A Look At Personality Profiles and the Personality-Attitude-Behavior Link in Predicting Consumer Behavior

Larry Percy, Ketchum, MacLeod & Grove
ABSTRACT - The lack of reliable empirical support for the intuitive connection between personality and behavior is discussed. Two explanations are presented to account for this low accountability: a) the use of personality traits rather than profiles; and b) a lack of cognizance of the personality-attitude-behavior link and the actions of intervening variables. Study results are discussed that address these points and indicate strong predictive ability.
[ to cite ]:
Larry Percy (1976) ,"A Look At Personality Profiles and the Personality-Attitude-Behavior Link in Predicting Consumer Behavior", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 119-124.

Advances in Consumer Research Volume 3, 1976      Pages 119-124


Larry Percy, Ketchum, MacLeod & Grove


The lack of reliable empirical support for the intuitive connection between personality and behavior is discussed. Two explanations are presented to account for this low accountability: a) the use of personality traits rather than profiles; and b) a lack of cognizance of the personality-attitude-behavior link and the actions of intervening variables. Study results are discussed that address these points and indicate strong predictive ability.


Despite an apparent lack of success in empirical explanation, personality as a predictor or mediator in communications and buyer behavior theory persists. The nexus is drawn over and over in the literature. Howard and Sheth (1969) describe personality traits as those characteristics that account for differences among people and that are predictive of their behavior (emphasis added). They go on to discuss a number of "significant" ways in which certain personality traits are important in their theory of buyer behavior, as an element of the buyer's frame of reference. Engel, Kollat and Blackwell (1973) talk about personality as the sum total of factors involved in each individual's way of thinking, behaving and responding that make their image. They proceed to devote an entire chapter to the relationship between personality and consumer behavior. Parr (1972) suggests that, other things permitting, the preferences and demands of an individual's personality will affect his choice of surroundings. From this one might easily infer that a person's purchase behavior and media mode selection are components of his surrounding, and thus a function of personality.

The importance of understanding personality from a communications standpoint is mentioned by Howard and Sheth (1969), and discussed extensively by McGuire (1968, 1972) among others. McGuire (1968) spends a great deal of time detailing the effects of specific personality traits (e.g. self-esteem) on the various stages of his informational processing paradigm and notions of attitude change. Particular attention is drawn to the mediational aspects of many variables in understanding the effect of personality on a behavioral or attitudinal change.

Perhaps this is a portent to why it has been so difficult to demonstrate in a straight forward manner a strong relationship between personality and behavior. As McGuire points out, this mediational principle in itself does not allow one to make any confident predictions, only to call our attention to the complexity of the undertaking. And complex it has proven. Typically only about five or ten percent of the total variance in buyer behavior has been accounted for by personality construct measures. But as Kassarjian (1971) concludes his review of research in personality and consumer behavior, to expect the influence of personality variables to account for a large portion of the variance is most certainly asking too much. He lays a good deal of this blame at the feet of the many other interrelated influences on the consumer decision process.

Other explanations for this low accountability have been offered. Nakanishi (1972) finds that both Kassarjian and Jacoby (1971) attribute this uncertain predictive ability to either the lack of a proper hypothesis to test or sloppy adaptation of an otherwise clinical measurement of personality. Returning to Engel et al., their feeling is that personality characteristics are better thought of as moderator variables, and hence one should expect a better prediction of behavior in one situation than another. Crane {1972) suggests that the personality of most individuals is so multifaceted that they can meet the demands of any situation they are likely to encounter. He therefore feels role tends to be more useful than personality in predicting and influencing human behavior. Yet one of the four roles he finds important to the marketing communicator (family decision making unit, those of husband or wife) Engel et al. suggest is influenced by the personality characteristics of the individuals involved in decision making. They offer the example of wives who have a strong need for love and affection generally having less influence in purchasing decisions. One seems to be traveling on a Moebius Strip.

A more rational explanation is perhaps to be found in Rosenblatt and Miller (1972): they concern themselves with experimental design. While many researchers evaluate response to some social event (e.g. an advertising message) among people who differ in personality, unfortunately personality is almost always correlated with pre-test differences on the communication topic. Personality becomes confounded with the message or source: And there are scale problems (other than those of the personality measurement instrument). As discussed below, in addition to confounding in communications, there may be confounding in attitude and behavior.

Beyond any possible problems in research design, however it is the purpose of this paper to discount many of the explanations offered above, and to suggest that the reasons for the low accountability and conflicting results to date of research in personality and behavior are: a) reliance on personality traits as opposed to personality profiles; and b) a failure to recognize the importance of attitude or disposition as a mediator of behavior.


