Effects of Consistency of Visible Consumption Patterns on Impression Formation

Russell W. Belk, University of Utah
ABSTRACT - This experiment tests the hypotheses that we evaluate those who display consistent consumption stereotypes more favorably than those who do not and that we like those whose consumption patterns are most like our own. The second (similarity) hypothesis is supported but the primary (consistency) hypothesis is only supported where the consistent consumption pattern is also similar to our own preferred consumption patterns. Two explanations of the failure of the consistency hypothesis are offered and an experiment to test these explanations is briefly described.
[ to cite ]:
Russell W. Belk (1980) ,"Effects of Consistency of Visible Consumption Patterns on Impression Formation", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 365-371.

Advances in Consumer Research Volume 7, 1980     Pages 365-371


Russell W. Belk, University of Utah


This experiment tests the hypotheses that we evaluate those who display consistent consumption stereotypes more favorably than those who do not and that we like those whose consumption patterns are most like our own. The second (similarity) hypothesis is supported but the primary (consistency) hypothesis is only supported where the consistent consumption pattern is also similar to our own preferred consumption patterns. Two explanations of the failure of the consistency hypothesis are offered and an experiment to test these explanations is briefly described.

There are at least two reasons why understanding the phenomenon of impression formation based on visible consumption information is important to the study of consumer behavior. One reason is that the way we perceive other consumers has an impact on our own consumption choices. More specifically, when we see known others using unknown products or brands, the natural tendency is to generalize from the traits we know the product user possesses to the traits we then presume that the product possesses. Such inferences in turn may affect our responses toward these products. Moreover, an inference in the opposite direction is equally plausible. When we see unknown others using known products or brands, the generalization goes from product traits to person traits. This sort of inference may affect our responses toward these people. Both types of inferences are on-going manifestations of the consumer's general expectations that others' consumption selections are expressive behaviors which may be interpreted in light of a shared system of product and service symbols. In order to fully understand one half of this mutual inference system we must also understand the other.

A second reason for studying impression formation based on visible consumption, is to determine the extent to which certain products and services actually do act as shared symbols. Blumberg (1974) and Felson (1974, 1978) have provided arguments and suggestive evidence that, at least for inferences of social status, some products and services may no longer serve as the strong and clear symbols they once were in this country. Whether or not the conclusion that the so-called status symbol is a disappearing entity is correct, the existence of broader consumption-based stereotypes is an assumption in need of testing. To the extent that consumption-based stereotypes do exist and we invoke them in forming our perceptions of and actions toward others, clearly this is a phenomenon of sufficient social significance to warrant a determined research effort.


One need not go as far back in the literature as Veblen's (1889) familiar hypothesis that conspicuous consumption patterns serve to communicate and reinforce social class distinctions, to be able to find theorizing about the effects of visible consumption patterns on impression formation. More accurately, there is much recent literature on the intended effects of visible consumption on impression formation. For instance, Goffman (1959) has spoken of the use of products and services as "sign equipment" in the attempt to create a particular "personal front" to present to others. Schlenker (1978) extended these notions by suggesting that we employ whatever consumption selections we believe will enhance our image so long as the projected image is thought to be credible to the intended "audience". Various researchers dealing with self concepts have found that consumers apparently intend to express themselves in their selections of a great many products and services. [E.G. Grubb and Grathwohl, 1967; Birdwell, 1968; Grubb and Hupp, 1968; Dolich, 1969; Green, et. al., 1969; Hamm and Cundiff, 1969; Mason and Mayer, 1970; Grubb and Stern, 1971; Ross, 1971; Dornoff and Tatham, 1972; Greeno, Sommers, and Kernan, 1973; Landon, 1974; Belch and Landon, 1977; Schewe and Dillon, 1978.] Cocanougher and Bruce (1971) demonstrated that our consumption choices may also sometimes be an attempt to display a consumption stereotype in order to emulate a socially distant reference group. And several researchers have extended Brock's (1968) commodity theory into a consumer realm and found that scarcer goods are more valued as symbols presumably intended to convey status or uniqueness to the consumer. [Fromkin, Olson, Dipboye, and Barnaby, 1971; Worchel, Lee and Adewole, 1975; Szybillo, 1975.]

