Brand Categorization and Product Involvement

ABSTRACT - The effects of product involvement on brand categorization are examined with respect to brand set size. Data collected from 487 female users of cosmetics did not confirm the effect of product involvement on the evoked, reject, foggy, and salient set sizes. Significant effects were found for the awareness, trial and hold sets. Although the high and the low involvement groups did not differ on socio-demographic variables, high involvement seemed associated with heavier users and higher brand loyalty.


Jacques E. Brisoux and Emmanuel J. Cheron (1990) ,"Brand Categorization and Product Involvement", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 101-109.

Advances in Consumer Research Volume 17, 1990      Pages 101-109


Jacques E. Brisoux, Universite du Quebec a Trois-Rivieres

Emmanuel J. Cheron, Universite Laval

[The authors gratefully acknowledge the financial support of FCAR and the assistance of Helene Hamel in the research project.]


The effects of product involvement on brand categorization are examined with respect to brand set size. Data collected from 487 female users of cosmetics did not confirm the effect of product involvement on the evoked, reject, foggy, and salient set sizes. Significant effects were found for the awareness, trial and hold sets. Although the high and the low involvement groups did not differ on socio-demographic variables, high involvement seemed associated with heavier users and higher brand loyalty.


The question of how consumers deal with the multitude of brands that are available in many product categories has attracted increasing attention by researchers in recent years. Several brand categorization models have been proposed in the marketing literature. The Brisoux and Laroche (1980) model is an expansion of earlier conceptualizations by Narayana and Markin (1975) and Howard (1963, 1977). The model suggests four stages in the consumer decision process: awareness, processing, consideration, and preference (Figure 1).

Brand Categorization

From the available set brands are first classified in either the awareness set or the unawareness set. The former is divided in a processed set and a unprocessed (or foggy) set. Brands in the foggy set are presumed to have not been processed on any of the salient attributes (the consumer has no specific brand comprehension). Brands belonging to the processed set are either evoked, rejected, or in a "hold" position. The evoked set is analogous to the Howard and Narayana & Markin conceptualizations. It includes the subset of (processed) brands that a consumer considers as acceptable purchase alternatives. The reject set contains those brands which the consumer considers to be unacceptable purchase alternatives. Brands in the hold set are not considered as purchase alternatives, and consumers may have positive, negative. or neutral attitudes towards them. For a more complete description of the basic framework and managerial implications of the brand categorization model, the reader is referred to Laroche et al. (1986).

The framework used by Brisoux and Laroche (1980), and later by Laroche et al. (1983, 1984) and Brisoux et al. (1986) differentiated the evoked, hold, reject, and foggy sets in terms of attitudes, intentions, confidence and brand evaluation, as well as quantity of information processed. The model has been partially tested and generally supported using various product classes, in routinized response behavior (Brisoux and Laroche 1980, Laroche et al. 1983, Brisoux et al. (1986) as well as extensive (Laroche et al. 1984) and limited problem-solving situations (Church et al. 1985).

Product Involvement

The concept of involvement has been a major center of interest in consumer research literature for the past 20 years. There is no commonly accepted definition of this hypothetical construct although a (temporary) generally acceptable generic definition and a few characteristics are proposed by Rothschild: "Involvement is a st',e of motivation, arousal or interest. This state exists in a process. It is driven by current external variables (the situation, the product, the communications) and past internal variables (enduring; ego; central values). Its consequences are types of searching, processing and decision making" (Rothschild, 1984, p. 217). The author argues that, for the time being, the field does not need any more theoretical papers or literature reviews on the topic. We will therefore limit our presentation to the main elements necessary to the understanding of our study.

Table 1 presents a summary of the global concept of involvement. This table is based on the Houston and Rothschild (1978) paradigm which subdivides the concept into three types: situational, enduring, and response involvement. The paradigm includes a listing of the stimuli which have the ability to arouse each type of involvement, their antecedents, the hypothesized effects and the proposed measures. Other well known recent studies on the construct are cited in Table 1 (Traylor and Joseph, 1984; Laurent and Kapferer, 1985; Zaichkowsky, 1985).

