A Structural Examination of Two Optimal Stimulation Level Measurement Models

ABSTRACT - The study examines the internal structure of two measurement models used to operationalize the optimal stimulation level (OSL) construct in consumer research. Using a causal modeling approach, the structure of Zuckerman's Sensation Seeking Scale was confirmed. However, the analysis raises questions regarding the internal specification of Mehrabian and Russell's Arousal Seeking Tendency instrument. The need for a theory-based definition of OSL in consumer behavior is identified. Implications for future OSL scale refinement research are offered

Citation:

Russell G. Wahlers and Michael J. Etzel (1990) ,"A Structural Examination of Two Optimal Stimulation Level Measurement Models", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 415-425.

Advances in Consumer Research Volume 17, 1990      Pages 415-425

A STRUCTURAL EXAMINATION OF TWO OPTIMAL STIMULATION LEVEL MEASUREMENT MODELS

Russell G. Wahlers, Ball State University

Michael J. Etzel, University of Notre Dame

ABSTRACT -

The study examines the internal structure of two measurement models used to operationalize the optimal stimulation level (OSL) construct in consumer research. Using a causal modeling approach, the structure of Zuckerman's Sensation Seeking Scale was confirmed. However, the analysis raises questions regarding the internal specification of Mehrabian and Russell's Arousal Seeking Tendency instrument. The need for a theory-based definition of OSL in consumer behavior is identified. Implications for future OSL scale refinement research are offered

INTRODUCTION

The OSL Construct and Adaptive Behavior

Behavioral scientists' interest in the concept of stimulation dates back at least a century. In the field of psychology, the concept has been incorporated in several frameworks that describe the link between one's psychological state and his/her reaction to environmental stimuli (Hebb 1955; Leuba 1955). A summary of the OSL literature can be found in Rogers (1979). A generally accepted conceptual treatment of the stimulation construct is that each individual has a uniquely determined, homeostatic degree of stimulation or an "optimal stimulation level" (OSL) with which he/she is comfortable. When the environment is deficient in providing stimulation at this level, one tends to seek complexity or novelty. Conversely, when the environment provides more stimulation than the desired, optimal level, the individual will engage in behavior to reduce stimulation. Thus, the individual is viewed as adapting to his/her environment so as to maintain a balance between actual and optimal levels of stimulation. This adaptive behavior will involve either stimulation seeking or avoidance depending on the individual's OSL and present profile of environmental input. Thus, three factors must be considered in the analysis of stimulation adaptive behavior: (a) the individual's OSL, (b) the stimulating or arousal potential of the object, person, or event within the environment which is confronting the individual, and (c) one's state at the time of exposure to the stimulus in question (Berlyne 1960).

Several alternative theoretical frameworks incorporating the OSL construct have been developed by psychologists to explain stimulation-related exploratory behavior in a clinical research setting (Berlyne 1960, 1963; Driver and Streufert 1965; Fiske and Maddi 1961; Hunt 1963). More pragmatically, the construct has obvious significance in a consumer behavior context as OSL may contribute to an explanation of differences across consumers relative to a range of diverse behaviors including brand switching, acceptance of innovations, shopping behavior, and response to advertising repetition as summarized by Raju (1980).

OSL Measurement Models In Consumer Research

Stimulation seeking as an influence on consumer behavior has been addressed in the marketing literature both conceptually (Faison 1977; Hirschman 1984; Howard and Sheth 1969; Kirby 1975; Markin 1974; Raju 1981; Raju and Venkatesan 1980; Venkatesan 1974) and empirically (Etzel and Wahlers 1984; Goodwin 1980; Grossbart, Mittelstaedt, and Devere 1976; Joachimsthaler and Lastovicka 1984; Mittelstaedt, Grossbart, Curtis, and Devere 1976; Raju 1980; Venkatraman and MacInnis 1985; Wahlers and Etzel 1985; Wahlers, Dunn, and Etzel 1986). The evidence provided by prior studies suggests that the tendency to seek stimulation is a significant behavioral influence that is manifest in varying degrees across a number of consumer activities. Research in marketing thus far has dealt with arousal seeking and has not focused specifically on stimulation avoidance behavior.