It will be noted that all of the preceding perspectives have focused on determinants of a consumer's consumption choices. For reasons noted earlier, the present focus is on the effect of these consumption choices on the impressions observers form of a given consumer. To date there have been only a few exploratory studies from this point of view. [Holman, 1976; Belk, 1977.] There are however two related streams of research which are relevant to this focus. One is the study of stereotypes, including their formation, veridicality, and the underlying cues relied upon in deciding whether a given individual should be considered in the context of a particular stereotype. [For a review, see Tajfel, 1969.] The other is the study of person perception, including determinants of the ability to accurately perceive others, integration of person attribute cues, and the effects of various cue presentation formats. [See for instance, Tagiuri, 1969; Warr and Knapper, 1965; Huston and Levinger, 1978.] While the hypotheses and findings in these areas are potentially of great relevance in considering the effects of visible consumption on impression formation, these research areas have almost totally ignored consumption cues in favor of other cues such as lists of personality traits, the stimulus person's sex, and various physionomic features. Since the procedure in much of this research has been to give subjects lists of stimulus person attributes and ask them to make further inferences about the person, these findings are based on cues which are normally less readily observable but perhaps more directly interpretable (i.e., less symbolic) than are consumption cues. Therefore it is not certain at this point in time that the processes and principles governing the use of consumption cues in impression formation are identical to those which apply to the nonconsumption cues which have been studied.


One area of previous research of particular interest in the present study is the positive effect of cue consistency on ratings of liking for a stimulus person. The underlying theoretical arguments are the cognitive consistency premise that inconsistency is unpleasant, and the response generalization premise that this negative affect will attach to the evaluation of the stimulus person possessing the inconsistent cues. Prior research has introduced inconsistency by providing conflicting adjectives to describe an unseen person (e.g., resourceful and helpless, or vigorous and withdrawn) and has obtained evidence that subjects evaluate persons described by consistent cues more favorably providing that cue patterns are balanced for their desirability (e.g., Anderson and Jacobson, 1965; Hendrick, 1972; Kaplan, 1973). If this finding were to hold when the cues are visible consumption choices, it would mean that we prefer those who conform to consumption-based stereotypes over those who instead show seemingly inconsistent consumption patterns. This seems plausible and is the primary hypothesis which motivated the present study, even though cue consistency may be harder to detect when the cues are visible consumption choices.

In addition to the ease of detecting cue inconsistency, another difference between impression formation based on consumption cues and impression formation based on other cues, is that the desirability of given product and service choices is not universal. While most people would agree that it is more desirable to be resourceful than non-resourceful, it is not generally agreed whether it is more desirable to own a luxury car or a sports car, to live in the city or suburbs, or to drink beer or scotch. However by utilizing an extension of Byrne's (1971) similarity-attraction paradigm, we may infer that people expect others whom they like to share a preference for the products and services which they themselves own or aspire to own. There is in fact some support for this extension of Byrne's hypothesis, at least when the others are also similar to the subject in other respects (Belk, 1976). Thus a secondary hypothesis guiding the present research is that stimulus persons who display consumption choices similar to those of the subjects will be evaluated more favorably than stimulus persons who display consumption choices unlike those of the subjects.



In order to integrate the testing of the consistency preference hypothesis and the similarity preference hypothesis, it was not possible to employ a completely crossed factorial experimental design. This is due to the fact that if subjects' own consumption patterns are normally seen as consistent, then an inconsistent consumption pattern will not exist which is also similar to subjects' consumption patterns. Therefore only three types of consumption patterns were sought:

A = Consistent and similar to perceivers' consumption desires,

B = Consistent and dissimilar to perceivers' consumption desires, and

C = Inconsistent and partially dissimilar to perceivers' consumption desires.

The degree of dissimilarity in condition C was to be less than the degree of dissimilarity in condition B so that a stronger test of the more tenuous consistency preference hypothesis would be provided. The predicted order of affect toward the stimulus persons represented by these three types of consumption patterns (to be called persons A, B, and C) then becomes: A> B> C

Person B is predicted to be evaluated more favorably than person C due to consistency preference, despite the fact that person C's consumption is somewhat more similar to the subjects' consumption than is person B's. Nevertheless the dissimilarity between person B's consumption and the subjects' consumption is the basis for predicting that person B will be evaluated less favorably than person A. It is in this sense that the experiment provides a more severe test of the consistency preference hypothesis, since this consistency effect would have to overpower a moderate increment in dissimilarity in order to be supported in the comparison of ratings of persons B and C.


A series of pretests were conducted using college student subjects similar to those used as subjects in the main experiment. The major purpose of these pretests was to find two sets of four products each, such that:

1.  One set was perceived to be representative of male college students similar to the pretest respondents [Both pretest and main experiment subjects were approximately 50 percent female. As in the main experiment, analyses by sex showed no significant differences.] and the other set was perceived to be representative of an older male college graduate who has become established in his career.

