An &Quot;Importance&Quot; Subscale For the Consumer Involvement Profile

ABSTRACT - After reviewing the structure (dimensionality) of two scales that have been proffered as measures of the involvement construct, Zaichkowsky's Personal Involvement Inventory (PII), and Laurent and Kapferer's Consumer Involvement Profile (CIP), the authors propose and provide initial support for a new subscale for the CIP; one designed to measure Importance, a construct not now encompassed by that scale. The relationship between Importance and the remaining CIP subscales designed to measure various involvement antecedents (ie., Interest-Pleasure, Sign, Risk Probability and Risk Importance) is then discussed.

Citation:

Kenneth C. Schneider and William C. Rodgers (1996) ,"An &Quot;Importance&Quot; Subscale For the Consumer Involvement Profile", in NA - Advances in Consumer Research Volume 23, eds. Kim P. Corfman and John G. Lynch Jr., Provo, UT : Association for Consumer Research, Pages: 249-254.

Advances in Consumer Research Volume 23, 1996      Pages 249-254

AN "IMPORTANCE" SUBSCALE FOR THE CONSUMER INVOLVEMENT PROFILE

Kenneth C. Schneider, St. Cloud State University

William C. Rodgers, St. Cloud State University

ABSTRACT -

After reviewing the structure (dimensionality) of two scales that have been proffered as measures of the involvement construct, Zaichkowsky's Personal Involvement Inventory (PII), and Laurent and Kapferer's Consumer Involvement Profile (CIP), the authors propose and provide initial support for a new subscale for the CIP; one designed to measure Importance, a construct not now encompassed by that scale. The relationship between Importance and the remaining CIP subscales designed to measure various involvement antecedents (ie., Interest-Pleasure, Sign, Risk Probability and Risk Importance) is then discussed.

INTRODUCTION

While the specific wording still changes from one author to the next, most consumer researchers would probably not argue with a characterization of product involvement as a consumer's "heightened motivational state" (Mittal 1989a, p. 697) toward some product or service category that derives from its relevance or importance to the consumer. As such, product involvement can be distinguished from involvement with other objects, including the purchase process (Slama and Tashchian 1985, Mittal 1989), advertising messages (Andrews and Durvasula 1991), and so on. This so-called relevance or importance, in turn, results from one or another personal or situational conditioners, referred to as involvement antecedents.

Consumer researchers have made considerable progress toward developing a scale to measure the product involvement construct in the past ten years or so. To date, however, attendant research has proceeded along two parallel, non-intersecting lines of inquiry. As a consequence, researchers now have available two rather disparate involvement scales, the Personal Involvement Inventory and the Consumer Involvement Profile.

The next section presents a brief discussion of the structure or dimensionality of these two scales. After that discussion, an extension which adds an "Importance" subscale to the Consumer Involvement Profile is proposed and analyzed. It is intended that the new subscale will allow the Consumer Involvement Profile to more fully measure the involvement construct.

COMPARING THE STRUCTURE OF TWO INVOLVEMENT SCALES

Personal Involvement Inventory. The twenty item, semantic differential scaled Personal Involvement Inventory (PII) was originally proposed and analyzed by Zaichkowsky (1985). This scale, conceived as a unidimensional scale to identify the personal relevance of a particular product to the consumer, has been further analyzed and refined by Celuch and Evans (1989), Flynn and Goldsmith (1993), Jain and Srinivasan (1990), McQuarrie and Munson (1987, 1992), Mittal (1989a), and Zaichkowsky (1986).

Of particular concern here, several authors have addressed the dimensionality of the PII. After strenuously arguing for a restrictive, unidimensional definition of product involvement, Mittal (1989a) suggested that the original PII contained at least three constructs. These include (1) product involvement itself (six scale items that measure the importance, relevance and significance of the product, plus "of concern," "matters" and "means a lot" to me), (2) a hedonic antecedent of product involvement (four scale items that measure interest in, fascination with, appeal of, and excitement generated by the product), and (3) an amalgam of scale items that probably measure something more akin to attitude than involvement.

