Consumers' Involvement Profile: New Empirical Results

ABSTRACT - A new stream of research on consumer involvement is reviewed. A cumulative body of empirical results, based on 22 product categories and more than 4400 observations, leads to three main implications : 1. One should stop measuring involvement by a single indicator, but substitute it by an involvement profile based on the antecedents of involvement. 2. Prediction of the consequences of involvement entails taking into account the full involvement profile. The consequences depend on the antecedent conditions of involvement. 3. In each market, it is possible to identify segments with contrasted involvement profiles. An illustrative application is presented. High and low involvement are just two situations among mans others.

Citation:

Jean-Noel Kapferer and Gilles Laurent (1985) ,"Consumers' Involvement Profile: New Empirical Results", in NA - Advances in Consumer Research Volume 12, eds. Elizabeth C. Hirschman and Moris B. Holbrook, Provo, UT : Association for Consumer Research, Pages: 290-295.

Advances in Consumer Research Volume 12, 1985      Pages 290-295

CONSUMERS' INVOLVEMENT PROFILE: NEW EMPIRICAL RESULTS

Jean-Noel Kapferer, HEC-ISA, 78350 JOUY-en-JOSAS (FRANCE)

Gilles Laurent, HEC-ISA, 78350 JOUY-en-JOSAS (FRANCE)

[The authors thank IFOP-ETMAR and Foundation Jours de France for funding data collection.]

ABSTRACT -

A new stream of research on consumer involvement is reviewed. A cumulative body of empirical results, based on 22 product categories and more than 4400 observations, leads to three main implications :

1. One should stop measuring involvement by a single indicator, but substitute it by an involvement profile based on the antecedents of involvement.

2. Prediction of the consequences of involvement entails taking into account the full involvement profile. The consequences depend on the antecedent conditions of involvement.

3. In each market, it is possible to identify segments with contrasted involvement profiles. An illustrative application is presented. High and low involvement are just two situations among mans others.

BACKGROUND

A survey of current uses and measures of consumers' involvement reveals two basic facts : a circular misusage and an hesitation concerning the dimensionality of the concept.

Measuring the Concept by its Consequences

When researchers to not rely on one single item of product "perceived importance," they fall often in a circular mis-usage : to operationally define involvement they make use of alleged consequences. For instance Engel and Blackwell (1982) suggest measuring involvement by the time spent during product search, the energy spent, the number of brands examined, the extensiveness of the decision process. Robertson (1976) uses brand commitment as an indicator of product involvement. Stone (1984) defines involvement as time and/or intensity of effort expanded in the undertaking of behaviors. But, as also noted by Cohen (1983) and Rothschild (1984), is it involvement or what results from involvement ? This reminds us of the fallacy of measuring attitudes by behavior thus preventing any test of the theory concerning the behavioral consequences of attitudes.

Hesitations about Involvement Dimensionality

Summarizing current involvement work, Rothschild (1984) regrets that the prototypical paper consists in a theoretical review where it is shown that involvement has many dimensions, with some disagreement among authors as to the number of such dimensions (Houston and Rothschild 1978 ; Bloch and Richins 1983 ; Muncy and Hunt 1984). This constant recompartmentalization of the concept seems also to block engaging in new empirical work and collecting fresh data.

These two basic facts are related. Looking at a recent review of indicators of product involvement (Antil 1984), one is struck by the overwhelming majority of measures that are in fact consequences of involvement. Since these consequences are distinct aspects of consumer behavior, it is not surprising that many authors -eel they are actually measuring different concepts, despite the common name : involvement (Muncy and Hunt 1984).

The stream of research we conducted since 1981 stemmed from the same diagnosis. At a theoretical level, in order to test empirical propositions concerning the consequences of involvement, a displacement of emphasis was needed : indicators of involvement should be found in the antecedents of involvement. Also, instead of saying this is the "real involvement" and this is not, one should move from the search for the elusive single dimension to the acceptance of an involvement profile : the relationship of a consumer to a product category is better described by a profile on a parsimonious number of facets. At a practical level, in order to provide managers with a tool to segment markets in subgroups with differing involvement profiles, a set of reliable and valid scales had to be created (one scale by involvement facet). This resulted in the creation of generalizable items that could be made product specific, thus meeting the need expressed by Mitchell (1979), Bloch (1981) and Antil (1984).

