The Generality of Typicality Effects on Preference and Comparison: an Exploratory Test

ABSTRACT - Consumer researchers have become interested in understanding typicality effects in product categories. But little research has explored factors that mediate the nature or existence of such effects. This paper focuses on two phenomena the positive relationship between typicality and preference and the tendency to use more typical category members as referents in comparisons-and explores the circumstances under which they might hold or not hold in product categories. Data from laboratory studies show that the relationships of interest may break down or reverse in product categories for which prestige, exclusiveness, or novelty are important purchase goals.

Citation:

James Ward and Barbara Loken (1988) ,"The Generality of Typicality Effects on Preference and Comparison: an Exploratory Test", in NA - Advances in Consumer Research Volume 15, eds. Micheal J. Houston, Provo, UT : Association for Consumer Research, Pages: 55-61.

Advances in Consumer Research Volume 15, 1988      Pages 55-61

THE GENERALITY OF TYPICALITY EFFECTS ON PREFERENCE AND COMPARISON: AN EXPLORATORY TEST

James Ward, Arizona State University

Barbara Loken, University of Minnesota

ABSTRACT -

Consumer researchers have become interested in understanding typicality effects in product categories. But little research has explored factors that mediate the nature or existence of such effects. This paper focuses on two phenomena the positive relationship between typicality and preference and the tendency to use more typical category members as referents in comparisons-and explores the circumstances under which they might hold or not hold in product categories. Data from laboratory studies show that the relationships of interest may break down or reverse in product categories for which prestige, exclusiveness, or novelty are important purchase goals.

INTRODUCTION

Consumer researchers have recently become interested in better understanding product categorization processes (Alba and Hutchinson, 1987). In particular, research has focused on understanding the determinants and consequences of differences in product "typicality" (Nedungadi and Hutchinson, 1985; Ward and Loken, 1986; Loken and Ward, 1987). This interest has been prompted by research which shows that the typicality of a category member is associated with a number of effects that are relevant to understanding consumer behavior. In particular, research shows (Mervis and Rosch, 1981) that more typical members of a category tend to be:

- named first in free recall of category instances;

- learned faster, classified more quickly, and classified more accurately;

- used as cognitive reference points in comparisons;

- perceived to possess greater amounts of valued attributes.

Studies of typicality effects in the consumer behavior literature have focused on demonstrating their existence in categories of products. Several studies have found positive relationships between typicality, recall, and preference across a variety of product categories (Nedungadi and Hutchinson, 1985; Loken and Ward, 1987). Research is now needed on the generality of these effects. In other words, what are the circumstances under which typicality effects might hold, not hold, or perhaps reverse in product categories?

This paper focuses on two phenomena--the positive relationship between typicality and preference and the tendency to use more typical category members as referents in comparisons--and discusses the types of product categories in which these effects should hold or break down. Both effects suggest that greater typicality confers an advantage to products when they are compared to others. In particular, they imply that typical products will tend to be better liked than less typical products and will tend to be used as "standards of comparison" in judgements.

The paper will discuss possible reasons for a positive relationship between typicality and preference, discuss some circumstances under which this relationship should break down or become negative, and test a set of two hypotheses resulting from these considerations.

TYPICALITY AND PREFERENCE

Studies in the marketing and psychology literature have demonstrated positive relationships between typicality and preference in a variety of product categories. Nedungadi and Hutchinson (1985) found a positive relationship between typicality and preference in four product categories--news magazines, business magazines, soft drinks, and beer. Loken and Ward (1987) also found a positive relationship between the typicality of shampoo brands, attitude toward the brands, and a measure of the degree to which the brands possessed attributes salient for purchasing category members. In the psychology literature, Barsalou (1985) found a positive relationship between the typicality of category members and the degree o which they possess "ideal" attributes in nine goal-derived categories (e.g., clothes to wear in the snow) and nine common taxonomic categories (e.g., vegetables, fruits, birds).

EXPLANATIONS FOR A POSITIVE RELATIONSHIP BETWEEN TYPICALITY AND PREFERENCE

The studies reviewed above raise two questions: Why are more typical products and services more preferred? Under what circumstances might this relationship break down or become negative? Several lines of research are relevant to these issues. The discussion below will focus on the first issue; a later section will focus on the second

Natural Selection

In categories of relatively inexpensive, widely affordable products, brands or types with preferred characteristics may become more typical by a process of natural selection. For example, a new entrant to a category might have preferable but atypical characteristics. As the entrant gains share, other brands will attempt to imitate its characteristics. Finally, the preferred entrant will become more typical of the category because it will share attributes with its imitators and will be more frequently encountered as an instance of its category. An exception to this process might be products that are initially preferred because of their prestige, exclusiveness, or novelty. As such products become more widely available and imitated, preference for them may tend to decline instead of rise (Snyder and Fromkin, 1979).

