Brand Categorization Strategies in Rrb Situations: Some Empirical Results

Michel Laroche, Concordia University
Jerry Rosenblatt, Concordia University
Jacques E. Brisoux, University of Quebec at Three Rivers
Robert Shimotakahara, Concordia University
ABSTRACT - This paper reviews the Howard, Narayana-Markin and Brisoux-Laroche paradigms of consumers' brand categorization strategies in a Routinized Response Behavior situation. A survey was conducted on toothpaste brands and the Brisoux-Laroche categories, i.e., evoked, hold, reject and foggy, were determined. After assessing the mean scores for attitude, intention, confidence and quantity of information processed, the authors conclude that the results tend to confirm the hypothesized profiles for the four sets. In addition, a discriminant analysis was performed on a brand basis, which yielded significant predictive power, and provided additional insights on various criteria for inclusion into one of the sets.
[ to cite ]:
Michel Laroche, Jerry Rosenblatt, Jacques E. Brisoux, and Robert Shimotakahara (1983) ,"Brand Categorization Strategies in Rrb Situations: Some Empirical Results", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 549-554.

Advances in Consumer Research Volume 10, 1983      Pages 549-554

BRAND CATEGORIZATION STRATEGIES IN RRB SITUATIONS: SOME EMPIRICAL RESULTS

Michel Laroche, Concordia University

Jerry Rosenblatt, Concordia University

Jacques E. Brisoux, University of Quebec at Three Rivers

Robert Shimotakahara, Concordia University

[The authors wish to thank an anonymous ACR reviewer for his most constructive comments and suggestions.]

ABSTRACT -

This paper reviews the Howard, Narayana-Markin and Brisoux-Laroche paradigms of consumers' brand categorization strategies in a Routinized Response Behavior situation. A survey was conducted on toothpaste brands and the Brisoux-Laroche categories, i.e., evoked, hold, reject and foggy, were determined. After assessing the mean scores for attitude, intention, confidence and quantity of information processed, the authors conclude that the results tend to confirm the hypothesized profiles for the four sets. In addition, a discriminant analysis was performed on a brand basis, which yielded significant predictive power, and provided additional insights on various criteria for inclusion into one of the sets.

INTRODUCTION

The question of how consumers deal with all the brands in a product category has attracted increasing attention by researchers in recent years. Howard (1963) first made passing reference to the individual's evoked set when he stated that, "when a buyer considers making a purchase, the number of alternatives that come to mind are probably less than the number that is objectively available" (p. &4). At that time, the concept was not really elaborated, nor was an explanation given as to how it was integrated into his buyer behavior theory. It was basically a "common sense" argument for a general cognitive efficiency to exist in consumers when faced with a purchase decision involving limited problem solving. In a later version of buyer behavior theory, the concept was re-introduced in a more formal manner. In the Howard-Sheth theory, evoked set was defined as, "the brands that the buyer considers as acceptable for his next purchase constitute his evoked set" (Howard and Sheth, 1969, p. 98). Explicit mention was made as to the nature of evoked set and its integration into the model of buyer behavior. Theoretical explanation from other disciplines was offered for its existence. Various factors were also postulated as being determinants of its size.

Howard's Conceptualization

In 1977, Howard offered the most precise definition of the concept. It was given as, "the subset of brands that consumer considers buying out of the set of brands that he or she is aware of in a given product class" (Howard, 1977, p. 306). At this time, evoked set was fully mentioned in the routinized response behavior stage of consumer behavior. Figure 1 diagrammatically illustrates Howard's brand categorization Process.

This model is essentially a simplification of the consumer's decision making process when motivated to purchase a product. Without further elaboration about the various constructs and variables involved with this model, it can be said that, over the course of time, an individual's evoked set develops as the decision making process is exercised numerous times as products within a given category are repeatedly purchased. Given that the consumer has become totally familiar with the product class and also familiar with the brands within that product class, he then exhibits "routinized response behavior" toward that purchasing situation. This type of purchasing situation is typically characterized by a small amount of time and required information necessary to make the actual purchase choice. With the absence of new product entries and modifications made to existing brands within the product class or no changes in preferences or motivations by the consumer, it is unlikely that the composition of that consumer's evoked set would change over time.

