Alternative Models of Cognitive Processes Underlying Consumer Reactions to Conjunction Categories

ABSTRACT - The effects of consumers' category knowledge on evaluations of products have been intensively studied in consumer behavior over the past few years. However, few studies to date have closely investigated how consumers make their evaluations of a product that belongs to a category, called a conjunction category, defined by two simple categories, the implications of which are often inconsistent with each other. Based on the existing literature, this research develops and tests alternative models of consumer evaluation of a conjunction category. The results of the research suggest that, when consumers are given two category labels of a product as well as specific attribute information, they select either one of the categories and use it as the basis for their overall evaluation of the product. Implications of the results for consumer information processing as well as for marketing strategy are discussed.

Citation:

Moonkyu Lee and Francis M. Ulgado (1994) ,"Alternative Models of Cognitive Processes Underlying Consumer Reactions to Conjunction Categories", in NA - Advances in Consumer Research Volume 21, eds. Chris T. Allen and Deborah Roedder John, Provo, UT : Association for Consumer Research, Pages: 483-488.

Advances in Consumer Research Volume 21, 1994      Pages 483-488

ALTERNATIVE MODELS OF COGNITIVE PROCESSES UNDERLYING CONSUMER REACTIONS TO CONJUNCTION CATEGORIES

Moonkyu Lee, University of Colorado at Denver

Francis M. Ulgado, Georgia Institute of Technology

ABSTRACT -

The effects of consumers' category knowledge on evaluations of products have been intensively studied in consumer behavior over the past few years. However, few studies to date have closely investigated how consumers make their evaluations of a product that belongs to a category, called a conjunction category, defined by two simple categories, the implications of which are often inconsistent with each other. Based on the existing literature, this research develops and tests alternative models of consumer evaluation of a conjunction category. The results of the research suggest that, when consumers are given two category labels of a product as well as specific attribute information, they select either one of the categories and use it as the basis for their overall evaluation of the product. Implications of the results for consumer information processing as well as for marketing strategy are discussed.

INTRODUCTION

The use of category knowledge to make an evaluation of an object is well-established in social psychology (Brewer 1988; Fiske and Neuberg 1990; Fiske and Pavelchak 1986; Lingle and Ostrom 1981) and in consumer behavior (Boush and Loken 1991; Lee forthcoming; Meyers-Levy and Tybout 1989; Sujan 1985; Sujan and Bettman 1989). Since the multiattribute models of product evaluation have been dominant in consumer behavior (see Shocker and Srinivasan 1979, and Wilkie and Pessemier 1973 for a review of multiattribute models), it is certainly an intriguing and challenging notion that consumers do not always rely only on the product information given during the situation at hand when evaluating a product; they can and do use their pre-existing knowledge about the product. The existing literature describes the evaluation process as a matching process - matching between category knowledge and target information at hand. Applied to the product evaluation situation, the literature implies that that when the product information matches the prior knowledge and expectations about the product category, the affect associated with the category is directly transferred to the final evaluation of the product (category-based or top-down process); when the information does not match the category knowledge, evaluations of specific attributes of the product are integrated into the final evaluation (piecemeal or bottom-up process; Fiske and Neuberg 1990; Fiske and Pavelchak 1986; Sujan 1985).

However, most of the studies in the area focused on conditions under which only one category label was available with some attribute information. For example, Fiske et al. (1987) examined situations where people made evaluations of others with their job category labels (e.g., doctor, artist, etc.) and some specific trait information which was either consistent or inconsistent with the labels. In a similar vein, Sujan (1985) investigated product evaluation situations where consumers were given product category labels (i.e., a 35 mm SLR versus a 110 camera) and some specific feature information. Thus, an important question yet to be answered is: How do consumers react to a more complex category defined by the intersection of two simpler categories? An issue of particular interest is how they react to a complex category from two simple categories that are inconsistent with each other.

Such complex categories have been termed conjunction or intersection categories (Hastie, Schroeder, and Weber 1990; Kunda, Miller, and Claire 1990; Murphy 1988; Smith, Osherson, Rips, and Keane 1988). A conjunction category consists of more than one "ingredient category" which can be represented as a single lexical or idiomatic item (Murphy 1988). A Polo t-shirt sold at $30 can be an example of a conjunction category, which is composed of two ingredient categories, one implying the brand name (Polo) and the other, the price ($30).

