Personality and Consumer Traits: the Beat Goes On

Jerome B. Kernan, University of Cincinnati
[ to cite ]:
Jerome B. Kernan (1979) ,"Personality and Consumer Traits: the Beat Goes On", in NA - Advances in Consumer Research Volume 06, eds. William L. Wilkie, Ann Abor, MI : Association for Consumer Research, Pages: 125-127.

Advances in Consumer Research Volume 6, 1979      Pages 125-127


Jerome B. Kernan, University of Cincinnati


Over the years, many of us have professed a vow to one of consumer research's cardinal articles of faith--that behavior is a (however convoluted) function of personality. Just as "there must be a God," a person's behavior "must" derive, in some meaningful way, from his/her idiosyncratic self. That this relationship has never been proved (that faith is the only basis for accepting it) has not deterred the personality disciples from their evangelical ministry; after all, this is the sort of thing that any right-headed (no reference to Krugman) person knows is correct. Besides, if "faith can move mountains" it surely can withstand the curious barbs of a few unwashed existentialists.

Personality cognoscenti are a clever lot; we know where and how we are culpable. So we repair to the inoculation game (McGuire, 1969; Bither et al., 1971), taking care to place our most credible sources in the vanguard (Wells, 1966; Jacoby, 1969; Kassarjian, 1971). The result, of course, is thoroughly predictable; we are thereafter left free to purvey our prestidigitation with impunity. Indeed, we are encouraged; the American Marketing Association--hardly known for its support of unpopular causes--has bestowed the ultimate imprimatur on us by issuing a bibliography on "personality research" (Twedt et al., 1977). The whole thing at times seems a bit like sinners defining evil.

But all this is innocent enough, and infinitely more interesting than teaching impatient students, running dreary crosstabs for persistent clients, or getting hassled by your section chief for deviating from OMB's latest procedural directive. And occasionally we even learn something.

So what have we here? The Hirschman/Wallendorf paper is not in the "Sociological Perspectives" session: Coney and Harmon are not in the "Situational Impacts...Gift Giving" session; and "Consumer Satisfaction/Dissatisfaction'' must survive without Fornell and Westbrook. Three displaced papers? Perhaps; but only if one focuses on their object sets. If we regard these papers contextually, they are disparate--adoption/diffusion, gift-giving, and complaining would hardly load on the same factor (unless it's "consumer behavior"). But if we look carefully, a common theme emerges; each of the papers essays a trait/behavior connection. Not each study designates its trait(s) as "personality" and the measures of behavior are not uniformly explicit. But at least by implication each of the researches proffers a trait + behavior relationship; each falls within the psychology-of-individual-differences domain.

Following are some comments to which a variety of valences can be attached. They are benign items--just matters potentially worth consideration.


There are two particularly encouraging things about this study. First, it advances the cause of replication, something sorely needed in our empirical work. And second, it emphasizes the importance of environmental circumstances on consumer behavior, demonstrating the obvious (but so often ignored) fact that people are not trans-situational automatons. Kissing your lover is not the same as kissing your sibling.

Some issues about which one might become a bit anxious:

* The experimental subjects--student volunteers-- and whether they behaved naturally.

T* he criterion measure (purchase intention) vis-a-vis purchase, per se; Fishbein notwithstanding, everybody realizes the two can be very different depending on (what else?) the equivalence of the situations under which each is measured.

* The innovation measure; this is a mystery unless one happens to be familiar with the relevant Jacoby (1971) and Coney (1972) studies. Presumably, the chance-alone mean value is 3.0--12 products x .25.

* Whether the most appropriate ANOVA model has been used; an argument might be made that dogmatism is nested within subjects while situations are not, hence the design is a mixed one and not the factorial as used.


This study is interesting for its attempt to carry an aspect of personality theory through a counseling-psychology paradigm to a consumer (complaining) behavior context. Whether it produces dramatic results is not so important as the fact that it is attempted; we need--and are likely to see--more such behavior-is-a-process work in the future. The study represents an interesting conceptualization, one that is likely to stimulate replications and extensions. And what might those researches attempt to establish? Two possibilities would seem obvious:

* Whether submissiveness/assertiveness/aggressiveness are distinct traits or simply differences in degree of a trait (presumably one born of frustration or anxiety) and, depending on that answer, whether effective complaining (measured non-hypothetically) is a non-monotonic function of such a continuum or of something(s) altogether different. Equity theory (Berkowitz and Walster, 1976) would seem to offer some useful insights here.