Goal-Oriented Categorization

Barsalou's work suggests that in goal-derived categories, a positive relationship should exist between typicality and preference. Barsalou (1985) shows that when people think of goal-derived categories they tend to judge exemplars that are more relevant to goal achievement as most typical of the category. Thus, if widely shared goals for buying a shampoo are cleaning and conditioning hair, shampoos that are perceived to perform well on these dimensions should not only be favorably evaluated but also perceived as typical. However, in product categories where prestige, exclusiveness, or novelty is a goal, the relationship between typicality and preference may not hold. This possibility will be discussed in a later section.

Familiarity

Zajonc (1968) and later researchers have demonstrated that greater familiarity with a stimulus leads to greater liking for the stimulus. This relationship could provide a partial explanation for the relationship between typicality and liking, but studies of the relationship between familiarity and typicality have produced conflicting results. Some studies have found a positive relationship between typicality and various measures of familiarity or frequency of encounter (Barsalou, 1985; Hampton and Gardiner, 1983; Malt and Smith, 1982; Ashcraft, 1978) and others have not (Rosch, Simpson, and Miller, 1976; Mervis, Catlin, and Rosch, 1976). Thus, it is unlikely that the influence of familiarity accounts completely for the typicality-preference relationship.

The reasons cited above for a positive relationship between typicality and preference suggest the following hypothesis:

H1: In categories of products for which prestige, exclusiveness, or distinctiveness are not important purchase goals, the relationship between prototypicality and preference should be positive.

This hypothesis is supported by past research findings (Nedungadi and Hutchinson, 1985; Ward and Loken, 1986; Loken and Ward, 1987) and thus is of interest principally in contrast to the next hypothesis to be presented.

REASONS WHY TYPICALITY AND PREFERENCE SHOULD NOT BE RELATED

Despite previous findings of a positive relationship between typicality and preference, at least three related perspectives suggest that typical products should sometimes be less preferred. These include literature on the value of scarcity, variety-seeking behavior. and attitude theory.

The Value Or Uniqueness

Consumers may value uniqueness itself, especially in more expensive, higher involvement products that are perceived as means of self-expression such as clothing and automobiles . Snyder and Fromkin's (1979) uniqueness theory asserts that people have a basic need to feel moderately dissimilar to others. Because of this need, people value rare and unusual products, ideas, and experiences. For example, a collector of jazz records may value an old, scratched recording of a forgotten artist because it is the only known copy, not because it is better or even more valuable than others. Uniqueness theory suggests that when consumers buy products important to their self-concept, they may value atypicality per se.

Variety Seeking/Innovativeness

Variety seeking theory also predicts that people will sometimes prefer the atypical to the typical. Researchers (McAlister and Pessemier, 1982; Raju, 1980) have proposed that consumers have a basic motivation to seek variety in their experiences. McAlister and Pessemier (1982) note that such arguments are based on Driver and Streufert's (1964) optimal stimulation level theory. According to this theory, as stimulation falls below an optimal level, people will seek out new, unusual, and exciting products, services, and experiences. But if the stimulation generated by these or other events is too high, people will seek less stimulation. For example, Raju (1980) found positive correlations between consumers' optimal stimulation level and their desire for unusual products and services.

Like variety seeking theory, the literature of innovativeness also predicts that certain consumers will prefer atypical products over currently accepted more typical products. For example, in his dissertation Szybillo (1973) demonstrated that fashion opinion leaders as compared to non-opinion leaders preferred pants suits perceived to be "scarce" over the same suits perceived to be "plentiful."

Attitude Theory

Fishbein and Ajzen's (1975) expectancy-value model of attitude theory does not by itself predict that consumers will prefer atypical products. But this prediction emerges if the theory is combined with insight into the structure of consumer product categories. The expectancy-value model of attitude formation suggests that consumers will have a better attitude (A object) toward products they perceive as more likely to have valued attributes. If a strong positive relationship between price and quality exists in a product category, and if better quality products tend to have smaller market shares because of their expense, then products that consumers perceive to have valued attributes may tend to be perceived as atypical of the product category.