Consumers' reactions to conjunction categories have important strategic implications for marketers. Many types of strategic changes for an existing product can create conjunction categories. Consider an example of a bi-national product: a Honda made in Mexico. In this case, the image associated with the brand name, Honda does not easily match the country image. Suppose Rolex watches are sold through discount stores such as K-Mart or Target. Obviously, in this case, the general impression implied by the brand and the image of the stores are not compatible, either. Such examples of conjunction categories in the marketplace are countless. However, little is known about cognitive processes underlying evaluations of conjunction categories.

Based on past literature, this study develops some alternative hypotheses regarding how consumers deal with the inconsistency created by conjunction categories. An experiment is then conducted to examine a situation where two ingredient category labels of a product are available as well as specific attribute information. The purpose is to investigate the reasoning process mediating consumers' responses to the dilemmas posed by conjunction categories.

THEORETICAL BACKGROUND

Categorization literature provides important insights into cognitive processes underlying product evaluations (Cohen and Basu 1987; Fiske and Neuberg 1990; Fiske and Pavelchak 1986; Smith and Medin 1981). Categories are defined as cognitive structures which contain instances that are perceived similar or equivalent (Mervis and Rosch 1981; Smith and Medin 1981). Over time and through experience, people develop categories of objects in an effort to organize and understand them. More importantly, people also develop (1) a set of expectations about the features of a typical category member with respect to a particular category, which is termed category schema or category knowledge, and (2) an affective reaction to the category, called category affect (Mandler and Parker 1976; Fiske and Pavelchak 1986; Fiske and Neuberg 1990). For example, suppose a consumer has a predefined category of a Sony stereo color TV set. The product schema, in this case, would involve the consumer's expectations about what features an average member of that category has to offer (e.g., stereo sound system, multi-color system, etc.) as well as what levels of performance it has to provide along those feature dimensions (e.g., good sound quality, good picture quality, etc.). In this sense, a category schema is often thought of as a bundle of expectations about the attributes of the category prototype, which is organized under a category label (a Sony stereo color TV set in this example). In addition, it is assumed that an overall affective reaction to the category (e.g., good, bad, etc.) is developed and associated with the category.

Generally, in the product domain, it is believed that the most basic types of categories chronically accessible to consumers in product evaluation situations include brand name, price, country-of-origin, and name of the store where the product is available, as a considerable amount of literature provides evidence of their effects on product evaluations (see Zeithaml 1988, and Rao and Monroe 1989 for a review; also see Bilkey and Nes 1982 for a review of the literature on country-of-origin effects). These are called extrinsic cues of a product since they are product-related, but not physical or intrinsic attributes of the product (Olson 1977). It is assumed that these extrinsic cues serve as category labels which are used to organize and understand the remaining product attributes.

TABLE 1

PREDICTED EFFECTS OF INGREDIENT CATEGORIES AND ATTRIBUTES UNDER ALTERNATIVE HYPOTHESES

As mentioned earlier, the use of category schema to make an evaluative judgment of a product is well-documented in consumer behavior (e.g., Sujan 1985; Sujan and Bettman 1989). However, few studies to date have examined conditions where consumers encounter an intersection category made from two simple categories that are either consistent or inconsistent with each other. The literature on categorization and on impression formation suggests at least three alternative models that could account for processes mediating evaluations of conjunction categories. These models are described in the sections that follow.