This paper will present in turn the thought process used in identifying the facets of the involvement profile, the measurement of the profile, evidence of nomological validity (data concerning the consequences of involvement), and an illustration of how the involvement profile is being currently used for market segmentation.

THE FACETS OF CONSUMERS' INVOLVEMENT

Consumers' degree of involvement in products or issues is to only held as a major mediating variable of consumer behavior. Despite differences in emphasis and preferences among researchers, a consensus seems to emerge as to its following generic definition (Mitchell 1979 ; Bloch 1981 ; Cohen 1983 ; Rothschild 1984) : "Involvement is an unobservable state of motivation, arousal or interest. It is evoked by a particular stimulus or situation. It has drive properties : its consequences are types of searching, information processing and decision making" (Rothschild, p. 217).

Since involvement is a hypothetical variable, it cannot be directly measured. Involvement is inferred from the presence or absence of its alleged determinants or antecedents. Actually, a review of the researchers' practices in operationally defiling involvement reveals a small number of such antecedents.

Early research on involvement conceptualized it as a permanent or at least enduring drive state, stemming from the product or issue being tied to one's ego, or central values or self-concept (Sherif and Cantril 1947 ; Ostrom and Brock, 1968). Typically, the Sherifs used W.C.T.U. members as high-involved subjects, hereby implicitly suggesting three possible antecedents of their involvement : having a deep interest in the issue, finding it extremely rewarding (enjoyable) to act about the issue, identifying oneself totally with the issue. Drawing from the literature on leisure behavior and hobbyists, Bloch and Bruce (1984) suggest the same three antecedent conditions. For them, enduring involvement at high levels is akin to product enthusiasm : it is evoked by a strong, abiding, hobby-like interest in the product class in question. Also, the involvement state is linked to the rewards generated by the usage of the product (enjoyment, pleasure). Finally, the actors identify themselves completely with the activity : the hobby acts as a message of who they are, vis-a-vis relevant others. The hobby expresses their self-concept. Interestingly, when trying to classify product categories in terms of involvement, Lastovicka and Gardner (1979) found three classes of products : high involvement and low involvement products and what they called special-interest products (related to the respondent's hobby).

A number of researchers have tried to experimentally manipulate people's degree of involvement. Thus, Zimbardo (1960) told the students in the "high involvement" group of a communication persuasiveness experiment, that their opinions would be made public. In doing so, Zimbardo suggested two possible antecedents of involvement (i) making one's intimate opinions overtly available to others, as a sign of the kind of person one is, and (ii) perceived risk, at least one of its two sub-components (Bauer 1967) : the perception of that the decision may lead to negative consequences. As noted by many authors, such experiments do not create an enduring involvement but a transitory, situational arousal state.

In a marketing context, this led to the distinction between product involvement and purchase-of-the-product involvement. The former is enduring, the latter may be only transitory or situational. The former entails the latter, but the latter does not entail the former : despite no interest nor pleasure in Champagne, buying a bottle of Champagne to receive good friends is an involving situation because one may be judged on the chosen brand, or because of the price. Actually perceived risk is the most frequently quoted antecedent condition of purchase involvement or lack of it. Whenever a wrong decision has important negative consequences and making a wrong decision may occur, involvement is present (Chaffee and McLeod 1973 ; Tyebjee 1979 ; Arora 1984 ; Muncy and Hunt 1984). For instance Antil (1984) considers that low involvement arises when consumers have high confidence in the expected benefits of the product choice (the subjective probability of making a mispurchase is absent). The risk may be physical, financial or psychological (if the brand choice is perceived as related to one's self-image).

To summarize, five antecedent conditions of involvement have been used by researchers : interest in the product category, enjoyment or pleasure derived from it, perception of self expression through product category, and the two components of perceived risk : the stake and subjective probability factors (Bauer 1967). The first facet (interest) is an antecedent of enduring involvement only. Pleasure and sign value may apply to both enduring and situational involvement. Perceived risk induces mostly situational involvement. In any case by measuring a consumer's position on each of these five facets, the exact nature of his/her involvement is specified.