* What "changes" when a non-student sample is used and assertiveness and complaining are not measured in the same setting? Granting all the reasons for using a student sample we all realize that this practice imposes severe limitations. Students "...chosen selectively to represent all class levels and both sexes" is virtually a contradiction in terms; all class levels simply can't be found at "...a major private university" and, even if they could, students are notoriously better connected to the social/industrial/political structure, have measurably superior communication skills and proclivities and, in general, are just more "together" than "plain folks." A priori, students can be expected to be more assertive (it's a cognitively complex activity) than a consumer population in general (an alarming proportion of whom are functionally illiterate, and that fact affects both absolute and relative associations (student scores have much less elevation).


This research, although approaching the matter from a slightly different perspective, is reminiscent of the Reynolds and Darden (1971) model, which also argued that the more socially-connected a person is, the more s(he) is likely to behave in innovative ways. Now, as in 1971, that contention seems a bit awkward to dispute. Let us therefore accept it and comment briefly on some methodological points. The ultimate implications of this work--role strain isn't so intolerable (Sieber, 1974) and societal organizations pose threats to gradual social change (Blau, 1974)--are so value-laden that to discuss them is to add yet another role, polemicist.

The methodological comments, if a guess is permissible, all hang on the same factor--reanalysis of already-collected data. Thus one can appreciate that certain constraints existed; yet some legitimate concerns remain:

* Is role accumulation per se the variable of interest or should we be concerned with that constraint as it is a perceived instrumentality? Which is more important--that one joins groups or why one joins?

* Can contamination occur in operationalizing role accumulation? Suppose I belong to a country club. Given that a choice between recreational and social organization can be made, do I "get credit for" golfing, tennis, bridge, and spectator-sports events (I do each, but with the same people)? Isn't an alternative (heaven forbid, we've been brainwashed by expectancy-value models) to regard organizational membership as a degree of institutionalization for an activity?

* Several curious judgments seem to be reflected in the operationalization of the variables. For example, in what sense is the "reduced set" of TV program types more "stimulation rich" than the original set; what happened to "variety" as a criterion for stimulation? Do demographic and socioeconomic variables count for naught; are the measured variables mediated equivalently regardless of who a person is?


At this juncture it is a capital offense to reiterate the do-and-don't litany of personality research in consumer behavior. Every graduate student has read and/or heard it, ad nauseaum. No harangues, then; just four modest proposals.

Beware of Operationalized Constructs

We must be realistic. Personality is a construct and, as such, has no real-world referent. Accordingly, it can be operationalized only according to someone's invention. Unlike the concept, age (which can be measured with both trans-situational and trans-personal consensus), personality "means" something different according to who measures it and how. As a predictor variable, then, "personality" is of questionable validity and reliability. And we don't reify it by multi-variately laundering a set of existing traits or ad hoc personality-like statements. Molar sets of traits "work better" than individual traits, but unless the predictor is "inclination to buy Brand X," one should not expect much of Brand X behavior to be explained by the predictor.

Back to Basics

Typically, personality/behavior associations are weak. That is, the measures imply weak associations. But what does that tell us? Of course, the association might indeed be weak. But often as not the 'Weakness" proceeds from unstable or irrelevant criterion variables (how much confidence can we place in self-report measures?), invalid or unreliable predictor variables (as noted above), a non-linear relationship, or an indirect one. We know about each of these possibilities, yet seem oblivious to them when the quick-and-dirty-study occasion presents itself. Would any of us really believe a high r-?

Who's Kidding Whom?