FIGURE 1

COMPARISON OF THREE CONSUMER CATEGORIZATION PROCESSES

The adjustment process most often precipitated by the introduction of new brands or significant modification of existing brands leads to a "limited problem solving" situation for the consumer's decision making process. This "limited problem solving" situation is characterized by the need for more information and time before a purchase choice can be made by the consumer. It will prevail until the consumer develops enough of a brand image of these new brands or altered existing ones to allow them to either become an evoked set member or another non-legitimate alternative. At this time, the consumer will once again exhibit "routinized response behavior" toward that purchasing situation. The tendency and reasons why the composition of the consumers' evoked sets change over time should now be apparent.

It would be useful to address the issue of why the evoked set exists at this point. Many researchers have offered various rationale toward its explanation although the theories of Miller and Wallace have received the most attention (Howard and Sheth, 1969; Campbell, 1969; Ostlund, 1973; Jarvis and Wilcox, 1973; Gronhaug, 1973). Miller (1956) argued that human beings possess learning or physiological limitations which severely constrain their capability to make distinctions between alternatives on the basis of unidimentional stimuli.

Wallace also has argued for cognitive limitations operating in all humans, be they culturally advanced or primitive. A study of various diverse cultures resulted in the finding that no relationship could be found between the complexity of kinship terminology used and the technological level of the culture. These findings suggest that all men, irrespective of their social and technological development, possess a cognitive limit as to the complexity of semantic spaces which they are able to handle (Wallace, 1961).

Narayana and Markin's Conceptualization

Narayana and Markin (1975) have expanded upon Howard's evoked set concept by suggesting that consumers may actually define their product classes more thoroughly by categorizing all brands of which they are aware. This mental categorization of all available brands is then the consumer's perception of the current marketing situation for that product class. Their conceptualization of the consumer's brand categorization process begins with the total set. This set is simply all of the available brands which exist within the product class at any given time. The size of the total set will fluctuate over time as new brands are introduced and unsuccessful ones are withdrawn. The total set is then divided into the awareness and unawareness sets in recognition of the fact that most consumers are not likely to be aware of all brands within the product category. It is from the awareness set that consumers categorize the alternative brands in terms of their viability as purchase alternatives. The outcome of the viability process results in a given brand to be categorized into one of three different sets.

The first is the familiar evoked set. It consists of those brands which the consumer actively considers when making a purchase decision. The brands within it are evaluated positively by the consumer for purchase and consumption purposes. The inept set contains those brands which the consumer has rejected from purchase consideration. For whatever reason, these brands have been evaluated negatively. All brands found within the inert set are those which, although the consumer is aware of them, are neither rejected from nor accepted for purchase consideration. It is believed that such brands are evaluated neutrally by consumers. A diagram of this categorization process is given in Figure 1. It must be mentioned that the author's model is dynamic in nature. The theory suggests that the marketing situation of a given product class and the consumer's categorization of aware brands is variable over time. The changes in the market situation are manifested in the size and composition of the total set of brands. New product entries are continually introduced and add to the size of this set. Likewise, unsuccessful brands are withdrawn thereby reducing the size of the total set. These inevitabilities, combined with the changing needs of consumers, forces the reevaluation of their evoked set over time.

Brisoux and Laroche's Conceptualization

Brisoux and Laroche (1980) have proposed a modification to the previous conceptualization by proceeding from the general standpoint that the greater the precision by which the categorization process can be specified, the more complete and rigorous will the theory be. In this section we review the proposed formulation and make some revisions in the initial conceptualization based on some recent findings in the literature.

The authors make the point that, in a routinized response behavior situation, as a result of the excessive number of brands in many product classes today, not all of the consumer's awareness set of brands would be completely processed due to limited cognitive capacity (Miller-Wallace argument). Another reason for incomplete processing is the strategy of simplification used by the consumer in a choice situation. Nicosia (1966) has conceptualized the final purchase act as "emerging from a funneling process." This process seems to be multi-stage (Bettman 1979): the consumer forms an evoked set by eliminating brands which fail to meet some acceptable minimum level on one or more evaluative criteria (Belonax and Mittelstaedt 1978, Green and Srinivasan 1978, Bettman and Park 1980, Johnson and Russo 1981). Empirical evidence for the use of a conjunctive decision rule in the formation of evoked set in a R.R.B. situation has been found by Brisoux and Laroche (1981). This conjunctive type of processing could explain why recall accuracy under nondirected learning (choice situation) is higher for chosen as opposed to rejected brands (Biehal and Chakravarti 1982). Some brands may be quickly eliminated as they do not meet some minimum attribute values.