MEASURING THE INVOLVEMENT PROFILE

Having identified five facets of involvement, at least theoretically speaking, a major question is left open : do some or all of these facets converge in fewer dimensions, or is there actually five facets, exhibiting specific variance, thus discriminant validity. There would be not point in keeping our a priori distinctions between facets, if empirically, for consumers, these facets converge.

To answer this question, we followed Churchill's (1979) iterative multi-step approach. A multi-item scale was created for each of the five facets. On the basis of a literature review, consumers' in-depth interviews and interviews with marketing experts, the scales were developed through successive small data collections. They can be applied in any product category : this result is far from being trivial, since many items that apply well to, say, dresses look ridiculous when asked about dishwashing liquid. The scales resulting from the pruning process are monodimensional and are, with one exception, fairly reliable. Table 1 presents the reliability coefficients of each scale in a large survey (n - 1568, measures done over 20 product categories, 800 housewives or so were interviewed, each one about two product categories).

Now the main question remains : are these five presumed facets empirically distinct or do they converge in fewer dimensions ? To answer this question, an oblique factor analysis of all the items in the five scales was performed. An oblique solution was retained since, all facets being tied to the same construct, no independence was postulated. Three independent data collection were undertaken :

- A 20 products survey on 1568 observations with female respondents (these products range from TV, mattresses, bras and dresses, to detergents, yoghurt, or perfumes) (Table 2).

- A survey of 588 female respondents, on bread.

- A survey of 2250 male and female respondents on a frequently purchased food product.

Table 2 presents the results derived from the first data set.

In the three data collections the results were similar : five factors have an eigenvalue superior to one. Also, as can be seen in Table 2, each scale loads on one factor and no two scales load on the same factor. The five presumed facets do have discriminant validity and seem to empirically tap five distinct aspects of involvement. This conclusion is based on 4400 observations and 22 different product categories. It may nevertheless happen that for some specific product categories two scales merge : for instance in the case of foot products, interest and pleasure may load on the same factor indicating that for consumers these two facets are close. Our experience indicates that when the eigenvalue criterion (above 1) leads to four facets, the fifth factor is just below 1 and appears exactly as expected in a confirmatory factor analysis.

TABLE 1

MEASURES OF THE PRESUMED FACETS OF INVOLVEMENT

TABLE 2

OBLIQUE FACTOR ANALYSIS

At this stage, empirical analyses corroborate the literature review : there seems to be five facets of involvement. Although they are correlated, each one has specific variance. Measuring the five facets allows to better grasp the nature of the involvement state.

Having a measure of the involvement profile, independent of the alleged consequences of the construct, it is now possible to empirically test these consequences. In this process, we shall also contribute to the construct validity of the involvement profile. In effect, it could be said that nothing more than face validity substantiates our claim that we are tapping the involvement construct. Examining the link with involvement consequences is a step towards nomological validity.

THE CONSEQUENCES OF INVOLVEMENT

According to theory, when a consumer is involved in a product category, he should, among other consequences, engage in a lengthy decision process, exhibit brand commitment (Robertson 1976 ; Assael 1981), pay attention to information concerning the product class, etc. In this paper, we shall concentrate on these three variables to demonstrate the nomological validity of the profile.

Two scales were specifically created to measure the extensiveness of the decision process and brand commitment. The propensity to read articles concerning the product category was measured by a single self-report item. Table 3 presents the linear regressions of these three variables on the facets of the involvement profile The analyses are based on 1568 observations and 20 contrasted product categories.

The extensiveness of the decision process is influenced by all facets of the profile, especially interest in the product category and perception of important negative consequences in case of mispurchase. Expecting a strong pleasure leads also to spend time and energy in the decision process. Brand commitment increases with interest in the category, perception that the product choice conveys one's image (sign), and the two components of perceived risk. It is independent of perceived pleasure value of the category. Finally, the propensity to read articles increases when one is interested in the product, or when one believes that the choice expresses one's image or when the product has a pleasure value. perceived risk is not related to the propensity to regularly read articles about the product category.

