Sensation Seeking Scale [PDF]

Citation preview

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/241843512



The need inventory of sensation seeking (NISS) Article in European Journal of Psychological Assessment · January 2011 DOI: 10.1027/1015-5759/a000085



CITATIONS



READS



11



766



2 authors, including: Philipp Hammelstein Psychiatrische Dienste Aargau AG 28 PUBLICATIONS 330 CITATIONS SEE PROFILE



All content following this page was uploaded by Philipp Hammelstein on 27 May 2014. The user has requested enhancement of the downloaded file.



M. Roth & P. European Hammelstein: JournalThe of PsychologicalA Need Inventory ssessment of Sensation © 2011 2012; Hogrefe Vol. Seeking 28(1):11–18 Publishing (NISS)



Original Article



The Need Inventory of Sensation Seeking (NISS) Marcus Roth1 and Philipp Hammelstein2 1



University of Duisburg-Essen, Essen, Germany, 2University of Düsseldorf, Germany



Abstract. Based on the conception of sensation seeking as a need rather than a temperamental trait (Hammelstein, 2004), we present a new assessment method, the Need Inventory of Sensation Seeking (NISS), which is considered to assess a motivational disposition. Three studies are presented: The first examined the factorial structure and the reliability of the German versions of the NISS; the second study compared the German and the English versions of the NISS; and finally, the validity of the NISS was examined in a nonclinical study and compared to the validity of conventional methods of assessing sensation seeking (Sensation Seeking Scale – Form V; SSS-V). Compared to the SSS-V, the NISS shows better reliability and validity in addition to providing new research possibilities including application in experimental areas. Keywords: sensation seeking, assessment, need theory



Introduction The concept of sensation seeking can be traced back to the works of Zuckerman (1971), yet its definition has changed over time. Zuckerman (1979) initially described sensation seeking as a “trait defined by the need for varied, novel and complex sensations and experiences and the willingness to take physical and social risks for the sake of such experience” (p. 10). Sensation seeking was conceptualized as an age- and gender-dependent trait (for details, see Roth, Hammelstein, & Brähler, 2007; Roth, Schumacher, & Brähler, 2005). In later years, however, Zuckerman rejected his own definition of sensation seeking as a need and instead referred to it as a trait involving the “seeking of varied, novel, complex and intense sensations and experiences” (Zuckerman, 1994, p. 27). This modification was made on account of the implication of “compulsion” (p. 26) conveyed by the term “need.” It is, however, almost impossible to avoid conceiving of sensation seeking as a need; an exclusively behavioral definition without a motivational component would result in a completely descriptive conceptualization without an explicative function. Even Zuckerman himself sometimes continues to speak of a “need” for sensations (1994, p. 374). Apart from this breach of logical explication (definition of “sensation seeking” as “seeking”), problems arise when defining and operationalizing sensation seeking based on highly specific behavior via the Sensation Seeking Scale – Form V (SSSV; Zuckerman, 1994). There has been much critique of both the conceptualization and the assessment of sensation seeking using the SSSV (e.g., Arnett, 1994; Hammelstein, 2004; Jackson & Ma© 2011 Hogrefe Publishing



raun, 1996; Roth, 2003; Roth et al., 2007). Major points of criticism include the lack of a theoretical foundation for the SSS-V items; the lack of a conceptual differentiation between sensation seeking and impulsivity; the use of items which are dependent on age, culture, and socioeconomic status (e.g., “skiing,” “mountaineering”); and the inclusion of items that themselves describe dysfunctional behavior (e.g., the use of marijuana) and that therefore potentially lead to a confounding of predictor (sensation seeking) and criterion (substance abuse). The present studies do not aim to repeat these critical issues. An alternative concept of sensation seeking was presented by Hammelstein and Roth (Hammelstein, 2004; Roth et al., 2007; Roth, 2009), and a theoretical discussion between Zuckerman and the exponents of an alternative model has been published elsewhere (Roth, Hammelstein, & Brähler, 2009). Rather, we intend only to outline our theoretical model inasmuch as necessary for an understanding of the development of an alternative assessment method. With respect to the theoretical foundations and a corresponding discussion of this alternative method, we refer to the publications cited above. Here, sensation seeking is defined as a need for novelty and intensity which develops its motivational power via the registration of discrepancies between the as-is and the to-be state. This definition focuses on the aim (or function) of a specific behavior (e.g., feeling of excitement in the body) rather than the behavior or behavior pattern itself (e.g., mountain climbing). In our view, it is thus essential that a given behavior pursue experiences of novelty or intensity before it can be spoken of as sensation seeking. The concrete form this behavior takes (e.g., partying, sexual behavEuropean Journal of Psychological Assessment 2012; Vol. 28(1):11–18 DOI: 10.1027/1015-5759/a000085