The theoretical perspectives reviewed above suggest the following hypothesis:

H2: In categories of products for which prestige, exclusiveness, or distinctiveness are important purchase goals, the relationship between prototypicality and preference should be negative instead of positive.

Another possibility should be mentioned. If the category includes products that are atypical but negatively evaluated, typical and more positively evaluated, and atypical but very positively evaluated, a curvilinear instead of negative relationship between typicality and preference should hold.

TYPICALITY AND REFERENTIAL JUDGEMENTS

Another cognitive advantage of typical members of a category is their tendency to be used as reference points that less typical members are compared to. For example, people tend to describe less typical members of a category as "sort of," "essentially" or "almost" typical members (Rosch, 1975). This phenomena is interesting from a consumer behavior perspective because it suggests that consumers will tend to use more typical members of a category as standards of comparison for less typical members.

Rosch's (1975) study is the most widely cited support for the tendency to use more typical category members as reference points. Rosch found that people tend to use more prototypical colors as reference points for the classification of less prototypical colors. In one experiment, the participants were presented with sentence frames such as "___ is almost ___." Above each frame the participants saw two numbered color chips from a single color category. One chip was a variant of the color (e.g., a purple red) and the other was a color that was pretested to be prototypical of the category. The participants were instructed to fill-in the sentence frames with the numbers for each color chip in the most appropriate order. The results showed that participants had a strong tendency to compare less prototypical category members to more prototypical category members.

This study raises two issues for consumer researchers. The first is whether the tendency to use more prototypical category members as reference points generalizes to product categories. In categories of products for which prestige, exclusiveness, or novelty are not purchase goals, consumers might tend to use typical products as referents for reasons suggested by Rosch (1975). In such categories typical members share more attributes than other members, are closer to the central tendency of their category, and are therefore more useful than atypical instances in classifying stimuli as members or non-members of the category. Typical members might also be used as reference points because they are preferred over atypical members.

The above considerations suggest hypothesis 3:

H3: Participants will use more typical instead of less typical products as referents in linguistic hedges constructed about a category of products for which prestige, exclusiveness, or novelty is not an important purchase goal.

The second issue is whether the tendency to use typical category members as reference points generalizes to categories of more expensive, higher involvement products. In such product categories, marketers and consumers often speak of the most prestigious products as "standards of comparison" or "standards of excellence." These products may be used as referents because they offer the best performance on desired attributes and therefore form endpoints useful for judging other products. Referent products in more expensive, high involvement categories are usually not the most typical because they tend to have more or different attributes than other brands, tend to rate better on shared attributes, and tend to be less frequently encountered. These considerations lead to hypothesis 4:

H4: Participants will use less typical instead of more typical products as referents in linguistic hedges constructed about a category of products for which prestige, exclusiveness, or novelty is a purchase goal.

Experimental studies, reported next, were conducted to test hypotheses one through four. These studies were designed to demonstrate the existence of the predicted exceptions to typicality--preference relationships but not to differentiate between different explanations for their existence or to estimate their generality across product categories.

METHODOLOGY

Overview

Three laboratory studies were designed to address the above hypotheses. In study 1, prototypicality ratings were collected from 21 participants for brands of new automobiles, clothing stores, and brands of cola. In study 2, 29 participants provided ratings of their global attitudes toward the same set of stimuli. The attitude and prototypicality ratings were collected between groups, instead of within subjects, to reduce the possibility that subjects would guess the study's hypothesis and to increase the comparability of the findings with previous studies of typicality relationships. In study 3, 24 participants filled-in hedged sentences (e.g., "A ___ is almost a ___.") with names of items from the categories noted earlier.

Participants

The participants were undergraduate students in marketing classes at a sunbelt university. They took part in the experiment during class time but did not receive course credit for their participation.