TABLE 3

INVOLVEMENT FACETS AND ASPECTS OF CONSUMER BEHAVIOR

These results provide nomological validity to the involvement profile. All facets contribute to the prediction of extensiveness of the decision process ; most of them influence the two other variables. Most important is the fact that all facets do not have the same influence on different aspects of consumer behavior for instance the importance component of perceived risk strongly influences the extensiveness of the decision process, far less brand commitment, but not at all regular reading of articles on the product class. Similarly the rewarding nature of the product (pleasure) affects the decision process and exposure but not brand commitment.

This simply means that limited predictions of behavior can be made from one single indicator of involvement. Only knowing the full involvement profile allows complete predictions. Also, involvement theory is probably oversimplified : as it stands now, it seems that involvement leads to a score of behavioral consequences. Our data indicate that, depending upon the antecedent condition of involvement, some behavioral consequences Sill take place or not. Further empirical data on other behavioral consequences and other data bases confirm the necessity of taking the antecedent conditions into account if one wants to predict the behavioral consequences of involvement. It is time to quit using single indicators of involvement.

SEGMENTING MARKETS BY INVOLVEMENT PROFILE

Since the exact nature of consumers' involvement can now be measured by means of the involvement profile, it is possible to segment markets on that basis. In a market, say Champagne, what are the prevalent types of involvement profiles ? The authors have performed several segmentation studies in the context of ad hoc surveys devoted to specific product categories. For proprietary reasons, we cannot present these studies here. However, we shall present an example based on a non-proprietary survey, funded by a marketing research foundation (Foundation Jours de France). The purpose of this section is merely illustrative of the way the involvement profile is currently used by firms to get a new understanding of their market : space constraints prevent any methodological development.

To show how markets can be segmented, we performed a cluster analysis of the 1568 observations available from this non-proprietary survey. These observations had been collected in 20 product categories. (Most product categories had about 50 interviews, four "large" product categories had about 200 interviews)O As a result of this richness of product categories, we retained a rather large number of clusters (ten). In commercial applications performed for one product category, the number of clusters retained has been much smaller (between three and six).

Given the number of observations, the cluster analysis was performed using a double dynamic clustering approach.

In this illustrative application, the clustering was performed on the basis of the involvement facets described earlier, as well as of two other variables : the consumer's perceived difference across brands and the consumer's self declared competence. Perceived difference was included following suggestions by De Bruicker (1979), Assael (1981) or Ray (1982), according to which involvement and perceived--difference are two distinct dimensions that should be examined together in order to analyze consumer behavior. Using perceived difference as an active variable allowed the possibility to define clusters corresponding to the consumer situations distinguished by these authors.

Similarly Day (1970) and Lastovicka and Gardner (1979) suggest that a consumer's feeling of competence in a product category may change his or her behavior. Four items were used to measure the consumer's competence in a product category (Cronbach's g - 0.74) and three items for perceived difference (g - 0.63).

Finally, because of the relative unreliability of its present scale, the probability facet of perceived risk was not included as an active variable in the typology.

The ten types finally retained in the cluster analysis were given intuitive names as follows.

1  Minimal involvement

2  Functional differentiation

3  Undramatized risk

4  Small pleasure

5  Conformist purchase

6  Riskless involvement

7  Functional involvement

8  Pleasure involvement

9  Need for help

10  Total involvement

Table 4 presents, for each type, the average scores on four facets of involvement, as well as on the dimensions of perceived difference and competence (for each one of these variables, standardized scores have been computed with a mean of 100 and a standard deviation of 50 over all 1568 observations). Table 5 indicates, for each of the 20 product categories, in which types the consumers have been allocated. For example, 43 x of the consumers of batteries have been classified in the "minimal involvement" type. (For clarity, we reprinted such percentages here only when they were above 10 %). As can be seen, if there are on the whole ten types, each market is generally segmented in no more than five or six main types.

We shall make a brief comment on six of our ten types.

The easiest types to comment on are the two extreme ones : "total involvement" and "minimal involvement," also called by many : high and low involvement.

