Must Consumer Involvement Always Imply More Information Search?

ABSTRACT - Current theory holds that high consumer involvement in a product's purchase will lead to extensive pre-choice information search. A modification in this theory is proposed, namely that information search is high only when the product is functional or utilitarian; when the product serves psycho-social or expressive goals, the consumer would not seek much information, a high level of involvement notwithstanding. Data from 231 housewives supported the hypotheses. Implications for future research are outlined.


Banwari Mittal (1989) ,"Must Consumer Involvement Always Imply More Information Search?", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 167-172.

Advances in Consumer Research Volume 16, 1989      Pages 167-172


Banwari Mittal, Northern Kentucky University

[Financial support for this work came in part from the marketing Science Institute under a grant to C. Whan Park and myself. The author is grateful to C. Whan Park, Jerry Zaltman, and Irene Frieze for guidance during the research. The present paper is excerpted from Mittal (1983a).]


Current theory holds that high consumer involvement in a product's purchase will lead to extensive pre-choice information search. A modification in this theory is proposed, namely that information search is high only when the product is functional or utilitarian; when the product serves psycho-social or expressive goals, the consumer would not seek much information, a high level of involvement notwithstanding. Data from 231 housewives supported the hypotheses. Implications for future research are outlined.

The consumer choice literature now recognizes the important role of involvement as a determinant of how much brand information the consumer will search. The more involved consumers are deemed to seek and utilize more information about brands (Lastovicka and Gardner 1979, Mitchell 1981, Bloch, Sherrel, and Ridgway 1986). This paper proposes and investigates one modification in the current research on this topic. We propose a broad classification scheme of products -- namely, functional and expressive, and hypothesize that pre-choice information search differs across these two product types, high levels of consumer involvement notwithstanding. Such a product classification is implicitly recognized (not necessarily under these labels) by marketing professionals, but its information search effects have not been- formally investigated before.

Differences between Functional and Expressive Products

It is proposed that the information search effects of involvement differ between what will be called here functional and expressive products. These products are best described, borrowing Udel's words, as those where a consumer derives satisfaction from the physical performance of the product (functional) and those where consumer satisfaction stems from his/her social and psychological interpretation of the product (expressive) (Udel 1964). A product is functional when its physical performance is of overriding concern. It is expressive when such physical performance is either of secondary importance or is assumed away, and when the consumer focus instead turns to the expressive quality of choice objects. Expressive quality refers to a product's ability to express its user's personality, or self-concept, or mood, as well as enable a consumer to experience these entities.

By these definitions, detergent, headache remedy, vacuum cleaner, etc., will be examples of functional products; perfume, designer clothes, souvenirs, etc., will be examples of expressive products. These are our expectations about averages; the meaning of a product for different consumers could of course range widely.

Our motivation for examining the functional/ expressive distinction in the context of involvement came from the fact that in the literature involvement has been seen as inevitably associated with a great deal of information seeking and high cognitive effort. To single out one example, Houston and Rothschild (1977) suggest involvement matrix as a measure of response involvement; according to this measure a highly involved person will examine many features of a product or many issues in a political race. That might well be, but the case of a blind-faith, non-issue based, emotional voter, and of a consumer purchasing perfume or a music album come to mind. The person in these instances may "care' a lot, (i.e., be involved) and yet manifest little "cognitive" activity; rather his choice may be dictated by emotional enchantment with an alternative. Accordingly, the involvement in the purchase of expressive products is here termed as expressive involvement, and that in the purchase of functional products as functional involvement. Although most products will-have a central tendency of being deemed functional or expressive by a majority of consumers, this involvement typology is not attached to products per-se. Rather, an individual consumer's orientation toward a product will determine the typology. Thus, it is a person-product dyad variable.

Previous researchers have linked involvement to the extent of information processing (e.g., Beatty and Smith, 1987). We hypothesize that although involvement will influence information search during brand decision process, such influence will be mediated by the functional/expressive character of involvement. Specifically, when involvement is expressive, the psychological interpretations of the product rather than its inherent features are of relevance. And, one can explicitly seek information about the latter, while the former, i.e., the psychosocial interpretation is largely idiosyncratic and less susceptible to explicit information search (c.f. Markus 1977, Holbrook and Hirschman 1982, Mittal 1983a, 1983b, 1987). Therefore, it was hypothesized here that in the case of expressive involvement compared to functional involvement, consumers will:

Ha: Use fewer sources of information

Hb: Do less extensive brand comparisons, and

Hc: Examine fewer brand features.



