Brand Categorization Strategies in an Extensive Problem Solving Situation: a Study of University Choice

ABSTRACT - This paper examines consumer brand categorization processes. Specifically, it deals with the Brisoux-Laroche brand categorization paradigm. In the context of university choice, the evoked, hold, foggy and reject set categorizations are determined. Moreover, the authors address the issues of order effect and of direct measurement of the set categorizations. The lack of concern for these methodological issues had previously been identified as potentia; sources of bias. The results of the present research suggest that: 1) the proposed paradigm is generally supported for a service product in an extensive problem solving situation; 2) order effects are negligible; and 3) consumers may not always categorize brands into mutually exclusive sets when self-administered questionnaires are utilized.

Citation:

Michel Laroche, Jerry Rosenblatt, and Ian Sinclair (1984) ,"Brand Categorization Strategies in an Extensive Problem Solving Situation: a Study of University Choice", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 175-179.

Advances in Consumer Research Volume 11, 1984      Pages 175-179

BRAND CATEGORIZATION STRATEGIES IN AN EXTENSIVE PROBLEM SOLVING SITUATION: A STUDY OF UNIVERSITY CHOICE

Michel Laroche, Concordia University

Jerry Rosenblatt, McGill University

Ian Sinclair, Concordia University

ABSTRACT -

This paper examines consumer brand categorization processes. Specifically, it deals with the Brisoux-Laroche brand categorization paradigm. In the context of university choice, the evoked, hold, foggy and reject set categorizations are determined. Moreover, the authors address the issues of order effect and of direct measurement of the set categorizations. The lack of concern for these methodological issues had previously been identified as potentia; sources of bias. The results of the present research suggest that: 1) the proposed paradigm is generally supported for a service product in an extensive problem solving situation; 2) order effects are negligible; and 3) consumers may not always categorize brands into mutually exclusive sets when self-administered questionnaires are utilized.

INTRODUCTION

A review of the literature concerning consumer research reveals that consumers' brand selection processes have been studied extensively (e.g., Bettman, 1979). How and why consumers move to simplify and limit the number of brands that are considered in a purchase decision is of major concern to marketing practitioners and theorists. Why consumers attempt to simplify their decision process has been addressed by Howard and Sheth (1969), camp-Dell (1969), Ostlund (1973), Jarvis and Wilcox (1973), Gronhaug (1973), in the field of marketing and by Miller (1956) and Wallace (1961) in the fields of psychology and anthropology respectively. However, how consumers move to simplify and limit the number of brands has been investigated to a lesser extent. Howard (1977) defined the evoked set as, "the subset of brands that a 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). This initial conceptualization, depicted in Figure 1, was integrated into the Howard-Sheth model of buyer behavior in the routinized response stage of consumer problem solving. This conceptualization has not been adequately tested for other problem solving situations such as limited and extensive problem solving. Questions relating to the stability and composition of the evoked set nave only recently been formulated (Belonax and Mittelstaedt, 1978; Green and Srinivasan, 1978; Bettman, 1979; Bettman and Park, 1980; Johnson and Russo, 1981; Biehal and Chakravarti, 1982; Laroche et al., 1983). The major shortcoming of the Howard (1977) conceptualization is that it is incomplete, in that it only identifies and categorizes those brands that are presently considers as purchase alternatives. There is no discussion of other brands.

Narayana and Markin (1975) expanded on the Howard conceptualization and identified those subsets of the awareness set --evoked, inert and inept (see Figure 1). The evoked set is analogous to Howard's, the inept set are those brands which are totally rejected, and the inert set contains those brands that are neither accepted nor rejected, and about which neither positive nor negative attitudes are held (Sherif et al., 1965).

Review of the Brisoux-Laroche Conceptualization

The definitions and hypothesized profiles of the evoked, hold, foggy and reject sets are summarized in Table 1 (Brisoux and Laroche, 1980; Laroche et al., 1983). The first set within the processed set is the evoked set. Attitudes, purchasing intentions, quantity or information processed and confidence with respect to the evaluation of these brands are all expected to be highest relative to brands in the other sets.