In the minimal involvement type, consumers are at the lowest on all active variables (interest, sign, pleasure, risk importance, competence and perceived difference). From all points of view, this corresponds to a very limited involvement of consumers in the product category. Products that have many consumers in this minimal involvement type include batteries, dishwashing liquid, detergents, pasta (Table 5).

On the other extreme, the total involvement type is at the highest on all six variables : interest, sign, pleasure, risk importance, competence, perceived difference. From all points of view, the consumers are very much involved into the product category. Products that have many consumers in the total involvement type include dresses, perfume, bras, coffee, Champagne.

Up to now the two types we have described (total involvement and minimal involvement) are reasonably similar to the stereotypes one would expect under those names. One must note, however, that these two types are two extreme cases that together represent only one fourth of all consumers in our sample (9 Z for minimal involvement. 16 X for total involvement).

The other eight types, which represent 75 Z of our sample, present contrasted profiles. This means that a type has high marks on certain facets of involvement, and low marks on other facets. The picture of these types is no longer all white or all black, it is a picture of nuances.

Consider type 5, the "conformist purchase." This subjective situation concerns almost half of the purchasers of Champagne. Here (Table 4), the consumer attaches only a limited interest to the product (91). However he perceives a high pleasure value in the product (123), a very high sign value (130); an important risk (120) attached to the brand choice, in relation with his belief that large differences between brand, exist (123). Unfortunately, he has very limited knowledge in the product class (57) : he does not know how to choose. This typical subjective situation of purchase concerns large segments of consumers in the TV or perfume markets

Consider type 3, undramatized risk. Here products (Table 4) have only one involving feature : the risk they entail (120). Interest (65), sign (65), pleasure (64) are low. The consumer feels there are very little differences across brands (50) and finds herself incompetent (56). The products that fall the most in that type are two durables : vacuum-cleaners and TV sets (Table 5).

Type 6 is that of riskless involvement. Risk is low (79), in relation to a small perceived difference (92). But there is relatively high involvement on the other facets : interest (113), sign (142) and pleasure (111). This breeds competence (127). The two products most represented in that type are jam and perfume.

Type 7 is that of functional involvement. Here risk (124) and interest (121) are high, but sign (73) and pleasure (47) are low. In this situation the consumer perceives differences between alternatives (121) and feels competent (114). Typical products here are shampoo, dishwashing liquid, detergents.

TABLE 4

TEN INVOLVEMENT TYPES: DESCRIPTION

TABLE 5

SEGMENTING MARKETS BY INVOLVEMENT TYPE

CONCLUSION

On tile basis of our cumulative body of empirical data, three- Implications can be drawn:

1) Using single indicators of involvement seems a dead end street. The phenomenon is not unidimensional. Any single index runs the risk of actually merging facets of involvement. For instance measuring involvement by a "purchase importance" item leaves unspecified the roots of this importance, that is the nature of the involvement. If the single index matches one of the five distinct facets, using it only would create another risk. We showed that no consequence of involvement could be predicted unless the full profile of its antecedent conditions was known : no single facet alone allows full prediction of the behavioral consequences.

2) Among practitioners and theoreticians, it is now a habit to think in terms of a mere high-low dichotomy. Actually, segmentation research shows that cases where consumers are low or high on all facets of the profile represent two extreme cases. Most of consumers have contrasted types of profile, with ups and downs on different facets. It is precisely the subjective situation created by the interaction of facets which leads to specific behavioral outcomes.

3) Researchers have now at their disposal five clean, conceptually clear and multi-product scales. We believe the thrust of future research should be to empirically test the so many alleged consequences of involvement. The authors are presently examining how the five facets of the involvement profile relate to involvement in communications, to sensitivity to different appeals and creative styles, and to the purchase of distributor's private brands and generic products.

REFERENCES

Antil, J.H. (1984), "Conceptualization and Operationalization of Involvement," in Advances in Consumer Research, Vol. 11, T.C. Kinnear, ed., Association for Consumer Research, 203-209.