Therefore, the brands which are attribute processed belong to the processed set while the others comprise the foggy set. The latter includes brands the consumer knows exist (with or without aided recall) and can be identified with the product class but for which he does not have any specific comprehension. Such brands do not have significant meaning as they cannot really be distinguished in terms of the evaluative criteria of that product class. Although the individual is able to recognize them, they would not be considered for purchase as no opinion is held about them. Specific reasons why a foggy set brand could exist are as follows: the consumer "has not seen any advertisement about them or does not remember seeing any, or if she/he does, it was not informative-enough to allow her/him to judge the brand; she/he has not tried some of these brands or if she/he had the personal experience with it was inconclusive; she/he does not remember whether anybody has mentioned c, consumed it or ordered it." (Brisoux and Laroche 1980).

Unlike the foggy set brands, those found in the processed set have all been processed by the consumer on at least one attribute. Russo and Johnson (1980) have proposed a taxonomy of knowledge composed of two central components: the level of inference, which refers to the relationship between external information and knowLedge in memory, and the inferential basis, which refers to the organization of knowledge by brand or by attribute. Johnson and Russo (1981) have shown that the differential recall between chosen and eliminated brands is also affected by the level of inference: the higher the inferential level, the greater the recall advantage of the chosen brand.

This finding when considered with the results of Biehal and Chakravarti (1982), which showed that more information is processed and stored in long term memory for chosen as opposed to rejected brands, suggests other hypotheses in addition to those proposed by Brisoux and Laroche (1980). Their initial conceptualization differentiated the reject set, the hold set, and the evoked set composing the processed set in terms of attitudes and intentions. These additional hypotheses deal with the quantity of information processed by the consumer and his/her confidence in brand evaluation.

The definitions and hypothesized profiles of the aforementioned sets are summarized in Table 1, and can be described in the following manner. The first set within the processed set is the evoked set. As with Narayana and Markin (1975), these brands are chose which the consumer would consider when making a purchase decision. Attitudes, purchasing intentions, quantity of information processed by the individual, and confidence in evaluation with respect to these brands are expected to be highest in comparison to brands in the other sets.

TABLE 1

SUMMARY OF HYPOTHESES

The second category of brands found within the processed set is the hold set. The concept of this set parallels that of Social Judgement Theory (Sherif et al. 1965) in that while such brands are neither acceptable purchase alternatives of the consumer, they could have either positive, negative, or neutral attitudes held about them. Emery's mapping model with respect to price and quality perceptions is cited by Brisoux and Laroche (1980) as offering a possible explanation of why a brand might be found within the hold set (Emery 1969). Such considerations include:

(a) The consumer may have a positive attitude toward the brand, but may not include it into the evoked set because it is not perceived as adequate for his motives; it is not appropriate for consumption situation or the price is too high in relation to quality, or no one in her/his reference group consumes it. If one of these conditions changes, these brands might move into the evoked set.

(b) Conversely, the consumer may have a somewhat negative attitude toward a brand in the hold set but may not reject it because its price is low enough that it may be considered a bargain.

(c) The consumer may be truly neutral toward the rest of the brands in the hold set. She/he neither likes them nor dislikes them, for they are judged mediocre and if nothing happens over a period of time the brand, through forgetting and lack of reinforcement, some might move into the foggy set (Brisoux and Laroche 1980, p. 2).

Attitudes, purchasing intentions, quantity of information processed by the individual, and confidence in evaluation with respect to these brands, are expected to be higher than those brands belonging to the reject and foggy sets, yet lower (average by comparison) than brands in the evoked set.

The next set is labeled the reject set. It is analogous to Narayana and Markin's inept set and contains those brands which the consumer finds unacceptable and would not consider when making a purchase choice. Attitudes in this set are expected to be at their lowest. Confidence and quantity of information processed are expected to be lower for brands in this set as compared to those in the evoked and hold sets, but higher than those in the foggy set, while purchasing intentions should be lowest (or null).

The fourth category of brands, which is not found within the processed set, is the foggy set. Attitudes are expected to be lower for brands in this set than for those in the evoked and hold sets, however, more positive than brands in the reject set. Confidence and quantity of information processed should be at their lowest, while purchasing intentions are expected to be slightly more positive for brands in this set than for those in the reject set (if not null). The entire conceptualization is found in Figure 1 and summarized in Table 1.