Category Integration Hypothesis

Consider a situation where a consumer is trying to evaluate an automobile branded Honda and made in Mexico. Apparently, the two ingredient category labels, the brand name and the country-of-manufacture, are inconsistent with each other. The consumer may have a favorable image of Honda products while s/he may have a moderately unfavorable impression of the products manufactured in Mexico. Concerning this case, the social judgment literature and the information integration paradigm (Anderson 1970, 1974, 1981; Bettman, Capon, and Lutz 1975; Lynch 1985; Shanteau and Ptacek 1983; Troutman and Shanteau 1976) suggest one possible hypothesis: the consumer uses some type of integration rule that would combine the affect associated with one category with that associated with the other into an overall affective reaction to the conjunction. This is termed category integration hypothesis in this study. Among a variety of integration rules that can be used, past empirical research has generally provided support for linear, adding and averaging types of integration rules. Although providing a complete description of adding and averaging integration rules is beyond the scope of the study (see Anderson 1974, for a review of the integration rules), some measures should be clarified that allow us to detect whether a consumer actually uses the integration rules in the evaluation situation. The literature suggests that, if a consumer indeed uses a linear integration rule to make an overall evaluation of a conjunction category, a factorial plot of the observed responses should exhibit parallelism, evidenced by a statistically significant main effect of one ingredient category, a significant main effect of the other, and a nonsignificant interaction between the two categories on the evaluation of the conjunction in an analysis of variance model (Anderson 1981; Lynch 1985). Under the information integration rules, there can be a special case where a zero weight is assigned to either one of the ingredient categories. In that case, the main effect of the category will become nonsignificant. This type of case will be dealt with separately in the next section because the underlying process can be different from the one hypothesized here.

The category integration hypothesis implies that the evaluative implications of the two simple categories are combined into the final judgment of the intersection category. The possible occurrence of integration of product attribute information into the overall evaluation of the conjunction category is explored in the third hypothesis discussed below. The predicted effects of element categories and attributes on dependent measures suggested by the category integration hypothesis are summarized in Table 1.

Subtyping Hypothesis

Consider again the same example given above: a Honda automobile made in Mexico. Fiske and her colleagues suggest that when the object being evaluated does not find the best-fitting category, it can be subtyped on the basis of the current categories (Fiske and Neuberg 1990; Fiske and Pavelchak 1986). The subtyping would entail accessing a category at the level subordinate to either one of the ingredient categories. Thus, the present product can be subtyped into either a category of "basically Honda automobiles, but manufactured in Mexico" or one of "essentially Mexican-made automobiles, but branded Honda."

Although the subtyping phenomenon has been observed in some studies (see, e.g., Fiske et al. 1987 and Sujan 1985), any predictions about their evaluative outcomes have not been made nor tested. It is postulated that, to the extent that a subtype is formed on the basis of its superordinate category, the direction and the extremity of the evaluation of the subtype would be consistent with those of the affect associated with the original category. In the present example, the automobile could be evaluated on the basis of either the Honda or the Mexican-made automobile category depending on which of the two is more salient and, thus, used for the subtyping. Consequently, only one of the two ingredient categories becomes relevant to the final evaluation of the conjunction category. What determines the relative salience of the element categories is certainly an important issue if not a primary concern of the present study. It can be anything which expedites category accessibility, e.g., physical manifestation, temporal primacy, contextual novelty, and so forth. More generally, the relative salience of the categories should depend on the cognitive category structure, i.e., which category is located at a higher level in the hierarchy of categories.

To summarize, the subtyping hypothesis implies that a consumer uses only one of two ingredient categories to make an evaluation of the conjunction category. Accordingly, the hypothesis would be supported by a significant main effect of one of the two ingredient categories, a nonsignificant main effect of the other, and a nonsignificant interaction between the two categories on the evaluation of the conjunction category in an analysis of variance design. It is assumed that the evaluative implications of specific attributes are not yet determined nor integrated into the overall evaluation of the conjunction category. Thus, the effects of product attribute information on the overall evaluation would be nonsignificant. The implications of the subtyping hypothesis are summarized in Table 1.

Piecemeal Elaboration Hypothesis

The final hypothesis is suggested by the possibility that the ingredient category labels are ignored; instead, the specific product attributes are elaborated upon. This process entails an effortful integration of available information in an attribute-by-attribute fashion. The rationale behind the hypothesis is that the two element categories, when contradictory to each other, lose their diagnosticity to the evaluation. The categorization literature suggests that when a category label becomes irrelevant to the evaluation task, a piecemeal or an attribute-by-attribute evaluation process is used (Fiske et al. 1987; Fiske and Neuberg 1990; Fiske and Pavelchak 1986; Pavelchak 1989; Sujan and Bettman 1989).

