Do Market Mavens Categorize Brands Differently?

Michael T. Elliott, University of Missouri-St. Louis
Anne E. Warfield, University of Missouri-St. Louis
ABSTRACT - A survey of 172 consumers studied the impact of market mavenism on brand categorization processes across several diverse product categories. Extending the Feick and Price (1987) seminal study, this research examines how market mavens differ from other consumers in categorizing seven brands sets based on the Brisoux and Laroche (1980) framework. Findings indicate that market mavens have (1) larger salient (unaided recall) sets, (2) larger aware (aided recall) sets, (3) larger trial sets, and (4) larger hold (undecided) sets than other consumer groups. In addition, the results were found to be generalizable over a wide range of products exhibiting varying levels of product involvement.
[ to cite ]:
Michael T. Elliott and Anne E. Warfield (1993) ,"Do Market Mavens Categorize Brands Differently?", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 202-208.

Advances in Consumer Research Volume 20, 1993      Pages 202-208

DO MARKET MAVENS CATEGORIZE BRANDS DIFFERENTLY?

Michael T. Elliott, University of Missouri-St. Louis

Anne E. Warfield, University of Missouri-St. Louis

ABSTRACT -

A survey of 172 consumers studied the impact of market mavenism on brand categorization processes across several diverse product categories. Extending the Feick and Price (1987) seminal study, this research examines how market mavens differ from other consumers in categorizing seven brands sets based on the Brisoux and Laroche (1980) framework. Findings indicate that market mavens have (1) larger salient (unaided recall) sets, (2) larger aware (aided recall) sets, (3) larger trial sets, and (4) larger hold (undecided) sets than other consumer groups. In addition, the results were found to be generalizable over a wide range of products exhibiting varying levels of product involvement.

INTRODUCTION

The market maven, described by Feick and Price (1987) as an influencer with general marketplace expertise, represents a truly attractive target segment to advertisers and other marketers. The authors found that market mavens display higher levels of marketplace involvement; are engaged in extensive search activities (e.g., reading Consumer Reports); and demonstrate greater participation in market activities such as couponing, store browsing and reading advertisements. Furthermore, mavens appear to be an accessible market segment in that they tend to read magazines and view television more than other consumers.

Of particular note, Feick and Price reported that market mavens display an earlier awareness of new products and brands than other consumers. Since the early 1980s, there has been a proliferation of new brands within certain product categories. Food and drug manufacturers, for instance, introduced an average of 11,000 new brands during this period (Fortune, 1985). As a result, consumer researchers have focused more attention on how individuals simplify their decision making heuristics through the process of brand categorization (Brisoux and Laroche 1980; Narayana and Marking 1975). This approach delineates brand categorization sets (i.e., awareness, salient, foggy, evoked, trial, reject, and hold) as they relate to the stages of the consumer decision making process (i.e., awareness, processing, consideration and preference).

The brand categorization paradigm offers an intriguing perspective of consumer involvement. Any given brand that the consumer knows about could move to his or her evoked set (i.e., it could become a brand that he or she considers acceptable for purchase), hold set (i.e., he or she would not have an opinion as to whether it should be accepted or rejected) or reject set (i.e., he or she would consider it unacceptable for consideration). Considering the proliferation of new product introductions and line extensions in recent years, the truly interesting question raised (and tentatively answered) by this research is: Do consumers with general marketplace expertise (i.e., market mavens) categorize their known brands differently than other consumers across a wide range of product types? If so, in what brand set categories do these differences exist?