Indeed, McQuarrie and Munson (1992) have recently presented a new version of the PII, more or less revised along these lines. Those authors suggest organizing the PII into (1) an Importance dimension of the product involvement construct using five items from the PII (Mittal's "involvement itself"), (2) an Interest dimension of the product involvement construct using three items from the PII and two new items (Mittal's hedonic antecedent), and (3) remaining items from the PII to be discarded (Mittal's "attitude-like" items, plus a potpourri of items (e.g., mundane, nonessential) thought to be complexly worded).

Consumer Involvement Profile. In an independent line of inquiry, Laurent and Kapferer (1985) proposed a quite different scale; the sixteen item, Likert scaled Consumer Involvement Profile (CIP). Unlike Zaichkowsky's unidimensional scale, the CIP is actually a series of subscales, each designed to measure an antecedent of product involvement. When using the CIP, level of involvement must be inferred from observed measurements on these antecedents. The five antecedents proposed by Kapferer and Laurent include Interest, Pleasure, Sign, Risk Probability and Risk Importance.

The CIP has also been further analyzed and refined by Celuch and Evans (1989), Jain and Srinivasan (1990), Kapferer and Laurent (1985a, 1985b, 1993), Mittal (1989a), Mittal and Lee (1988), and Rodgers and Schneider (1993). Of particular concern here is that two of Laurent and Kapferer's antecedents, Interest and Pleasure, seem to merge into a single factor in studies using American consumers, a conclusion that was reached independently by Jain and Srinivasan (1990), and by Rodgers and Schneider (1993). Thus, in U.S. domestic use anyway, the CIP seems to contain four dimensions, (1) Interest-Pleasure (Mittal's hedonic antecedent), (2) Sign, (3) Risk Probability, and (4) Risk Importance. Indeed, in their most recent set of replication studies, Kapferer and Laurent (1993) often found a similar merging of the Interest and Pleasure factors in studies using French consumers.

Comparing the Scales' Structures. Consequently, two separate lines of scale development research have led to scales with only one overlapping dimension; level of interest in or pleasure derived from the product. In motivating their Interest subscale, McQuarrie and Munson (1992) refer to it variously as emotional involvement (see, also, Park and Mittal 1985, Vaughn 1986, and Zaichkowsky 1987) and as a measure of the enjoyment derived from a product. Indeed, the two new items added to the PII and included in the Interest subscaleCfun, and neat (as opposed to dull)Care direct measures of pleasure. Thus, the Interest component of McQuarrie and Munson's revised version of Zaichkowsky's PII and the merged Interest-Pleasure antecedent subscale in the CIP should readily converge. (In fact, Celuch and Evans (1989) have already found fairly large correlations between the original PII and the original Interest and Pleasure subscales in the CIP.)

Whether this Interest-Pleasure subscale found to reside in the CIP (or, equivalently, this new Interest subscale in the revised PII) constitutes a second dimension of involvement or an involvement antecedent will be a matter of considerable debate for some time yet to come. The authors here are inclined to support Mittal's characterization of Interest-Pleasure as a hedonic antecedent of involvement, for several reasons. First, like all constructs, product involvement can be as narrowly or broadly conceived as a researcher chooses; it is hoped that continued research and discussion eventually leads to definitional convergence. Whenever possible, however, parsimony suggests that several rather narrowly defined constructs are preferable to few rather broadly defined ones. As noted earlier, the core of product involvement concerns the importance or relevance that a product assumes for a given consumer. Everything else can be just as easily considered as either leading to, paralleling, or flowing from the importance that is involvement.

Thus, like Mittal (1989a), the authors here contend that product involvement itself should be narrowly conceived, encompassing only the importance or centrality of the product to the consumer. Other facets, including the extent to which a consumer finds a product or service category interesting or pleasurable, can be linked to involvement without necessarily being involvement.

