y., 194 Pan, W., 261, 267 Papageorgis, D., 266 Pappas-DeLuca, K. A., 220 Parenteau, A., 196 Park, H. S., 122–123, 157, 163, 241 Parker, D., 109, 118–119 Parker, R., 111 Parrott, R. L., 116 Parschau, L., 114–115, 143 Parson, D. W., 17 Parsons, A., 243 Parvanta, S., 60 Pascual, A., 240, 249, 256 655



Pasha, N. H., 258, 266 Passafaro, P., 117 Patel, D., 187, 232 Patel, S., 212 Patnoe-Woodley, P., 219 Pattenden, J., 139 Patzer, G. L., 205–206 Pauker, S. G., 244 Paulson, R. M., 10, 12 Paulussen, T., 116 Paunesku, D., 212 Payne, J. W., 181 Pearce, W. B., 189, 191, 193, 210 Peay, M. Y., 61, 74 Pechmann, C., 193, 223 Peck, E., 249 Pedlar, C., 256 Penaloza, L. C., 266 Peng, W., 241 Pennings, J., 72 Perez, M., 88 Perez-Stable, E. J., 116, 253 Perloff, R. M., 237 Perron, J., 109 Pertl, M., 123, 223 Perugini, M., 113, 118 Peters, G.-J. Y., 231, 233, 248 Peters, M. D., 67 Peters, M. J., 221 Peters, R. G., 193 Peterson, M., 253 Petkova, Z., 218 Petosa, R., 260 Petraitis, J., 127 Petrova, P. K., 240 Pettigrew, J., 261 Petty, R., 119 Petty, R. E., 9–10, 15, 17–18, 43–44, 54, 67, 75, 148–155, 157–161, 163–165, 168–169, 171–175, 181, 191, 195–197, 204, 208, 210–211, 255, 259–260, 262, 267 Pfau, M. W., 186, 258, 266 656



Phillips, A. P., 186 Phillips, C., 260 Phillips, W. A., 189, 193, 210 Piccinin, A. M., 261 Pichot, N., 212 Piercy, L., 217 Pierro, A., 82, 97, 131, 166, 168 Pieters, R., 82 Pietersma, S., 262 Pietrantoni, L., 217 Pinckert, S., 250 Pinkleton, B. E., 207 Piotrowski, J. T., 117 Pitts, S. R., 242 Pizarro, J., 174, 226 Plane, M. B., 221 Plessner, H., 66 Plotnikoff, R. C., 60, 116 Plummer, F. A., 221 Poehlman, T. A., 9 Pollard, K., 220 Polonec, L. D., 108 Polyorat, K., 217 Popova, L., 232, 249 Poran, A., 219, 257 Pornpitakpan, C., 197 Porticella, N., 60, 218 Porzig-Drummond, R., 233 Posavac, E. J., 204 Potts, K. A., 256 Povey, R., 10, 109, 118 Powers, P., 17 Powers, T., 114 Prabhakar, P., 41 Prapavessis, H., 111 Prati, G., 217 Pratkanis, A. R., 46, 55, 181, 237 Praxmarer, S., 205 Preacher, K. J., 185 Preisler, R. M., 155, 159–160, 172, 181 Preiss, R. W., 89, 96, 241 657



Prelec, D., 182 Prentice, D. A., 38 Prentice-Dunn, S., 137, 228–229, 242, 246 Presnell, K., 88 Press, A. N., 64, 70 Preston, M., 234, 250 Prestwich, A., 111, 113, 115 Pretty, G. M., 103 Price, L. L., 204 Priester, J. R., 67, 148, 161, 172, 210 Primack, B. A., 241 Prince, M. A., 107 Prislin, R., 10, 262 Pritt, E., 232 Prochaska, J. O., 132–135, 144–145 Pronin, E., 226 Prothero, A., 256 Pruyn, A. T. H., 158 Pryor, B., 265 Pryzbylinski, J., 181 Psyczynski, T., 181 Puckett, J. M., 160 Queller, S., 204 Quick, B. L., 187, 233, 254, 256, 264 Quine, L., 130 Quinlan, K. B., 138 Quinlan, M. R., 233 Quinn, J. M., 157, 259–260, 265–267 Qureshi, S., 139 Raaijmakers, J. G. W., 186 Raats, M. M., 121 Radecki, C. M., 72, 113 Rademaker, A. W., 224, 243 Rae, G., 104 Raghubir, P., 89 Rains, S. A., 256, 258, 265–266 Rakowski, W., 134 Ramirez, A., 88 Randall, D. M., 130 658



Rasanen, M., 103 Ratchford, B., 41 Ratcliff, C. D., 240 Rauner, S., 249 Rea, C., 260 Read, S. J., 93 Reading, A. E., 110, 128 Real, K., 232 Reardon, R., 260 Redding, C. A., 132–135, 144–145 Reed, M., 249 Reed, M. B., 262 Reeve, A., 2 Regan, D. T., 11 Reger, B., 117 Reichert, T., 186 Reid, A. E., 107 Reid, J. C., 114 Reidy, J. G., 240 Reilly, S., 196 Reimer, T., 210 Reinard, J. C., 191 Reinhart, A. M., 218, 256 Reis, H. T., 97 Reiter, P. L., 244 Remme, L., 114 Rempel, J. K., 67 Rendon, T., 122 Renes, R. J., 219–220 Resnicow, K., 253–254 Reubsaet, A., 115 Reuter, T., 114, 115 Reynolds, G. L., 139 Rhine, R. J., 158 Rhoads, K., 108 Rhodes, N., 15, 172, 254, 264 Rhodes, R. E., 60, 75, 113, 116, 123 Rich, M., 241 Richard, R., 14, 118–119 Richardes, D., 220 Richardson, C. R., 223 659