One predicted outcome of market mavenism would be the demonstration of a greater awareness of brands across diverse product categories. Similarly, we would expect informed consumers, like market mavens, to exhibit more trial purchase behaviors. During the consideration stage (i.e., evoked, reject, and hold sets), however, we do not have an intuitively superior expectation. If we were to apply Social Judgement Theory (Sherif and Hovland 1961) without modification to answer this question, we would expect that higher levels of general marketplace involvement would cause the evoked set (analogous to latitudes of acceptance) to decrease, the hold set (analogous to latitudes of neutrality) to remain constant, and the reject set (analogous to latitudes of rejection) to increase. However, past research has not supported this view. What Brisoux and Cheron (1990) found was that, with increased product involvement, the sizes of both the evoked set and the reject set remained relatively constant while the size of the awareness and hold sets dramatically increased. This suggests that while involved consumers possess higher "top of mind" awareness, their processing of brand evaluation information is not superior to others.

In order to extend the literature on the market maven concept and provide additional empirical support for its existence, this study investigates how market mavens differ in their brand categorization processes from nonmavens. Specifically, we focus on three important questions:

1) Do market mavens categorize brands differently relative to the brand categorization process (measured in terms of brand set size)?

2) Can any differences found in brand set size among market maven groups be generalized across product categories?

3) Do differences in brand set size among market maven groups remain after taking product-specific involvement into account?

MARKET MAVENISM

The introduction of the market maven concept to marketing has precipitated a great deal of research interest (Higie and Feick 1987; Price Feick, and Higie 1987; Slama and Williams 1990). Drawn from the consumer behavior, communications and social science literatures, the market maven concept refers to "individuals who have information about many kinds of products, places to shop, and other facets of markets, and initiate discussion with consumers and respond to requests from consumers for market information" (Feick and Price, 1987, p. 85). In validating their market maven scale, Feick and Price were able to show that the tendency to be a maven, the tendency to be an opinion leader and the tendency to be early purchasers represent three related but distinct concepts.

Arguably, the most salient characteristic of market mavens is their possession of a wide variety of market information. That is, these consumers may possess information regarding places to shop, be aware of innovative products and more brands, or have in-depth knowledge of specific brand performance criteria through activities such as reading Consumer Reports (Kotler and Zaltman 1976). One other indicator of this generalized marketplace knowledge, as reported by Feick and Price (1987), is the market maven's early awareness of a large number of brands across a broad range of product categories.

These preliminary findings raise two interesting issues. First, the effects of market mavenism may well extend beyond the awareness stage of the consumer decision process. We hypothesize that the brand set sizes during the awareness, processing, consideration and preference stages differ based on degree of market mavenism. Since mavens are posited to have higher levels of marketplace knowledge and expertise, it seems plausible that they might be aware of, try, and perhaps reject more brands while at the same time consider, be unaware, and be unsure (i.e., hold and foggy) about fewer brands than other consumers. Variations in brand set sizes, if detected, would provide support for the usefulness of the brand categorization model in understanding consumer involvement and information processing activities.

Second, the theoretical underpinnings of the market maven concept depend, to a large extent, on its generalizability across a wide range of product types. Unlike the opinion leader, who exhibits mostly product-specific knowledge, the market maven's expertise is purported to exist across a wider range of product categories. Slama and Williams (1990) provided some evidence of the generalizability of the information provision tendency of the market maven by investigating twenty categories of products and services. However, the question of whether market mavenism is truly a global characteristic or whether it is simply an artifact of a more generalized form of opinion leadership manifested is still largely unresolved.

BRAND CATEGORIZATION

The manner in which consumers handle information regarding the multitude of brands that are available in many product categories has been viewed as the "psychology of simplification" in choice behavior (Narayana and Markin 1975, p. 1). Several models representing this cognitive simplification process have been proposed in the marketing literature. Topics such as category formation processes (e.g., Cohen and Basu 1987), measurement of consideration set or evoked set (Gruca 1989), as well as elements of the evoked set (Brisoux and Laroche 1980), have been addressed.

Brisoux and Laroche's (1980) brand categorization framework is an expansion of earlier models by Narayana and Markin (1975) and Howard (1977). It suggests four stages in the consumer decision process: awareness, processing, consideration and preference (see Figure). This model successfully links a consumer decision process perspective to brand set formation research.