In terms of measuring the OSL construct, Raju (1980) has identified several paper-and-pencil instruments found in the psychology literature as potential OSL operationalizations. Among these are the Sensation Seeking = scale (Zuckerman 1979; Zuckerman, Kolin, Price, and Zoob 1964), the Arousal Seeking Tendency scale (Mehrabian and Russell 1974), the Change Seeker Index (Garlington and Shimota 1964), the Stimulus Variation Seeking scale (Penney and Reinhr 1966), and the Similes Preference Inventory (Pearson and Maddi 1966). The Sensation Seeking (SS) scale and the Arousal Seeking Tendency (AST) scale have been the two predominant instruments employed in consumer behavior- studies. The principal studies involving an empirical treatment of OSL in a marketing context are chronologically summarized in Table 1.

An inspection of the Sensation Seeking scale and the Arousal Seeking Tendency scale and their respective subscales proposed by the scale developers suggests that these two instruments tap somewhat different aspects of OSL. The SS instrument consists of forty pairs of statements in a forced-choice format. Zuckerman (1979) has identified four components which underlie SS: (a) Thrill and Adventure Seeking, (b) Experience Seeking, (c) Boredom Susceptibility, and (d) Disinhibition. Alternatively, the AST scale consists of forty items combined with nine-point Likert agreement-disagreement scales. Mehrabian and Russell (1974) assert that the instrument taps five underlying factors relating to OSL: arousal associated with (a) Change, (b) Unusual Stimuli, (c) Risk, (d) Sensuality, and (e) New Environments.

TABLE 1

MEASUREMENT OF OSL IN SELECTED CONSUMER BEHAVIOR STUDIES

While it appears intuitively obvious that these factors are related both internally and across scales, the two scales have not been compared in a consumer behavior study. In addition, the degree of interscale congruence with respect to scale components is unclear. Thus far these two scaling approaches have been used interchangeably in consumer research.

Research Objective

Prompted by the observation that Sensation Seeking and Arousal Seeking Tendency have been the scaling methods of choice to operationalize OSL in the consumer behavior literature -- and noting that the scales have different subscale descriptions, the present study was conducted to examine both the measurement congruence between the scales and to assess the dimensionality of the measures. The results should provide consumer behavior researchers some direction in OSL scale selection -- particularly when dimensionality is relevant.

THE STUDY

Method

To examine the Sensation Seeking and Arousal Seeking Tendency as alternative measures of OSL, data were collected via mail questionnaire from a Midwestern consumer panel. The questionnaire contained Form V of the Sensation Seeking scale (Zuckerman 1979) and the Arousal Seeking Tendency instrument (Mehrabian and Russell 1974). A pretest of the questionnaire found no item order bias (either between or within the two OSL instruments). Thus, the two instruments included in the final questionnaire were presented in the same form that they appear in the psychology literature. The results reported in this paper are based upon 697 usable responses to the questionnaire.

In an initial examination of the data, SS and AST scores were calculated for each respondent after making the appropriate item scale reversals described in the literature. In addition, subscale scores for SS and AST were generated for each subject. The subscale structure of the Sensation Seeking scale is fully detailed by Zuckerman (1979). However, since the Arousal Seeking Tendency instrument's item-to-subscale pairings are only partially disclosed in the literature (Mehrabian and Russell 1974, pp. 4243), the complete structure (see Figure 2) was reconstructed with the assistance of the scale developers.

Reliability estimates for SS (0.851 overall) and each of the four SS subscales were calculated in addition to between-subscale correlations. Both reliabilities and correlation coefficients associated with the SS scale were found to be in agreement with parallel statistics reported in previous studies by Zuckerman (1979, pp. 110-113) and Ridgeway and Russell (1980, p. 663). Similarly, reliability estimates and between-subscale correlations were calculated for AST. While no previous reports of subscale reliabilities were found in the literature, the reliability estimates for the total scale (0.853) was consistent with AST reliability (0.881) reported earlier by Wahlers, Dunn, and Etzel (1986). In addition, between-subscale correlations were judged to be in reasonable agreement with those disclosed by Mehrabian and Russell (1974). Thus, the consistency in the measurement models' reliabilities and subscale intercorrelations observed between the present study's subjects and other previous samples used by the scale developers and other researchers was interpreted as evidence of sample adequacy.