Data for this research came from a mail survey of 300 consumers randomly sub-sampled from the Consumer Panel of the Bureau of Business Research University of Pittsburgh. 240 responses were returned within 3 weeks, ar.d no follow-up was considered necessary. Of these, 231 questionnaires were usable.

In this survey, female members of the household, who were identified by name, were asked to rate statements of involvement and choice processes for each of the 15 specified products selected to represent a wide range of involvement: detergent, jeans, automobile, stereo, perfume, headache remedy music album, dry cell batteries, facial tissue. canned peas, salt, light bulbs, pen, and dry milk.


Independent Variables. Level (i.e., magnitude) of involvement and its type (viz., functional/expressive) were measured as two independent variables. The level was used to delineate high involvement products, and the type was used to further classify high involvement products into functional and expressive groups.

Involvement (Level) Our interest was in identifying a measure which is more specific to the brand purchase decision (as opposed to a general product class involvement). This measure was deduced from Kassarjian's (1978) argument: "Because we care about brands, or packaging, or information processing, we assume that the average consumer also cares but the average consumer who blithely purchases, consumes and discards the product most likely couldn't care less (emphasis added)." This view of involvement implies that consumers are casual in buying some products, caring little about which of several available brands they buy; in other instances consumers are concerned about and careful in choosing their brand. in the research to be reported, involvement was operationalized, therefore, as the 'degree of caring' for what a consumer may choose from among several alternatives of a product.(See figure 1.) The measure in Figure 1 should not be confused with brand-commitment. It is not necessary that the consumer already have a brand in mind to care about; rather the measure taps a "mind-set" that the consumer brings to the brand-choice task.

This rating was obtained for each of the 15 products. Additionally, respondents were also asked to directly indicate whether they considered their "degree of caring" for each specified product as "high" or "low."

Functional/Expressive Type First, the following introductory statement was provided. "People may not purchase these products for the same reason. They may buy some products mainly to fulfill psychological or emotional needs such as expressing their personal image or personality type. On the other hand, they may buy other products expecting, mainly, a satisfactory performance of its physical function. For those products, they would be very concerned about whether the product works as it is supposed to." The respondents were then asked to assign the "high" involvement products (i.e., the products which they assigned to "high" category in the previous question) to one of the two categories: I) where functional needs are more important than psychological (or emotional) needs, and ii) where psychological or emotional needs are more important than functional needs. Suggestions for refinement of this scale in future research are made later.

Dependent Variables Information Seeking was measured by the statement, "I normally seek information from various sources about this product before buying it"; Brand Comparison was measured by the statement: "I normally compare several brands of this product before making my selection." Each statement was rated on a 5 point scale (1 = not at all, 5 = very much so). To measure Number of Features, respondents were asked to write down he product features they would consider in their brand selection.


Measure Reliability and Validity.

Reliability Test. Reliability of the involvement (i.e., "degree of caring") measure was assessed through a test-retest procedure by eliciting the same measure (for the same 15 products) from the respondents after 3 months via a follow-up questionnaire. 147 of the 231 respondents returned the questionnaire in this second study. The reliability coefficients for individual products ranged between 0.41 and 0.70, all statistically significant (p < 0.001). The reliabilities are somewhat lower than desired, and would seem to be so due to the use of a single item measure.

Face Validity. Mean involvement scores (standard deviations) on the "degree of caring" scale for products, in the decreasing order were: automobile 6.32 (1.23), perfume 5.89 (1.61), music album 5.70 (1.80), detergent 5.60 (1.54), stereo 5.43 (1.79), headache remedy 5.29 (1.93), wine 4.88 (2.02), facial tissue 4.35 (1.86), jeans 4.30 (2.03), canned peas A.!4 (2.03), dry cell batteries 4.08 (1.82), salt 3.55 (2.17), light bulb 3.47 (1.73), pen 3.28 (1.88), and dry mill; 2.96 (2.08). In relative terms, therefore, salt, light bulb, pen, dry milk, etc., were deemed by our respondents as low involvement products; automobiles, stereos, perfume, headache remedies, etc., were considered high involvement products.



Convergent Validity In another study, 138 consumers rated 12 products on the "degree of caring measure and two other statements thought also to tap the involvement construct. These latter statements concerned "the importance of the right selection of the brand," and "concern with your selection outcome." The "degree of caring" measure correlated well with these two items: r=.75 and .67, respectively. (These results indicate convergent validity not in terms of Campbell and Fisk's Multitrait, Multimethod procedure, but rather in that the employed measure correlated well with other items purported to measure the same concept.)