According to their model, a consumer is either aware or unaware of all of the available brands in a given product class. The brands for which the consumer is aware can be distinguished as either processed or foggy brands. Brands presumed to be in the consumer's foggy set have no salient attributes by which the consumer attains comprehension of the brands. This may occur because of a lack of exposure to the brands' advertisements or a lack of experience with the product category. As an example, most consumers would likely have large foggy sets for a product such as scuba diving equipment because of a lack of experience with the product class.

Those brands possessing the salient attributes which allow for processing are further subdivided into three sets: evoked, hold and reject sets. The evoked (consideration) set contains those brands which the consumer has a positive attitude toward and actually considers as purchase alternatives (Howard and Sheth, 1969; Brisoux and Laroche 1980). The existence of evoked sets has been confirmed in the choice process for nondurables (grocery products - Campbell, 1969), consumer durables (cars - Gronhaug, 1974), and services (retailing - Spriggle and Sewall, 1987).

The trial set is viewed as a subset of the evoked set and consists of those brands that the consumer has actually purchased and consumed. Though not explicitly referred to in the Brisoux and Laroche model, it represents brands within the consideration stage possessing a higher probability of selection than other evoked brands because of prior usage experience. Trial sets also represent an indicator of the degree of brand switching and experiential behavior by the consumer.

The hold set consists of those brands that the consumer has perceived no advantage in purchasing. Brands exist in a consumer's hold set for various reasons. These reasons include: (1) viewing the brand as not being adequate for the buying motives or purchase situation, (2) possessing no reference group influences or (3) perceiving brands as being too high in price in relation to quality. As a result, the consumer generally has either a positive or neutral feelings toward these brands. Lastly, the reject set is defined as those brands that the consumer has a decidedly negative attitude toward and has completely rejected from consideration as a purchase alternative (Brisoux and Laroche 1980).

RESEARCH OBJECTIVES AND HYPOTHESES

Previous studies have demonstrated the existence of market mavens and suggest that they possess and are active providers of information about new products and brands across a wide variety of product categories (Feick and Price 1987; Slama and Williams 1990; Lichtenstein and Burton 1990). Despite this, the market maven's general awareness of brands within a wide range of product categories has not been rigorously investigated. An equally important issue concerns the disposition of those brands that the market maven has at least partially processed.

To address these empirical questions, the present study investigates differences in brand set sizes - awareness, evoked, reject, foggy, salient, trial, and hold - relative to a consumer's level of market mavenism. Several diverse product categories are used to assess the generalizability of the findings. Because of its impact on brand awareness and knowledge, the impact of product-specific involvement (Zaichkowsky 1985; Brisoux and Laroche 1990) on brand set size is adjusted for in determining the specific effects of the market maven trait. Put more generally, the purpose of this research is to assess the validity and generalizability of the purported market maven trait of general marketplace knowledge and expertise vis-a-vis the brand categorization process.

Six hypotheses pertaining to the relationship between market mavenism and brand set size are proposed. Because the conceptualization of product involvement is closely aligned with that of market mavenism, it is utilized as a covariate to better assess the impact of market mavenism separately.

H1 Consumers with a higher level of market mavenism will exhibit smaller evoked sets.

H2 Consumers with a higher level of market mavenism will exhibit larger reject sets.

Previous research has been inconclusive regarding the relationship between marketplace involvement and the evoked and reject set sizes. For instance, Jarvis and Wilcox (1973) found that individuals with greater ego-involvement in a product class (i.e., car enthusiast) have smaller evoked sets. Conversely, Brisoux and Cheron (1990) in their study of perfume purchasers, found no relationship between product involvement and either evoked or reject sets.

H3 Consumers with a higher level of market mavenism will exhibit smaller foggy sets.