Indeed, in a temporal context, the authors strongly believe that Interest-Pleasure is more properly considered a hedonic antecedent of involvement (as in the CIP) than as a dimension of involvement (as in the PII). Certain products (e.g., hobbies, many participative sports) come to mind as ones that are relevant and important to the consumer precisely because of the pleasure they bring. It is difficult to imagine, for example, the model railroader's or the amateur golfer's intense concern for their respective associated products as being primarily derived from anything but the sheer enjoyment of the activity (though it is stipulated that there may be contributory symbolic value or risk related antecedents at work as well). For some product categories and some consumers, then, importance or personal relevance is driven primarily by the pleasure or interest inherently found in the product. For such products, Interest-Pleasure is, indeed, an antecedent.

Also, certain other products (e.g., "life's little indulgences," some leisure activities) can be thought of as being a source of interest or pleasure yet not particularly important to many consumers. It is difficult to imagine, for example, that the consumption of a candy bar or a VCR rental, while admittedly a source of at least temporary pleasure or interest, would have any great relevance to the individual. In such situations, even a relatively high level of perceived interest or pleasure does not result in the sorts of behavioral outcomes normally associated with high involvement products. For instance, one rarely sees more than a modicum of information seeking attendant to the purchase of a chocolate bar or the latest video release. Thus, not all interesting or pleasurable products seem to be involving, as would be the case if Interest-Pleasure were classified as a second dimension of the product involvement construct.

In general, then, a compelling case can be argued for thinking of Interest-Pleasure as an antecedent that can (but, not always must) lead to product involvement, and for thinking of product involvement itself as a unidimensional concept that signifies the importance or relevance of the product to the individual.

What is clear is that current versions of both the PII and the CIP are incompletely structured. Specifically, the PII measures personal relevance or importance (ie., product involvement itself), and one of several antecedents (ie., Interest-Pleasure). On the other hand, the CIP measures a richer array of antecedents (ie., Interest-Pleasure, Risk Importance, Risk Probability and Sign), but has no direct measure of personal relevance or importance (ie., product involvement itself).

One acceptable solution for the researcher intent on measuring both product involvement and its antecedents is to utilize a combination of the PII and CIP scales. In doing so, one need only recognize that the two scales employ different response formats; the PII uses a semantic differential format and the CIP uses a Likert format. A second solution entails expanding either scale so that both product involvement and its possible triggering antecedents can be measured under the same format. The next section sets forth a proposed new subscale for the CIP; one designed to measure product importance itself.

AN IMPORTANCE SUBSCALE FOR THE CIP

Exhibit 1 presents a seven item importance subscale that was crafted as a potential addition to the CIP. These seven items were derived from several sources. One (IMP02) is an item reassigned from the five item merged Interest-Pleasure subscale that resulted from U.S. applications of the CIP (see Rodgers and Schneider 1993). (While IMP02, an item from the original CIP Interest subscale, consistently loaded with four other items measuring interest taken in or pleasure derived from a product, its wording ("I attach great importance to selecting a C") is such that it should be realigned with other items purporting to directly measure product importance instead of those that measure Interest-Pleasure.) One item (IMP04) is an adaptation of an item that appeared in an Importance subscale in an early version of the CIP (see Laurent and Kapferer 1985). Three items (IMP03, IMP06, and IMP07) are unabashed Likert scaled versions of items from the Importance subscale in the revised PII (see McQuarrie and Munson 1992). Finally, two items (IMP01 and IMP05) were especially written for this new CIP subscale. As with other items in the CIP, each of the seven items in the proposed Importance subscale is measured with a five-point Likert scale.

The dimensionality and reliability of the proposed Importance subscale for the CIP were investigated in two separate studies. Results from both studies appear in Exhibit 2.

Study 1: Health Care Clinic. Study 1 was conducted in conjunction with a survey of customers and noncustomers of a major health care clinic in a mid-sized, upper midwest MSA. Responses to the seven item Importance subscale only were obtained from 514 adult participants in the survey. Interviews were conducted by telephone, and queried participants as to the importance attached to selecting a "health care clinic". Specific responses (and associated numerical scores) were "totally agree" (1), "somewhat agree" (2), "neither agree nor disagree" (3), "somewhat disagree" (4), and "totally disagree" (5).