Criterion Validity Three criterion variables (namely, information seeking, brand comparisons, a: number of features used were analyzed with T-test (pairs) across "low" and "high" involvement product groups. (Products within each group may have varied from respondent to respondent. Products were placed in the "low" or "high" group according to the respondent-specific direct rating of the product on the "degree of caring" measure.) The results presented in Table I show that the "degree of caring" measure had implications, as expected, for the 3 criterion variables.

Validity Checks for Functional/Expressive Involvement Classification

Face Validity Detergent, automobile, stereo, and headache remedy were assigned to the "functional category" by 99.5%, 90.%, 86.3%, and 87.7% of the respondents. Perfume, music album, and wine were assigned to the expressive category by 85.5%, 83.2%, and 61.5% of the respondents, respectively. Jeans were assigned about equally to both categories.

Convergent Validity A test of convergent validity for the Type of Involvement measure was furnished by data from the follow up study (mentioned earlier). In that study, Type of Involvement was measured using the same statement as in the main study. In addition, respondents were asked to rate each product wit respect to the extent to which: a) the expected functional performance (Statement 'A') and b) "the type of person they were and what others might think of them" (Statement 'B') were the main considerations in their brand selection. If the Type of Involvement measure was validly answered, then the scores on Statement 'A' should be higher for product person dyads (PPD's) categorized as functional than for product-person dyads categorized as expressive. The converse should be the case for scores on Statement 'B'. These expectations were borne out by the data. On Statement 'A', functional PPD's scored higher than did expressive PPD's (M = 4.62 and 3.90, respectively, t = 5.65, p<.001). On Statement 'B', Functional PPD's scored lower than did Expressive PPD's (M = 1.71 and 2.51 respectively, t = 10.31, p<.001).

Test of Hypotheses

From among the products assigned to "high" involvement category, two further categories were created:

- Product-Person Dyad Group A (High Involvement Functional Group) This consists of those products which were assigned by the respondents to the "Functional" category.

- Product-Person Dyad Group B (High Involvement Expressive Croup) This consists of those products which were assigned by the respondent to the "Expressive" category.

For each of the two product-person dyad (PPD) groups identified above, we computed an average score for each respondent for the level of involvement measure and the three dependent measures. To illustrate, if for a particular respondent, PPD Group 'A' contained three products and PPD Group 'B' contained 5 products then, we summed up, for that person, the 3 individual product scores for Group 'A' and 5 individual product scores for Group 'B', on, say, "extent of information seeking" measure, and divided by 3 for PPD Group 'A' and by 5 for PPD Group 'B'. All subsequent analyses employ the average scores within a PPD group.



Note that specific products (and also their number) in the above groups may have differed from respondent to respondent. This procedure for creating the functional and expressive PPD's was employed to allow within-product individual variations. Pooling whatever products a respondent deemed functional and likewise pooling products that were deemed expressive was a preferred procedure to the alterative of performing a product by product analysis. This was because our hypotheses concern the conceptual categories of functional and expressive PPD groups, not about "concrete" products.

To test the hypotheses, repeated measure ANOVAs were run on each of the three dependent variables with PPD Groups 'A' and 'B' considered as replications. Although these groups were distinguished by the type of involvement, and although both groups were high on level of involvement, the precise magnitude or level of involvement could vary nonetheless. Therefore, level of involvement was used as a cc variate.

The results furnished in Table 2 show that, as hypothesized, the extent of information seeking, the extent of brand comparisons, and the number of features used are more for functional than for expressive PPD groups. All these differences are statistically significant at p < .01. Notably, however, the covariate, viz., involvement, itself did not show statistical significance (See Table 2). These results are discussed below.


Based on respondent ratings, products were assigned to functional or expressive involvement categories. On the aggregate level, product assignments were congruent with intuitive expectations. Moreover, for products assigned to the functional group, expected functional performance was the dominant basis of brand choice; and, for expressive products, personality associations dominated.

We tested the functional/expressive distinction only for those products which the respondents classified as "high" on involvement. Our thinking presently is that because at the low level of involvement, consumers did not "care," it did not matter what type of involvement the consumer experienced. An alternative research design would be a 2x2 design with two levels of involvement and two types of involvement; one could then test for involvement magnitude X type interaction. This remains a task for future research.

In the present research, the level of involvement (used as a covariate in Table 2) did not have any statistical significance for the 3 information search variables. This result does not refute the effects of involvement over its entire range; it does show nonetheless, that in a fairly wide range (say, 4.30 for jeans, to 6.32 for auto) at the upper end. not the level of involvement but its type (functional/expressive) influences the information processing activity.