There are several process/outcome measures that have been used to detect piecemeal processes (see Boush and Loken 1991, Lee forthcoming, Fiske et al. 1987, and Sujan 1985, for the different measures). Among them, verbal protocol and correlation measures are used in this research. It is expected that if a consumer indeed engages in piecemeal processing, s/he will generate more thoughts and verbalizations related to product attributes (called piecemeal responses) than if s/he does not. As a result, there will be a high positive correlation between the overall evaluation of the conjunction category and the attribute evaluations. Table 1 shows the predicted effects of the ingredient categories and those of the attributes on the overall evaluation of the conjunction category for the piecemeal elaboration hypothesis.

METHOD

An experiment was conducted to find out how consumers deal with inconsistency created by conjunction categories. Subjects were given specific attribute information about products under two ingredient category labels which were either consistent or inconsistent with each other. Brand name and country-of-manufacture were the selected levels of ingredient categories in the study. Subjects reactions to the conjunction categories were measured. More details about the experiment are described below.

Pretests

For a successful manipulation of consistency between two ingredient categories, it was critical to select product classes that were well-known to the population under study. A pretest was conducted to choose appropriate product categories for the experiment. Ninety-four students at a major metropolitan state university completed a questionnaire asking about their levels of familiarity with several product classes. It was found that TV sets and athletic shoes were the ones that they were most familiar with.

Another pretest was conducted with 82 students to determine brand names and countries-of-manufacture that were associated with either favorable or unfavorable image. The students were asked to list the best and the worst brand names in each of the product classes as well as the best and the worst countries that would manufacture each of the products. The results showed that Sony was perceived to be the best and Emerson, the worst brand name in the TV set category; Germany was perceived to be the best and Taiwan, the worst country that would manufacture TV sets. The results also revealed that Nike was seen as the best and Converse, the worst brand name in the athletic shoe category; the U.K. was regarded as the best, and Mexico, the worst country that would manufacture athletic shoes.

The stimulus materials were constructed for the main study on the basis of these brand and country names representing either favorable or unfavorable category image as well as 9 pieces of attribute information, drawn from Consumer Reports. The pieces of attribute information were put together such that they varied in terms of favorableness and relevance to the evaluation task. The intent was to ensure that the overall configuration of the information was not suggestive of any particular direction on the favorableness and relevance dimensions, and thus, to create evaluation situations that were closer to reality. The attribute information used for the study is presented in Appendix.

Design and Procedure

Based on pretest results, two levels of brand name (favorable vs. unfavorable) were crossed with two levels of country-of-manufacture (favorable vs. unfavorable). Each subject evaluated two products, a TV set and a pair of athletic shoes, with brand name and country information. Thus, the design was a 2 (ingredient category 1: brand name) X 2 (ingredient category 2: country-of-manufacture) X 2 (product) factorial design, with brand and country names as between subject factors and product as a within subject factor. The same attribute information was given within the same product conditions.

A total of 93 students enrolled in business courses participated in the experiment. They were randomly assigned to each of the experimental conditions; cell sizes ranged from 22 to 25. At the beginning of each session, they were told the study was designed to measure their perceptions and feelings about international products. They were first asked to give their best estimates of the average prices of the products under study. They were also asked to provide their perception of the brand names and the countries-of-manufacture on 9-point bipolar scales ranging from "very bad" to "very good." Then, under a brief scenario (starting with "Suppose you run into the following brand name of a product manufactured in the following country...") they were given stimulus information (A Sony TV set manufactured in Taiwan, for instance, and its specific attribute information), and asked to read the information with the purpose of forming an evaluation of the product. After reading it, they rated the quality of the product on a 9-point bipolar scale ranging from "very bad" to "very good." They also indicated their importance weights for each of the attributes (on 9-point scales from "not-at-all important" to "very important"), their likings for the attributes (on 9-point scales from "strongly like" to "strongly dislike"), and their overall impression of all the attributes considered in totality (on a 9-point scale from "strongly like" to "strongly dislike). They were then asked to write down all the thoughts and ideas that they experienced while reading the description. It was requested that they report all thoughts no matter how simple, complex, relevant, or irrelevant they might seem. Finally, subjects were debriefed and thanked.