Selection of Stimuli

The product categories selected for the experiment were colas, new automobiles, and stores that people from the metropolitan area could buy clothes at. The categories were selected by several criteria. One goal was to choose categories that students would have some knowledge of. Judgement suggested that most students at the university would have some knowledge of colas, automobiles, and clothing stores. Another goal was to select categories that varied in the extent that prestige, exclusiveness, or distinctiveness were purchase goals for college students. The cola category was chosen intuitively to represent a category of inexpensive, relatively low involvement products which consumers do -hot particularly value for their exclusivity. The categories of new automobiles and clothing stores were chosen intuitively to represent categories in which the prestige, exclusiveness, or distinctiveness of the brand or store are important goals for evaluation, purchase, or patronage. Prior studies support these judgements. For example, in a survey of the residents of a southern metropolitan area, King and King (1980) found that almost 80 percent of respondents under 30 perceived the automobile as a status symbol to all U.S. socioeconomic classes. Studies also suggest that status or class image have a significant influence on consumers' image of stores that sell clothing (Lindquist, 1975; Jacoby and Mazursky, 1984). To further assess students' perceptions of these product categories, three groups of ten students in a marketing class were asked to list either reasons for purchasing or not purchasing automobiles, reasons for purchasing or not purchasing colas, or reasons for patronizing or not patronizing clothing stores. The lists revealed that a majority of ten participants listed prestige, status, or image as reasons for buying automobiles and a majority of ten participants noted status, upscale merchandising, or quality of clothing as reasons for patronizing a clothing store. Participants did not note prestige, upscale image, or premium quality as reasons for buying colas.

As an aid to selection of the specific items presented within each category, a pre-test sample of students was asked to list members of the various categories. The stimuli (see Table 1) were chosen from the items they listed with the objectives of choosing items that would be recognizable, would vary in typicality, and in the case of the new automobile and store categories, would vary in prestige, exclusiveness, and distinctiveness.

Prototypicality Ratings

In study 1, 21 participants rated the prototypicality of brands of automobiles, stores that people in the metropolitan area could buy clothes at, and brands of cola The instructions and measure used were adapted from Rosch and Mervis (1975). Specifically, the participants rated the prototypicality of each category member on a 0-10 point scale with endpoints of extremely poor example (0) and extremely good example (10).

To reduce any tendency of the participants to provide attitude instead of typicality ratings on the "extremely good"C"extremely poor" scales, the experimenter noted that a good example meant a more typical example, not a more preferred or favorite example, and that a poor example meant a less typical example, not a less preferred or least favorite example. The experimenter also noted that a more typical example does not necessarily occur more frequently and a less typical example does not necessarily occur less frequently. To help compensate for possible differences in the participants' knowledge about automobiles, small (1" by 2") black and white pictures of each automobile were reproduced above its name and the rating scale.

TABLE 1

PROTOTYPICALITY AND ATTITUDE SCORES BY CATEGORY

Participants saw only the names of the stores and colas since pretest data indicated that almost all students were knowledgeable about these categories.

Automobiles were rated first, followed by stores and then colas. The instructions were briefly repeated before participants rated the latter two categories. Within each category, the category members were presented in a random order and its reverse with the two orders counter-balanced across participants.

Attitude Ratings

In study 2, 29 participants were asked to rate their attitudes toward the previously specified sets of automobile brands, stores that people in the metropolitan area could buy clothes at, and cola brands. The instructions asked participants to "rate your attitude toward each (category member) on the three scales that appear below its name." The measures were three 0-10 point evaluative scales with endpoints high-low quality, good-bad, and satisfactory-unsatisfactory. The participants saw black and white pictures of the automobiles reproduced on the questionnaire (the same pictures used in study 1) but no pictures of the other stimuli.

As before, the categories were presented in the same order (automobiles, stores, and colas) to ensure that the instructions could be briefly repeated before each category. Within each category, the stimuli were presented in a random order and its reverse, counterbalanced across participants.

Linguistic Hedges

In study 3, the participants were presented with sentence frames that formed linguistic hedges. The hedge terms used, "is almost" and "is sort of," were borrowed from Rosch (1975). The resulting sentence frames were: "___ is almost ___" and "___ is sort of ___." Two items from one of the previously studied categories appeared above each hedge. The participants were instructed to "please fill-in the two blanks with the two given words (one per blank) in whatever order seems most appropriate to you."

In total, the participants were presented with six sentence frames (three "almost" and three "sort of") and six pairs of items for each of the three categories studied (see Table 2). The items were chosen by reference to previously collected prototypicality data so one item was more prototypical of the category than the other. The frames were presented in a random order, and its reverse, counter-balanced across participants.

RESULTS

Hypotheses One and Two

Mean prototypicality and mean summed attitude ratings for the categories of automobiles, clothing stores, and colas are shown in Table 1. Pearson correlations between the prototypicality and attitude scores of items within each category are shown at the bottom of the table.

Hypothesis one predicted that the relationship between prototypicality and global attitude should be positive among the brands in a category of inexpensive consumer products. The results confirm this hypothesis. The Pearson correlation between prototypicality and attitude for colas is .83, which is significantly different from zero, t (16) = 5.95, p < .05.