Richert, J., 114–115, 139, 143 Richter, T., 217, 219 Richterkessing, J. L., 237 Ricketts, M., 218 Ridge, R. D., 39, 42, 46–47 Rieh, S. Y., 190 Riemsma, R. P., 139 Riesz, P. C., 213 Rietveld, T., 186 Rimal, R. N., 232 Ringwalt, C. L., 261, 267 Rise, J., 120, 130–131 Risen, J. L., 81, 95 Rittle, R. H., 235 Ritvo, P., 221 Rivers, J. A., 88 Rivis, A., 103, 112, 116–117, 127, 129, 131 Robertson, C. T., 192 Robertson, K., 155 Robins, D., 208 Robinson, J. K., 111 Robinson, N. G., 103 Rodewald, L., 221 Rodgers, H. L., Jr., 25 Rodgers, W. M., 123 Rodriguez, R., 151, 172 Roehrig, M., 88 Roels, T. H., 220 Roetzer, L. M., 241 Rogers, R. W., 69, 74, 166, 205–206, 228–229, 242, 246 Rogers, T., 114, 116 Rohde, P., 88 Rokeach, M., 50 Rolfe, T., 103 Rollnick, S., 256 Romero, A. A., 260 Ronis, D. L., 181 Rose, S. L., 192 Roseman, M., 115, 240 Rosen, B., 220 Rosen, C. S., 133 660



Rosen, J., 174 Rosen, S., 84 Rosenberg, M. J., 72, 74 Rosenbloom, D., 134 Rosenbloom, S. T., 221 Rosen-Brown, A., 240 Rosenzweig, E., 82 Roskos-Ewoldsen, D. R., 12, 198, 262–263 Rosnow, R. L., 258 Ross, K. M., 192 Rossi, J. S., 133–135 Rossi, S. R., 134 Rossiter, J. R., 232 Rothbart, M., 205–206 Rothman, A. J., 140, 226–227 Rothmund, T., 9 Rothstein, H. R., 96, 130, 182–183, 250, 264–265, 267 Rouner, D., 210, 219–220, 241 Royzman, E. B., 74, 225 Rozelle, R. M., 216 Rozin, P., 74, 225 Rozolis, J., 249 Ruberg, J. L., 241 Rubin, D. L., 186 Rubin, Y. S., 221 Ruch, R. S., 204 Rucker, D. D., 75, 174, 255 Ruder, M., 197 Ruiter, R. A. C., 69, 226, 229, 231–233, 244, 248, 254, 267 Ruiz, S., 75 Rusanen, M., 203 Russell, C., 108 Rutherford, J., 26 Rutter, D. R., 61, 72, 127 Ryerson, W. N., 219 Saba, A., 121 Sabogal, F., 116, 253 Sadatsafavi, M., 221 Sagarin, B. J., 81, 108 Sailors, J. J., 42 661



Saine, T. J., 197 Sakaki, H., 26 Salgueiro, M. F., 122, 261 Sallis, J. F., 261 Salovey, P., 174, 226–227, 254 Salter, N., 114 Sampson, E. E., 198 Sampson, J., 230 Samuelson, B. M., 265 Sanaktekin, O. H., 254 Sandberg, T., 118–119, 127 Sanders, D. L., 221 Sanders, J., 218 Sanders-Thompson, V., 253 Sandfort, T. G. M., 115 Sandman, P. M., 138, 142 Sarge, M. A., 223, 232 Sarma, K. M., 247 Sarnoff, I., 40 Sarup, G., 25 Saucier, D., 254 Sauer, P. L., 207 Saunders, L., 256 Savage, E., 110, 128 Sawyer, A. G., 240 Sayeed, S., 106, 116–117, 131 Scarberry, N. C., 240 Schaalma, H. P., 69, 116 Schepers, J., 112, 121 Scher, S. J., 97 Schertzer, S. M. B., 253 Scheufele, D. A., 70 Schlehofer, M. M., 233 Schmidt, J., 196 Schmitt, B., 54, 222 Schneider, I. K., 15, 174 Schneider, S. L., 244 Schneider, T. R., 110, 174, 226 Schneier, W. L., 152 Scholl, S. M., 262 Schönbach, P., 81 662



Schoormans, J. P. L., 218 Schott, C., 232 Schreibman, M., 220 Schrijnemakers, J. M. C., 186 Schroeder, D. A., 249 Schulenberg, J., 260–261 Schulman, R. S., 261 Schultz, A., 253–254 Schultz, P. W., 10, 122, 127 Schulz, P. J., 241 Schulz-Hardt, S., 81 Schumann, D. W., 151–153, 157, 159 Schünemann, H., 244 Schüz, B., 139, 143, 262 Schüz, N., 139, 143, 262 Schwartz, S. H., 50 Schwarz, N., 5, 9, 150, 152, 158, 255 Schwarzer, R., 114–115, 137, 139, 142–143 Schweitzer, D., 189 Schwenk, G., 112 Scileppi, J. A., 158, 211 Scott, M. D., 259 Sears, D. O., 84, 153, 260 See, Y. H. M., 172 Seeley, S., 220 Segall, A., 50 Segan, C. J., 133 Segar, M. L., 223 Seibel, C. A., 256 Seibring, M., 108 Seifert, A. L., 121 Seignourel, P. J., 158, 196 Seiter, J. S., 2, 17 Seo, K., 233, 255 Sereno, K. K., 200, 203 Sestir, M., 218, 219 Settle, J. E., 108 Severance, L. J., 158 Severson, H., 261 Sexton, J., 12 Shabalala, A., 219 663