2.  The products in each set were consistent in the sense that they were perceived as likely to be owned by someone owing any of the other products in the set, but unlikely to be owned by someone owning any of the products in the other set.

In addition to seeking two consistent product sets which differed in their similarities to the products owned by the student subjects of the pretest, these guidelines made it possible to form mixes of products drawn from the two sets which would then be perceived as inconsistent. Thus consistency was operationalized in terms of the likelihood that the consumption choices would be made by the same person, as perceived by college student subjects. After several pretests with a total of 40 different products, the item sets shown in Table 1 were selected. Set 1 formed the consistent/similar set, set 2 formed the consistent/dissimilar set, and sets 3 and 4 formed the inconsistent/partially dissimilar sets. Sets 3 and 4 are comprised of two products from each of the other sets. The average correlations in desires were based on the respondents' desires to own each of the items in a set. As intended, those items in each of the consistent item sets (1 and 2) had positive and significantly correlated desire scores, but not so strong that these products could be viewed as complimentary goods. The average correlation between pairs where one item was drawn from set 1 and one from set 2 was nonsignificant. The mean desire scores for the two sets of products show that the products in the similar set (1) were indeed desired by subjects significantly more than those in the dissimilar set (2). However both sets of products were at least moderately appealing to these subjects as intended. Finally, the mean co-ownership ratings show that the pairs of items within each of the first two sets were seen as very likely to be owned by the same person, while the pairs of items drawn from both of the two sets were seen as somewhat unlikely to be owned by the same person. While the items in sets 1 and 2 may not be entirely balanced in such characteristics as the amount of activity versus passivity implied by their use, the pretests suggest that these differences are inseparable aspects of the lifestyles implied by the college and post-college consumption stereotypes. The measures shown in Table 1 were also collected in the main experiment as a manipulation cheek. The results of these checks were comparable to those shown in Table 1.




In the main experiment subjects were 58 male and 62 female undergraduate business students. A between-subjects design was employed with a total of 30 subjects per cell. For convenience and clarity the design was treated as a completely randomized design with 4 treatment levels rather than a nested 2 x 2 factorial design. Initial analyses showed no significant effect for subject sex, so this blocking variable is ignored in subsequent analysis.

The study was run without an elaborate disguise as a study of "interpersonal insight". Subjects received a consumption profile consisting of one of the four item sets and were asked to describe "Mr. X" who owned these items, as best they could in a series of structured questions.

The major dependent variables were derived from a factor analysis of responses to the 15 stimulus person descriptors show in Table 2. This set of descriptors was selected based on Pervin's (1976) insightful work in eliciting open-end person descriptions. Each potential descriptor was applied by the subjects to the stimulus person in the sentence, "Mr. X is ________"and rated on a 5- point "strongly disagree" to "strongly agree" response scale. Factor analyses were also performed separately by cell and separately for males and females. Since these analyses never failed to obtain reasonable approximations of the two factors shown here, the more stable total sample analysis was the one retained. The first factor (rotated) is clearly affect toward the stimulus person being rated. This was taken as the dependent variable for hypothesis testing, although subsequent tests substituting the simple scores on variable seven lead to the same conclusions. The second factor might be described as dominance, success, or power. Factor scores on both variates were used in further analysis although there were no a priori hypotheses concerning the dominance factor. Several other dependent measures were obtained including estimations of the stimulus per- son's demographic characteristics and judgements of how the stimulus person would make the forced consumption choice shown in Table 3. It was reasoned that due to limited time, money and scope of consumption interests, people do make such consumption choice tradeoffs. The data were intended to provide additional depth to subjects' descriptions of the stimulus person by probing inferences about his lifestyle. The only a priori hypothesis was that these attributed choices would differ between the four stimulus person consumption patterns.





Following the subjects' rating of the stimulus person, they also provided self descriptions on the characteristics on which the stimulus person was rated (except for the lifestyle forced choices) and provided the data for the manipulation checks. The self descriptions were employed in supplementary analyses which will be discussed following a consideration of the primary results.