According to Exhibit 2(A), a factor analysis led to a one factor solution (with Eigenvalue of 3.860) that explained 55.2% of the response variance. (Note that the factor analytic routine did not prespecify a one factor solution; using a criterion minimum Eigenvalue of 1.0, only one emerged.) An internal consistency analysis led to an alpha of .846. Note, however, that IMP07 might be superfluous to the scale. Its factor loading and item to total correlation were both lower than for the remaining six items.

Study 2: Financial Institution. Study 2 was conducted in conjunction with a survey of customers and noncustomers of a financial institution serving a tri-county, nonurbanized, upper midwest area. Responses to the seven item Importance subscale, and to the four CIP antecedent subscales, were obtained from 267 adult participants in the survey. Responses were obtained through a mailed questionnaire follow-up to a preliminary telephone interview, and queried participants as to issues concerning the selection of a "financial institution". Specific responses (and associated numerical scores) were "totally disagree" (1), "disagree" (2), "neither agree nor disagree" (3), "agree" (4), and "totally agree" (5).

According to Exhibit 2(B), results for Study 2 were virtually identical to those of Study 1. A single factor solution again emerged from the analysis (with Eigenvalue of 3.654), and the full seven item subscale's alpha was .847.

EXHIBIT 1

PROPOSED "IMPORTANCE" ITEMS FOR CIP SCALE

EXHIBIT 2

PSYCHOMETRIC EVALUATION OF THE CIP "IMPORTANCE" SUBSCALE

Overall, then, this new Importance subscale for the CIP does appear to be unidimensional with reasonable internal consistency. (In the interest of succinctness, IMP07 can probably be deleted from the subscale; it adds little to the psychometric properties of the subscale.) In addition, the respective item means suggest that far more consumers agree than disagree that decisions about financial institutions are important ones, and even more agree than disagree that decisions about medical clinics are important ones. While far from even a minimally acceptable assessment of the Importance subscale's validity, these results conform with what a reasonable person would have expected a priori, and do at least tentatively establish that the subscale is measuring what it purports to measure; perceived importance attached to a product, or in this case, a service.

The Relationship Between Importance and its Antecedents

As previously noted, Laurent and Kapferer originally intended to include an Importance subscale in the CIP. Ultimately, they gave up that goal, presumably because the Importance subscale could not be isolated from other subscales in the CIP. But, if Importance flows from one or another antecedent (or a combination thereof), one would expect such a relationship to its antecedents to exist. Indeed, Importance is more properly considered a dependent variable to the set of independent variable antecedents. This would, of course, suggest searching for a regression relationship involving Importance rather than attempting to isolate an independent Importance factor.

Exhibit 3 presents the results of such a regression analysis using the full set of CIP data from Study 2 (with IMP07 deleted from the new Importance subscale). The complete model was highly significant (F= 56.6; p < .001), with R-Squared of 53.6%.

There are significant relationships between Importance and each of the four CIP antecedents in the model. Three of those relationships are linear. Consequently, it was concluded that, at least so far as the choice of a financial institution is concerned, product importance is linearly related to Interest-Pleasure, to Sign and to Risk Importance. Among the three linear relations, that to Risk Importance was especially strong. Based on the regression coefficients (which can be directly compared here because all data was input as arithmetic "means" of five point scales), product importance is particularly responsive to increases in risk importance. The choice of a financial institution tends to become especially involving as, from one consumer to another, the perceived consequences of a poor decision increase.