Correlation Between SS and AST

The first objective of the study was to determine the overall comparability of the two scales. To do that a simple correlation between SS and AST was computed. The results (r=.951) indicate that the scales produce very similar scores. Thus researchers interested in an overall, general OSL measure could feel comfortable with either scale.

TESTING THE INTERNAL STRUCTURE OF SS AND AST

Sensation Seeking as a Measure of OSL

The internal structure of the SS measurement model was examined using a second-order factor model shown in Figure 1. The items comprising each of the four subscales are identified by number for the sake of brevity using the numbering scheme consistent with Zuckerman's item list (1979, pp. 397-401).

FIGURE 1

SENSATION SEEKING MEASUREMENT MODEL

TABLE 2

PARAMETER ESTIMATES FOR SENSATION SEEKING MEASUREMENT MODEL

Parameters were estimated using LISREL (Joreskog and Sorbom 1978). In an initial analysis, the model produced unsatisfactory fit. (Because of the large sample size and model complexity, absolute fit in terms of the x2 statistic was not expected (Burt 1973, p. 148).) Therefore in order to continue the analysis, the model's fit was adjusted by sequential modifications indicated by the modification indices. It is important to note that our goal was to minimize unnecessary modifications so as to preserve the structural integrity of the SS measurement model relative to the form specified by the developers and used in past consumer behavior studies. Hence, modifications were continued only to a point at which the x2 statistic was approximately equal to twice the number of degrees of freedom (Carmines and McIver 1981, p. 80). The fit statistics for the SS measurement model were X2(731)=1508.61, GFI 0.883, AGFI=0.868, and RMSR=0.050. Parameters are included in Tables 2 and 3.

With the exception of Item 6 which crossloaded on the first subs ale (Thrill and Adventure Seeking), the remaining thirty-nine items exhibited the hypothesized loadings structure described by Zuckerman. In addition, only four correlated error terms were uncovered in the analysis which is not unexpected given the total number of indicators comprising the SS model. Thus, given the large sample size and small number of data-based modifications required to achieve approximate fit (i.e., x2 =2(df)), one can conclude that the SS model is reasonably specified both conceptually and empirically.

Arousal Seeking Tendency as a Measure of OSL

The second-order factor model associated with the Arousal Seeking Tendency framework is shown in Figure 2. Again, items comprising each of the five AST subscales are identified by number consistent with the item numbering scheme used by Mehrabian and Russell (1974, pp. 218-219).

Parameters were estimated using LISREL adjusting the model to fit the same X criterion described above. The fit statistics for the AST measurement model were X2(707)=1693.19, GFI=0.859, ACFI=0.837, RMSR=0.054. Parameters are included in Tables 4 and 5.

Results of this analysis point to several problems. First, the third subfactor (113) purportedly representing Risk exhibited an unsatisfactory parameter estimate (standardized 91=0.042, p>0.05). An inspection of the item loadings included in Table 5 further shows sixteen items specified for subscales one (Change), two (Unusual Stimuli), four (Sensuality), and five (New Environments) incorrectly cross loaded on the third subscale (Risk); and one item specified on the third subscale incorrectly cross loaded on subfactor one. In addition, seven items were found to exhibit parameters having nonsignificant t-values on their respective hypothesized subscales. A total of eight statistically significant intercorrelated error terms were further uncovered. Thus, although the overall goodness-of-fit statistics are encouraging, the abundance of cross loadings, nonsignificant parameter estimates, and a fair number of intercorrelated error terms leads to questions regarding the internal specification of the AST measurement model's structure.