APPENDIX

RESULTS

The success of the manipulations of brand and country levels was checked through a 2 (brand name) X 2 (country-of-manufacture) analysis of variance procedure for each of the product classes. For the TV set category, the main effect for brand name on brand perception reached significance in the expected direction (F(1,92) = 194.83, p < .000; Ms = -1.34 vs. 3.08), and the main effect for country name on country perception was also significant (F(1,92) = 33.29, p < .000; Ms = -1.17 vs. 1.09). In addition, for the athletic shoe category, there was a significant main effect for brand name on brand perception (F(1,92) = 139.31, p < .000; Ms = -.94 vs. 3.24) as well as a significant main effect for country name on country perception (F(1,94) = 97.88, p < .000; Ms = -1.13 vs. 1.94), both in the expected direction. Thus, the manipulation worked as intended.

The data were analyzed through a 2 (brand) X 2 (country) analysis of variance procedure performed for each of the two product classes. Subjects' estimates of average prices of the products were used as a covariate in the analysis. Table 2 shows the mean ratings of product evaluations across the conditions. For the TV set category, a significant main effect was found for brand name (F(1,92) = 81.86, p < .000). However, neither the main effect for country information nor the brand X country interaction effect reached significance (F(1,92) = .26, p > .61 and F(1,92) = .05, p > .82 for the two effects, respectively). The data for the athletic shoe category revealed the same pattern of results. While the main effect for brand name was found to be significant (F(1,92) = 63.39, p < .000), the main effect for country name and the interaction effect between brand and country name were not (F(1,92) = 1.24, p > .26 and F(1,92) = 3.05, p > .08 for the two effects, respectively). For both of the product classes examined in the experiment, only the main effect for one ingredient category, brand name, was significant.

Additional tests were performed to see whether the product attribute information influenced the overall evaluations of the conjunction categories. The expectation was that, if the attribute information played any role in resolving the inconsistency created by the conjunction categories, more attribute-oriented thoughts, or piecemeal responses, would be generated under the conditions involving the inconsistency, i.e., the favorable brand/unfavorable country and unfavorable brand/ favorable country conditions. Thus, the expectation would be supported by a statistically significant brand X country interaction effect on piecemeal responses in the current analysis of variance design. Subjects' retrospective verbal protocols were coded into four categories: brand-oriented thoughts (e.g., "Sony makes good electronic products." in the T.V. set category), country-oriented thoughts (e.g., "I am not sure if T.V. sets made in Taiwan last long."), attribute-oriented or piecemeal thoughts (e.g., "Background noise from a T.V. set can be annoying."), and other thoughts (e.g., "There is too much violence on T.V."). Two independent judges coded the verbalizations into the four categories and their initial agreement was 86%. Any disagreement was resolved by discussion. Among the types of thoughts generated, the attribute-oriented or piecemeal thoughts accounted for a dominant portion of the total thoughts generated; they averaged a frequency of 3.52 per subject (or 61% of total thoughts) for the TV set category and a frequency of 2.95 per subject (or 57% of total thoughts) for the athletic shoe category. However, the results of the analysis of variance procedure showed that the brand X country interaction was not significant on piecemeal responses (F(1,89) = 1.68, p > .20 for the TV set category; F(1,89) = 2.71, p > .10 for the athletic shoe category). In addition, the correlations between the overall evaluations and the attribute evaluations were not significantly greater than zero in any of the conditions in the study, which suggests that the evaluative implications of the product attributes were not determined, nor considered (if determined), in making an overall evaluation of the products.

TABLE 2

CELL MEANS FOR THE DEPENDENT MEASURE

DISCUSSION

The pattern of results, considered in totality, is compatible with that implied by the subtyping hypothesis. It seems that the categories identified with the brand names dominated and incorporated the categories implied by the country-of-manufacture information. Apparently, the products were subtyped within the categories defined by the brand names.