Shaeffer, E. M., 240 Shaffer, D. R., 42 Shaikh, A. R., 254 Shakarchi, R. J., 255, 265 Shamblen, S. R., 267 Shani, Y., 81 Shanteau, J., 145, 218 Shantzis, C., 261 Shapiro, M. A., 60, 218 Shapiro-Luft, D., 60 Sharan, M., 220 Sharot, T., 81 Sharp, J., 62, 72 Shavitt, S., 35, 37–41, 43, 46, 50, 53–54, 70 Shaw, A. S., 249 Shaw, H., 88 Shea, S., 221 Sheeran, P., 68, 74, 95, 103, 109, 112–119, 123, 127, 129–131, 216, 229, 246, 262 Shelton, M., 108 Shen, F., 226, 241 Shen, L., 233, 237, 247, 255–256, 264 Shepherd, G. J., 11 Shepherd, J. E., 103 Shepherd, R., 10, 65, 109, 121, 190 Sheppard, B. H., 103, 112 Shepperd, J. A., 262 Sherif, C. W., 19–20, 22–25, 30, 33–34, 215 Sherif, M., 19–20, 22–25, 29–30, 33–34, 215 Sherman, D. K., 93, 181, 227, 262–263 Sherman, R. T., 240 Sherrell, D. L., 195, 211 Shi, F., 81, 95 Shi, Y., 227 Shillington, A., 108 Shiota, M. N., 255 Shiv, B., 181 Shongwe, T., 219 Shulman, H., 152, 255 Shuptrine, F. K., 42 Sia, C.-L., 208 664



Sicilia, M., 75 Siebler, F., 166 Siegel, J. T., 186 Siegrist, M., 190 Siemer, M., 186 Siero, F. W., 25, 111, 246 Sieverding, M., 103, 126 Sigurdsson, S. O., 221 Silberberg, A. R., 194 Silk, K. J., 116 Silva, S. A., 122, 261 Silvera, D. H., 253 Silvera, S. A. N., 174, 254 Silverthorne, C. P., 159 Silvia, P. J., 212, 256, 265 Simmonds, L. V., 262 Simon, L., 97 Simoni, J., 212 Simons, H. W., 198, 200–201 Simons-Morton, B. G., 110 Simpson, P., 220 Sims, L., 226 Simsekoglu, O., 118 Sinclair, R. C., 171, 208, 255 Singh, A., 116 Singhal, A., 219 Sirgy, M. J., 38 Sirsi, A. K., 192 Sivaraman, A., 257 Six, B., 10–11, 112 Sjoberg, L., 11 Skalski, P. D., 194, 212 Sklar, K. B., 181 Skowronski, J. J., 74, 81, 145 Slade, P., 114 Slama, M., 42, 50 Slater, M. D., 30, 163, 171, 186, 207, 210, 219–220, 241 Slaunwhite, J. M., 108 Slemmer, J. A., 240 Slocum, J. W., Jr., 118 Smerecnik, C. M. R., 229 665



Smit, E. G., 207 Smith, A., 233 Smith, D. C., 261 Smith, J. K., 95 Smith, J. R., 10, 97, 103, 116, 120 Smith, L., 109 Smith, L. M., 88 Smith, M. A., 159 Smith, M. B., 37, 53, 55 Smith, M. C., 174 Smith, M. J., 26 Smith, N., 120 Smith, R. A., 34, 219, 220 Smith, R. E., 11, 193 Smith, R. J., 95 Smith, S., 122–123, 194 Smith, S. M., 10, 84, 108, 152, 154 Smith, S. W., 29, 108, 122, 232 Smith-McLallen, A., 69, 165, 166 Smucker, W. D., 230 Snelders, D., 218 Sniehotta, F. F., 111, 143 Snyder, J., 221 Snyder, M., 10, 12–13, 37, 39, 42–44, 46–47, 49, 53, 205–206, 253 Soldat, A. S., 171, 208, 255 Soley, L. C., 160 Solomon, S., 181 Sorell, D. M., 256 Sorrentino, R. M., 181, 229 Southwell, B. G., 233 Sowden, A. J., 139 Sox, H. C., Jr., 244 Spangenberg, E. R., 95, 130 Sparks, P., 10, 65, 72, 103–104, 106, 109, 262 Spears, R., 257 Speelman, C., 116 Spencer, C. P., 106, 114, 127 Spencer, F., 241 Spencer, S. J., 36, 185 Spencer-Bowdage, S., 256 Sperati, F., 244 666



Spiegel, S., 166–168 Spoor, S., 88 Spreng, R. A., 151 Spring, B., 224, 243 Sprott, D. E., 95 Spruyt, A., 8 St. Lawrence, J. S., 107, 220 Staats, H., 120, 131 Stagner, B., 255 Stambush, M. A., 88 Stangor, C., 174 Stansbury, M., 208 Stanton, T., 221 Stapel, D. A., 267 Stapleton, J., 111 Stark, E., 38 Stayman, D., 67 Stayman, D. M., 207 Stead, M., 104, 117 Steadman, L., 61, 72, 127, 130 Steblay, N. M., 234, 250 Stebnitz, S., 196 Steele, C. M., 93, 262 Steele, L., 21 Steele, R. G., 256 Steenhaut, S., 118 Steffen, V. J., 11 Steiner, G. A., 142 Steinfatt, T. M., 65, 265 Stephenson, M. T., 187, 233, 254 Sternthal, B., 197, 211, 212 Stevenson, R., 233 Stevenson, Y., 107 Steward, W. T., 174, 226 Stewart, C., 108 Stewart, R., 31, 212 Stewart-Knox, B., 104 Stice, E., 88, 93 Stiff, J. B., 237 Stillwell, A. M., 93 Stoltenberg, C. D., 171 667



Stone, J., 13, 78, 90–91, 93 Stone, K., 220 Storey, D., 241 Stormer, S., 88 Strack, F., 95, 255 Stradling, S. G., 109, 118–119 Strathman, A. J., 148, 222 Straughan, R. D., 192 Strauss, A., 221 Strecher, V. J., 254 Strickland, B., 159 Stroebe, W., 72, 115, 142, 229–230, 232, 242, 247 Stroud, N. J., 84–85 Struckman-Johnson, C., 215 Struckman-Johnson, D., 215 Struttmann, T., 217 Strutton, D., 207 Studts, J. L., 241 Stukas, A. A., 39, 42, 47 Suchner, R. W., 25 Suci, G. J., 5 Sunar, D., 254 Sundar, S. S., 208 Sundie, J. M., 75 Sunnafrank, M., 201 Supphellen, M., 57, 121 Sussman, S., 261 Sutton, S., 104 Sutton, S. R., 74, 103, 109, 116, 120–121, 133–134, 139–140, 142– 143, 230 Swartz, T. A., 202–203 Swasy, J. L., 181 Sweat, M., 212 Swedroe, M., 108 Sweeney, A. M., 262 Swinyard, W. R., 11 Sykes, B., 111 Syme, G. J., 91, 106, 127 Symons, C. S., 25 Szabo, E. A., 258 Szilagyi, P., 221 668