This hypothesis assumes that the maven's involvement with the marketplace results in a larger number of processed brands (i.e., the converse of foggy brands). Due to the unprocessed nature of the foggy set, a larger foggy set should be associated with a lower level of market mavenism.

H4 Consumers with a higher level of market mavenism will exhibit larger salient (unaided recall) sets and larger awareness (unaided recall) sets.

H5 Consumers with a higher level of market mavenism will exhibit larger hold sets.

H6 Consumers with a higher level of market mavenism will have larger trial sets.

FIGURE

THE BRAND CATEGORIZATION PROCESS

Hypothesis four is proposed based on the positive relationship between brand awareness and market mavenism shown by Feick and Price (1987). Similarly, the salient set (measured as unaided recall) should be larger among market mavens than other individuals. In addition, Brisoux and Laroche (1990) reported that awareness sets, trial sets, and hold sets were larger for highly involved perfume purchasers.

METHODOLOGY

Sample

A simple random sample of 1000 residents was drawn from a Midwest metropolitan statistical area (MSA) utilizing a recent voter registration list. A five-page questionnaire mailed to respondents contained an accompanying cover letter and instructions. To better ensure the validity of the brand set measures, special emphasis was place on the importance of completing each section in proper sequence. The first section of the questionnaire asked respondents to provide an unaided recall of brands within four product categories (athletic shoes, potato chips, toothpaste, and stereo receivers). Section two assessed product involvement for each product category by using a 5-item modified version of the Personal Involvement Inventory developed by Zaichkowsky (1985). In the third section, the brand set sizes (evoked, reject, trial, salient, foggy, hold, and awareness sets) for the four selected product categories were obtained by providing a current list of available brands. Respondents were asked to indicate which brands within a specified product category belonged to each brand set. In the last section the market maven measure (Feick and Price 1987) and general demographic information such as gender, age, income, education, and household size were gathered.

Of the 1000 questionnaires mailed, 182 were returned, resulting in a response rate of 18.2 percent. After removing 10 unusable responses, 172 cases were retained for analysis. The sample was primarily female (76%) and white (85%). In addition, the median age category was 40-49, the median education level was some college, and the median household income was $40,000-$55,000. Nonresponse error was assessed by comparing early respondents (i.e., first one-third) with late respondents (i.e., last one-third). Chi-square tests revealed that the two groups do not differ on any of the demographic variables or market mavenism (p > .05).

Product Selection

One limitation of previous brand categorization studies is that only a limited range of products have been investigated (i.e., only one or two categories) and justification for products chosen is usually not provided. The product categories selected for the present study include athletic footwear, toothpaste, stereo receivers, and potato chips.

These products were chosen because they: (1) represent both durable (stereo receivers, athletic shoes) and nondurable (toothpaste, potato chips) goods; (2) tend to be gender-neutral; (3) represent limited to extensive decision making processes (as opposed to routinized); and (4) represent a range of low to high involvement product categories. A pretest indicated a Product Involvement Inventory (see Zaichkowsky 1985) score ranging from 64 (potato chips) to 112 (toothpaste), indicative of low and high involvement products, respectively. Though an investigation of an even larger inventory of product categories would be desirable, questionnaire length was considered a methodological constraint.

Definition of Variables

The operational concepts of the brand sets (dependent variables) and of the market maven (independent variable) are defined here. The conceptualization and operationalization of the six brand sets shown in the Figure were borrowed from the previous work of Brisoux and Laroche (1980).

Salient Set: Consists of those brands of the product class that the respondent can recall without the aid of a list or other type of prompt.

Awareness Set: Expanding on the salient set, it consists of those brands that the respondent can identify from a list of brands.

Trial Set: Consists of those brands that the respondent has already bought and used for him or herself (selected from brands listed in the awareness set).

Evoked Set: Consists of those brands that the respondent would consider as a purchase alternative at that point in time (selected from brands listed in awareness set).

Reject Set: Consists of those brands that the respondent deems unacceptable for purchase at that point in time (selected from brands in the awareness set).