Hypothesis two predicted that in categories of products for which prestige, exclusiveness, or distinctiveness are more important purchase goals, the relationship between prototypicality and preference should be negative. The results for the automobile and clothing store categories confirm the hypothesis. The correlation between prototypicality and attitude in the automobile category was -.74, different from zero at p < .05, t (16) = 4.40. The same correlation for clothing stores was -.49, also different from zero at p < .05, t (16) = 2.25. The correlations between prototypicality and attitude in both the automobile and clothing store categories were also significantly different from the correlation in the cola category. The lower negative correlation in the clothing store category appears to result from at least two factors. First, as predicted, certain stores were rated as relatively typical places to buy clothes and were also favorably evaluated. Second, certain stores such as K-Mart were judged relatively atypical places to buy clothes, and were unfavorably evaluated. As noted earlier, when atypical but unfavorably evaluated exemplars exist in a category along with atypical but favorably evaluated exemplars the linear correlation between typicality and preference may be reduced.

The possibility of a curvilinear relationship between the prototypicality and attitude measures in the clothing store and automobile categories was tested by polynomial regressions. In both categories, the contribution to R2 of predicting prototypicality ratings with a third degree polynomial model for the attitude measure was compared to the R2 for a first degree model. The third degree polynomial model resulted in no significant (alpha = .05) contribution to R2 for the automobile category (F(1,14) = 1.45) at alpha = .05. The contribution to R2 for the clothing store category was not significant at .05 but was significant at an alpha of .10 tF(1,14) = 3.32).

Hypotheses Three and Four

Table 2 shows how often the two category members presented with each sentence frame were used in the referent and non-referent positions in the sentence. If the choice between the positions was random, the chance that the two category members would be used in the referent position would be .50 for each.

The significance of departures from this expectation in the results for each sentence frame was tested by chi-square. Hypothesis three predicted that participants would use more typical instead of less typical category members as referents in linguistic hedges about colas. The results of the chi-square tests, also shown in Table 2, show that the more prototypical category member was significantly more likely to be used as a referent in 5 of 6 sentence frames about colas.

Hypothesis four predicted that participants would use less typical products as referents in linguistic hedges .*out automobiles and clothing stores. The results in Table 2 show that less typical category members were significantly more likely to be used as referents than more typical members in 6 of 6 frames about automobiles and 5 of 6 frames about clothing stores.

DISCUSSION

The results suggest that typicality effects on preference and comparison are not generalizable across all product categories. In particular, if consumers regard prestige, exclusiveness, or novelty as important goals for purchasing the products in a category, they appear to perceive a negative relationship between typicality and preference, and tend to use less typical but more preferred products as cognitive reference points. Although these results appear intuitive, they are the first to demonstrate limits to the positive relationship between typicality and preference and the tendency to use more typical products as referents.

These findings have important implications for both marketing practitioners and consumer researchers. Marketing practitioners might have assumed from previous research that greater typicality would generally confer a variety of cognitive advantages on their product. These results show that this assumption needs to be carefully evaluated before it is applied to specific product categories.

To consumer researchers, the findings of the study suggest a need to focus research on categorization processes to investigating factors that moderate typicality effects and their limits of generalizability. Such efforts will help build a theory of categorization tailored to consumer decision processes.

The results should be regarded as exploratory, to be confirmed by later studies, for several reasons. First, only three product categories were studied. Second, the categories of automobiles and clothing stores were included in the study because consumers, particularly college students, are usually thought to be especially conscious of how cars and clothes influence their self-image and the image they project to others. The negative relationship between typicality and attitude might be weaker for other categories of more expensive, higher involvement goods and other populations of consumers. Finally, the participants rated their overall attitude toward the products--not their attitude toward the act of purchase. If the participants had been asked to rate their attitude toward purchase, their scores might have been substantially more negative toward, e.g., a Ferrari, and more positive toward, e.g., a Toyota. Nevertheless, the more expensive, less typical car would likely remain their ideal.

TABLE 2

RESULTS FOR HEDGED SENTENCES

Future Research

The relationships between typicality, preference, and comparison should be explored in other types of categories. For example, do art collectors perceive a positive relationship between a work's prototypicality as a representative of a particular genre and their preference for the piece? Whitman and Slater (1979) suggest that such a tendency may exist. But creativity, atypicality, and uniqueness are also valued in art. Turning to non-product categories, typicality and affect might have interesting relationships in categories comprised of undesirable items like complaints about service in a restaurant. Perhaps more typical complaints tend to be evaluated less negatively than atypical complaints or vice versa.