FIGURE 1

COMPARISON OF THREE CONSUMER CATEGORIZATION PROCESSES

TABLE 1

SUMMARY OF HYPOTHESES

The second category in the processed set is the hold set. While such brands are neither acceptable nor unacceptable alternatives at the present, consumers may have either positive, negative or neutral opinions about them: 1) the consumer may have a positive attitude toward the brand, but it is not presently perceived as being adequate to fulfill the consumer's motives; 2) the consumer may hold a negative attitude about the brand, but would consider it if its price, for example, was to drop low enough to make it a real bargain; and 3) the consumer may truly be neutral toward brands in the hold set; they are neither liked nor disliked, and over time through forgetting and lack of reinforcement, may move into the foggy set. Attitudes, purchasing intentions, quantity of information processed and confidence with respect to the evaluation of these brands, are expected to be at least as large, if not larger, than those brands belonging to the reject and foggy sets, and lower than those brands in the evoked set.

The next set in the processed set is the reject set. It contains those brands which the consumer would not consider when making a purchase decision. Attitudes and intentions are expected to be lowest for brands in this set compared to all other sets. 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, and higher than those in the foggy set.

The last category of brands is labelled the foggy set. The consumer is aware of brands in this set, and can identify these brands with the particular product class. However, the consumer has no specific comprehension about the brand. Such brands usually do not have significant meaning as they cannot really be distinguished in terms of the evaluative criteria of the product class. By definition, these brands have not been attribute processed. Attitudes 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 reject set. Confidence and quantity of information processed should be at their lowest for brands in the foggy set, since virtually no information on product attributes has been processed. Purchasing intentions are expected to be low relative to evoked and hold sets, but marginally higher than for brands in the reject set.

All these conceptualizations are presented in Figure 1. Narayana and Markin (1975) expanded the Howard model and made it more complete. The Brisoux-Laroche hold set is similar to the Narayana-Markin inert set in that the consumer is aware of and has an opinion of these brands. The major difference is that attitudes toward brands in the inert" set are completely neutral (i.e., = 0), while brands in the "hold" set may be positive or negative. Brisoux and Laroche (1980) expanded the conceptualization further by identifying another set of brands (foggy set) of which the consumer is aware, yet has no relevant information with respect to those brands, and therefore no opinions concerning those brands. This paper uses the Brisoux-Laroche paradigm to serve as the conceptual framework since it possesses the greatest degree of precision concerning the identification of consumer brand categorization processes.

University Selection As An Extensive Problem Solving Situation

Most research attempting to validate consumer brand categorization processes has focused on durable or non-durable products, mostly in routinized or limited problem solving situations. Little research has been conducted concerning the existence or composition of an evoked set in complex or extensive problem solving situations (Thompson & Cooper, 1979; Maddox, 1977). Similarly, there has been limited research to explain the purchases of services. Moreover, virtually no work has been published on the composition of the reject, hold and foggy sets in extensive problem solving situations. Consequently, the present research investigates the purchase of a service, i.e., university selection. in a complex problem solving situation.

In recent years there has been a substantial increase in the number of universities in Canada, in the number and sophistication of available programs and in the financial support available. This creates an information processing problem (Bettman, 1979) for junior college students who must thus simplify their selection dilemma.

While there has been extensive research into student motivation and decision behavior in selecting a university, there is little agreement on specific choice factors or models. Holland (19583 asked high school seniors why they selected the university in which they planned to enroll. Major reasons given included their perception of the university's academic reputation, cost, distance from home, and recommendations by others. Douvan and Kaye (1962) reported similar influences and also found that motivation to attend college is different for students of varying socio-economic levels. Richards and Holland (1965) reported four influential factors centering on intellectual emphasis, practicality, advice of others and social emphasis. Stordahl (1970) observed that motivational influences on a young adult to attend university involved a complex interplay of forces such as personal goals, ability, personality, parental values, socio-economic status and other environmental factors

Spekman, Harvey and Bloom (1980) conducted a mail survey of high school students considering attending university and found that university selection was a highly complex problem; that students selecting a university probably use a "process by brand" approach (Jacob r et al., 1976) rather than by attribute; that is, while major factors may mediate school choice, t-ne importance of these attributes is likely to change, depending on which university is being evaluated; that the university evaluation process may differ markedly across segments of students, with different decision rules and attribute saliences being used depending on a student's background, level of involvement, grade-point average and time away from university entry.