The principal hypothesis tests were obtained from the analysis of variance on affect factor scores summarized in Table 4. As hypothesized, the similar/consistent college student consumption set resulted in the highest affect ratings. These college student subjects rate someone conforming to a consistent college student consumption stereotype as most highly liked. The "mixed" third and fourth product sets were evaluated as equally well liked, also as hypothesized. However, instead of the dissimilar/consistent set resulting in the second highest affect scores as predicted, this consistent post post-college set was evaluated the least favorably of the four conditions. Explanations of this departure from hypothesis will be considered shortly.







Table 5 presents the results of the analysis of variance on the dominance factor scores. Although there were no a priori hypotheses here, the results are readily interpretable. Both consistent product sets resulted in higher attributions of dominance than the inconsistent sets. It appears that more ambiguous (inconsistent) consumption patterns are attributed to consumers who are unable or unwilling to make stereotypical consumption choices and are judged to be ineffectual as a result. The fact that the dissimilar (post-college) set of products also led to higher dominance scores than the similar (college) set, may be interpreted to be a function of age or life experience. That is, a stimulus person judged to be older and more experienced may also be seen as more dominant. This interpretation is supported by an analysis of the estimated ages of the stimulus person showing that the post-college stimulus person's mean age was estimated to be 37.0 years versus 26.2 years for the college stimulus person consumption profile.

The effect of the four treatment levels on the lifestyle forced choices is summarized in Table 6. Most interesting here is the comparison of the choices attributed to the stimulus person with the college student consumption profile (set 1) versus the post-college consumption profile (set 2). It may be seen that the consumption-evoked stereotypes of these two consumers differs greatly and in ways which are consistent with the lifestyle stereotypes which might well have been obtained if he stimulus cues read directly, "college student" and "post college". These results also demonstrate that the impressions formed based on the visible consumption patterns examined, extend well beyond the general affect and dominance effects and include differences in expectations regarding specific behaviors by the stimulus person.

Some additional insight into the similarity-attraction hypothesis may be gained from examining the data provided by subject self descriptions. Although the hypothesis is supported by the greater affect scores toward the person with the similar/consistent consumption profile versus the dissimilar/consistent consumption profile, it may or may not hold when extended to liking of the consumption items reportedly owned by the stimulus person across all treatment cells. In order to test this extension, an index of subject/stimulus person similarity was constructed by taking the reciprocal of the summed absolute differences between a subject's ratings of the stimulus person on the 15 attributes shown in Table 2 and this subject's self description on these same attributes. This similarity index was then correlated with an index of desire for the items the stimulus person reportedly consumed, calculated as the sum of the subject's desire ratings (see Table 1) for the four items in the profile they were given. The resulting correlation coefficient was a significant +.30, lending support to the extended hypothesis that not only do we like those people better who are more like us, but we also like the consumption choices of these people better. Due to the nature of the data underlying the similarity index, we must interpret this finding cautiously, but it provides tentative support for the premise that impressions of people affect our desire for products associated with these people.


Let us first consider the conclusions which clearly emerge from the study. First it is evident that we like those who like the things we do (i.e., who consumer as we do). Secondly, those who display consistent consumption stereotypes are judged, at least on the basis of first impressions, as more dominant. However these same people are not necessarily more liked. While the college student subjects liked those who consistently consume like college students, those displaying inconsistent consumption patterns are seen as more likeable than those displaying consistent consumption patterns which are dissimilar to the college students' preferences while they are in college.

Several explanations of this unexpected finding are possible. One intuitively appealing explanation is that the similarity of consumption patterns to those preferred by the perceiver is the strongest determinant of liking for the stimulus person and a compensatory model of cue integration is utilized in processing cues which differ in their appeal due to varying amounts of similarity to the subject's preferences. In this explanation since either of the two inconsistent sets contain two out of four products which are similar (appealing) to the consumption of the college student sample, these sets (and their owners) are more attractive than the consistent/dissimilar set in which none of the four products reflect college student consumption preferences. Norman Anderson (1974) provides much evidence for such a cue integration model in studies of non-consumption cues in impression formation. It will be noted however that the present findings with dominance judgements based on visible consumption cues seem to support a non-compensatory cue integration model in which the consistency of the total cue configuration affects dominance inferences.

An alternative explanation of the failure of the consistency hypothesis for affect judgements has to do with the life cycle stage of the subjects. In this explanation it is recognized that while most people gradually acquire new roles and new value-expressive symbols as they go through life, college juniors and seniors are nearing a disruptive transition from student roles to markedly different career roles and perhaps new family roles as well. Since they are still students, the consumption items which are most similar to college student consumption patterns are still the most attractive. However because they can soon anticipate a transitional stage in which they are shedding old symbols and acquiring new ones which are more appropriate to their change in roles, the mixed (inconsistent) consumption set is the next most attractive to them and results in the next highest affect scores toward the presumed owner of these items.