Table 2 presents a comparison of three consumer involvement measuring instruments that have been recently proposed (Traylor and Joseph, 1984; Laurent and Kapferer, 1985; Zaichkowsky, 1985). The comparison is done on the prominent underlying concept, the number of product categories used for testing each measuring instrument and corresponding sample sizes, the category of respondents, the type of scale, the number of dimensions being measured, the number of items in each scale, and the observed Cronbach alphas. The table also indicates the type of validity testing conducted on each scale (concurrent, trait, discriminant, content, construct and convergent).

Rosenblatt (1985) has investigated the relationship between involvement and its effects on brand categorization but has not specifically examined the effect of involvement on brand set size and content.




The aim of this research is to test the relationship between the level of product involvement and the structure of brand categorization. More specifically, the research examines differences in brand set sizes for low, medium and high involvement consumer groups.

This research can be related to the validity network schema proposed by Brinberg and Hirschman (1986). The focus of this research is first to combine the conceptual and the empirical domains to form a theory (Figure 2).

The theory is then submitted to test using specific methodological approaches. The conceptual domain comprises the construct of brand categorization and the involvement construct. The empirical domain refers to consumer characteristics and make-up products for eyes, lips and the face. The methodological domain includes the structure of brand categories proposed by Brisoux and Laroche (1980) and the measure of product involvement developed by Zaichkowsky (1985).

Since the emphasis of this research is to test a theory, greater attention will be invested in concept definitions and methodological procedures. Although every effort will be made to get a relevant sample, its convenient nature is acceptable with this type of research (Calder et al., 1981).


Following the literature review and the research objectives, a series of specific hypotheses will be tested. The hypotheses are stated for the limited problem-solving (LPS) situation and will be supported or rejected in the empirical context of cosmetics purchase. Four hypotheses pertaining to the relationship between product involvement level and brand set structure are proposed:

H1 Consumers with higher level of product involvement will be associated with smaller evoked sets.

H2 Consumers with higher level of product involvement will be associated with larger reject sets.

These above hypotheses are justified by previous research confirming such relationships for other products (Brisoux, 1980; Ho, 1980; Traylor and Joseph, 1984; Rosenblatt, 1985).

H3 Consumers with higher product involvement will be associated with smaller foggy sets.

This hypothesis is the result of the unprocessed characteristic of the foggy set, which will be associated with a low level of involvement.

H4 Consumers with higher product involvement will be associated with larger salient (unaided awareness) sets, larger awareness sets, larger trial sets and larger hold sets.

Hypothesis 4 is proposed on since the relationship between product involvement and trial set size has been previously examined by Zaichkowsky (1985). And, as an extension, higher involvement will be tested in association with a larger salient, trial and hold sets.

Table 3 presents a detailed summary of the four hypotheses between product involvement and brand set size.








Product choice

The conceptual domain of the research specifies that the chosen product be associated with a limited problem-solving situation and that at least seven distinct brands be included in the product class (Miller, 1956). In general, make-up for women refers to a limited problem-solving situation except for a few highly brand loyal customers with routinized buying behavior. A total of 80 different brands of make-up are listed in Cosmetics (1985). The total market for cosmetics is relatively stable with a rather high competitive activity, each manufacturer fighting for his market share.

Furthermore, access to users of the product is easy, since most women are consumers of make-up products. Managers in the industry also believe that the brand is important in the buying decision (Wood, 1986). Frequent buying and using of the product make it easier for consumers to evaluate brands. According to a convenience sample of beauty consultants, special occasions do not have an influence on the choice of products. Therefore controlling for the buying situation was not necessary.

In this research, make-up products include all products that can be used on the eyes, the lips, or the face. To the authors' knowledge, brand categorization has not been previously examined for this product class.

Definition of variables

This section specifies the operational definitions of the different brand sets and the product involvement variable. In addition to the four sets conceptualized in figure 1, Brisoux and Laroche (1980) measured the salient and trial sets and used the following definitions:

Salient set: Those brands of the product class that come to the mind of the consumer without the help of a list of brands.

Awareness set: Those brands that the consumer can recognize from a list of brands.

Evoked set: From the awareness set, those brands that the consumer would certainly consider to buy at the time of the survey.