This finding is explainable in terms of other related ideas in the literature. Expressive products, as we employ this term here, do not as easily lend themselves to content or feature discriminations. When brands need to be assessed on personality or image associations, the evaluation is likely to be holistic (Zajonc 1980, Zajonc and Marcus 1982, Mittal 1983b, Holbrook and Hirschman 1982). It is therefore reasonable to expect reduced levels of information "seeking" and fewer brand features considered. As regards brand comparisons, although one can and may certainly compare several brands, the personality images are created entirely by marketer-communications or social communications. If a consumer finds a particular image appealing and self-congruent (e.g., Sirgy 1982), he/she would develop a liking, even enchantment with that brand (Hirschman and Holbrook 1982). The brand preferences can thus be developed by day to day exposure to product communications, making pre-choice comparisons quite superficial if not unnecessary.

Two probable counter-explanations must be addressed. First, it might be argued that there should be differences in the nature rather than in the number of features considered. For example, while a functionally involved tennis racket buyer would use product-based evaluative criteria, an expressively involved buyer might be concerned with such criteria as: Does my tennis hero John McEnroe use this brand? Or, would my teammates be impressed by my choice? Our point when would be that for expressive PPD's at least the product-based features will be used to a lesser degree. Even this limited assertion would have important implications. Moreover, product evaluations based upon considerations other than a product's intrinsic features will not require much explicit, episodic information acquisition; such appraisals will more likely emerge from everyday social observations.

Secondly, it may appear that expressive products (e.g., perfume) by nature have fewer features to consider and seek information about. Objectively, this is not true. Perfumes for example have ingredients, floral/herbal/woodsy/musk/spicy types of essences, "notes, " "composition: " "character," "intensity," "persistence," etc. (Cf. Jesse 1951). If consumers do not know or do not appraise these features, it is because they do not perceive them to be relevant to their goals in the product's purchase.

Most other research on involvement has been situated within a levels-of-involvement framework and has accordingly been focussed on effects due to magnitude differences in involvement (e.g., Petty, Cacioppo, and Schumann 1983; Beatty and Smith 1987). Laurent and Kapferer (1985) took a more inclusive view of the domain of involvement and did find-hedonic and "sign" value aspects of involvement to be distinct from its "performance-risk-importance" aspects. They found further that the extensive decision processes were influenced more strongly by the risk-importance facet than by sign value or hedonic facets. Our results support this pattern of findings.

An important limitation of our research was the use of a direct self-report measure to classify products into functional or expressive PPD groups. A suggestion for future research is that these PPD groups may be created based upon respondent rating of the importance of a specified list of goals. These goals may be classified as predominantly functional or expressive by expert judges. Observations may be placed into functional or expressive categories depending on whether the importance scores summed and averaged across functional goals exceed the similar scores for expressive goals. Although our classification showed face validity, improved measures (multi-item, as for example in Zaichkowsky 1985, and also less direct) are no doubt desirable in replication studies.

Despite the use of single item measures (necessitated by our need to employ a large number of products) our data are clear in showing that the functional or expressive nature of involvement matters. Involvement is "caring," the opposite of nonchalance, so it is only logical that one asks what in the nature of products (more precisely, in the nature of product-person dyad) is that "caring" due to. Depending upon whether the orientation of one's caring is functional or expressive, one's pre-choice brand decision process would be accompanied by more or less information processing. Our results are corroborated by Park and Young's recent (1986) results where they found that "orienting" the subjects into a "cognitive" or "affective" mood influenced the subjects' processing of the information in an advertisement. Our results also provide a theoretical basis for understanding the efficacy of different advertising approaches. In a "quit-smoking" campaign, for example, the use of popular teenage model Brooke Shields would be expected to be more effective than a "harmful-effects" message for expressively involved teenagers. And the use of the "McDonald's and You" theme in McDonald's response to Burger King's comparative offensive a few years ago would be expected to influence consumers by orienting them into an expressive mode of involvement in the brand selection. These are, we believe, sufficient grounds for further study of the functional/expressive orientation of involvement with more refined measure development


While the causal influence of level of involvement on information processing is not questioned, it seems over-simplistic to assume that hi ,her involvement will always lead to more information seeking. Rather, whether a consumer deems a product as primarily functional or primarily expressive will be a significant moderator. In the case of expressive orientation of consumer involvement, the pre-choice information seeking that a consumer will engage in is likely to be less.