Szybillo, G. J., 259 Tabb, K. M., 261 Tagg, S., 104, 117 Talbot, T. R., 221 Talibudeen, L., 104, 107 Tal-Or, N., 219, 257 Tam, S. F., 111 Tamborini, R., 194 Tan, A., 60 Tanjasiri, S. P., 220 Tannenbaum, P. H., 5, 76 Tao, C.-C., 184, 247 Taylor, L., 133 Taylor, N., 111 Taylor, N. J., 103, 106, 116–117, 127–128 Taylor, S. E., 227 Taylor, S. F., 123 Taylor, V. M., 106 Teel, J. E., 42 Teffera, N., 219 Teige-Mocigemba, S., 8 Telaak, K., 25 Telesca, C., 42, 44 Tellini, S., 40, 42 Terrenato, I., 244 Terry, D. J., 10, 97, 103, 120, 130–131 Terwel, B. W., 195, 211 Tetlock, P. E., 9 Teufel, J., 260 Tewksbury, D., 70 Thabane, L., 221 Thau, S., 196 Theodorakis, Y., 103, 109 Thibodeau, R., 91 Thimons, E. D., 232 Thomas, E., 233 Thomas, J. C., 240 Thomas, K., 123, 223 Thompson, B., 106 Thompson, D., 241 669



Thompson, E. P., 166–169, 173, 175 Thompson, J. K., 88 Thompson, R., 219, 261 Thompson, S. C., 233 Thomsen, C. J., 30, 171 Thuen, F., 130 Thurstone, L. L., 7 Thyagaraj, S., 260 Till, B. D., 213 Ting, S., 186 Tittler, B. I., 25 Tobler, N. S., 261 Todorov, A., 149 Tokunaga, R. S., 218–219, 241 Tollefson, M., 196 Tolsma, D., 254 Tom, S., Jr., 88, 89 Tormala, Z. L., 15, 18, 75, 161, 173–174, 191, 196–197, 208, 211– 212, 255 Törn, F., 213 Tost, L. P., 256 Towles-Schwen, T., 12 Trafimow, D., 68, 74, 103, 109, 123, 127, 130 Traylor, M. B., 200 Trembly, G., 216 Trompeta, J., 212 Trope, Y., 149 Trost, M. R., 173 Trudeau, L., 218 Trumbo, C. W., 152, 172 Tryburcy, M., 111 Tseng, D. S., 221 Tuah, N. A. A., 139 Tufte, E. R., xvii Tukachinsky, R., 218–219, 241 Tung, P. T., 97, 120 Tuppen, C. J. S., 189–190 Turner, G. E., 261 Turner, J. A., 26 Turner, M. M., 152, 233, 249, 255–256 Turrisi, R., 111 670



Tusing, K. J., 237, 266 Tversky, A., 244 Twyman, M., 190 Tykocinski, O. E., 9 Ubel, P. A., 217 Udall, A., 154 Uhlmann, E. L., 9 Ullen, H., 103 Underhill, J. C., 233 Updegraff, J. A., 181, 223, 226–227, 244–245 Uribe, R., 217, 241 Usdin, S., 219 Ussher, M., 103 Uzzell, D., 97, 103, 131 Vakratsas, D., 142 Valdez, R. B., 111 Valente, T. W., 219–220 Valentine, J. C., 182 Valentino, N. A., 85 Valois, P., 109 van ’t Riet, J., 226, 233, 244, 254, 267 van Assema, P., 139 van Baak, M. A., 221 Van Bavel, J. J., 9 van den Berg, P., 88 van den Hende, E. A., 218 van den Putte, B., 116, 121, 241, 244 van der Linde, L. A. J. G., 267 van der Pligt, J., 14–15, 57, 62, 70, 72, 118–119, 174 van Dijk, A., 182 van Dijk, L., 221 van Enschot-van Dijk, R., 165 van Griensven, G. J. P., 115 van Harreveld, F., 15, 62, 70, 72, 174 van Herpen, E., 122 van Hout, R., 186 van Ittersum, K., 72 Van Kenhove, P., 118 Van Koningsbruggen, G. M., 262–263 671



van Laer, T., 218–219, 241 van Leeuwen, L., 219 Van Loo, M. F., 130 van Mechelen, W., 116 van Meurs, L., 207 Van Osch, L., 115 Van Overwalle, F., 93 van Trijp, H. C. M., 72, 122 van Weert, J. C. M., 221 van Woerkum, C., 220 Vann, J., 221 Vardon, P., 103 Vassallo, M., 121 Vaughn, L. A., 218–220 Vaught, C., 187 Velicer, W. F., 133–134, 139 Venkatesh, V., 121 Venkatraman, M. P., 181 Verplanken, B., 151, 233 Vervloet, M., 221 Vet, R., 107 Viachopoulos, S. P., 103 Villagran, P. D., 157 Vincent, J. E., 235–236 Vincus, A. A., 267 Vinkers, C. D. W., 114 Visconti, L. M., 218–219, 241 Visser, P. S., 15, 157, 171, 174 Vist, G. E., 244 Vitoria, P. D., 122, 261 Voas, R. B., 108 Vohs, K. D., 158, 250 von Hippel, W., 149 Vonkeman, C., 171, 255 Voss-Humke, A. M., 121 Waalen, J., 221 Wachtler, J., 199 Wagner, A. K., 81 Wagner, W., 200–201 Wakefield, M., 218–219 672