12



M. Roth & P. Hammelstein: The Need Inventory of Sensation Seeking (NISS)



ior, gambling, etc.; items of the SSS-V) is not essential to defining sensation seeking. While, in our view, specific behaviors such as mountain climbing can represent a search for experiences of intensity, they might also represent aims that have nothing to do with the construct of sensation seeking. An instrument is therefore needed to assess the underlying motivational aims and not the concrete behavior, as is the case with the SSS-V and even the Arnett Inventory of Sensation Seeking (AISS; Arnett, 1994). Temperamental traits are defined by a specific behavior pattern and can therefore be best measured based on concrete behavior (e.g., introversion by reserved, less outgoing, and less sociable behavior). In contrast, needs are defined by the aims that drive behavior (stimulation, sex, control, attachment, etc.) and therefore cannot be measured via the concrete behavior itself. Grosse Holtforth and Grawe (2000) developed a German questionnaire for clinical purposes to measure different needs. In this questionnaire, participants are asked how important it is for them to experience specific target states (e.g., having control, being complimented on something, etc.). The basis of this questionnaire is constituted by 94 items from the statements of people in treatment. Everyday goals, such as avoiding criticism and seeking the admiration by others, are grouped into avoidance and approach motives. This questionnaire aims to assess interpersonal motives that are relevant for treatment planning. Similar to this questionnaire, the Need Inventory of Sensation Seeking (NISS) was developed as an assessment method to ask respondents to state how pleasant a specific stimulation situation (target state) is to them. The higher number of situations, or the more pleasant the instance of stimulation is rated, the greater the need for stimulation. The following studies aim to test the reliability and validity of the NISS. Study 1 examines the factorial structure and the reliability of the German version. Study 2 presents the English version of the NISS and examines its comparability with the German version. Study 3 tests the validity of the German version and examines the possibility of using the NISS to predict the willingness of individuals to change their job.



1994) and Arnett (1994). Items should reflect a physical or psychological sensation as a goal state rather than a concrete behavior. This is the distinction between the goal (i.e., sensation) and the means (i.e., the concrete behavior or the concrete kind of stimulation to reach this sensational goal). In the first step the authors generated 27 items that were then sent to four colleagues who had to rate whether the items reflected a concrete behavior/stimulation or a state of experience. At least the behavior had to be described very broadly so that different kinds of behavior could lead to this state (e.g., “I enjoy doing exciting and stimulating activities with others”). Only if all four subjects rated the items as sensation oriented the item had not been cancelled (e.g., the item “Finding myself in unfamiliar situations is exciting” was rated as reflecting concrete behavior and was omitted, whereas the item “I like feeling totally charged” was rated as reflecting a state of sensation). A total of 10 items were omitted during this step, so that 17 items remain (see Appendix). Respondents are required to rate how often they have felt this way in the past 6 months on a 5-point Likert scale (1 = almost never, 5 = almost always). Thus, the first study examined the factorial structure as well as the reliability of the NISS comprising the items generated as described above.



Methods Participants An e-mail was sent to students and employees of a German university. It contained an internet link to the questionnaire, which in turn was linked to a MySQL database. Since it was not possible to record e-mail addresses or IP addresses, anonymity was guaranteed. A total of 555 respondents completed the questionnaire. Participants’ mean age was M = 29.40 years (SD = 9.42) with a range of 17 to 68 years. 313 participants were women (54.6%). The sample was almost completely characterized by a high level of education, with 519 (93.5%) having attended school for 13 years (Abitur).