Beatty, Sharon E.. and Scot M. Smith (1987), "External Search Effort: An Investigation Across Several Product Categories," Journal of Consumer Research, 14 (June), 83-95.

Bloch, Peter, Daniel L. Sherrel, and Nancy M. Ridgway (1986), "Consumer Search: An Extended Framework," Journal of Consumer Research, 13 (1), 1 19-126.

Hirschman, E. C. and M. B. Holbrook (1982), "Hedonic Consumption: Emerging Concepts, Methods and Prepositions," Journal of Marketing, 46, 92-101.

Holbrook, M. B. and E. C. Hirschman (1982), "The Experiential Aspects of Consumption: Consumer Fantasies, Feelings, and Fun," Journal of Consumer Research, 9 (September), 13240.

Houston, M. J. and M. L. Rothschild (1977), "A Paradigm for Research on Consumer Involvement," WP 11-7746, University of Wisconsin-Madison.

Jesse, J.E. (1951), Perfume Album, New York: Perfume Production Press.

Kassarjian, H. H. (1978), Presidential Address, 1977: "Anthropomorphism and Parsimony", in Advances in Consumer Research, 11 K. Hunt, ed. Ann Arbor, MI: Association for Consumer Research, XII-XIV.

Lastovicka. J. L. and D. M. Gardner (1979), "Low Involvement versus High Involvement Cognitive Structures," in Advances in Consumer Research, H K. Hunt, ed. Ann Arbor, MI: Association for Consumer Research, 87-92.

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

Markus, H (1977), "Self-Schemata and Processing Information about the Self," Journal of Personality and Social Psychology, 35 (FebruarY). 63-78.

Mitchell, A. A. (1981), "The Dimensions of Advertising Involvement," in Advances in Consumer Research, K. Monroe, ed. Ann Arbor, MI.: Association for Consumer Research.

Mittal, Banwari (1983a), "Understanding the Bases ar.d Effects of Involvement in the Consumer Choice Process," (Doctoral Dissertation, University of Pittsburgh, 1982). Ann Arbor, MI: University Microfilms International.

Mittal, Banwari (1983b), "Consumers' Cognitive Journey through the Product Forest," in R. P. Bagozzi &nd Alice Tybout (eds.), Advances in Consumer Research, Vol. X, Provo, Utah: Association for Consumer Research, pp. 464468.

Mittal, Banwari (1987), "A Framework for Relating Consumer Involvement to Lateral Brain Functioning," in M. Wallendorf and P. Anderson (Eds.), Advances in Consumer Research. Vol XIV, Provo, Utah: Association for Consumer Research., 4145.

Park, C. W. and S. M. Young (1986), "Consumer Response to Television Commercials: The Impact of Involvement and Background Music on Brand Attitude Formation," (February), 11-24.

Petty, R. C., J. T. Cacioppo and D. Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10 135 -4 6.

Rothschild, .M. L. (1979), "Advertising Strategies for High and Low Involvement Situations," in Attitude Research Plays for High Stakes, J C Maloney and B. Silverman, eds. Chicago: American Marketing Association, 74-93.

Sirgy, M. J (1982), "Self-Concept in Consumer Behavior: A Critical Review," The Journal of Consumer Research, 9 (December), 287-300.

Udel, J. G. (1964), "A New Approach to Consumer Motivation," Journal of Retailing, 40, 6-10.

Zaichkowsky, J. L. (1985), "Measuring the Involvement Construct," Journal of Consumer Research, 12, 341-52.

Zajonc, R. B. (1980), "Feeling and Thinking: Preferences Need No Inferences," American Psychologist, 35 (February), 151-75.

Zajonc, R. B. and H. Markus (1982), "Affective and Cognitive Factors in Preferences," Journal of Consumer Research, 9 (September), 123-31.



Banwari Mittal, Northern Kentucky University


NA - Advances in Consumer Research Volume 16 | 1989

Share Proceeding

Featured papers

See More


C8. Can Packaging Imagery Fill Your Stomach? Effects of Product Image Location on Flavor Richness, Consumption Quantity, and Subsequent Choice

Taku Togawa, Chiba University of Commerce
Jaewoo Park, Musashi University
Hiroaki Ishii, Seikei University
Xiaoyan Deng, Ohio State University, USA

Read More


E6. The Effect of Crowding Perception on Helping Behavior ——Is Squeeze Warmer than Isolation?

Qingqing Guo, Shanghai Jiao Tong University

Read More


Data-Driven Computational Brand Perception

Sudeep Bhatia, University of Pennsylvania, USA
Christopher Olivola, Carnegie Mellon University, 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.