Waks, L., 66 Walker, A., 139 Wall, A.-M., 123 Wallace, D. S., 10, 12 Walster, E., 82, 192, 204 Walters, L. H., 186 Walther, E., 95 Walther, J. B., 95, 208 Wan, C. S., 88 Wang, X., 48, 53, 121 Wang, Z., 208 Wansink, B., 72 Warburton, J., 120, 131 Ward, C. D., 210 Wareham, N. J., 74, 104, 121 Warnecke, R. B., 85 Warner, L., 220 Warren, W. L., 152 Warshaw, P. R., 103, 112, 121 Wasilevich, E., 232 Wathen, C. N., 211 Watson, A. J., 221 Watson, C., 197 Watt, I. S., 139 Watt, S. E., 9, 35 Wearing, A. J., 91 Webb, T. L., 114–115 Webel, A. R., 212 Weber, R., 231 Webster, R., 254 Wechsler, H., 108 Wegener, D. T., 15, 18, 26, 43–44, 150, 152, 157, 160–161, 164, 168, 171–175, 197, 208 Weigel, R. H., 11 Weil, R., 95 Weilbacher, W. M., 142 Weinberger, M. G., 211 Weiner, J., 116 Weiner, J. L., 223 Weinerth, T., 152 Weinstein, N. D., 64, 130, 138, 140, 142 673



Weisenberg, M., 199 Weissman, W., 261 Wells, G. L., 154–155, 181 Wells, J., 111 Wenzel, M., 108 Werrij, M. Q., 226, 244 Wessel, E., 194 West, M. D., 190 West, R., 140 West, S. K., 261, 267 Westermann, C. Y. K., 157, 163 Westfall, J., 186 Wetzels, M., 112, 121, 218–219, 241 Whately, R., 147 Wheeler, D., 235–236 Wheeler, S. C., 43–44, 54, 168–169, 175 Whitaker, D. J., 113 White, G. L., 80 White, K., 257 White, K. M., 97, 103 White, M., 139 White, R. W., 37, 53, 55 White, T. L., 227 Whitehead, J. L., Jr., 190–191, 210 Whitelaw, S., 140 Whittaker, J. O., 26 Whittier, D. K., 220 Wicherts, J. M., 182 Wicker, A. W., 17 Wicklund, R. A., 81 Widgery, R. N., 204 Wieber, F., 114 Wiedemann, A. U., 114, 143 Wiegand, A. W., 91 Wiener, J. L., 195 Wildey, M. B., 261 Wilke, H. A. M., 120, 131 Wilkin, H. A., 219–220 Williams, E. A., 221 Williams, P., 95, 253–254 Williamson, P. R., xviii, 96 674



Williams-Piehota, P., 174, 254 Willich, S. N., 221 Willms, D., 131 Wilmot, W. W., 22, 30, 31 Wilson, C. P., 111 Wilson, E. J., 195, 211 Wilson, M., 50 Wilson, S. R., 232 Wilson, T., 72 Windschitl, P. D., 181 Winslow, M. P., 13, 90–91 Winter, P. L., 108 Winterbottom, A., 220 Winzelberg, A., 203 Wise, M. E., 241 Witte, K., 139, 145, 187, 219–220, 229–233, 242, 246–249 Wittenbrink, B., 5, 9 Wogalter, M. S., 69 Wohn, D. Y., 95 Wojcieszak, M. E., 84–85 Wolff, J. A., 130 Wolfs, J., 103 Wong, N. C. H., 232 Wong, S., 240 Wong, Z. S.-Y., 109, 123 Wood, M. L. M., 266 Wood, M. M., 139 Wood, W., 17, 112, 115, 155, 157, 159–160, 172, 181, 190, 192–193, 254–255, 259–260, 262, 264–267 Woodruff, S. I., 261 Woodside, A. G., 95, 200–201 Wooley, S., 241 Worchel, S., 158, 240 Worth, L. T., 255 Wreggit, S. S., 110 Wright, A., 114 Wu, E. C., 249 Wu, R., 256 Wyer, N., 204 Wyer, R. S., Jr., 66, 76, 155, 217, 257 Wynn, S. R., 260–261 675



Xie, X., 66 Xu, A.J., 257 Yang, V. S.-H., 266 Yard, S., 212 Yarsevich, J., 254 Yates, S., 154, 159 Ybarra, O., 130 Yew, W. W., 111 Yi, M. Y., 208 Yi, Y., 89, 130 Yoo, J., 89 Yoon, J. J., 208 Young, A. M., 256 Young, R. M., 103 Young, T. J., 189, 259 Yu, X., 81 Yzer, M. C., 60, 104, 111, 116–117, 123, 131, 233, 246 Zani, B., 217 Zanna, M. P., 38–39, 46, 67, 72, 74, 89, 91, 115, 185, 219 Zebregs, S., 241 Zeelenberg, M., 81–82 Zehr, C. E., 115 Zenilman, J., 138 Zhang, G., 254 Zhang, X., 253 Zhao, X., 207, 262 Zhao, X. S., 185 Ziegelmann, J. P., 114, 139 Ziegler, R., 44, 152, 171, 208, 255 Ziel, F. H., 221 Zikmund-Fisher, B. J., 217, 223 Zimbardo, P. G., 199 Zimet, G. D., 244 Zindler, D., 196 Zinman, J., 221 Zirbel, C., 196 Ziv, S., 257 Zubric, S. J., 258, 266 Zucker, R. A., 260–261 676