Research Objectives

The objectives of this research are twofold: 1) to further examine the Brisoux and Laroche (1980) conceptualization of consumer brand categorization strategies. More specificaLly, to test the hypothesized profiles for the evoked, reject, foggy and hold sets, in terms of consumer's attitudes, intention, confidence and quantity of information processed; and, 2) to further explore the nature of the evoke g hold, reject and foggy sets.

METHODOLOGY

The research reported herein, is part of a larger project examining the toothpaste market in Quebec. The operational universe was restricted to residents of two predominantly English-speaking suburbs of Montreal. A nonprobabilistic area sampling technique was employed which j yielded 197 completed and usable questionnaires.

Data Collection Method

All data used in the study was collected over the months of November and December of 1981 and January of 1982. All interviews were conducted by one of the authors. The data was collected in two phases. Initially a structured personal interview was conducted at the respondent's residence. Each interview lasted approximately fifteen minutes. Upon its completion, a second questionnaire was left with the respondent, at which time instructions were given so as to ensure its proper completion. The respondent was told that the interviewer would return in one week to collect the completed questionnaire. Most respondents noted it required approximately fifteen to thirty minutes to complete.

During the personal interview, the following information was obtained by using a set of eleven cards, in random order, each depicting a different toothpaste package:

(1) Measurement of the Awareness Set

(2) Measurement of brands previously tried

(3) Measurement of brands purchased within the last six months

(4) Measurement of the Evoked Set

(5) Measurement of the Reject Set

(6) Measurement of the Foggy Set

The measurements of the evoked, reject and foggy sets were adapted from those used by Brisoux and Laroche (1980) and introduced in Appendix 1. Identification of the hold set was indirect (i.e., by deduction from the other four sets).

In the questionnaire left with the respondent, the consumer's perceptions of toothpaste brands on several product class and personal dimensions, as well as the consumer's purchasing habits and decision criteria (importance variables) were measured. The respondents' perceptions of specific brands as well as the importance variables were measured on nine-point bipolar semantic differential type scales. The dimensions chosen to study the purchasing habits of toothpaste users were those employed by Sheth and Talarzyk (1972). Introduction preambles were used on all scales to ensure that the respondent would have clear understanding of what the dimensions represented. Finally, intention was measured by asking the respondent to list the brand(s) s/he would select in the next 10 purchases (see appendix 2 for all operational definitions).

Results and Interpretation

In-order to test the hypothesized profiles, the mean and standard error were calculated for set size, attitude, intention, confidence and quantity of information processed for each of the four categories. The results are summarized in Table 2.

TABLE 2

MEAN ATTITUDE, INTENTION, CONFIDENCE & INFORMATION OF THE BRANDS IN THE EVOKED, HOLD, REJECT AND FOGGY SETS

The hypothesized profile of brands in evoked set was well supported. The measures of mean attitude, intention, confidence and quantity of information processed were all significantly higher for brands in the evoked set than brands in all other sets (all significant at p < .01). This essentially supports the accepted belief that consumers have a highly developed concept and well formulated opinions of brands in their evoked set.

The hypothesized profile of brands in the hold set was generally supported. The measures of mean attitude, confidence and quantity of information processed for brands in the hold set were all lower as compared to brands in the evoked set, and higher than brands in the foggy and reject sets (all significant at p < .01). Moreover, as expected, mean intention for hold set members was found to be significantly lower than evoked set members (p < .01), higher than foggy set members (p < .01), but not significantly different than reject set members. These findings suggest that consumers of toothpaste have reasonably well developed brand concept of their hold set members. While the measure of mean attitude for brands in the hold set was approximately neutral (4.92), it appears that attitudes for these brands range across the moderately positive to moderately negative spectrum. In addition, the finding of significant differences (all significant at p < .01) in the measures of mean confidence, quantity of information processed and attitude, supports the hypothesis of the Brisoux-Laroche paradigm that there are basic distinctions between the hold and foggy sets.