While the difference between these two explanations may seem subtle, it is important in terms of understanding the effect of consumption cue consistency. In order to test the possible explanations of the cue consistency results, an exact replication of the study just reported was conducted using subjects from a community sample of 72 college graduates in Champaign-Urbana (35 male, 37 female; median age 43). These older college graduates selected because they were believed to have references most favorable to the products in set 2 (consistent/ dissimilar in the study reported here). Since the set 2 and set 1 products were indeed found to have reversed similarities for this subject group, the prediction under either of the ex post facto explanations is that set 2 would yield higher person affect than would set 1. Furthermore if the compensatory cue integration explanation is correct, this group should again judge the owners of the mixed sets to be more liked than the owner of the set 1 products which are for them consistent/dissimilar. However since these older subjects did not anticipate a forthcoming transitional life cycle stage combining college and post-college symbols, if the second explanation is instead correct, the college products (set l) should result in stimulus person affect no different than that from the mixed sets.

While space prohibits a full presentation of results from the post-college sample, these results favored the compensatory cue integration explanation. In other words, these subjects showed the highest affect toward the stimulus person when they were shown the post-college consumption objects list and the lowest affect toward the person represented by the college consumption objects. The affect ratings in both of these conditions were significantly different from ratings under the "mixed" conditions.


The results presented do not provide entirely compelling evidence that there is no consistency preference effect in person perception based on visible consumption cues. They merely suggest that if such an effect exists it is not sufficiently strong to overcome the similarity effect obtained here. In order to test whether the consistency effect exists at all, further study is needed holding similarity constant. This is feasible only under conditions of dissimilarity since the similar/ inconsistent combination cannot normally occur.

Before pursuing additional research of this type, it was deemed important to validate the basic measurement procedure. This was done in a second replication in which the nature of the task was reversed. The mean significant person attribute scores and the model significant life style choices from the initial results, were presented for the four treatment levels to four new samples of college subjects. They were then asked to effectively reconstruct the present object stimuli by assessing the likelihood that the person described owned each of several products, including those manipulated in the present experiment.

Space again precludes detailed results, but they strongly support the present methods. In almost every instance the person attribute and lifestyle descriptions led to ratings of the products comprising particular treatment level as more likely owned than products not in the corresponding treatment level. The only exceptions occurred for three of the eight products in the "mixed" cells where ambiguity would be expected. This retrieval lends support to the non-antifactual existence of the effects detected in the present study. In addition these results provide additional confirmation of the existence of clear consumption-based stereotypes.


Anderson, Norman H. (1974), "Cognitive Algebra: Integration Theory Applied to Social Attribution," in Advances in Experimental Social Psychology, 7, ed. L. Berkowitz, New York: Academic Press, 1-101.

Anderson, Norman H. and Jacobson, Ann (1965), "Effect of Stimulus Inconsistency and Discounting Instructions in Personality Impression Formation, Journal of Personality and Social Psychology, 2, 531-539.

Belch, George E. and Landon, E. Laird, Jr. (1977), "Discriminant Validity of a Product-Anchored Self-Concept Measure," Journal of Marketing Research, 14, 252-256.

Belk, Russell W. (1976), "It's the Thought that Counts: A Signed Digraph Analysis of Gift-giving," Journal of Consumer Research, 3, 155-162.

Belk, Russell W. (1977), "Assessing the Effects of Visible Consumption on Impression Formation," Advances in Consumer Research, 5, 39-47.

Birdwell, Al E. (1968), "A Study of the Influence of Image Congruence on Consumer Choice," Journal of Business, 41, 76-88.

Blumberg, Paul (1974), "The Decline and Fall of the Status Symbol: Some Thoughts on Status in a Post-Indus-trial Society," Social Problems, 21, 480-498,

Brock, Timothy C. (1968), "Implications of Commodity Theory for Value Change," in Psychological Foundations of Attitudes, eds. A. G. Greenwald, T. C. Brock, and T. M. Ostrom, New York: Academic Press.

Byrne, Daniel (1971), The Attraction Paradigm, New York: Academic Press.

Cocanougher, A. Benton and Bruce, Grady D. (1971), "Socially Distant Reference Groups and Consumer Aspirations,'' Journal of Marketing Research, 4, 379-381.