Zuckerman, C., 187, 232 Zuckerman, M., 40, 42, 254 Zwarun, L., 218



677



Subject Index Accessibility of attitude, 10 Advertising, 193, 207, 224 Advocated position ambiguity of, 25, 27–28 counterattitudinal vs. proattitudinal, 156, 197, 215 discrepancy of, 25–26, 34(n5) expected vs. unexpected, 192–193 influence on credibility, 192–193 influence on elaboration valence, 156 Affect anticipated, 13, 118–20, 233 as attitude basis, 67–68, 74(n13), 75(n20) Age, 253 Ambiguity (of position advocated), 25, 27–28 Ambivalence, 10, 174(n16), 226 Anger, 233, 256 Anticipated feelings, 13, 118–20, 233 Appeals affective vs. cognitive, 75(n20) consequence-based, 165–166, 221–223 fear, 228–233 function-matched, 41–44, 49–51, 54(n6) gain-framed and loss-framed, 225–228 normative, 106–108 one-sided and two-sided, 79, 193, 211(n4), 223–225, 265(n10, n12), 266(n16) scarcity, 75(n18) threat, 228–233 Approach/avoidance motivation (BAS/BIS), 54(n5), 227 Arguments consequence-based, 165, 166, 174(n19), 221–223 discussion of opposing, 79, 193, 211(n4), 223–225, 265(n10, n12), 266(n16) gain-framed and loss-framed, 225–228 number of, 159 strength (quality) of, 156–157, 163–166 Assimilation and contrast effects, 24–25, 27, 33(n3), 34(n4, nn6–7), 678



215 Attitude accessibility, 10 ambivalence, 10, 174(n16), 226 bases of, 56–59, 66–68 certainty (confidence), 174(n16) concept of, 4–5 functions of, 35–38 measurement, 5–9 relation to behavior, 9–12 strength, 15, 18(n12), 171(n3) toward behavior. See Attitude toward the behavior Attitude toward the behavior assessment of, 99 determinants of, 105 influencing, 105 relation to norms, 112 Attitude-behavior consistency factors affecting, 9–12 influencing, 12–14 Attitude-toward-the-ad, 67 Attitudinal similarity, 201 Attractiveness, 160, 204–206 Attribute importance, 61–62, 72(n3), 78 Audience. See Receiver Audience adaptation, xv–xvi elaboration likelihood model (ELM) and, 162, 174(n17) functional attitude approaches and, 41–44, 49–51 individual differences and, 252–255 reasoned action theory (RAT) and, 117 social judgment theory and, 29 stage models and, 134–141 summative model of attitude and, 59–61 Audience reaction (as peripheral cue), 159 Averaging model of attitude, 65–66 Aversive consequences and induced compliance, 97(n16) Balance theory, 76, 95(n1) Behavior relation to attitude, 9–12 relation to intention, 112–116, 129(n21), 130(n22) 679



Belief content, 62–63 evaluation, 57, 69, 75(n18), 104. See also Consequence desirability importance, 61–62, 72(n3), 78 lists, 63–64 salience, 56, 61, 69–71 strength (likelihood), 57, 63–64, 68–69, 104, 127(n8). See also Consequence likelihood Belief-based models of attitude description of, 56–59 implications for persuasion, 59–61, 68–71 sufficiency of, 66–68 Bias, knowledge and reporting, 190, 192 Bipolar scale scoring, 64, 73(nn8–9) Central route to persuasion, 150 Character identification, 218 Choice effects in induced compliance, 89, 96(n12), 200 Citation of evidence sources, 191 Cognitive bases of attitude, 56–59 Cognitive dissonance and decision making, 78–83 defined, 77 factors influencing, 77–78 and hypocrisy induction, 90–92 and induced compliance, 85–90, 96(nn12–13), 97(n16), 199 means of reducing, 78 and selective exposure, 83–85 Communicator credibility, 188–198 ethnicity, 206–207 liking, 198–200 physical attractiveness, 160, 204–206 self-interest, 192 similarity to receiver, 200–204 Competence (credibility dimension), 189 Conclusion omission, 214–216 Congruity theory, 76, 95(n1) Consensus heuristic, 159 Consequence desirability, 165, 174(n19), 221–223. See also Belief, 680



evaluation Consequence likelihood, 166. See also Belief, strength (likelihood) Consideration of future consequences (CFC), 54(n5), 222, 264(n1) Contrast and assimilation effects, 24–25, 27, 33(n3), 34(n4, nn6–7), 215 Correspondence of attitude and behavior measures, 10–11, 17 of intention and behavior measures, 113 Counterarguing, 256, 257, 259–260. See also Elaboration Counterattitudinal vs. proattitudinal messages, 156, 197, 215 Credibility dimensions of, 188–190 effects of, 194–198 factors affecting, 190–194 heuristic, 158 relationship to similarity, 202–203 Cultural background, 253 Cultural truisms, 266(nn17–18) DARE, 261 Decisional balance, 134–136 Decision-making and dissonance, 78–83 Defensive avoidance, 261 Delivery dialect, 202 nonfluencies, 191 Descriptive norm (DN), 29, 261 assessment of, 100 determinants of, 107–108 influencing, 108 relation to attitude, 112 Dialect, 202 Direct experience, 11–12 Discrepancy, 25–26, 34(n5) Disgust, 233 Dissonance and decision making, 78–83 defined, 77 factors influencing, 77–78 and hypocrisy induction, 90–92 and induced compliance, 85–90, 96(nn12–13), 97(n16), 199 681



means of reducing, 78 and selective exposure, 83–85 Distraction, 154–155 Door-in-the-face (DITF) strategy, 235–237 explanations, 236–237 moderating factors, 236 Dual-process models, 149 Ego-involvement concept of, 22, 30–31, 153 confounding in research, 29–30 measures of, 23–24, 31, 34 relationship to judgmental latitudes, 22–23 Elaboration ability, 154–155 assessment of, 149 continuum, 149, 150 definition of, 149 factors affecting amount of, 152–156, 197, 255 factors affecting valence of, 156–157, 197 motivation, 152–154 Elaboration likelihood model (ELM), 148–175 Emotions anger, 233, 256 anticipated, 13, 118–20, 233 fear, 229–232 disgust, 233 guilt, 93, 97(n16), 233, 237 regret, 82–83, 118 Entertainment-education (EE), 219–220 Ethnicity of communicator, 206–207 Even-a-penny-helps strategy, 249(n50) Evidence, citation of sources of, 191 Examples vs. statistics, 241(n9) Expectancy confirmation and disconfirmation, 192–193, 211(n3) Expectancy-value models of attitude, 72(n1) Experimental design, 176–178 Expertise (credibility dimension), 189 Explicit measures of attitude, 5–7 Explicit planning of behavior, 114–115 Explicit vs. implicit conclusions, 214–216 682