The hypothesized profile of the reject set was also generally supported. Mean attitude, confidence and quantity of information processed were all lower for brands in the reject set than those in the hold and evoked sets (all significant at p < .01). The measure of mean intention were as hypothesized - i.e., hold more positive than reject - however no significant difference in mean intention for brands in the hold and reject set was observed. The findings - i.e., negative attitudes and low purchasing intentions - tend to confirm that consumers do not consider these brands as purchasing alternatives. The hypothesized low degree of confidence and quantity of information processed was also found, which supports the findings of Biehal and Chakravarti (1982). It seems that a consumer's decision to reject a brand of toothpaste can be made with little information and does not require great confidence. This suggests that the categorization of brands into the reject set involves a limited degree of problem-solving.

The hypothesized profile of the foggy set was partially supported. As expected, quantity of information processed and confidence were lower for brands in the foggy set as compared to those in all other sets. However, mean attitude expected to be lowest in the reject set, was found to be lowest in the foggy set (however, no significant difference was found between mean attitude in these sets). Similarly, mean intention was found to be lower for brands in foggy set than those in the reject set, which was not expected, yet no significant difference in mean intention for brands in these sets was found. It is interesting that the measures of mean confidence and quantity of information processed for brands in the foggy set were lower than those in the reject set. This tends to support the Brisoux and Laroche (1980) hypothesis that brands in the "processed set" have all been evaluated on at least one attribute, and process more information than brands in the foggy set. This finding also supports the Russo and Johnson (1980) taxonomy of knowledge which refers to the organization of information by brand or by attribute. As expected, attitudes toward brands in the foggy set were found to be lower than brands in the hold and evoked sets. Contrary to expectations, attitudes toward foggy set brands were also found to be lower than reject set brands, however, the difference between mean attitude for reject and foggy set brands was not significant.

In order to reach our second objective and further explore the nature of the evoked, hold, reject and foggy sets, a multiple discriminant analysis was performed. All discriminating variables were entered into a stepwise discriminant procedure to classify the brands into one of the four sets. The discriminating variables were classified as either product class, brand specific or personal. The results of the discriminant analysis are summarized in Table 3.

The discriminant analysis once again tends to support the hypothesized profiles of the Brisoux-Laroche paradigm. Attitude, confidence, quantity of information processed, intention and whether an individual has tried a particular brand, all seem to be relatively good discriminating variables (p < .05). Table 3 reveals that the classificatory and predictive power of all the canonical discriminant functions were highly significant (p < .01) compared to the proportional chance criterion (Morrison, 1969 p. 156-163). The proportional chance criterion represents the conditional probability of classifying an individual correctly, given the relative set size of each of the groups.

Other brand specific variables which tend to be significant are perception of cavity protection, breath freshening and teeth whitening capabilities, as well as pleasant taste, dentist's approval and availability (p < .05).

TABLE 3

RESULTS OF DISCRIMINANT ANALYSIS ON A BRAND BASIS

Several product class variables were also found to be good predictor variables. Importance of taste, teeth whitening and breath-freshening, as well as frequency or purchase were all significant (p < .05). Finally, two personal variables, age and sex, tended to be able to discriminate between the groups for fifty percent of the brands studied.

In order to more fully understand the multiple discriminant analysis, it is necessary to analyze the results of the product class (importance variables) and brand specific (product attributes) variables separately. Upon inspection, it seems curious that "importance of cavity protection" is never found to be a significant predictor variable. The explanation is that this importance variable scores very high for most individuals (across all brands) and cannot discriminate between the groups. The managerial implication is that this criterion is a crucial factor for most individuals, even though it will not discriminate between the groups at this level. At the product attribute level, "perception of cavity protection" is shown to discriminate between the groups for three brands (Aim, Colgate and Ultrabrite). It seems then that for these brands when the perception of the evidence of cavity protection ingredients (i.e., Fluoride) is high, individuals who placed these brands in the evoked set did so in part due to this factor. Conversely, those who claimed these brands to be in their reject set, did so because of the perception of low cavity fighting protection.

For illustrative purposes, let us examine one brand in which "perception of cavity importance" is able to in part discriminate between the groups. For Aim the mean score on the product class variable "importance of cavity protection" for all four sets was 7.897 indicating a very high degree of importance for this factor. Since there were no significant differences in the mean scores on this variable between the groups (e.g., for evoked 7.90, xreject = 7.79, Xhold 7.85, Xfoggy = 7.93), this variable could not discriminate between the groups. However, at the brand attribute level, on the variable "perception of cavity protection" there are significant differences between the mean scores. For example, the mean score for "perception of cavity protection" for those who placed Aim in the evoked set was 7.18 (" = 50), and for those who placed Aim in their reject set, 5.42 (n = 19). This difference is significant beyond the .005 level. Testing the difference between x k d = 7.18 (n = 50) and xhold = 6.04 (n = 83) yields significant results, but slightly less significant than the difference between xevoked and xreject. This analysis is also supported by the classification results.