Materials



Study 1: Factorial Structure, Item Analysis, and Preliminary Evidence of Validity In the first step, items were generated that reflect pleasure in the target state of stimulation or in the opposite state (rest, calmness). Example items include “I like the feeling of excitement in my body” or “I enjoy it when there’s nothing going on for a while.” The background of the generation of items was shaped by the descriptions of “high sensation seekers” in the publication of Zuckerman (1979, European Journal of Psychological Assessment 2012; Vol. 28(1):11–18



Participants were administered the German version of the NISS comprising the 17 items (see Appendix).



Results Exploratory Analysis of the Dimensional Structure of the NISS The factorial structure of the NISS was examined using a principal component analysis (PCA) with oblimin rotation. Since the Kaiser criterion (eigenvalues greater than © 2011 Hogrefe Publishing



M. Roth & P. Hammelstein: The Need Inventory of Sensation Seeking (NISS)



13



Table 1. Results of PCA with oblimin-rotation and item analysis concerning the NISS of the German Sample A (n = 555) Descriptives



PCA F1



Item analysis



Item no.



M



SD



F2



h²



ris



rit



05



2.67



0.98



.77



–.16



.59



.69



.58



04



2.63



1.06



.75



–.31



.59



.67



.66



13



2.66



0.93



.74



–.03



.57



.65



.49



16



2.92



0.89



.73



–.20



.53



.65



.57



07



2.30



1.00



.71



–.10



.50



.62



.51



14



2.62



1.02



.67



–.16



.44



.58



.52



06



2.90



0.98



.66



–.17



.44



.57



.52



03



2.76



1.06



.65



–.06



.43



.56



.46



01



2.70



0.97



.60



–.11



.37



.52



.47



15



2.64



0.92



.56



–.22



.32



.48



.48



02



2.81



1.12



.54



–.14



.29



.46



.45



11



3.13



1.01



–.14



.82



.68



.70



.63



09



3.47



0.98



–.31



.80



.66



.67



.70



08



3.80



0.90



–.26



.75



.57



.62



.64



10



3.08



0.99



–.04



.74



.56



.60



.52



17



3.09



0.95



–.21



.72



.53



.58



.60



12



2.74



1.02



–.07



.52



.27



.38



.43



λ 5.2 3.5 Note. λ = Eigenvalue; ris = item to scale correlation; rit = item to total correlation. Highest loadings are presented in bold.



1) frequently leads to an overestimation of the number of factors to be retained, a parallel analysis (PA) was performed in the sense of Horn (1965). Using PA, the PCA results in two factors which explain 49.0% of the total variance (for two factors: λempirical = 2.8332 > λrandom = 1.2627; for three factors: λempirical = 1.2071 < λrandom = 1.2072). Loadings and communalities (h²) are presented in Table 1. The first factor contains items describing stimulation as aims of approaching according to Grawe (2002); this factor is referred to as “need for stimulation” (NS). The second factor consists of all negatively poled items and is therefore called “avoidance of rest (or tranquility)” (AR). The factor structure also remains stable when applying common factor analysis instead of PCA and varimax instead of oblimin rotation.



Discussion Reliability of the German NISS-version is considerably higher than that found for other measures of sensation seeking (SSS-V: e.g., Ball, Farnhill, & Wangeman, 1984; Ferrando & Chico, 2001; AISS: e.g., Ferrando & Chico, 2001; Roth, 2003). The two-dimensional structure is in line with Grawe’s (2002) theoretical considerations concerning fundamental needs. In his view, schemes of approach and avoidance emerge for every fundamental need. However, it should be noted that all items of the second factor (referred to as “avoidance of rest and tranquility”) are negatively poled. It is therefore questionable whether this factor actually reflects the aim of avoidance or rather a different aim of approach (see General Discussion).



Item Analysis



Study 2: Confirmatory Factor Analysis and Validity



Table 1 shows the Item-scale correlations (ris) and item-total correlations (rit). Internal consistency is high with α = .88 for the subscale NS and α = .82 for AR, and α = .86 for the total scale. The two scales have a correlation of r = .25. The correlation between the NS and the total score is r = .78, and the correlation between the AR and the total score is r = .80.