EXHIBIT 3

REGRESSION RESULTS FOR SIX ITEM CIP IMPORTANCE SUBSCALE VERSUS FOUR INVOLVEMENT ANTECEDNTS STUDY 2: FINANCIAL INSTITUTIONS

The fourth relationship, that between Importance and Risk Probability, is particularly intriguing. The preliminary linear regression model resulted in a significant negative coefficient being assigned to Risk Probability. Because such an inverse relationship would have been counter to the expected linkage between perceived risk and involvement, further analysis was undertaken. Indeed, the scatterplot reproduced in Exhibit 4 suggested a quadratic relationship involving the Risk Probability antecedent. So, squared terms for Risk Probability, and for each of the other three antecedents as well, were constructed and allowed to enter the regression model. Only the quadratic term for Risk Probability did so.

Apparently, then, there is a curvilinear relationship between the importance attached to financial institutions and the perceived chance of making a poor decision. Consumers who are quite confident of their ability to make a good choice, and those who are not at all confident of their ability, tend to be more involved in the selection of a financial institution than are consumers between these extremes.

Among the several plausible, if speculative, explanations for such a curvilinear relationship, the authors offer the following. What if Exhibit 4 depicts a "point in time" view of a dynamic process that includes, as an intervening variable, information acquisition over time? That is, suppose consumers who perceive a given product or service category as fraught with considerable potential for making purchase errors also devote themselves to acquiring information and, ultimately, expertise, about that product over time.

The extremes in Exhibit 4, then, might represent consumers at different points in the acquisition of that expertise. Some consumersCperhaps younger onesCwho think of a product category as particularly risky in terms of the chance of making a poor choice, become very involved with those choices, including acquiring some level of expertise about the product. Over time, these consumers, who maintain a consistent if not growing level of involvement, nonetheless find themselves eventually with considerable self-confidence in their decision making ability.

Meanwhile, other consumers, who perceive only a moderate level of risk probability attendant to the decision, do not tend to become as involved and, hence, do not tend to acquire information and expertise about the product. Consequently, their beliefs about risk probability would tend to remain the same over time.

It is true that the acquisition of product expertise would also tend to characterize consumers who become involved with a product via other antecedents; Interest-Pleasure, Risk Importance or Sign. However, is it likely that product expertise would alter neither the extent to which those consumers find the product category inherently interesting, pleasurable or self-expressive, nor the perceived magnitude of external negative consequences of a poor choice. Thus, the linear relationship between involvement and these antecedents would be preserved, the acquisition of product expertise notwithstanding.

At any rate, all of this is currently little more than pure speculation, except that product importance remains curvilinearly related to the perceived risk probability attendant to choosing a financial institution. Interesting extensions of this study might include (a) determining whether the nature of this relationship holds across other product and service categories, and (b) testing, perhaps in a structural equations environment, the possibility that information acquisition serves, over time, as an intervening variable between product importance and risk probability.

CONCLUDING OBSERVATIONS

The research reported herein attempted to "complete" the CIP scale developed by Kapferer and Laurent by adding a subscale designed to measure the personal relevance or importance of a product to the consumer. While in need of additional psychometric evaluation, especially with respect to tests of validity, the CIP is now capable of measuring product involvement itself (through this new Importance subscale), as well as four crucial antecedents to such involvement.

The study also presented preliminary evidence that three of the antecedents (Interest-Pleasure, Sign and Risk Importance) are linearly related to involvement, with Risk Importance the most strongly related. The fourth antecedent, Risk Probability, appears to be related to involvement in a curvilinear fashion. But, considerably more research needs to be undertaken to replicate those relationships with other populations, and to extend them to product and service categories other than financial institutions.

EXHIBIT 4

MEANS PLOT-IMPORTANCE SUBSCALE VERSUS RISK PROBABILITY

Finally, it appears as if researchers are inching ever closer to a generally accepted involvement measurement scale. With the PII now expanded to encompass Interest-Pleasure (and perhaps, eventually, other involvement antecedents as well), it and the CIP move nearer and nearer to one another's dimensionality. Indeed, it is hoped that at some point there is little to differentiate the two scales save for their response formats. It was the authors intent here to encourage such a convergence.