As these examples suggest, the possibilities for further research are diverse. Future studies should explore the relationships between typicality and preference across a variety of categories, populations, and mediating factors.

REFERENCES

Alba, J. and W. Hutchinson (1987), "Dimensions of Consumer Expertise," Journal of Consumer Research, 13, 411-455.

Ashcraft, M. H. (1978), "Property Norms for Typical and Atypical Items from 17 Categories: A Description and Discussion," Memory and Cognition, 6, 227-232.

Barsalou, L. W. (1985), "Ideals, Central Tendency, and Frequency of Instantiation as Determinants of Graded Structure," Journal of Experimental Psychology: Learning, Memory, and Cognition, 11 (October), 627-654.

Driver, M. J. and S. Streufert (1964), "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, Krannert Graduate School of Management, Purdue University, West Lafayette, IN.

Fishbein, M. and I. Ajzen (1975), Belief, Attitude, Intention, and Behavior: An Introduction to Theory and Research, Reading, MA: Addison-Wesley.

Hampton, J. A. and M. M. Gardiner (1983), "Measures of Internal Category Structure: A Correlational Analysis of Normative Data," British Journal of Psychology, 74, 491-516.

Jacoby, J. and D. Mazursky (1984), "Linking Brand and Retailer Images Do the Potential Risks Outweigh the Potential Benefits," Journal of Retailing, 60 (Summer), 105-122.

King, A. and J. King (1980), 'The Changing Impact of Market Conditions on Consumer Perceptions of the Automobile as a Status Symbol," in Developments in Marketing Science, 11, ed. VenKataKrishna Bellur, 14-18.

Lindquist, J. (1975), "Meaning of Image," Journal of Retailing, 50 (Winter), 29-38.

Loken, B. and J. Ward (1987), "Measures of the Attribute Structure Underlying Product Typicality," in Advances in Consumer Research, 14, ed. Paul Anderson, Association for Consumer Research, 22-26.

Malt, B. C., and E. E. Smith (1982), "The Role of Familiarity in Determining Typicality," Memory and Cognition, 10, 69-75.

McAlister, L. and E. Pessemier (1982), "Variety Seeking Behavior: An Interdisciplinary Review," Journal of Consumer Research, 9, 311-322.

Mervis, C. and E. Rosch (1981), "Categorization of Natural Objects," in Annual Review of Psychology, 32, eds. Rozenweigh and L. W. Porter, 89-115.

Mervis, C. B., J. Catlin, and E. Rosch (1976), "Relationships Among Goodness-of-Example, Category Norms, and Word Frequency," Bulletin of the Psychonomic Society, 7, 283-294.

Nedungadi, P. and J. Hutchinson (1985), "The Prototypicality of Brands: Relationships with Brand Awareness, Preference, and Usage," in Advances in Consumer Research, 12, eds. Elizabeth Hirschman and Morris Holbrook, Association for Consumer Research, 498-503.

Raju, P. S. (1980), "Optimum Stimulation Level: Its Relationship to Personality, Demographics, and Exploratory Behavior," Journal of Consumer Research, 7 (December), 272-202.

Rosch, E. (1975), "Cognitive Reference Points," Cognitive Psychology, 7, 532-547.

Rosch, E. H. and C. B. Mervis (1975), "Family Resemblances: Studies in the Internal Structure of Categories," Cognitive Psychology, 7, 573-605.

Rosch, E. H., C. Simpson and R. S. Miller (1976), "Structural Bases of Typicality Effects," Journal of Experimental Psychology: Human Perception and Performance, 2, 491-502.

Snyder, C. R. and H. L. Fromkin (1979), Uniqueness: The Human Pursuit of Difference, New York, NY: Plenum Press.

Szybillo, G. J., The Effects of Price and Scarcity on the Valuation of Fashions by Fashion Opinion Leaders and Non-opinion Leaders, Doctoral Dissertation, Purdue University, 1973.

Ward, J. and B. Loken (1986), "The Quintessential Snack Food: Measurement of Product Prototypes," in Advances in Consumer Research, 13, ed. Richard Lutz, Association for Consumer Research, 126-131.

Whitfield, T. W. A. and P. C. Slater (1979), "The Effects of Categorization and Prototypicality on Aesthetic Choice in a Furniture Selection Task," British Journal of Psychology, 70, 65-75.

Zajonc, R. B. (1968), "Attitudinal Effects of Mere Exposure," Journal of Personality and Social Psychology Monograph Supplement, 9, 1-27.

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