When considered overall, this body of research (only a few examples have been cited) illustrates the considerable complexity of the university selection problem facing young adults. This places it clearly within Howard's (1977) description of extensive problem solving, that is, one in which the product class is not well known and for which brand selection criteria are not yet well formulated. As a consequence much new information is necessary which requires a lengthy time to process and ultimately select a brand.

All of the above research has focused on "why" students choose to attend certain universities; this research focuses on 'Mow" those decisions are mate.

Methodological Issues

Previous experience has suggested that the order in which an individual is asked questions, with respect to how brands are categorized, may lead to an order bias. Since this research utilizes relatively complicated questions, with only minor differences in the wording of some of the questions (see Appendix 1), brands which are borderline between two sets (e.g., between hold and evoked) may be placed in the set to which the consumer responds first. This research employed the questionnaire format as opposed to card selection. The card selection procedure eliminates order bias since all categorizations are made simultaneously. In order to assess the degree of measurement error in the survey, twelve different versions of the questionnaire were developed, and are presented in Appendix 1.

In addition, all previous measures of the four sets have been derived by asking respondents to view a set of cards and have them place each card in one of the various sets. Also, never before have there been direct measures of all four sets. In the past, either the hold or foggy set has been deduced from the others. This research uses the questionnaire method to obtain the desired information. Moreover, direct measures for all four sets were obtained. The authors recognize the problem of using single-item measures for such complex constructs, however lengthening the questionnaire to obtain multiple measures would have only served to make a long and taxing questionnaire even longer.

METHODOLOGY

The objectives of this research are three-fold: 1) to further examine the Brisoux-Laroche brand categorization paradigm. Specifically, to test the hypothesized profiles of the evoked, hold, reject and foggy sets for a service product in an extensive problem solving situation; 2) to determine if the ordering of questions significantly effects the manner in which respondents categorize brands; and 3) to evaluate the use of the questionnaire method to obtain direct measures of the evoked, hold, reject and foggy sets.

Sample

The research reported herein was conducted in a Montreal post secondary junior college in February, 1983. A quota sampling technique was employed in order to test one of the basic research questions concerning order effect. It was hypothesized that the order in which respondents answered the four questions measuring the four sets would not significantly effect the overall results. In order to test this hypothesis, twelve versions of the questionnaire were utilized. The ordering sequence along with all operational definitions of all the variables reported in this paper are presented in Appendix 1.

Data Collection

Questionnaires were handed out to junior college students in ten different classes. An attempt was made to assure that all respondents were at roughly equivalent stages in their decision-making processes. To this end, only first year junior college students enrolled in a two-year business program were asked to participate in the study. After a two week period of handing out questionnaires, 392 completed and usable questionnaires were obtained. Often in marketing research studies students are used as subjects, due to limited budgets and time constraints. Here, the use of students is most appropriate, since it is their attitudes and opinions that are in fact desired.

The measurement of the evoked, reject, hold and foggy sets was adapted and modified from those used by Brisoux and Laroche (1980) and Laroche et al. (1983), and are presented in Appendix 1. In this research, however, all sets were measured directly. The first section of the questionnaire measured the awareness set, as well as the four sets of prime importance. The second section measured the student' perceptions of the various universities in question across a number of dimensions (brand specific variables in Appendix 1). All dimensions, attributes and universities were selected after an initial pilot study had identified the major issues and concerns.

FINDINGS AND INTERPRETATIONS

Test of Hypothesized Profiles

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

TABLE 2

MEAN SIZE, ATTITUDE, INTORMATION, INTENTION, CONFIDENCE FOR BRANDS IN THE EVOKED, REJECT, HOLD AND FOGGY SETS

The hypothesized profile of the evoked set was well supported. The measures of mean attitude, intention, confidence and quantity of information processed were all significantly higher for universities in the evoked set than those in all other sets (all significant at p < .01) . This supports the notion that consumers have a highly developed concept and well formulated opinions of; as well as evaluate positively, brands in the evoked set. In addition, the smallest set size was found for the evoked set, and the largest was the foggy set. This illustrates that the number of universities being considered from the available set was small in comparison to the others, and significantly smaller (p < .01) than those universities about which students had either rejected or were uncertain about.

