Attribution Theory in Marketing Research: Problems and Prospects

Citation:

R. E. Burnkrant (1975) ,"Attribution Theory in Marketing Research: Problems and Prospects", in NA - Advances in Consumer Research Volume 02, eds. Mary Jane Schlinger, Ann Abor, MI : Association for Consumer Research, Pages: 465-470.

Advances in Consumer Research Volume 2, 1975      Pages 465-470

ATTRIBUTION THEORY IN MARKETING RESEARCH: PROBLEMS AND PROSPECTS

R. E. Burnkrant, University of California, Berkeley

[R. E. Burnkrant is Assistant Professor, Schools of Business Administration, University of California, Berkeley.]

Attribution theories address the issue of how people infer, from limited available evidence, unobservable attributes or dispositions about the objects and organisms in their environment. As such, they are theories about how people go beyond the directly observable "data" to infer further elements, that is, to complete a partial representation of some focal object. Attribution theories, then, are theories about how people make attributions. They attempt to account for the conditions in which and the extent to which people are able to infer dispositions or states in other organisms or objects from limited available evidence.

The bulk of the attribution theory research may be traced back to Heider's (1958)-work on causal inference in which he addressed the issue of how people attribute causes to the events or occurrences in their environment. This work, was extended by Jones and Davis' (1965) theory of correspondent inferences and Kelley's (1967) presentation of attribution theory. Both of the latter contributions sought to further specify the conditions under which observers are able to infer unobservable dispositions or attributes in the organisms or objects they encounter. These two contributions have had a direct impact on the consumer behavior applications of attribution theory.

ATTRIBUTION RESEARCH IN MARKETING

The bulk of the marketing research on attribution theory published to date has focused on communications issues. This work, which we might lump together as representing the Settle school, has sought to specify the conditions under which a person's statements about an object will tend to be informative about that object.

In a study entitled "Attribution Theory and the Acceptance of Information," Settle (1972) found that the type of source which would be most informative about an object depends on the type of object about which information is sought. His independent variables were product type (i.e., complex, visible, durable, or multipurpose) and source type (i.e., expert, close friend, own experience over time, or own experience in a variety of ways). His dependent variable was the subject' 2 confidence in a "good" product choice. No attempt was made to deal with, explicate, or measure any part of an attribution process. The only relevance of the study at all to attribution theory is in the rather loose correspondence of his source types to three of Kelley's inference validating criteria (i.e., consistency over time, consistency over modality, and consensus).

In one of this group's earliest studies, Settle, Faricy, and Warren (1971) manipulated the consistency of the responses actors made to each of several movies. By varying this consistency the researchers found that subjects were able to predict how a movie would be rated if the movie were consistently evaluated by several actors and, conversely, how an actor would rate a fourth movie if he rated all previous movies consistently and uniquely. They also measured the subject's confidence in these ratings.

It was suggested that people make attributions about others on the basis of the others' consistency of responding. Implicit in this is the assumption that, if a person gives the same response to several objects and if this response is unique to that person in the sense that it is not also made by others, some sort of biasing disposition such as the motivation to sell the objects is attributed to the person. On the other hand, if he makes a distinct or different response to each object and if these responses are consistent with those made by others the person would not tend to be seen as having some biasing disposition; and, hence, his responses should be informative about the object being evaluated. In this latter instance the individual would be seen as an objective observer and what he says would tend to be believed.

This early study led to the prediction that a two-sided message may lead to more attitude change in the recommended direction than a one-sided message. Settle and Golden (1974) attempted to provide support for this conclusion by exposing subjects to messages which were either one-sided or two-sided. They then measured the degree of confidence the subjects had in each statement that the focal brand was superior to brand X on a specific attribute. They combined the confidence rating with the importance the individual assigned to that attribute to obtain; derived measure which they called "expected value." They interpret their results as supporting the proposition that a two-sided message is more effective than a one-sided message.