Extended parallel process model (EPPM), 231–232 Fairness norms, 84 Familiarity of topic, 10, 155 Fear, 229–232 Fear appeals, 228–233 Follow-up persuasive efforts, 82 Foot-in-the-door (FITD) strategy, 233–235 explanation of, 234–235 moderating factors, 234 Forewarning, 157, 259–260 Formative basis of attitude, 11–12 Framing gain vs. loss, 225–228 issue, 70 Function-matched appeals, 41–44, 49–51, 54(n6) Functions of attitude assessing, 38–40, 46–48 vs. functions of objects, 45–46 influences on, 40–41 matching appeals to, 41–44, 49–51, 54(n6) typologies of, 35–38, 44–45, 48–49 Gain-framed vs. loss-framed appeals, 225–228 Games, 241(n14) General vs. specific recommendations, 216 Group membership, 203 Guilt, 93, 97(n16), 233, 237 Habit, 115–116 Health Action Process Approach (HAPA), 142 Heuristic principles, 157–159 Heuristic-systematic model, 149 Hierarchy-of-effects models, 142 Humor, 194 Hypocrisy, 13, 18(nn9–10) Ideal credibility ratings, 212(n10) Identification of source, timing of, 196 Identification with narrative characters, 218 683



Imagined behavior, 240(n6) Immersion (in narratives), 218 Implementation intentions, 114, 216 Implicit measures of attitude, 8–9 Implicit vs. explicit conclusions, 214–216 Importance of beliefs, 61–62, 72(n3), 78 of topic, 152–153, 163, 195, 199 Incentive effects in induced compliance, 85–87 Indirect experience, 11–12 Individual differences, 252–255 approach/avoidance motivation (BAS/BIS), 54(n5), 227 consideration of future consequences (CFC), 54(n5), 222, 264(n1) intelligence, 215, 253 need for cognition (NFC), 153–154, 174(n17), 226 regulatory focus, 54(n5), 227 self-esteem, 253 self-monitoring, 39–40, 42, 53(n2, n4), 54–55(n10), 222, 253 sensation-seeking, 254 Individualism-collectivism, 54, 222 Individualized belief lists, 63 Induced compliance, 85–90, 96(nn12–13), 97(n16), 199 choice effects in, 89, 96(n12), 200 incentive effects in, 85–87 Information exposure, influences on, 83–85 Information integration theory, 73(n10) Information utility, 84, 96(n11) Injunctive norm (IN) assessment of, 100 determinants of, 105–106 influencing, 106–107 relation to attitude, 112 Inoculation, 257–259 Intelligence, 215, 253 Intention, relation to behavior, 112–116, 129(n21), 130(n22) Involvement ego-involvement. See Ego-involvement personal relevance, 152–153, 163, 195, 199 Knowledge about topic, 10, 155 684



Knowledge bias, 190, 192 Latitudes, judgmental, 22 Legitimizing paltry contributions, 249(n50) Length of message, 159, 160 Likert attitude scales, 7 Liking for the communicator, 193, 198–200 heuristic, 158 relationship to similarity, 201–202 Loss-framed vs. gain-framed appeals, 225–228 Low price offer, 88 Low-ball strategy, 249(n50) Message characteristics, 214–251 Message tailoring. See Audience adaptation Message variable definition, 163–165, 183–184, 187(nn9–10), 247(n39) Meta-analysis, xviii, 186(n6), 187(n7) Metacognitive states, 18(n11), 174(n16) Modal belief lists, 57, 63 Modeling, 111 Mood, 157, 171(n4), 217, 226, 255 Moral norms, 120 Motivation to comply, 105 Motivational interviewing, 256 Multiple roles for variables, 160–162, 173(n16), 208–210, 217, 254 Multiple-act behavioral measures, 10 Multiple-message designs, 181 Narratives, 216–220 Need for cognition (NFC), 153–154, 174(n17), 226 Negativity bias, 74(n11), 225 Noncognitive bases of attitude, 66–68 Nonfluencies in delivery, 191 Nonrefutational two-sided messages, 223, 245(n32) Normative beliefs, 105 Norms descriptive. See Descriptive norm fairness, 84 685



injunctive. See Injunctive norm. moral, 120 Nudges, 220–221 Null hypothesis significance testing, xviii Number of arguments, 159 One-sided vs. two-sided messages, 79, 193, 211(n4), 223–225, 265(n10, n12), 266(n16) Optimal scaling, 73(n8) Ordered Alternatives questionnaire, 20–21 Own Categories procedure, 23, 33(n2) Paradigm case, 2 Peer-based interventions, 204 Perceived behavioral control, 216, 221 assessment of, 101 conceptualization of, 100, 123–124, 131(n33) determinants of, 108–110 influencing, 110–111 relation to attitudes and norms, 102, 122–123, 129(n19) relation to stages of change, 136–139 Peripheral cues, 150 Peripheral route to persuasion, 150 Persistence of persuasion, 151 Personal (moral) norms, 120 Personal relevance of topic, 152–153, 163, 195, 199 Personality characteristics, 252–255 approach/avoidance motivation (BAS/BIS), 54(n5), 227 consideration of future consequences (CFC), 54(n5), 222, 264(n1) intelligence, 215, 253 need for cognition (NFC), 153–154, 174(n17), 226 regulatory focus, 54(n5), 227 self-esteem, 253 self-monitoring, 39–40, 42, 53(n2,n4), 54–55(n10), 222, 253 sensation-seeking, 254 Persuasion, concept of, 2–4 Persuasive effects, assessing, 14–16, 18(n13), 185(n2) Physical attractiveness, 160, 204–206 Planned behavior, theory of, 126(n1) 686