Foggy Set: Consists of those brands for which the respondent has not formed an opinion and will not consider buying at that point in time (selected from brands listed in the awareness set).

Hold Set: Of the remaining brands from the awareness set, those brands for which the respondent has formed an opinion but will neither consider nor reject (selected from brands listed in the awareness set).

The market maven scale, a six-item, seven-point likert scale, measures a person's tendency to be a general provider of many types of market information to others on an informal basis (Feick and Price 1987). Example items include "People ask me for information about products, places to shop, or sales," and "My friends think of me as a good source of information when it comes to new products or sales" (7=strongly agree, 1=strongly disagree). The market maven scale was shown to be reliable in this study (alpha =.89).

The definition of product involvement, taken from Zaichkowsky (1985), states: "A person's perceived relevance of the object based on inherent needs, values and interests". This definition contains the general viewpoints of several researchers (e.g., Krugman 1967; Rothchild 1984). The operational measurement of product involvement is derived from the Personal Involvement Inventory (PII) scale developed by Zaichkowsky (1985). A modified 5-item, seven point semantic differential scale (important-unimportant, of no concern for me-of concern to me, very meaningful to me-means nothing to me, interesting-not interesting, significant-insignificant) was selected after studies by Nowak (1986) and Nowak and Salmon (1987) found that shorter versions of the scale were reliable. The coefficient alpha of the PII scale in this study ranged from .83 to .92.

Data Analysis

Following the practice of several earlier studies (i.e., Feick and Price 1987; Lichtenstein and Burton 1990; Slama and Williams 1990), the respondents was divided into three categories based on their scores on the market maven scale. The top scoring one-third of respondents was designated as market mavens or the high scoring group, the next two-thirds were the medium and low scoring groups.

A series of hierarchical ANOVAs were used to test the effect of market mavenism on specific brand set sizes as specified in hypotheses one through six. This approach allows for the testing of the effect of market mavenism alone before removing the effects of the covariate, product involvement. Subsequently, multiple comparisons were examined using the Scheffe procedure, which is appropriate when samples are of different sizes and allows testing for an overall significance level of group differences.

RESULTS

Table 1 shows the average brand set sizes for the four selected product categories. The brand set size proportions appear to be both similar across product categories and are consistent with past studies. For instance, the evoked set/awareness set proportion (ranging from .23 to .46) is comparable to that reported by Crowley and Williams (1991). Also, the current study found the average evoked set size of 3.72 to be similar to the average salient set size of 4.08. This is consistent with research by Brisoux and Cheron (1986) who reported similar evoked and salient set sizes (4.70 and 4.57, respectively) for the various brands studied. The average evoked set size (3.72) in this study is also comparable to the evoked set sizes of laundry detergent and margarine (3.67 and 3.78, respectively) reported by Reilly and Parkinson (1985). Lastly, the reject set size of 1.89 is similar to the reject set size (1.98) found with Brisoux and Cheron's study of perfume purchasers.

As shown in Table 2, three groups were formed from the market maven scale using the 33rd and 67th percentiles. The low maven group included scores of 26 or below, the medium level group scored from 27 to 32 and the high level group scored above 32. Out of the 172 valid cases, the average market maven scale score was 28.36 (standard deviation = 7.19, range = 8-42). The market maven scale parameters reported here compare favorably with those originally cited by the Feick and Price (1987) study (i.e., mean = 25.6; standard deviation = 8.5; range = 6-42).

To provide evidence on the generalizability of the market maven relationships across the four product categories, a series of two-way ANOVAs were conducted in which the dependent variables were the seven brand sets and the independent variables were the four product categories and the three levels of market mavenism (low, medium, high). Of the seven ANOVAs conducted, none resulted in significant product type X market maven interaction. These results indicate that market mavenism does not have a differential effect on the brand set size for different product types. Based on these results, within-subject responses on brand set size were aggregated across product categories.