REFERENCES

Andres, J. C. and S. Durvasula (1991), "Suggestions for Manipulating and Measuring Involvement in Advertising Message Content," in R. H. Holman and M. S. Solomon, eds., Advances in Consumer Research, 18, Association for Consumer Research, 194-201.

Celuch, K. and R. Evans (1989), "An Analysis of the Convergent and Discriminant Validity of the Personal Involvement Inventory and the Consumer Involvement Profile," Psychological Reports, 65, 1291-1297.

Flynn, L. R. and R. E. Goldsmith (1993), "Application of the Personal Involvement Inventory in Marketing," Psychology & Marketing, 10 (July-August), 357-366.

Jain, K. and N. Srinivasan (1990), "An Empirical Assessment of Multiple Operationalizations in Involvement," in M. E. Goldberg G. Gorn and R. W. Pollay, eds., Advances in Consumer Research, 17, Provo, UT, Association for Consumer Research, 594-602.

Kapferer, J. N. and G. Laurent (1985a), "Consumer's Involvement Profile: New Empirical Results," in E. Hirschman and M. Holbrook, eds., Advances in Consumer Research, 12, Provo, UT, Association for Consumer Research, 290-295.

Kapferer, J. N. and G. Laurent (1985b), "Consumer Involvement Profiles: A New Practical Approach to Consumer Involvement," Journal of Advertising Research, 25 (December-January), 48-56.

Kapferer, J. N. and G. Laurent (1993), "Further Evidence on the Consumer Involvement Profile: Five Antecedents of Involvement," Psychology & Marketing, 10 (July-August), 347-355.

Laurent, G. and J. N. Kapferer (1985), "Measuring Consumer Involvement Profiles," Journal of Marketing Research, 22 (February), 41-53.

McQuarrie, E. F. and J. M. Munson (1987), "The Zaichkowsky Personal Involvement Inventory: Modification and Extension," in P. Anderson and M. Wallendorf, eds, Advances in Consumer Research, 14, Provo, UT, Association for Consumer Research, 36-40.

McQuarrie, E. F. and J. M. Munson (1992), "A Revised Product Involvement Inventory: Improved Usability and Validity," in J. F. Sherry and B. Sternthal, eds, Advances in Consumer Research, 19, Provo, UT, Association for Consumer Research, 108-115.

Mittal, B. (1989a), "A Theoretical Analysis of Two Recent Measures of Involvement," in T. K. Srull, ed., Advances in Consumer Research, 16, Provo, UT, Association for Consumer Research, 697-702.

Mittal, B. (1989b), "Measuring Purchase-Decision Involvement," Psychology and Marketing, 6 (Summer), 147-162.

Mittal, B. and M. Lee (1988), "Separating Brand-Choice Involvement from Product Involvement Via Consumer Profiles," in M. J. Houston, M. Wallendorf and P. Anderson, eds, Advances in Consumer Research, 15, Provo, UT, Association for Consumer Research, 43-49.

Park, W. C. and B. Mittal (1985), "A Theory of Involvement in Consumer Behavior: Problems and Issues," in J. N. Sheth, ed., Research in Consumer Behavior, JAI Press, Greenwich, CT, 201-231.

Rodgers, W. C. and K. C. Schneider (1993), "An Experimental Evaluation of the Kapferer-Laurent Consumer Involvement Profile Scale," Psychology & Marketing, 10 (July-August), 333-345.

Slama, M. E. and A. Tashchian (1985), "Selected Socioeconomic and Demographic Characteristics Associated with Purchasing Involvement," Journal of Marketing, 49 (Winter), 72-82.

Vaugn, R. (1986), "How Advertising Works: A Planning Model Revisited," Journal of Advertising Research, 26 (February-March), 57-66.

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

Zaichkowsky, J. L. (1986), "Conceptualizing Involvement," Journal of Advertising, 15 (2), 4-14.

Zaichkowsky, J. L. (1987), "The Emotional Aspect of Product Involvement," in P. Anderson and M. Wallendorf, eds., Advances in Consumer Research, 14, Provo, UT, Association for Consumer Research, 32-35.

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