They suggest that subjects in the one-sided condition attribute to the source the motivation to "sell his particular brand." Presumably then, subjects in the two-sided condition attribute to the source the motivation to objectively inform or provide information about the brands. It is suggested that these attributions affect the believability of the message and, hence, the resultant belief scores about the focal object.

It seems to me that there are substantial problems with this research. The first problem concerns their failure to measure the attributions made by the subjects or to provide other validating evidence to support their contention that their treatments manipulated these attributions. The obtained dependent measures (i.e., predicted ratings or "expected values") were assumed to be the result of attributions made by the subject. These attributions, however, were not verified by any direct measurement. They were merely assumed to mediate the independent-dependent relationship under consideration. In the absence of substantiating evidence their contention remains merely an interesting hypothesis or possible explanation for the obtained results.

Attributions have frequently been measured in the literature. For instance, Calder and Burnkrant (1973) asked subjects-to describe an actor on a set of bi-polar adjective scales which yielded three personality dimensions. McArthur (1972) asked subjects to assign the cause of an actor's behavior by indicating whether it was due to something about the person, something about the stimulus object, something about the particular circumstances, or some combination of these causes. It is heartening in this regard to see that Mizerski (1974) attempted to provide a somewhat similar measure in his study when he asked subjects to estimate the extent to which "other reasons--reasons having nothing to do with" the stimulus abject, affected the actor's opinion.

A second and somewhat related problem is that, in the absence of direct measures of the attribution process, the contribution to our understanding of communications processes is minimal. They deal with the same variables treated in the extensive Yale Communications Research Program (e.g., Hovland, Janis, and Kelley, 1953) although they propose an attributional explanation for the observed phenomena. Thus, the finding that a two-sided message is more effective than a one-sided message is not all that unique. In their classic study in this area Hovland, Lumsdaine, and Sheffield (1949) found that, with an educated audience (by which they meant high school graduates), a two-sided message is more effective than a one-sided message at inducing belief change in the recommended direction Since the college students used by Settle and Golden (1974) are clearly consistent with what Hovland, Lumsdaine, and Sheffield consider to be an audience high in education their finding is not very surprising.

PROSPECTS FOR FUTURE RESEARCH

While an attempt to actually measure the attributions people make about a message source and the effects these attributions have on attitude change certainly seems warranted, there appear to be other marketing oriented issues to which attribution theories can offer a more unique contribution.

Since people make attributions about others in all situations every day as a normal part of coping with their environment, these attributions seem likely to exert an important influence on the purchase decisions people make. Therefore, research which addresses itself more directly to the verification and further specification of the inference process and its role or roles in consumer behavior seems to be badly needed.

There, no doubt, are many directions which this research could take. In the following material I would like to specify some areas to which I think attribution theories can contribute by looking at research in which I have been involved and some of the research questions related to this.

One approach which seems to merit investigation is to focus on the extent to which people base their purchase decisions on the attributions others are expected to make about them on the basis of their use of particular product brands in specific situations. In this regard Calder and Burnkrant (1973) --found that observers are able to attribute personality dispositions to others on the basis of their use of particular brands in a given social context. They found, further, that the attributions made by an observer are dependent upon the social context in which the brand is used.

This study may be viewed as the first step in a research program directed at uncovering the role played by attributions about users in product purchase decisions. In other words, it may be that buyers and users of products are aware that others will make attributions about them on the basis of their product usage in particular situations. This knowledge may influence their own purchase decisions. It may be, for instance, that people buy products and use them in particular situations in order to reinforce or vary the concept or perception others have about them. This seems consistent with the literature which has attempted to relate the self-concept to the perception people have of the typical user of their favorite brand.

If these conjectures are supported by research evidence it would then seem appropriate to try to specify the conditions, both in terms of situational and product dimensions, under which these attributions would be most important in determining purchase. One may hypothesize, for instance, that the expected attributions of others would be important determiners of purchase where the individual plans to use a conspicuous product in a highly involving situation.