Hierarchical analysis of variance was utilized to test the equivalence of mean brand set sizes for the low, medium and high scoring market maven groups. The data shown in Table 2 displays the univariate F tests for the seven brand set sizes. The first column of F values presents the F values for a single factor design testing the effects of market mavenism alone. The second column, FPI, presents conditional F values on the effect of market mavenism after removing or adjusting for the effects of product involvement. Significant differences were observed for the salient (H4), awareness (H4), hold (H5), and trial sets (H6), thereby confirming these hypotheses. After adjusting for product involvement, however, no differences were detected in the hold set (suggesting only partial support for hypothesis five). Conversely, results for the evoked set, reject sets, and foggy sets (i.e., H1-H3) indicated no difference in market maven group means. Finally, the Scheffe post-hoc contrast test, robust to non-normality and to unequal sample sizes, was conducted. Where overall differences existed, the Scheffe test indicated that the low scoring group consistently exhibited smaller brand set sizes than the medium and high scoring groups. These findings lend support to the hypothesized directionality of brand set size with respect to market maven groups.

TABLE 1

MEAN (STANDARD DEVIATION) OF BRAND SET SIZE BY PRODUCT CATEGORY

TABLE 2

ANOVA RESULTS: EFFECTS OF MARKET MAVENISM ON BRAND SET SIZE

DISCUSSION

This research investigated how "market mavens" differ from other consumers (nonmavens) in categorizing the seven brand sets of the Brisoux and Laroche (1980) framework and whether these results can be generalized over diverse product categories. The findings indicate that market mavens indeed have larger salient, awareness, trial, and hold sets. The market maven processes more brands across a range of products as evidenced by his larger salient and awareness sets. Therefore, the market maven appears to have better "top of mind" awareness of more brands across different product categories. These findings support the initial Feick and Price (1987) conceptualization of the market maven. In addition, the results demonstrate that mavens have larger trial sets, indicative of their experiential buying behavior and general interest in gather marketplace information. Contrary to expectations, differences were not detected in the foggy, evoked, and reject sets.

However, an increase in overall brand awareness does not result in more acceptable (evoked) brands or unacceptable (reject) brands but in more hold brands (note that this increases the probability that any processed brand will end up in the hold set). In attempting to explain this contradiction, perhaps the maven views many product categories as being more complex and thus sees the process of categorizing any given brand as acceptable or unacceptable as more difficult. Perhaps he or she does not process as much information about any given brand and thus, though he or she is processing more brands, he or she is not processing them as deeply. Or perhaps the market maven faces information overload and thus cannot carry out the processing needed to judge brands as acceptable or unacceptable. As Munch (1990, p. 145) offers: "The limitations of our processing capabilities may cause the involved consumer to be no better informed than the uninvolved consumer." This reasoning may provide some insight into the divergent findings of the hold set as opposed that of the evoked and reject sets.

In general, what do these results suggest about the attractiveness of the market maven segment? As evidenced by their higher level of brand awareness and propensity to try more brands, market mavens appear to mirror the profile of opinion leaders and early purchasers. However, unlike the opinion leader who may serve as an influencer primarily for high involvement products, the market mavens may be an ideal target for diffusing information on low involvement products (e.g., razor blades, laundry detergent) where the communication generally concentrates on a wide array of related goods (such as in the case of supermarket advertising).

In spite of these observations, several concerns remain about targeting the market maven. First, since mavens tend to have larger hold sets and therefore are "undecided" about a greater proportion of brands than other consumers, their marketplace expertise may be either limited or narrowly defined. It is plausible that information provided by these consumers is related more to marketing mix factors such as price changes, product availability and promotional deals than with brand evaluation information. Second, the market maven, while appearing to be an active information gatherer may be a passive diffuser of new product information if he or she has not adopted a particular brand.