To ensure that the structure of the German NISS, as revealed by factor analyses in the German sample, is equivalent to that of the English version of the instrument, we crossvalidated the factor model (Study 1) with the data of an English sample. We furthermore examined the relationship between the NISS and the conventional method of measuring sensation seeking (SSS-V). Correlations be-



© 2011 Hogrefe Publishing



European Journal of Psychological Assessment 2012; Vol. 28(1):11–18



14



M. Roth & P. Hammelstein: The Need Inventory of Sensation Seeking (NISS)



Figure 1. Final path diagram (standardized coefficients) of the NISS by sample C (n = 146). NS = Need for stimulation, AR = Avoidance of rest.



tween the two scales were expected to be midrange; higher correlations were not expected because of the different aspects (aims of behavior vs. concrete behavior patterns) assessed by the instruments.



Methods Participants Participants were recruited via different internet sport forums. This online method of recruitment would seem appropriate for obtaining a homogeneous, nonclinical, English-speaking sample (methodical limitations associated with this recruitment strategy are referred to in the discussion). Internet users were informed that they had to be English native speakers to take part in the study. The sample consisted of 146 participants (51 female, 34.9%) aged from 14 to 63 years (M = 33.1, SD = 11.7). 103 subjects came from the United States (70.5%), 12 from Canada (8.2%), 8 from the United Kingdom (5.5%), 8 from New Zealand (5.5%), and 15 from other countries (10.3%). The average level of education was rather high, ranging from 4 to 20 years of education (M = 16.0, SD = 3.8).



Materials Participants were administered the English version of the NISS as well as the original version of the SSS-V (Zuckerman, 1994). The SSS-V is a 40-item, forced-choice questionnaire that measures four factors: thrill and adventure seeking (TAS), experience seeking (ES), disinhibition (Dis), and boredom susceptibility (BS). European Journal of Psychological Assessment 2012; Vol. 28(1):11–18



Results Data were analyzed using EQS 5.7 (Bentler, 1995). Since χ² values are highly susceptible to variations in sample size (Bentler & Bonett, 1980; Browne & Cudeck, 1993), we used the comparative fit index (CFI), which is widely independent of sample size. Furthermore, the root mean square error of approximation (RMSEA; Jöreskog, 1993) and the standardized root mean square residual (SRMR) were provided as indices. As recommended by Bentler (1995), we interpreted a CFI value of 0.90 or above as showing adequate fit and a value of 0.95 as showing excellent fit. In line with Browne and Cudeck (1993), we considered values of RMSEA below 0.08 to reflect adequate fit and values below 0.05 good fit. Values of the SRMR are expected to stay below 0.10. Parameters were estimated using the robust maximum-likelihood (ML) estimation method. The tested NISS model was specified according to the results of Study 1, which revealed two dimensions of sensation seeking (see in detail Figure 1). The fit indices associated with this model indicate a suboptimal fit when applied to the English sample (χ² = 240.26, p < .001, df = 118; CFI = 0.89, RMSEA = 0.08; SRMR = 0.09). However, when allowing correlations between the unique variance of items 4 and 5 (see Figure 1), as was indicated by the modification indices (MIs) and standardized expected parameter changes (SEPCs), the fit indices indicate an reasonable fit to the data of the English sample: While χ² is significant (χ² = 220.61, df = 117, p < .001), the other fit indices now indicate reasonable model fit (CFI = 0.92, RMSEA = 0.07, SRMR = 0.09). Based on content analysis, it would seem appropriate to allow the unique variances of the both items to correlate: the common variance of items 4 and 5 may be interpreted as constituting a factor described as “stimula© 2011 Hogrefe Publishing



M. Roth & P. Hammelstein: The Need Inventory of Sensation Seeking (NISS)



Table 2. Correlations of the subscales of the SSS-V and the NISS (n = 146) TAS



Dis



BS



NS



.54**



.40**



.34**



AR



.08



.03



.06



ES .15 –.19



tor assesses something other than the need for sensation seeking.