The hypothesized profile of the hold set was also well supported. The measures of mean attitude, quantity of information processed, intention and confidence were all significantly higher (p < .01) for universities in the hold set as compared to those in the reject and foggy sets. This finding tends to confirm the hypothesis of the Brisoux-Laroche paradigm that there is a need to break up the sets originally proposed by Narayana and Markin (1975) into the processed set (consisting of the evoked, hold and reject sets) and the foggy, or unprocessed set.

The hypothesized profile of the reject set was generally supported. Measures of mean attitude and intention were significantly lower (p<.01) for universities in the reject set as compared to all other sets. As hypothesized, quantity of information processed was higher for universities in the reject set as compared to those in the foggy set, however no significant differences between the two sets were observed. Moreover, confidence for universities was found to be slightly lower for universities in the reject set as compared to those in the foggy set, however no significant difference was found. Biehal and Chakravarti (1982) reported that brands can be rejected with relatively little information and low confidence. Laroche et al., (1983) inferred that their findings of low confidence and little information for brands in the reject set suggests that the categorization of brands in the reject set involves a limited degree of problem solving. The present research found confidence for brands in the reject set to be much higher than previous results (Laroche et al., 1983), which suggests that there may be a difference in the role and effect of confidence on the decision making process across various problem solving situations. This research suggests that slightly more confidence is required to reject brands when a complex decision is at hand as compared to limited or routinized problem solving situations. In addition the reject set reflects virtually the least amount of information processing. This suggests that the categorization of the reject set brands may be according to a disjunctive, as opposed to a conjunctive decision rule which has been found for evoked set brands (Brisoux and Laroche, 1981).

The hypothesized profile of the foggy set was also generally supported. Mean attitude and intention were significantly higher for universities in the foggy set as compared to those in the reject set. As hypothesized, quantity of information processed was lowest for brands in the foggy set, which tends to support the notion that brands in the foggy set have not been attribute processed (Brisoux and Laroche, 1980).

Test of Order Effect

Previous research had not considered the issue of the ordering of the questions measuring the evoked, hold, foggy and reject sets. To test for any order effect, twelve versions of the questionnaire, each with a minimum of thirty respondents were utilized. The results are presented in Table 3 which reveals the range of means for set size, attitude, information, intention and confidence for each of the four sets, for all twelve versions. In only five of the enumerable instances was any order effect detected. (Scheffe Multiple Range Test at p < .05) . An investigation into the specific cases where an order effect was found revealed no consistent pattern, and thus the hypothesis that there is no order effect could not be rejected.

TABLE 3

RANGE OF MEAN ATTITUDE, INFORMATION, INTENTION, CONFIDENCE, SIZE FOR BRANDS IN THE EVOKED, REJECT, HOLD AND FOGGY SETS FOR 12 VERSIONS

The use of questionnaires and direct measures of all four sets had some unpredicted side effects. In some instances where were inconsistencies in the responses between the consumers' original unawareness set, in that some universities which were initially classified as 'unaware' by virtue f not having been placed in the awareness set, were evaluated and later placed in one of the tour sets. in addition, there was a high degree of duplication of responses. or example, some universities were placed both in the reject and foggy sets. The greatest degree of duplication was found between the evoked and hold sets, and the foggy and reject sets. Due to the nature of the research design, duplication between the hold and foggy sets was minimal.

The occurrence of duplication may not be a methodological

problem resulting from an insensitive measurement instrument. Rather, it may be that consumers do not possess sharply defined evoked, hold, foggy or reject sets. This lack of precision in set composition has been addressed by Bellman and Zadeh (1970). These authors claim that, "... a fuzzy set is a class of objects in which there is no sharp boundary between those objects that belong to the class and those that do not." (P. B-143). This suggests that the categorization process is probably more dynamic than hitherto assumed, and that if categorization criteria are not firmly established, as would be expected in an extensive problem-solving situation, brands may not be classified definitely into mutually exclusive sets.