What are the strategic marketing implications of the market maven concept raised in this article? Clearly, a major challenge for brand managers in utilizing the word-of-mouth communication potential of the market maven in moving processed brands from less favorable brand sets (i.e., hold, foggy, reject) to more desirable categories (i.e., trial, evoked). Past research has indicated that market mavens are especially attentive to magazine, direct mail and local classified newspaper advertising (Price and Feick 1987; Higie, Feick, and Price 1987). Advertisers and other marketers might well be advised to use more informative and comparative promotional messages when using these media. Effective comparative advertising can be achieved, for example, by linking brands likely to be in consumer's evoked set (i.e., market leaders) with the advertised brand, thereby facilitating that brand's movement from the hold set.

LIMITATIONS AND FUTURE RESEARCH

It is important to note the limitations of this research in interpreting the findings. First, the sample was somewhat atypical of the sampled MSA and the U. S. population as a whole in that it was primarily female (76%), affluent (i.e., median income of $40-50,000), and older (median age of 40-49). A national sample with a demographic profile more representative of the U.S. population would have been desirable. A second concern has to do with the reliability and validity of the self-report measures of brand set size. That is, the survey method provides little control over the sequence of responses or the uniformity of time spend by each respondent. Third, it would have been desirable to investigate an even larger inventory of product or service categories. Studies of professional (e.g., legal, medical) and generic services (banks, dry cleaners) have shown that consumers exhibit very distinct search behaviors when selecting and purchasing services (e.g., Zeithmal 1981). We suspect that these differences would affect the brand (or service provider) categorization process as well.

Though this research contributes to the notion of the market maven as a knowledgeable marketplace influencer, several unanswered questions remain concerning the brand categorization process. For example:

- During the brand categorization process, do mavens require more or less time in moving brands to the evoked and trial sets?

- Can the brand categorization process be applied to services and other intangible market offerings?

- Can distinct brand categorization patterns be detected between market mavens, opinion leaders, and early purchasers?

In sum, this study of how informed consumers (market mavens) categorize brands differently both enriches our understanding of information processing and presents possible strategic opportunities to marketers. At a time when process oriented models of consumer behavior are sought, our research contributes to this end.

REFERENCES

Brisoux, Jacques E. and Michel Laroche (1980), "A Proposed Consumer Strategy of Simplification for Categorizing Brands, " in John D. Summey and R. D. Taylor Eds., Proceedings of the Annual Meeting of the Southern Marketing Association, Carbondale, IL: Southern Marketing Association, 112-114.

Brisoux, Jacques E., E. J. Cheron and M. Ferne (1986), "Elements de validation d'un modele de categorisation des marques," Marketing Research, Analyzing Consumer Behavior Proceedings 13th International Research Seminar in Marketing, I.R.E.T., LaLonde Les Maures (France), 91-112.

Brisoux, Jacques E. and Emmanuel J. Cheron (1990), "Brand Categorization and Product Involvement," Advances in Consumer Research, 17, Marvin Goldberg, Gerald Gorn and Richard Pollay Eds., Association for Consumer Research, 101-109.

Campbell, Brian M. (1969), "The Existence and Determinants of Evoked Set in Brand Choice Behavior," unpublished doctoral dissertation, Columbia University.

Cohen, Joel B. and Kunal Basu (1987), "Alternative Models of Categorization: Toward a Contingent Processing Framework," Journal of Consumer Research, 13 (March), 455-472.

Crowley, Ayn E. and John H. Williams (1991), "An Information Theoretic Approach to Understanding the Consideration Set/Awareness Set Proportion," Advances in Consumer Research, 19, Rebecca H. Holman and Michael R. Solomon, Eds., Association for Consumer Research, 780-787.

Feick, Lawrence F., and Linda L. Price (1987), "The Market Maven: A Diffuser of Marketplace Information," Journal of Marketing, 51 (January), 83-97.

Fortune, "Products of the Year," December 9, 1985, 106-112.