It is somewhat surprising that the attribute information, despite its availability, did not have any effects on evaluation of the conjunction categories. There are at least two possible explanations for the nonsignificant attribute effect. The first possibility is concerned with general cognitive category structure. In this research, it was assumed that the category identified with brand name and that implied by country-of-manufacture are at the same level in the product category structure in consumers' memory. However, as the data suggest, the brand name category might be at a higher level in the hierarchical structure of categories, and thus, might have been able to take a dominant position when it was presented with the country-of-manufacture information. The second possibility is that the attribute information was read, but used only to make a confirmation check on the brand category membership of the product. An interesting point to be made here is that despite the availability of the same attribute information across different conditions, only the brand name effect was found to be significant. This implies that the same attribute information was interpreted either positively or negatively depending on the favorableness of the brand names. This type of phenomenon has been observed in consumer information processing research and elsewhere (e.g., Deighton 1984; Gilovich 1981). According to the literature, consumers' expectation about a product has a profound effect on their subsequent interpretation and evaluation of the product information. In the present study, it is speculated that the brand names created expectations, which served as a hypothesis, and the subjects selectively processed the attribute information in a manner supporting the hypothesis.

The results of the present study have an important implication for marketers who make strategic changes, e.g., change of price, channel outlet, location of plant etc. for their current product. Marketers who make such changes can often anticipate whether these changes will induce favorable or unfavorable reactions from consumers. Then, the most important objective of their communication strategy should be to emphasize desirable changes and/or to de-emphasize undesirable changes. The results of the study suggest that a company can downplay the change by reminding consumers of the brand name and the specific attributes of the product through an effective promotion campaign. As an example of a company that actually puts the idea into practice, Sony emphasizes its established brand name (e.g., the "It's a Sony" slogan) and the innovative product features, while the product could very well be made in countries outside Japan.

The present study has taken an initial step toward understanding cognitive processes underlying the resolution of conflicting categories in the product domain. It has examined a situation where product attribute information is available as well as two simple category labels. Thus, the generalizability of the study results is limited to such a situation. Future extensions of the study would be to investigate situational and individual differences in processing conjunction categories.

REFERENCES

Anderson, Norman H. (1970), "Functional Measurement and Psychological Judgment," Psychological Review, 77, 153-170.

Anderson, Norman H. (1974), "Information Integration Theory: A Brief Summary," in Contemporary Developments in Mathematical Psychology (Vol. 2), ed. David H. Krantz, R. Duncan Luce, Richard C. Atkinson, and Patrick Suppes, San Francisco: W.H. Freeman, 236-305.

Anderson, Norman H. (1981), Foundations of Information Integration Theory, New York: Academic Press.

Bettman, James R., Noel Capon, and Richard J. Lutz (1975), "Cognitive Algebra in Multivariate Attitude Models," Journal of Marketing Research, 12 (May), 151-164.

Bilkey, Warren J. and Erik Nes (1982), "Country-of-Origin Effects on Product Evaluations," Journal of International Business Studies, 13 (Spring/Summer), 89-99.

Boush, David M. and Barbara Loken (1991), "A Process-Tracing Study of Brand Extension Evaluation," Journal of Marketing Research, 28 (February), 16-28.

Brewer, Marilynn B. (1988), "A Dual-Process Model of Impression Formation," in Advances in Social Cognition, ed. Thomas K. Srull and Robert S. Wyer Jr., Hillsdale, NJ: Erlbaum, 1-36.

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

Deighton, John (1984), "The Interaction of Advertising and Evidence," Journal of Consumer Research, 11 (December), 763-770.

Fiske, Susan T., and Steven L. Neuberg (1990), "A Continuum of Impression Formation, From Category-based to Individuating Processes: Influence of Information and Motivation on Attention and Interpretation," Advances in Experimental Social Psychology, 23, 1-74.

Fiske, Susan T., Steven L. Neuberg, Ann E. Beattie, and Sandra J. Milberg (1987), "Category-based and Attribute-based Reactions to Others: Some Informational Conditions of Stereotyping and Individuating Processes," Journal of Experimental Social Psychology, 23, 399-427.

Fiske, Susan T. and Mark A. Pavelchak (1986), "Category-based versus Piecemeal-based Affective Responses: Developments in Schema-Triggered Affect," in The Handbook of Motivation and Cognition: Foundations of Social Behavior, ed. Richard M. Sorrentino and E. Tory Higgins, New York: Guilford Press, 167-203.