CONCLUSION

The results reported herein tend to support the Brisoux-Laroche brand categorization paradigm. The conceptualization of brands categorized into evoked, hold, reject and foggy sets has been well supported for a service product and in an extensive problem solving situation. This investigation also suggests that concern over the ordering of questions should be minimal, at least under these conditions. The consistency of these results with previously published reports suggests that the use of self-administered questionnaires (rather than lengthy and costly personal interviews), while not without its problems, is appropriate and has several advantages. Future research might use mail surveys, which would increase the types of products and product classes that could be studied.

There are some limitations to the research presented herein. The testing of the order effects was not a complete design, and thus in the future complete designs could be employed; and, of much greater importance, the high degree of duplication that was found must be further investigated. Moreover, if one were to consider the duplication phenomenon as a methodological problem, it could be reduced by questionnaire format redesign which prevents recategorization of brands already classified. However this might artificially force discrete judgements which might not be reflective of reality. It is suggested that the dynamic nature of set formation, stage and type of decision-making processes, as well as the possible multidimensional aspects of set composition be studied in the future.

APPENDIX 1

REFERENCES

Bellman, R.E. and Zadeh, L.A. (1970), "Decision-Making i Fuzzy Environment,' Management Science, Vol. 17, No. 4, December. B-141-164.

Bettman, James R. (1979), An Information Processing Theory of Consumer Choice, Addison-Wesley.

Bettman, James R. and Park, C. Whan (1980), "The Effects of Prior Knowledge and Experience and Phase of the Choice Process on Consumer Decision Processes: A Protocol Analysis," Journal of Consumer Research, Vol. 7, 234-248.

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. Jonn 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 Common Research, Vol. VIII, ed. Kent 8. 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 Ph.D. Dissertation. Columbia University.

Douvan, Elizabeth and Kaye, Carol (1962), 'Motivational Factors in College Entrance,'l in N. Sanford (ed.), The American College, New York: Wiley, 199-224.

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

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

Holland, John L. (1958), Student Explanations of College Choice and Their Relation to College Popularity, College Productivity and Sex Differences, College and University, Vol. 33, 313-320.

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

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

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

Jacoby, J.R., Weigl, K. and Fisher, W. (1976), "Pre-Purchase Information Acquisition: Description of a Process Methodology, Research Paradigm and Pilot Investigation," in Advances in Consumer Research, Vol. III, ed. Beverlee B. Anderson, Association for Consumer Research, 424-428.

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

Kohn, Meir G., Manski, Charles and Mundel, David (1976), "An Empirical Investigation of Factors Which Influence College-Going Behavior," Annals of Economic and Social Measurement, Vol. 5, 4, 391-419.

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

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

Narayana, them L. and Markin, Rom (1975), ';Consumer Behavior and Product Performance: An Alternative Conceptualization," Journal of Market , 35, October, 1-6.

Punj, G.N. and Staelin, Richard (1978), "The Choice Process for Graduate Business Schools," Journal of Marketing Re- search. Vol. 15. November, 588-593.

Richards, James and Holland, John (1965), A Factor Analysis of Student Expectations of Their Choice of a College, ACT Research Reports, No. 8, ACTP.

Russo, J. and Johnson, Eric (1980), "What Do Consumers Know About Familiar Products," in Advances in Consumer Research, Vol. VII, ed. Jerry Olson, Ann Arbor: Association for Consumer Research. 417-423.

Sherif, C.W., Sherif, M. and Nebergall, R.E. (1965), Attitude and Attitude Change: The Social Judgement Involvement Approach, Philadelphia, Saunders.

Spekman, R.E., Harvey, James and Bloom, Paul (1980), "The College Choice Process: Some Empirical Results," in Advances in Consumer Research, Vol. VII, ed. Jerry Olson, Ann Arbor: Association for Consumer Research, 700-704.

Stordahl, Kalmer E. (1970), "Student Perceptions of Influences on College Choice," The Journal or Educational Research. Vol. 63, 5, January, 209-212.

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

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