Gronhaug, Kjell (1974), "Some Factors Influencing the Size of the Buyer's Evoked Set," European Journal of Marketing, 8 (Winter), 232-41.

Gruca, Thomas S. (1989), "Determinants of Choice Set Size: An Alternative Method for Measuring Evoked Sets," Advances in Consumer Research, 16, Rebecca H. Holman and Michael R. Solomon, Eds., Association for Consumer Research, 515-521.

Higie, Robin A., Feick, Lawrence F. and Price, Linda L. (1987), "Types and Amount of Word-of-Mouth Communications About Retailers", Journal of Retailing, 63 (Fall), 260-278.

Howard, John A. (1977), Consumer Behavior: Application of Theory. Homewood, IL: Richard D. Irvin, Inc.

Jarvis, Lance P and Wilcox, James B. (1973), "Evoked Set Size - Some Theoretical Foundations and Empirical Evidence," Combined Proceedings, Thomas V. Greer, Ed., American Marketing Association, 18-37.

Kotler, Philip (1988), Marketing Management. Englewood Cliffs, NJ: Prentice-Hall, Inc.

Kotler, Philip and Gerald Zaltman (1976), "Targeting Prospects for a New Product," Journal of Advertising Research, 16 (February), 7-18.

Laroche, Michel, Jerry A Rosenblatt, Jacques E. Brisoux, and Robert Shimotakahara (1983), "Brand Categorization Strategies in RRB Situations: Some Empirical Results," Advances in Consumer Research, 10, Richard P. Bagozzi and Alice M. Tybout Eds., Ann Arbor, MI: Association for Consumer Research, 549-554.

Laroche, Michel, Jerry A. Rosenblatt and Ian Sinclair (1984), "Brand Categorization Strategies in an Extensive Problem Solving Situation: A Study of University Choice," Advances in Consumer Research, 11, Thomas C. Kinnear Ed., Ann Arbor, MI: Association for Consumer Research, 175-179.

Lichtenstein, Donald R. and Scot Burton (1990), "An Assessment of the Moderating Effects of Market Mavenism and Value Consciousness on Price-Quality Perception Accuracy," Advances in Consumer Research, 17, Marvin Goldberg, Gerald Gorn and Richard Pollay, Eds., Association for Consumer Research, 53-59.

Narayana, Chem L., Markin, Ron J. (1975), "Consumer Behavior and Product Performance: An Alternative Conceptualization," Journal of Marketing, 39, (October), 1-6.

Nowak, G. and C. Salmon (1987), "Measuring Involvement with Social Issues," Paper presented at the Association for Education in Journalism and Mass Communication, San Antonio.

Price, Linda L., Lawrence F. Feick, and Robin A. Higie (1987), "Information Sensitive Consumers and Market Information," Journal of Consumer Affairs, 21, No. 2, 328-341.

Reilly, Michael and Thomas L. Parkinson (1985), "Individual and Product Correlates of Evoked Set Size for Consumer Package Goods," Advances in Consumer Research, 12, Elizabeth C. Hirschman and Morris B. Holbrook, Eds., Association for Consumer Research, 101-109.

Sherif, Carolyn W. and Carl I. Hovland (1961), Social Judgement: Assimilation and Contrast Effects in Communication and Attitude Change. New Haven, CT: Yale University Press.

Slama, Mark E. and Terrell G. Williams (1990), "Generalization of the Market Maven's Information Provision Tendency Across Product Categories," Advances in Consumer Research, 17, Marvin Goldberg, Gerald Gorn and Richard Pollay Eds., Association for Consumer Research, 48-52.

Springle, Susan and Murphy A. Sewell (1987), "A Choice Set Model of Retail Selection," Journal of Marketing, 51, (April), 97-111.

Zaichkowsky, Judith Lynne (1985), "Measuring the Involvement Construct," Journal of Marketing Research, 12 (December), 341-352.

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