Arora, R. (1982), "Validation of an S-O-R Model for Situation, Enduring and Response Components of Involvement," Journal of Marketing Research, XIX, 505-516.

Assael, H. (1981), Consumer Behavior, New York: Wadsworth.

Bauer, R.A. (1967), "Consumer Behavior as Risk Taking," in Risk Taking and Information Handling in Consumer Behavior, D.F. Cox, ed., Boston, Harvard University Press. Division of Research, Graduate School of Business Administration, Harvard University, 23-33.

Bloch, P.H. (1981), "Involvement Beyond the Purchase Process: Conceptual Issues and Empirical Investigation," in Advances in Consumer Research, Vol. 8, R. Monroe, ed., Association for consumer Research, 61-65.

Bloch, P.H. and G.D. Bruce (1984), "Product Involvement as Leisure Behavior," in Advances in Consumer Research, Vol. 11, T.C. Kinnear, ed., Association for Consumer Research. 197-202.

Bloch, P.H. and M.L. Richins (1983), "A Theoretical model for the Study of Product Importance Perceptions," Journal of Marketing, 47, 3 (Summer), 69-81.

Chaffee, S.H. and J.M. Mcleod (1973), "Consumer Decisions and Information Use," in Consumer Behavior: Theoretical Sources, S. Wart -and S.S. Robertson, eds., Englewood Cliffs, N.J.: Prentice-Hall, 385-415.

Churchill, G.A. (1979), "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research, 16, 64-73.

Cohen, J.B. (1983), "Involvement: Separating the State from Its Causes and Effects," Working Paper N¦33, Center of Consumer Research, University of Florida.

Day, G.S. (1970), Buyer Attitudes and Brand Choice Behavior, New York: The Free Press.

De Bruicker, F.S. (1979), "An Appraisal of Low-Involvement Consumer Information Processing," in Attitude Research Plays for High Stakes, J.C. Maloney and B. Silverman, eds., American Marketing Association, 112-130.

Engel, F. and R.D. Blackwell (1982), Consumer Behavior, 4th Edition, New York: The Dryden Press.

Houston, M.J. and M.L. Rothschild (1977), "A Paradigm for Research on Consumer Involvement," Working Paper no 11-77-46, University of Wisconsin-Madison.

Lastovicka, J.L. and D.M. Gardner (1979), "Components of Involvement," in Attitude Research Plays for High Stakes, J.C. Maloney and B. Silverman, eds ., American Marketing Association, 53-73.

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

Muncy, 3.A. and S.D. Hunt, "Consumer Involvement: Definitional Issues and Research Directions," in Advances in Consumer Research, Vol. 11, T.C. Kinnear, ed., Association for Consumer Research, 193-196.

Ostrom, T.M. and T.C. Brock (1968), "A Cognitive Model of Attitudinal Involvement," in Theories of Cognitive Consistency: A Source Book, R.P. Abelson & al., eds., Chicago: Rand McNally, 373-383.

Ray, M. (1982), Advertising and Communication Management, New York: Prentice-Hall.

Robertson, T.S. (1976), "Low-Commitment Consumer Behavior," Journal of Advertising Research, 16 (April), 19-24.

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

Sherif, M. and H. Cantril (1947), The Psychology of Ego- Involvement, New York: Wiley.

Sherif, M. and C.L. Hovland (1961), Social Judgement: Assimilation and Contrasts Effects in Communication and Attitude Change, New Haven, Conn.: Yale University Press .

Stone, R.N. (1984), "The Marketing Characteristics of Involvement," in Advances in Consumer Research, Vol. 11, T.C. Kinnear, ed., Association for Consumer Research, 210-215.

Tyebjee, T.T. (1979), "Refinement of the Involvement Concept," in Attitude Research Plays for High Stakes, J.C. Maloney and B. Silverman, eds., American Marketing Association, 94-111.

Vaughn, R. (1980), "How Advertising Works: a Planning Motel," Journal of Advertising Research, 20 (october), 27-33

Zimbardo, P.G. (1960), "Involvement and Communication Discrepancy as Determinants of Opinion Conformity," Journal of Abnormal and Social Psychology, 60, 86-94.

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