Personality implies a whole, not a string of independent personality traits. Why should it be surprising that a single personality trait, which may co-vary in a multitude of ways with other traits, fails to contribute on its own much to the prediction of a specific consumer behavior such as a brand or product choice. It makes far better sense to consider a profile of personality, one that reflects the interactions of various personality traits. As McGuire (1972) has pointed out, any chronic personality trait that tends to make an individual extremely resistant (say) or extremely receptive to a particular influence probably becomes embedded over time in a matrix of other traits that serve as correctives in moving the person towards an intermediate level of behavior.

Sparks and Tucker (1971), after noting essentially weak or spotty relationships between specific personality traits and the use of particular products, went on to demonstrate significant relationships between profiles of these same traits and bundles of various product usage. Building on these results Alpert (1972) studied the relationship between personality structure and product attributes. Although he cited a few limitations, he none-the-less found strong relationships in matching personality profiles of product attributes' determinance. It seems that at least some support is developing for believing that the use of personality profiles, rather than specific personality traits are more helpful indicators of consumer behavior.


In seeking an appropriate instrument for personality measurement, initially an adaptation of Edward's EPPS was used. Fifty of the original paired statements ware retained after a great deal of testing, providing measures of nine personality traits. At issue was a need to provide an instrument that would not be overly taxing within the context of an otherwise long interview, and one which would enjoy a high level of participation on the part of subjects.

An early application in a study of banking behavior provided a very interesting set of profiles contrasting those who have savings accounts but no checking account vs. those who have checking accounts but no savings account. As detailed in Figure 1 below, one sees that the two groups exhibit almost completely opposite personality profiles. The importance of this finding was critical to the development of advertising copy: in talking to the saver who doesn't check, the creative tone should be soft and the message coddled in a sort-of "we know what's best" manner. Quite the contrary of course would apply to the checker who doesn't save; they would be ,uninterested in any soft-selling copy. As Engel et al. (1973) suggest, advertising copy can be much more effective if the artists and writers have a rich understanding of the total lifestyle of those to whom they are writing. Here is an example of a strong relationship between personality, expressed as a profile rather than a specific trait, and consumer banking behavior. Understanding this relationship resulted in more effective advertising strategy.




Attention is now turned to an examination of personality (expressed as a profile) as a common denominator among consumers who tend to hold similar attitudes toward a product-market, and in relating them, through these commonly held attitudes to specific behavior as shown in Figure 2. Personality is considered a determinant of behavior, but only through an underlying attitudinal construct. Personality relates to commonly held attitudes and perceived behavioral orientations, and it is these factors that mediate specific behavior.



Being able to describe a target market group in terms of their gross personality characteristics, which are in turn known to be good discriminators of a particular attitude or disposition, will provide a much clearer picture of whom one is dealing with for purposes of developing advertising and marketing strategies directed toward them. This connection with attitudinal disposition is important from a strategic viewpoint, for it is attitude that the marketer or advertiser seeks to exploit, change or modify in effecting positive behavior (i.e. usage and trial) toward his product.

However, it must be recognized that even given the above chain, intervening factors may effect behavior contrary to that predicted by a given attitudinal disposition. As an example, 'Imagine the man who hates lawn work, but finds himself with an unruly hedge on his hands. He may reluctantly purchase a hedge trimmer even though his personality profile matches that of a non-owner of lawn and garden tools. This fact of ownership does not make him a good prospect for additional lawn and garden tools. The intervening factors of an overgrown hedge and complaining neighbors effected his behavior. Figure 3 illustrates the possible action of intervening factors.

Two studies are discussed below: the first illustrating the Personality-Attitude-Behavior link; a second the results of intervening factors.



In the first study, a large scale inquiry into the meal among a nationally projectable sample of over 1,200 homemakers, personality measurements via the modified EPPS were gathered in addition to attitudinal and behavioral measures. The attitudinal data, gathered from five-point Likert scale measurements of 20 statements (which were developed from earlier research on a much larger list), provided the basis of an attitudinal segmentation of the market. Following Cattell (1966a, b) a Q-mode factor analysis revealed six groups of homemakers with different attitudes toward the meal. Personality profiles over nine traits were developed for each group.

These profiles were intercorrelated for each of the attitude groups, with the results reported in Table 1.



One notices a good correlation between attitude groups Q1 and Q3 of .769 and between attitude groups Q2 and Q4 with .713 (and possibly Q6). The thesis of this paper would suggest that because the personality profiles of these groups are similar (i.e. correlate well), this could be predictive of a common underlying attitude. In fact, this is precisely what one finds. The first "cluster" of Q1 and Q3 reflect two groups of homemakers whose attitudes might be described as traditional, centering on a consumer attitude of heavy personal involvement with the preparation and serving of the meal. They enjoy spending time in the kitchen. The second "cluster" of Q2 and Q4 express a strong interest in convenience foods, and little desire to be heavily involved in the meal. The positive correlations with this "cluster" by Q6 may be explained by their lack of interest in meal involvement; however, it is not as strongly correlated for they do not share the same attitude toward convenience foods.