Attribution theory seems also to have considerable potential for integrating and adding meaning to the substantial body of literature on product perception and social influence. One of the important criteria of any good theory, however, is that it leads us to ask more insightful questions capable of advancing our knowledge; and, in this regard too, attribution theory seems relevant to this literature.

Commonly, an individual in a purchasing context is faced with the problem of evaluating the worth of a product which he cannot possibly objectively evaluate from physical manipulation and observation of its characteristics alone. In these situations the individual would be expected to turn to other items of information which he has learned to use as a basis for inferring the characteristics of the object. Among the cues that have been found effective in various situations are price (e.g., Valenzi and Andrews, 1971), brand (e.g., Allison and Uhl, 1964), and the behavior of others (e.g., Venkatesan, 1966).

Therefore, we might expect these cues to affect the individual's perception of the product to the extent that the individual is unable to directly observe an objectively evaluate the product through the use of cues intrinsic to the product. It should, perhaps, be noted that Olson and Jacoby (1972) found that what they call "intrinsic cues" were seen as better predictors of quality than extrinsic cues such as price.

Attribution theory indicates that the confidence in an inference made about an object would be stronger if the attributional implications of all cues associated with the object were consistent than if they were inconsistent. Further, the theory would suggest that a cue would be discounted to the extent that reasons other than those related to the characteristics of the product could have accounted for the cue's presence. Both derivations seem to be directly applicable to the area of product perception.

Many other applications of attribution theory could, no doubt, be suggested. These applications, however, should consider, in much more direct terms than has been recognized in the previously-published applications of the Settle school, the extent to which inferences mediate aspects of (buyer) behavior, the types of inferences which mediate this behavior, and the conditions under which these inferences may be reliably obtained. These, after all, are the considerations which provide the central focus and unique contribution of the attribution theories.

REFERENCES

Allison, R. I., and Uhl, K. P. Influence of beer brand identification on taste perception. Journal of Marketing Research, 1964, 1, 36-39.

Calder, B. J., and Burnkrant, R. E. Attributions based on consumer behavior. University of Illinois, College of Commerce and Business Administration. Working paper number 130, 1973.

Heider, F. The psychology of interpersonal relations. New York Wiley, 1958.

Hovland, C. I., Lumsdaine, A. A., and Sheffield, F. D. Experiments on mass communication. Princeton: Princeton University Press, 1949.

Jones, E. E., and Davis, K. E. From acts to dispositions The attribution process in person perception. In L. Berkowitz (Ed.), Advances in experimental social psychology. Vol. II, New York Academic Press, 1965.

Kelley, H. H. Attribution in social psychology. Nebraska Symposium on Motivation, 1967, 15, 192-238.

Mizerski, R. W. A test of the relationship between trait and causal attribution, Paper presented at the Fifth Annual Conference of the Association for Consumer Research, Chicago, November, 1974.

Olson, J. C., and Jacoby, J. Cue utilization in the quality perception process. Paper presented at the Third Annual Conference of the Association for Consumer Research, ChicaRo, November 1972.

Settle, R. B. Attribution theory and acceptance of information. Journal of Marketing Research, 1972, 9, 85-88.

Settle, R. B., Faricy, J. H., and Warren, G. T. Consumer information processing: Attributing effects to causes. Paper presented at the Second Annual Conference of the Association for Consumer Research, College Park, Maryland, September 1971.

Settle, R. B., and Golden, L. L. Attribution theory and advertiser credibility. Journal of Marketing Research, 1974, 11, 181-185.

Valenzi, E. R., and Andrews, I. R. Effect of price information on product quality ratings. Journal of Applied Psychology, 1971, 55, 87-91.

Venkatesan, M. Experimental study of consumer behavior, conformity and independence. Journal of Marketing Research, 1966, 3, 384-387.

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