Reject set: From the awareness set, those brands that the consumer would refuse to buy as they seem unacceptable to his or her needs.



Foggy set: From the awareness set, those brands for which the consumer has not formed an opinion (no attribute processing has occurred) and which he or she is not ready to buy.

Hold set: From the awareness set, those brands for which the consumer has formed an opinion (some attribute processing has occurred) but for which he or she cannot say whether he or she would accept or reject them.

Trial Set: From the awareness set, those brands that the consumer has already bought and used for him or herself.

The definition of product involvement is taken from Zaichkowsky (1985): "A person's perceived relevance of the object based on inherent needs, values and interests". This definition contains the general viewpoints of several researchers (Krugman, 1967; Clarke and Belk, 1978; Mitchell, 1979; Greenwald and Leavitt, 1984; Rothschild, 1984). The operational measurement of product involvement is derived from the Personal Involvement Inventory (PII) scale developed by Zaichkowsky (1985). A presentation of the measuring instrument follows in the next section.

Measuring instrument

Personal interviews were used in previous research on brand categorization (Brisoux and Laroche, 1980; Rosenblatt, 1985; Brisoux, Cheron and Fernet, 1986). However for this research, a self-administered questionnaire was designed in order to obtain a fairly large sample despite a tight budget. The average time to complete the questionnaire on 45 brands was 30 minutes.

In the first part of the questionnaire respondents were asked to classify brands following the order of brand set definitions presented earlier. The second part was a French adaptation of the PII. This measuring instrument was chosen due to better consistency and validity (see Table 2). A back-translation procedure was used twice with three translators.

The French adaptation of the PII met most of the standards of the original scale. Internal reliability with a Cronbach alpha of 0.90 was comparable to results (above 0.90) reported by Zaichkowsky (1985). However, the unidimensionality of the adapted scale was not confirmed. A factor analysis produced four factors with eigenvalues above one. All factors explained up to 57.7% of variance instead of one factor accounting on the average for 70% of variance in the original study. The four factors could be interpreted as: 1) Derogative items (37.5%), 2) Product importance (8.6%), 3) Personal values (6.6%), 4) Visible appearance (5.0%). A multidimensional structure of the adapted scale of product involvement is supported by Kapferer and Laurent (1985).

Although with a larger mean (109 vs 90) and median (112 vs 96) the distribution of PII scores (N = 481) for cosmetics was skewed towards higher scores such as the overall distribution (N = 751) for various products obtained by Zaichkowsky (1985). The mean PII score of 109 for cosmetics was comparable to mean scores of 99 for jeans and 112 for calculators in the original study. Using the cutting points of the original study, three groups were formed with low product involvement level (PII score of 69 or below) medium level (PII score between 70 and 110) and high level (PII score of 111 or above).



The third part of the questionnaire contained classification variables related to make-up consumption such as: brand loyalty, self-designated reference group (homemaker, working woman, career woman), frequency of beauty services, allergy problems, seasonal amount spent for cosmetics and age group.

Sampling and data collection.

The target population included women 18 years on older using make-up products at the time of the survey. The observed population was limited to mothers with children in primary schools of the Quebec City metropolitan area and women listed in the Quebec Association of Business Women.

Overall, 487 completed questionnaires were obtained from a total of 696 questionnaires distributed. According to the standard definition suggested by Wiseman and Billington (1984), the response rate of eligible respondents was 70%.

Data analysis

Standard descriptive analyses were first conducted to get frequency distributions of the sample classification variables. Frequency analyses were also performed for brand sets and personal involvement scores. The sample classification variables were also examined for low versus high involvement. Figure 3 indicates the overall data analysis structure.

Three groups of personal involvement levels were formed using the 25%, 50% and 75% percentiles. The low-involvement group included PII scores of 99 or below, the medium level group scores from 100 to 122 and the high level group scores above 122. Out of 481 valid cases, the average involvement score was 109 with a minimum value of 44 and a maximum of 139 (possible scale range is from 20 to 140).