There are clearly strong personality-attitude relationships evident in these data to support the first premise, but what of the next link, to specific behavior. While the attitudes measured are perhaps a little too general to expect brand usage associations, Table 2 none-the-less illustrates a good relationship between the "traditional" groups (Q1 and Q3) and nonusers of the brand studied. This makes sense when it is understood that the brand in question is identified with main meal entree convenience foods, something a real "traditional" homemaker would be wont to use. While some associations might have been expected among the other attitude groups and brand users, it should be pointed out that only about five percent of all homemakers are users of this particular brand. The disposition to purchase may be positive, but marketing realities have thus far precluded it. It would appear that the personality-to-attitude-to-behavior chain enjoys a certain level of credibility.



In a second study both attitudinal groups and groups clustered on the basis of perceived behavior were profiled and compared with specific product usage groups. Unlike the above study, these data were collected from a national probability sample of men, and a modified version of the Gough and Heilbrun Adjective Check List (1965) replaced the modified EPPS. The nine scales utilized in the EPPS earlier are retained, along with the addition of six scales. Reasons for the instrument change centered largely on the reduction of administration time from about twenty minutes to five with the ACL with no apparent loss in discrimination.

The product category studied was light durables of male interest where ownership is known to be widely distributed over a number of different products, with individual subjects likely to own none, one, a few, or many items, each item available in a broad range of price, quality and sophistication. Examining the personality profiles between non-owners and sophisticated owners (so-called because of multiple product ownership coupled with ownership of more sophisticated items) revealed opposite profiles (correlating -.595). These profiles are displayed below in Figure 4 for the nine personality traits comparable with the EPPS profiles shown earlier.



Interpreting these profiles, one could say that those who own no items An this category may be described as unsure of themselves and their abilities, seeking stability and continuity in their environment, and apprehensive of ill-defined or risk-involving situations. They try to avoid situations calling for choice and decision-making, and are dubious about the results of expending effort or becoming involved with their labor. On the other hand, the sophisticated owners are determined to do well, and are usually successful. They have a quiet confidence in their own abilities and worth, welcoming the challenges to be found in disorder and complexity. Sincere and dependable, they are adaptable, resourceful, and comprehend problems and situations rapidly and incisively.

There is clearly a relationship, and a significant one, between the two extremes of category behavior and personality profiles. The study goes on to examine the predictive ability of these profiles in correctly classifying ownership with and without a mediator.

Personality profiles were developed for attitudinal groups generated by a Q-mode factor analysis of 36 attitude statements and the clusters generalized by the Singleton-Kautz (1969) minimized within cluster variance algorithm on 26 perceived behavior statements, and the results correlated with the above profiles.

These results along with the incidence of non-owners and sophisticated owners in each cluster and Q-group are shown in Table 3. While the personality profile correlation between non-owners and Cluster 1 is strong (.933), predicting the heavy incidence of non-owners in that group (81%), equally strong correlations of non-owners with Cluster 4 and Q-groups 3 and 4 (.899, .656, and .917) would also suggest heavy incidences of non-owners in these groups, when in fact non-ownership is quite low [9%, 12%, and 33%}.



This seems to debunk the notion of the predictive ability of personality profiles that has underlain a great deal of this paper. However, this is only the case if one fails to address the question of intervening factors as a mediator in the predictive process. In the detailed findings of this study reported elsewhere (Percy, 1975), the cluster and Q-groups were themselves clustered together according to similarity of personality profile and considered for compatibility of meaning. The results provide a three step construction representing negative behavior and attitude within the product market (P1), positive behavior and attitude (P2), and a benign group (P3). This accounted for 100% of the behavior component and 90% of the attitude component under study.

Table 4 details the personality profile correlations of these three groupings with non-owners and sophisticated owners. It is clear that the personality profile of non-owners is strongly related to that of P1, the negative behavior and attitude grouping. Each of the four components of this grouping reflected negative attitude toward usage, even though for other reasons certain of these subjects none-the-less own some items from the category. This is unimportant, for this usage cannot be construed as positively motivated --their attitude and motivation is against usage, and it is this disposition which the personality profiles have successfully predicted. The same follows for the relationship between the sophisticated owners and P2' It would appear that personality, although not a unique predictor of ownership, still discriminates well the underlying behavior and attitude factors that originate ownership.