Dolich, Ira J. (1969), "Congruence Relationships Between Self-Images and Product Brands," Journal of Marketing Research, 6, 80-84.

Dornoff, Robert J. and Tatham, Robert L., (1972), "Congruence Between Personal Image and Store Image, Journal of the Market Research Society, 7, 45-52.

Felson, Marcus (1976), "The Differentiation of Material Life Styles: 1925 to 1966," Social Indicators Research, 3, 397-421.

Felson, Marcus (1978), "Invidious Distinctions Among Cars, Clothes and Suburbs," Public Opinion Quarterly, 42, 49-58.

Fromkin, Howard L., Olson, Jerry C., Dipboye, Robert L., and Barnaby, David (1971), "A Commodity Theory Analysis of Consumer Preference for Scarce Products," Proceedings, 79th Annual Convention, APA, 653-654.

Goffman, Erving (1959), The Presentation of Self in Everyday Life, Garden City,: Anchor Books.

Green, Paul E., Maheshwari, Arun, and Rao, Vithala R. (1969), "Self-Concept and Brand Preference: An Empirical Application of Multidimensional Scaling," Journal of the Market Research Society, 11, 343-360.

Greeno, D., Sommers, Montrose S. and Kernan, Jerome B. (1973), "Personality and Implicit Behavior Patterns," Journal of Marketing Research, 10, 63-69.

Grubb, E. L. And Grathwohl, Harrison L. (1967),"Consumer Self-Concept, Symbolism and Market Behavior: A Theoretical Approach," Journal of Marketing, 31, 22-27.

Grubb, E. L. and Hupp, Greg (1968), Perception of Self, Generalized Stereotypes, and Brand Selection," Journal of Marketing Research, 5, 58-63.

Grubb, E. L. and Stern, Bruce L. (1971), "Self-Concept and Significant Others," Journal of Marketing Research, 8, 382-385.

Hamm, Bruce C. and Cundiff, Edward W. (1969), "Self-Actualization and Product Perception," Journal of Marketing Research, 6, 470-472.

Hendrick, Clyde (1972), "Effects of Salience of Stimulus Inconsistency on Impression Formation," Journal of Personality and Social Psychology, 22, 219-222.

Holman, Rebecca C. (1976), "Communicational Properties of Women's Clothing: Isolation of Discriminable Clothing Ensembles and Identification of Attributions Made to One Person Wearing Each Ensemble," Unpublished dissertation, University of Texas.

Huston, Ted L. and Levinger, George (1978), "Interpersonal Attraction and Relationships," Annual Review of Psychology, 29, 115-156.

Kaplan, Martin L. (1973), "Stimulus Inconsistency and Response Dispositions in Forming Judgements of Other Persons," Journal of Personality and Social Psychology, 25, 58-64.

Landon, E. Laird, Jr., (1974), "Self-Concept, Ideal Self-Concept, and Consumer Purchase Intentions," Journal of Consumer Research, 1, 44-51.

Mason, Joseph B. and Mayer, Morris L. (1970), "The Problems of the Self-Concept in Store Image Studies," Journal of Marketing, 34, 67-69.

Pervin, Lawrence A. (1976), "A Free-Response Description Approach to the Analysis of Person-Situation Interaction,'' Journal of Personality and Social Psychology, 34, 465-474.

Ross, Ivan (1971), "Self-Concept and Brand Preference," Journal of Business, 44, 38-50.

Schewe, Charles D. and Dillon, William R. (1978), "Marketing Information System Utilization: An Application of Self Concept Theory," Journal of Business Research, 6, 67-79.

Szybillo, George (1975), "A Situational Influence on the Relationship of a Consumer Attribute to New-Product Attractiveness," Journal of Applied Psychology, 60, 652-655.

Tagiuri, Renato (1969), "Person Perception," in Handbook of Social Psychology, 3, 2nd Edition, ed., G. Lindzey and E. Aronson, 395-449.

Tajfel (1969), "Social and Cultural Factors in Perception,'' in Handbook of Social Psychology, 3, 2nd Edition, ed. G. Lindzey and E. Aronson, 315-394.

Warr, Peter B. and Knapper Christopher (1968), The Perception of People and Events, London: Wiley.

Worchel, Stephen, Lee, Jerry, and Adewole, Akanbi (1975) "Effects of Supply and Demand on Ratings of Object Value," Journal of Personality and Social Psychology, 32, 906-914.

Veblen, Thorstein (1899), The Theory of the Leisure Class: An Economic Study of Institutions, New York: Macmillan.