Unless one is a bona fide personality researcher (i.e., personality qua personality) and/or wishes to opt for the sanctuary of the ivory tower (where "exploratory" is code for "anything goes"), it's patently obvious that personality research is a means to an end--personality mediates, but doesn't cause behavior. Accordingly, personality should be assessed in loco, within the context its effect is suspect. This means testing respondents like those whose behavior is under inquiry and in real situations. (It also implies that the typical student-sample study, replete with its demand characteristics, is fatuous.)

If You Want to Play Hardball. . .

Beyond the world of hassling with journal editors there is a land in which the name of the game is predictive validity. In that world, personality either works (meaning it has an instrumental value in identifying behavioral patterns) or it is discarded (it's true relationship to the behavior in question is so camouflaged as to be useless). Although the success rate varies widely, the research paradigm seems to be fairly well established.

* First, the policy issue or administrative problem is identified and some prospective renditions of criterion variables as well as research hypotheses are specified.

* Next, a series of focus-group or depth interviews is conducted to (1) ascertain the validity of the notions originally specified and (2) translate the phenomenon into respondent vernacular. The interviewers must be qualified and the participants representative of whatever population whose behavior ultimately is to be explained. For instance, one might inquire at this stage: "What sorts of problems do you encounter with (product)? What do you do when you feel you've been treated unfairly? Whose fault is it when a product doesn't work? Etc."

* Next, a set of variables--some of which are personality-like--is gleaned from the previous interviews and pilot-tested, in a structured format, on a group of perhaps 1002200 people who reflect all the important characteristics of the target population. It should be emphasized that situational, demographic, socioeconomic and product (service)-specific variables are tested simultaneously at this stage.

* Another set of qualitative interviews might then be necessary to verify any surprises uncovered in the pilot survey.

* The results are finally used to construct a standardized questionnaire for a statistical survey (n = 1000-1500). The output usually includes a small portion directed to the creative group and it says something like: "given these palpable measures of the various population (market) segments, the following personality data suggest which appeals will be most effective for each segment."


There is increasing skepticism being expressed about personality research in consumer contexts. Even psychographics is questioned. This is all a matter of personal opinion, of course but one view is that there is more to the world than r2s and that the study of personality can be evaluated apart from a criterion of prediction. Another way to pose that is to ask yourself whether sex that does not result in conception is functional.


Leonard Berkowitz and Elaine Walster, eds. Equity Theory: Toward a General Theory of Social Interaction (New York: Academic Press, 1976).

Stewart W. Bither, Ira J. Dolich and Elaine B. Nell, "The Application of Attitude Immunization Techniques in Marketing," Journal of Marketing Research, 8 (February 1971), 56-61.

Peter M. Blau, "Presidential Address: Parameters of Social Structure," American Sociological Review, 39 (October 1974), 615-635.

Kenneth A. Coney, "Dogmatism and Innovation: A Replication," Journal of Marketing Research, 9 (November 1972), 453-455.

Jacob Jacoby, "Personality and Consumer Behavior: How Not to Find Relationships," Purdue Papers in Consumer Psychology, No. 102, 1969.

Jacob Jacoby, "Personality and Innovation Proneness," Journal of Marketing Research, 8 (May 1971), 244-247.

Harold H. Kassarjian, "Personality and Consumer Behavior: A Review," Journal of Marketing Research, 8 (November 1971), 409-418.

William J. McGuire, "The Nature of Attitude and Attitude Change," in Handbook of Social Psychology, Vol. 3, Gardner Lindzey and Elliot Aronson, eds. (Reading, Mass.: Addison-Wesley, 1969), 136-314.

Fred D. Reynolds and William R. Darden, "Mutually Adaptive Effects of Interpersonal Communication," Journal of Marketing Research, 8 (November 1971), 449-454.

Sam D. Sieber, "Toward A Theory of Role Accumulation," American Sociological Review, 39 (August 1974), 567-578.

Dick Warren Twedt, Lyndon E. Dawson, Jr., Hugh G. Wales, and Gary N. Bunner, eds., Personality Research in Marketing: A Bibliography (Chicago: American Marketing Association, 1977).

William D. Wells, "General Personality Tests and Consumer Behavior, in On Knowing the Consumer, Joseph Newman, ed., (New York: John Wiley & Sons, Inc., 1966), 187-189.