The role of personality in the study of consumer behavior is frequently discussed. As we have seen, it often finds its way into the literature, but rarely with any empirical certainty as to its actual association with specific behavior. Numerous explanations have been ventured as to why this low accountability is experienced, yet none seem satisfying. This paper advances a two-fold explanation, illustrated and supported by specific examples.

First, personality traits are indicated to be less effective than personality profiles in predicting specific consumer behavior. Evidence is cited from Sparks and Tucker (1971) and A1pert (1972) showing the increased predictive power of the personality profile. An example of strong discrimination in banking behavior is presented, illustrating where personality profiles offered critical guidance in the development of targeted advertising strategy.

Second, and more importantly, it is suggested that a personality-attitude-behavior link exists, and that while personality (expressed as a profile) is a consistent predictor of attitudinal disposition, intervening factors often effect actual behavior in a manner contrary to that suggested by an individual's personality and attitude. Two studies confirm the relationship between commonly held attitudes toward a product-market and personality; one continuing the chain and predicting brand behavior, the other demonstrating the effect of intervening factors on prediction.

From an advertising or marketing strategic viewpoint, it should be pointed out, the prediction of a specific behavior (e.g. brand choice) is not as important as the prediction of a target group's attitudinal disposition. What one is generally interested in is potential future behavior, usually as a consequence of some advertising or marketing effort. To the extent that one may use personality to better understand a target group holding a particular set of attitudes (selected because their attitude presages a disposition toward positive behaviors, one should be in a better position to more effectively communicate to that group. One is maximizing potential through the exploitation, modification or changing of specific attitudes in one's advertising and marketing efforts. This personality-attitude-behavior link plays an important part in understanding the best way to go about it.

In the past, personality has been less than successful in predicting specific consumer behavior because: a) it has used traits rather than profiles; and b] it did not take cognizance of the personality-attitude-behavior link and the actions of intervening factors. This analysis does not obviate the intuitive supposition that behavior should be predictable knowing an individual's personality; rather it offers an explanation of the connection.


M. I. Alpert, "Personality and the Determinants of Product Choice," Journal of Marketing Research, Vol. IX (February, 1932) 89-92.

R. B. Cattell, "The Data Box," in R. B. Cattell, ed. Handbook of Multivariate Experimental Psychology. Chicago: Rand McNally Company, 1966a.

R. B. Cattell, "The Meaning and Strategic Use of Factor Analysis," in R. B. Cattell, ed. Handbook of Multivariate Experimental Psychology. Chicago: Rand McNally Company, 1966b.

E. Crane, Marketing Communication 2nd ed. New York: John Wiley & Son, 1972.

J. F. Engel, Do T. Kollat and R. D. Blackwell, Consumer Behavior, 2nd ed. New York: Holt, Rinehart & Wilson, 1973.

H. G. Gough, A. B. Heilbrun, Jr., The Adjective Checklist Manual. Palo Alto, California: Consulting Psychologists Press, 1965.

J. A. Howard, and J. N. Sheth, The Theory of Buyer Behavior. New York: John Wiley & Son, 1969.

J. Jacoby, "Personality and Innovation Proneness," Journal of Marketing Research, Vol. VIII (May, 1971) 244-47.

H. H. Kassarjian, "Personality and Consumer Behavior: A Review." Journal of Marketing Research, Vol. VIII (November, 1971) 409-18

W. J. McGuire, "Attitude Change: The Information-Processing Paradigm," in C. G. McClintock, ed. Experimental Social Psychology. New York: Holt, Rinehart & Winston, 1972.

W. J. McGuire, "The Nature of Attitudes and Attitude Change," in G. Lindzey and E. Aronson, eds. Handbook of Social Psychology/ Vol. 3. Reading, Mass. 1968.

M. Nakanishi, "Personality and Consumer Behavior: Extensions," Proceedings. Third Annual Conference Association for Consumer Research, 1972, 61-65.

A. E. Parr, "In Search of Theory," in H. M. Proshansky, W. H. Itterson and L. G. Rivlin, eds. Environmental Psychology. New York: Holt, Rinehart & Winston, 1970.

L. Percy, "Relating Personality Profiles and Consumer Behavior via Underlying Attitude and Perceived Behavior" paper presented at the American Psychological Association 83rd Annual Convention, Chicago, 1975.

R. C. Singleton, "A Fortran Program for Minimum Variance Clustering of Multivariate Data," Mathematics and Statistics Division Memorandum, Office of Naval Research, October, 1969.

D. L. Sparks and W. T. Tucker, "A Multivariate Analysis of Personality and Product Use," Journal of Marketing Research, Vol. VIII (February, 1971) 68-70.