TABLE 3

MAXIMUM LIKELIHOOD ESTIMATES FOR SENSATION SEEKING MEASUREMENT SCALE

FIGURE 2

AROUSAL SEEKING TENDENCY MEASUREMENT MODEL

TABLE 4

PARAMETER ESTIMATES FOR AROUSAL SEEKING TENDENCY MEASUREMENT MODEL

CONCLUSIONS

A second-order causal model analysis of the Sensation Seeking measurement model confirmed that its framework appears to be conceptually and empirically well specified. In contrast, however, a similar examination of Arousal Seeking Tendency uncovered serious problems in the structure of the paradigm's subfactors. Specifically, seventeen items were inappropriately cross loaded, seven items failed to load on their respective hypothesized subfactors, and eight statistically significant correlated error terms were discovered. These results suggest that researchers interested in selecting a measurement model to investigate the dimensions of the OSL construct would be advised to employ Sensation Seeking rather than Arousal Seeking Tendency. While both scales have been developed to tap a similar aggregate conceptualization of OSL, on a subscale level the two scales not only are comprised of somewhat different dimensions in terms of numbers and meaning, but additionally the AST structure could not be empirically confirmed using the present panel data.

Thus far the SS and AST paradigms have been used interchangeably in consumer research to represent a unidimensional, summed-scale measure of OSL. On this aggregate basis the two measurement models may tap a similar conceptualization of OSL. However, in cases in which the dimensionality of stimulation seeking is of interest, the structural differences between SS and AST may produce inconsistent findings.

Future research needs to address several issues. First, given the inability of the present data to confirm the structure of AST, this model requires further scrutiny before it is widely used in an applied marketing research context. Additionally, opportunity exists to compare the dimensional differences and similarities of SS and AST with respect to a range of actual behaviors reflecting stimulation seeking and/or avoidance. Finally, noting that both the SS and AST instruments were developed in a clinical context, there is an opportunity to develop a measurement instrument more directly representing the OSL construct in a consumer context. Given the apparent relevance of exploratory consumer behavior, this would be a logical research priority.

TABLE 5

MAXIMUM LIKELIHOOD ESTIMATES FOR AROUSAL SEEKING TENDENCY MEASUREMENT MODEL

REFERENCES

Berlyne, D. E. (1960). Conflict, arousal, and curiosity. New York, NY: McGraw-Hill.

Berlyne, D. E. (1963). Motivational problems raised by exploratory and epistemic behavior. In S. Koch (Ed.), Psychology: A Study of Science. New York: McGraw-Hill.

Burt, R. S. (1973). Confirmatory factor-analytic structures and the theory construction process. Social Methods and Research, 2, 131 -190.

Carmines, E. G., & McIver, J. P. (1981). Analyzing models with unobservable variables. In G. W. Bohrnstedt and E. F. Borgatta (Ed.), Social Measurement: Current Issues, (pp. 65-115), Beverly Hills: Sage Publications.

Driver, M. J., & Streufert, S. (1965). The 'general incongruity adaptation level' (GIAL) hypothesis: an analysis and integration of cognitive approaches to motivation. Paper No. 114, Institute for Research in the Behavioral, Economic, and Management Sciences, Herman C. Krannert Graduate School of Industrial Administration, Purdue University, West Lafayette, IN.

Etzel, M. J., & and Wahlers, R. G. (1984). Optimal stimulation and consumer travel preferences. In R.W. Belk, et al.(E 1.),1984 AMA Educators' Proceedings, 50, Chicago, IL: American Marketing Association, 92-95.

Faison, E. W. J. (1977). The neglected variety drive: a useful concept for consumer behavior. Journal of Consumer Research, 4, 172- 175.

Fiske, D. W., & Maddi, S. R. (1961). Functions of varied experience. Homewood, IL: The Dorsey Press.

Garlington, W. K., & Shimota, H. E. (1964). The change seekers index: a measure of the need for variable stimulus input. Psychological Reports, 14, 919-924.

Goodwin, S. A. (1980). Impact of stimulus variables on exploratory behavior. In J. C. Olson (Ed.), Advances in Consumer Research, 7, Ann Arbor, MI: Association for Consumer Research, 264269.

Grossbart, S. L., Mittelstaedt, R. A., & Devere, S. P. (1976). Consumers stimulation needs and innovative shopping behavior: the case of recycled urban places. In B.B. Anderson (Ed.), Advances in Consumer Research, 3, Chicago, IL: Association for Consumer Research, 30-35.

Hebb, D. O. (1955). Drives and the CNS (conceptual nervous system). Psychological Review, 62, 243-254.