CONCLUSION

There are some limitations to this research which must be explicated: the categorization of sets or groups was defined a priori by the researchers, in terms of responses to the answers in the personal interview (see Appendix 1). The hold set was measured by deduction from the other three sets; the fact that respondents were to order their responses from the most to least preferred brands may have generated some consistency bias; and, due to small sample sizes, the discriminant analysis results could not be cross-validated against a hold-out sample.

In conclusion the results reported herein tend to generally support the Brisoux and Laroche brand categorization paradigm. However, additional research is necessary particularly on improving measurement, identifying rules for inclusion in the sets and the changes which take place over time. Finally, this paradigm needs to be tested for other product classes such as durables, and other situations such as extensive problem solving (Howard. 1977).

APPENDIX 1

APPENDIX 2

REFERENCES

Bettman, James R. (1379), An Information Processing Theory of Consumer Choice, Addison-Wesley.

Bettman, James R. and Park C. Whan (1980), "Effects of Prior Knowledge and Experience and Phase of the Choice Process on Consumer Decision Processes: A Protocol Analysis", Journal or Consumer Research, Vol. 7, 234-268.

Biehal, Gabriel and Chakravarti, Dipankar (1982), "Information-Presentation Format and Learning Goals as Determinants of Consumers' Memory Retrieval and Choice Processes," Journal of Consumer Research, Vol. 8, 431-441.

Brisoux, Jacques E. and Laroche, Michel (1980), "A Proposed Consumer Strategy of Simplification for Categorizing Brands," in Evolving Marketing Thought for 1980, eds. John D. Summey and R.D. Taylor, Southern Marketing Association, 112-114.

Brisoux, Jacques E. and Laroche, Michel (1981), "Evoked Set Formation and Composition: An Empirical Investigation Under A Routinized Response Behavior Situation," in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor: Association for Consumer Research, 357-361.

Campbell, Brian M. (1969), "The Existence and Determinants of Evoked Set in Brand Choice Behavior," Unpublished ?h.D. Dissertation, Columbia University.

Emery, Fret (1969), "Some Psychological Aspects of Price," in Pricing Strategy, eds. B. Taylor and G. Wills, London: Staples Press, 98-111.

Green, Paul E. and Srinivasan, V. (1978), "Conjoint Analysis in Consumer Research: Issues and Outlook," Journal of Consumer Research, Vol. 5, 103-123.

Howard, John A. (1963). Marketing Management, Analysis and Planning, Irwin.

Howard, John A. (1977), Consumer Behavior: Application of Theory, McGraw-Hill.

Howard, John A. and Sheth, Jagdish N. (1969), The Theory of Buyer Behavior, New York: Wiley.

Jarvis, Lance P. and Wilcox, James B. (1973), "Evoked set size-Some Theoretical Foundations and Empirical Evidence," in Combined Proceedings, ed. T.V. Greer, Chicago: American Marketing Association, 236-240.

Johnson. Eric J. and Russo, J. Edward (1981), "Produce familiarity and Learning New Information," in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor: Association for Consumer Research, 151-155.

Miller, F.A. (1956). "The Magic Number Seven, Plus or Minus Two: Some Limits on our Capacity for Processing Information," The Psychological Reviews Vol. 63 (March), 81-97.

Morrison. Donald G. (1969), "On the Interpretation of Discriminant Analysis." Journal of Marketing Research, Vol. 6. May. 156-163.

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

Nicosia, Francesco M. (1966), Consumer Decision Processes, Marketing and Advertising Implications, Prentice-Hall.

Russo, J. Edward and Johnson, Eric J. (1980), "What do consumers know about familiar products?," in Advances in Consumer Research, Vol. 7, ed. Jerry C. Olson, Ann Arbor: Association for Consumer Research, 417-423.

Sherif, Caroline W., Sherif, Muzafer and Nebergall, Robert E. (1965), "Attitude and Attitude Change: The Social Judgement Involvement Approach, Philadelphia: Saunders.

Wallace, Anthony F.C. (1961), "On Being Just Complicated Enough*" Anthropology, Vol. 47, 458-464.

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