Planning of behavior, 114–115 Position advocated. See Advocated position Postdecisional spreading of alternatives, 80 Prior knowledge (of topic), 10, 155 Proattitudinal vs. counterattitudinal messages, 156, 197, 215 Product trial, 11 Prompts, 220–221 Prospect theory, 244(n28) Protection motivation theory (PMT), 228–229 Quality (strength) of arguments, 156–157, 163–166 Quantity of arguments, 159 Quasi-explicit measures of attitude, 7–8 Reactance, 255–257 Reasoned action theory (RAT) determinants of intention, 99–102 influencing attitude toward the behavior, 104–105 influencing descriptive norms, 107–108 influencing injunctive norms, 105–107 influencing perceived behavioral control, 108–111 influencing relative weights, 111–112, 128(nn17–18) Receiver factors, 252–267 Reciprocal concessions, 236 Recommendation specificity, 216 Refusal skills training, 260–261 Refutational inoculation treatments, 258 Refutational two-sided messages, 223, 265(n10), 266(n16) Regret anticipated, 118 postdecisional, 82–83 Regulatory focus, 54(n5), 227 Relevance of attitude to behavior, 10, 12–13 of topic to receiver, 152–153, 163, 195, 199 Reminders, 220–221 Reporting bias, 190, 192 Resistance to persuasion from different persuasion routes, 151 Role models, 111 Routes to persuasion, 150–152 687



Salience of beliefs, 56, 61, 69–71 Scale scoring procedures, 64–65, 127(n10), 128(n11) Scarcity appeals, 75(n18) Selective exposure, 81, 83–85 Self-affirmation, 261–264 Self-affirmation theory, 93 Self-efficacy, 216, 221, 226. and stage-matching, 136–139 See also Perceived behavioral control Self-esteem, 253 Self-identity, 131(n30) Self-monitoring, 39–40, 42, 53(n2, n4), 54–55(n10), 222, 253 Self-perception, 234 Self-prophesy effects, 95(n3) Semantic differential evaluative scales, 5 Sensation-seeking, 254 Sequential-request strategies, 233–237 Sidedness (of messages), 79, 193, 211(n4), 223–225, 265(n10, n12), 266(n16) Similarity of communicator and receiver relationship to credibility, 202–203 relationship to liking, 201–202 Single-act behavioral measures, 10 Single-item attitude measures, 6 Single-message designs, 178–181 Social judgment theory, 19–34 Source credibility, 188–198 ethnicity, 206–207 liking, 198–200 physical attractiveness, 160, 204–206 self-interest, 192 similarity to receiver, 200–204 Specific vs. general recommendations, 216 Stage models, 132–147 distinctive claims of, 140–142 stage assessment, 139 transtheoretical model (TTM), 132–140 Stage-matching and decisional balance, 134–136 and self-efficacy, 136–139 688



vs. state-matching, 141 Standardized belief lists, 63 Statistical significance testing, xviii Statistics vs. examples, 241(n9) Stories, 216–220 Strength argument, 156–157, 163–166 attitude, 15, 18(n12), 171(n3) belief, 57, 63–64, 68–69, 104, 127(n8). See also Consequence likelihood Subjective norm. See Injunctive norm Summative model of attitude, 56–59 Supportive treatments (for creating resistance), 258 Tailoring (of messages). See Audience adaptation That’s-not-all strategy, 249(n50) Theory of planned behavior, 126(n1) Threat appeals, 228–233 defining variations of, 247(n39) effects of, 230–231 explanations of, 231–233 Thurstone attitude scales, 7 Time decay of persuasive effects, 151 follow-up persuasive efforts, 82 interval between DITF requests, 236 interval between FITD requests, 234 interval between intention and behavior measures, 130(n22) interval between warning and message, 260 stability of intentions, 113–114 timing of communicator identification, 196 Topic knowledge of, 10, 155 personal relevance of, 152–153, 163, 195, 199 Transgression-compliance effects, 237 Transportation (into narratives), 218 Transtheoretical model, 132–140 decisional balance, 134–136 processes of change, 133, 144(n2) self-efficacy, 136–139 stages of change, 132–133 689



Trustworthiness (credibility dimension), 189 Two-sided vs. one-sided messages, 79, 193, 211(n4), 223–225, 265(n10, n12), 266(n16) Unexpected position, 192–193 Unimodel of persuasion, 166–169 Unipolar scale scoring, 64, 73(n8) Utility of information, 84, 96(n11) Variable definition, 163–165, 183–184, 187(nn9–10), 247(n39) Video games, 241(n14) Warning, 157, 259–260 Websites as sources, 207 Weights of reasoned action theory components, 101–102, 117, 127(n6), 130(n26), 131(n28)



690



About the Author Daniel J. O’Keefe is the Owen L. Coon Professor in the Department of Communication Studies at Northwestern University. He received his Ph.D. from the University of Illinois at Urbana-Champaign and has been a faculty member at the University of Michigan, Pennsylvania State University, and the University of Illinois. He has received the National Communication Association’s Charles Woolbert Research Award, its Golden Anniversary Monograph Award, its Rhetorical and Communication Theory Division Distinguished Scholar Award, and its Health Communication Division Article of the Year Award; the International Communication Association’s Best Article Award and its Division 1 John E. Hunter Meta-Analysis Award; the International Society for the Study of Argumentation’s Distinguished Research Award; the American Forensic Association’s Daniel Rohrer Memorial Research Award; and teaching awards from Northwestern University, the University of Illinois, and the Central States Communication Association.



691



692



目录 Half Title Publisher Note Title Page Copyright Page Brief Contents Detailed Contents Preface Chapter 1 Persuasion, Attitudes, and Actions Chapter 2 Social Judgment Theory Chapter 3 Functional Approaches to Attitude Chapter 4 Belief-Based Models of Attitude Chapter 5 Cognitive Dissonance Theory Chapter 6 Reasoned Action Theory Chapter 7 Stage Models Chapter 8 Elaboration Likelihood Model Chapter 9 The Study of Persuasive Effects Chapter 10 Communicator Factors Chapter 11 Message Factors Chapter 12 Receiver Factors References Author Index Index About the Author Advertisement



693



2 3 4 5 7 8 17 23 48 71 102 131 163 212 234 275 291 327 385 409 609 678 691 692