A one-way multiple analysis of variance (MANOVA) was conducted. Although the seven dependent variables did not appear to be independent, examination of the data indicated substantial overlapping among sets. Therefore, the awareness set is not a linear combination (summation) of the evoked, reject, foggy, hold and trial sets. Hence, the assumption of independence for the MANOVA was not considered violated. Subsequently a series of ANOVA were used to test the effect of product involvement level on brand set size as specified in hypotheses 1 to 4. Multiple comparisons were examined using the Scheffe procedure, appropriate when samples are of different sizes and allowing to test for an overall significance level.


Since some parts of the questionnaires were not completed by all respondents, data analysis will be based upon 456 individuals. As for age groups, 58.5% of the respondents were in the 30 to 39 category. Self-ascribed reference groups were: 36.9% homemaker, 28.2% working woman and 34.6% career woman. Beauty services were "occasionally" required 47.6% of the time (vs "never", "once", and "regularly"). Some 66.4% of the respondents had no allergy problems and 73.4% purchased annually less than $100 of make-up products.

Table 4 indicates the average brand set sizes. The average evoked set size of 4.70 is similar to the salient set size (4.57) and is comparable to 4.83 found for jeans in a previous study by Fernet (1986). The average awareness set size of 24.12 is however larger than for jeans (15.65). A previous study by Reilly and Parkinson (1985) reports average evoked set sizes of 3 to 4 (hand lotion: 3.10, toothpaste: 3.11, shampoo: 3.42, toilet soap: 4.09).



To test the effect of product involvement on the seven brand set sizes taken together as dependent variables, a one-way MANOVA was performed. Results indicated an overall significant statistical difference between the three levels of product involvement (Hotellings' approximated F(14,892) = 14.0 with p = 0.000). The F values for each univariate ANOVA to test hypotheses 1 to 4 are shown in Table 4. Significant differences are observed for the awareness, trial and hold sets. Hypotheses 1, 2, 3 and 4 (for the salient set) are not confirmed. Although not statistically significant, the hypothesized direction of brand set size is correct for the reject set and the salient set.


The effect of product involvement on brand set size was not found to be significant for evoked, reject, foggy and salient sets. The range and skewness of observed involvement scores may account for the failure to confirm most hypotheses. Product involvement had a significant effect on awareness, trial and hold sets. Although, low and high involvement groups did not differ on sociodemographic variables, high involvement seemed associated with heavier users and higher brand loyalty .

The main contributions of this paper lie in conceptual definitions and methodological procedures. Brand categorization was measured with a self-administered questionnaire and a French adaptation of the PII scale was developed and tested.

Further research results will be presented elsewhere with the available data. Specifically, the brand categorization structure will be examined for mutually exclusive sets. Brand set contents and brand set relative position will be examined. Data were also collected on product attributes and perceived elements of the marketing mix of various brands. Managerial results will also be presented comparing the marketing mix profiles of brands belonging to different set categories.


Brinberg, David and Elizabeth Hirschman (1986), "Multiple Orientations for the Conduct of Marketing Research: An Analysis of the Academy/Practitioner Distinction," Journal of Marketing, 50 (October), 161-173.

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

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

Calder, Bobby, Lynn Phillips and Alice Tybout (1981), "Designing Research for Applications," Journal of Consumer Research, 8 (September), 197-207.

Church, Nancy J., Michel Laroche and Jerry A. Rosenblatt (1985), "Consumer Brand Categorization for Durables with Limited Problem Solving: An Empirical Test and Proposed Extension of the Brisoux-Laroche Model," Journal of Economic Psychology, 6, 231-253.

Clarke, Keith, Belk, Russell W. (1978), "The Effects of Product Involvement and Task Definition on Anticipated Consumer Effort," Advances in Consumer Research, 5, H. Keith Hunt. Ed., Ann Arbor. MI., Association for Consumer Research, 313-318.

Cosmetics (1985), "Make-up Chart," Ronald A. Wood. Ed., 13, 6 (July-August), 84-95.

Greenwald, Anthony G., Leavitt, Clark (1984), "Audience Involvement in Advertising: Four Levels," Journal of Consumer Research, 11 (June), 581 -592.