Hirschman, E. C. (1984). Experience seeking: a subjectivist perspective of consumption, Journal of Business Research, I 2, 115-136.

Howard, J. A., & Sheth, J. N. (1969). The theory of buyer behavior. New York, NY: John Wiley and Sons, Inc.

Hunt, J. M. (1963). Motivation inherent in information processing and action. In O. J. Harvey (Ed.), Motivation and social interaction: cognitive determinants. New York, NY: The Ronald Press, 35-94.

Joachimsthaler, E. A., & Lastovicka, J. L. (1984). Optimal stimulation level - exploratory behavior models. Journal of Consumer Research, 11, 830835.

Joreskog, K. G., & Sorbom, D. (1978). LISREL IV: analysis of linear structural relationships by the method of maximum likelihood. Chicago: National Educational Resources.

Kirby, J. K. (1975). Consumer behavior: conceptual foundations. New York, NY: Dun-Donnelley Publishing Corporation.

Leuba, C. (1955). Toward some integration of learning theories: the concept of optimal stimulation. Psychological Reports, 1, 27-33.

Markin, R. J. (1974) Consumer behavior: a cognitive orientation. New York, NY: MacMillan Publishing Co., Inc.

Mehrabian, A., & Russell, J. A. (1974). An approach lo environmental psychology, Cambridge, MA: The MIT Press.

Mittelstaedt, R. A., Grossbart, S. L., Curtis, W. W., & Devere,,S. P. (1976). Optimum stimulation level and the adoption decision process. Journal of Consumer Research, 3, 84-94.

Pearson, P. H., & Maddi, S. R. (1966). The similes preference inventory: development of a structural measure of the tendency toward variety. Journal of Consulting Psychology, 30, 301-308.

Penney, R. K., & Reinhr, R. C. (1966). Development of a stimulus-variation seeking scale for adults. Psychological Reports, 18, 631-638.

Raju, P. S. (1980). Optimum stimulation level: its relationship to personality, demographics and exploratory behavior. Journal of Consumer Research, 7, 272-282.

Raju, P. S. (1981). Theories of exploratory behavior: review and consumer research implications. In J. Sheth (Ed.), Research in Marketing, 4, Greenwich, CT: JAI Press, 243-249.

Raju, P. S., & Venkatesan, M. (1980). Exploratory behavior in the consumer context: a state of the art review. In J. C. Olson (Ed.), Advances in Consumer Research, 7, Ann Arbor, MI: Association for Consumer Research, 258-263.

Ridgeway, D., & Russell, J. A. (1980). Reliability and validity of the sensation-seeking scale: psychometric problems in form V. Journal of Consulting and Clinical Psychology, 48, 662-664.

Rogers, R. D. (1979). Commentary on the neglected variety drive, Journal of Consumer Research, 6, 88-91.

Venkatesan, M. (1974). Cognitive consistency and novelty-seeking. In Consumer Behavior: Theoretical Perspectives. Englewood Cliffs, NJ: Prentice Hall.

Venkatraman, M. P., & MacInnis, D. J. (1985). The epistemic and sensory exploratory behaviors of hedonic and cognitive consumers. In E. C. Hirschman and M. B. Holbrook (Ed.), Advances in Consumer Research, 12, Provo, UT: Association for Consumer Research, 102-107.

Wahlers, R. G., Dunn, M. G., & Etzel, M. J. (1986). The congruence of alternative OSL measures with consumer exploratory behavior tendencies. In R. J. Lutz (Ed.), Advances in Consumer Research, 13, Provo, UT: Association for Consumer Research, 398402.

Wahlers, R. G., & Etzel, M. J. (1985). A consumer response to incongruity between optimal stimulation and lifestyle satisfaction. In E. C. Hirschman and M. B. Holbrook (Ed.), Advances in Consumer Research, 12, Provo, UT: Association for Consumer Research, 97-101.

Zuckerman, M. (1979). Sensation seeking: beyond the optimal level of arousal. Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

Zuckerman, M., Kolin, E. A., Price, L., & Zoob, I. (1964). Development of a sensation seeking scale. Journal of Consulting Psychology, 28, 477-482.

----------------------------------------