Gilovich, Thomas (1981), "Seeing the Past in the Present: The Effect of Associations to Familiar Events on Judgments and Decisions," Journal of Personality and Social Psychology, 40, 5, 797-808.

Hastie, Reid, Colin Schroeder, and Renee Weber (1990), "Creating Complex Social Conjunction Categories from Simple Categories," Bulletin of the Psychonomic Society, 28, 3, 242-247.

Kunda, Ziva, Dale T. Miller, and Theresa Claire (1990), "Combining Social Concepts: The Role of Causal Reasoning," Cognitive Science, 14, 551-577.

Lee, Moonkyu (forthcoming), "Informational and Motivational Influences on Consumer Evaluations of Line and Brand Extensions," Journal of Business and Psychology.

Lingle, John H. and Thomas M. Ostrom (1981), "Principles of Memory and Cognition in Attitude Formation", in Cognitive Responses in Persuasion, ed. R. E. Petty, T. M. Ostrom, and T. C. Brock, Hillsdale, NJ: Erlbaum, 399-420.

Lynch, John G., Jr. (1985), "Uniqueness Issues in the Decompositional Modeling of Multivariate Overall Evaluations: An Information Integration Perspective," Journal of Marketing Research, 22, 1-19.

Mandler, Jean M. and Richard E. Parker (1976), "Memory for Descriptive and Spatial Information in Complex Pictures," Journal of Experimental Psychology: Human Learning and Memory, 2 (January), 38-48.

Mervis, Carolyn B. and Eleanor Rosch (1981), "Categorization of Natural Objects," Annual Review of Psychology, 32, 89-115.

Meyers-Levy, Joan and Alice M. Tybout (1989), "Schema Congruity as a Basis for Product Evaluation," Journal of Consumer Research, 16 (June), 39-53.

Murphy, George L. (1988), "Comprehending Complex Concepts," Cognitive Science, 12, 529-562.

Olson, Jerry C. (1977), "Price as an Information Cue: Effects in Product Evaluation," in Consumer and Industrial Buying Behavior, ed. Arch G. Woodside, Jagdish Sheth, and Peter D. Bennett, NY: North Holland Publishing Company, 267-286.

Rao, Akshay R. and Kent B. Monroe (1989), "The Effect of Price, Brand Name, and Store Name on Buyers' Perceptions of Product Quality: An Integrative Review," Journal of Marketing Research, 26 (August), 351-357.

Shanteau, James, and C. H. Ptacek (1983), "Role and Implications of Averaging Process in Advertising," in Adverting and Consumer Psychology, ed. L. Percy and A. G. Woodside, MA: Lexington Books, 149-68.

Shocker, Alan D. and V. Srinivasan (1979), "Multiattribute Approaches for Product Concept Evaluation and Generation: A Critical Review," Journal of Marketing Research, 16 (May), 159-180.

Smith, Edward E. and Douglas L. Medin (1981), Categories and Concepts, Cambridge, MA: Harvard University Press.

Smith, Edward E., Daniel N. Osherson, Lance J. Rips, and Margaret Keane (1988), "Combining Prototypes: A Selective Modification Model," Cognitive Science, 12, 485-527.

Sujan, Mita (1985), "Consumer Knowledge: Effect on Evaluation Strategies Mediating Consumer Judgment," Journal of Consumer Research, 12 (June), 31-46.

Sujan, Mita and James R. Bettman (1989), "The Effect of Brand Positioning Strategies on Consumers' Brand and Category Perceptions: Some Insights from Schema Research," Journal of Marketing Research, 26 (November), 454-467.

Troutman, C. Michael and James Shanteau (1976), "Do Consumers Evaluate Products by Adding or Averaging Attribute Information?" Journal of Consumer Research, 3 (September), 101-106.

Wilkie, William L. and Edgar A. Pessemier (1973), "Issues in Marketing's Use of Multiple-Attribute Attitude Models," Journal of Marketing Research, 10 (November), 428-441.

Zeithaml, Valarie A. (1988), "Consumer Perceptions of Price, Quality, and Value: A Means-End Model and Synthesis of Evidence," Journal of Marketing, 52 (July), 2-22.

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