Ho, E. (1980), 'The Relationship Between Evoked Set Size and Ego Involvement in Product Class," Unpublished Manuscript, Cleveland State University. Cleveland, Ohio.

Houston, Michael J., Rothschild, Michael L. (1978), "Conceptual and Methodological Perspectives on Involvement," in 1978 Educators' Proceedings, Ed., S.C. Jain, Chicago: American Marketing Association, 184- 187.

Howard, John (1963), Marketing Management, Analysis and Planning, Homewood Illinois: Richard D. Irwin, Inc.

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

Kapferer, Jean-Noel, Laurent, Gilles (1985), "Consumers' Involvement Profile: New Empirical Results," Advances in Consumer Research, 12, Elizabeth C. Hirschman and Morris B. Holbrook Eds., Association for Consumer Research, 290295.

Laurent, G. and J.N. Kapferer (1985), "Measuring Consumer Involvement Profiles," Journal of Marketing Research, XXII, 41-53.

Krugman, Herbert E. (1967), "The Measurement of Advertising Involvement." Public Opinion Quarterly, 30 (Winter), 583-596.

Laroche, Michel, Jerry A. Rosenblatt, Jacques E. Brisoux, and Robert Shimotakahara (1983), "Brand Categorization Strategies in RRB Situations: Some Empirical Results," in Richard P. Bagozzi and Alice M. Tybout Eds., Advances in Consumer Research, Ann Arbor, MI: Association for Consumer Research, 10, 549554.

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

Laroche, M., J.A. Rosenblatt and J.E. Brisoux (1986), "Consumer Brand Categorization: Basic Framework and Managerial Implications," Marketing Intelligence & Planning, 4, No 4, 6074.

Miller, F.A. (1956), "The Magic Number Seven, Plus or Minus Two," The Psychological Review, 63, 81 -97.

Mitchell, A. (1979), "Involvement: A Potentially Important Mediator of Consumer Behavior," Advances in Consumer Research, 6, W.L. Wilkie, Ed., Association for Consumer Research, 191-196.

Narayana, Chem L., Markin, Rom J. (1975.), "Consumer Behavior and Product Performance: An Alternative Conceptualization," Journal of Marketing, 39, No 4 (October), 1-6.

Reilly, M., Parkinson, T.L. (1985), "Individual and Product Correlates of Evoked Set Size for Consumer Package Goods," Advances in Consumer Research, 12, Elizabeth C. Hirschman and Morris B. Holbrook Eds, Provo, UT, Association for Consumer Research, 492496.

Rosenblatt, Jerry Allan (1985), "Interactive Effects in Brand Categorization: The Relationship of Involvement and the Evoked Set Phenomenon," A Thesis in the Department of Marketing, Concordia University, Montreal (April).

Rothschild, M.L. (1984), "Perspectives on Involvement: Current Problems and Future Directions," Advances in Consumer Research, 11, T.C. Kinnear Ed., Association for Consumer Research, 216-217.

Traylor, Mark B. and W. Benoy Joseph (1984), "Measuring Consumer Involvement in Products, Developing a General. Scale," Psychology and Marketing, 1, No 2 (Summer), 65-77.

Wiseman, F. and M. Billington (1984), "Comment on a Standard Definition of Response Rate," Journal of Marketing Research, 21 (August), 336338.

Wood, R.A. (1986), "CCTFA's Beauty Retailing Seminar," Cosmetics, (May-June), 106.

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



Jacques E. Brisoux, Universite du Quebec a Trois-Rivieres
Emmanuel J. Cheron, Universite Laval


NA - Advances in Consumer Research Volume 17 | 1990

Share Proceeding

Featured papers

See More


How the Unconstructed Identity Relieves Consumers of Identity-Relevant Consumption

Tracy Rank-Christman, University of Wisconsin - Milwaukee, USA
Lauren Poupis, Iona College

Read More


How Numeric Roundness Influences Probability Perceptions

Julio Sevilla, University of Georgia, USA
Rajesh Bagchi, Virginia Tech, USA

Read More


Visualizing Price Magnitude: How Slider Scales Change Willingness-to-Pay

Manoj Thomas, Cornell University, USA
Ellie Kyung, Dartmouth College, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.