SSS-V total .57** –.02



NISS total .41** .28* .27* –.02 .37** Notes. TAS = Thrill and Adventure Seeking; Dis = Disinhibition; BS = Boredom Susceptibility; ES = Experience Seeking; SSS-V total = total score of the SSS-V; NS = Need for stimulation; AR = Avoidance of rest; NISS total = total score of the NISS. *p < .05, **p < .01 (after Bonferroni adjustment).



tion of the body” that goes beyond the congeneric factor of a general need for stimulation. Furthermore, measurement equivalency was tested using multiple group comparisons with nested models. As proposed by Raju, Lafitte, and Byrne (2002), invariant factor loadings between the English and German groups were used as criteria for measurement invariance. Since Δχ² has been criticized on account of its dependence on sample size (e.g., Hu & Bentler, 1998), we used ΔCFI as an indicator in the comparison of models. According to Cheung and Rensvold (2002), values greater than .01 indicate a significant drop in fit. Multigroup CFA indicate that the structure and the factor loadings are invariant across the two countries (unconstrained model: χ² = 793.77, df = 234; model with equal factor loadings: χ² = 824.71, df = 249; ΔCFI = 0.004) showing the stability of the NISS over the two samples. Internal consistency was similar to that found for the original German version. Cronbach’s α ranged from .79 (AR) to .91 (NS; .88 for the total scale). The correlation coefficients between the SSS-V and the NISS are described in Table 2. As can be seen, the subscale NS moderately correlates with the subscales of the SSS-V and corresponds to the correlations between the intensity subscale of the AISS and the subscales of the SSS-V (see Arnett, 1994; Ferrando & Chico, 2001). There is, however, no relationship between the AR subscale and the SSS-V.



Discussion Overall, the results of the confirmatory factor analyses support the bidimensional structure of the English instruments. This appears to imply a clear crosscultural and crosslingual stability of the factorial structure of the NISS. Furthermore, as indicated by the Cronbach’s α coefficients, we can assume that the psychometric properties of the English adaptations of the NISS are adequate to excellent. These results support the applicability of the English versions. Correlations between NISS and SSS-V are midrange, as expected. However, the subscale AR does not correlate with any of the SSS-V scales, which might suggest that this second fac© 2011 Hogrefe Publishing



15



Study 3: Validity in a Nonclinical Context By conceptualizing sensation seeking as a basal need for stimulation that may be expressed in multiple areas of a person’s life, the present study examined the validity of the NISS in comparison with the SSS-V by applying a criterion other than various types of risk behavior. In keeping with Arnett’s suggestion (1994) to explore the sensation seeking trait in relation to work, we employed the willingness for occupational change as criterion with the following manifestations: “intention to change workplace” and “wanting to learn or work in a new job.” It was expected that, compared to low sensation seekers, the general need for stimulation of high sensation seekers would be manifested in a comparatively high willingness to change workplace. This assumption is also supported by Zuckerman’s (1994) description of high sensation seekers as preferring changing surroundings and low sensation seekers as preferring environments that are highly stable and predictable. Because the intention to change jobs is influenced by the current situation of the employment market, we chose to examine a sample of hairdressers, as hairdressing is one of the few occupations in Germany in which vacancies exist.



Methods Participants The participants in this study were female employees of 60 hairdressing salons in Germany. Altogether, 146 people work for this company, who were asked to complete a questionnaire. The scales were administered by the same person in all cases (including an introduction). Of the 146 employees invited to take part in the study, 118 participated. Responses from participants who had more than three missing answers were excluded (n = 11). Male hairdressers (N = 2) were excluded due to their small number. This left a final sample of 105 female hairdressers aged between 19 and 65 years (M = 40.83, SD = 13.39).



Measures In addition to questions relating to demographic information, participants completed the German form of the NISS and the German version of the SSS-V by Beauducel, Strobel, and Brocke (2003; see description above). Respondents were also requested to indicate how often they intended to change their work place during their professional life and whether they would like to learn or switch to a new European Journal of Psychological Assessment 2012; Vol. 28(1):11–18



16



M. Roth & P. Hammelstein: The Need Inventory of Sensation Seeking (NISS)



Table 3. Means (M) and standard deviations (SD) of NISS and SSS-V scores in participants who do not intend to change their workplace (No-ICW; n = 71) vs. participants who intent to do so (ICW; n = 34), and results of MANOVA and MANCOVA (covariate: age) No ICW ICW MANOVA MANCOVA M SD M SD F ε² F ε² NISS total 41.17 6.61 47.01 8.41 14.53*** .13 10.66** .10 NS 23.26 5.57 28.80 7.41 17.68*** .15 13.27*** .12 AR 15.92 3.62 16.04 3.19