Attitude Function: Is It Related to Attitude Structure?

Mark P. Zanna, University of Waterloo
[ to cite ]:
Mark P. Zanna (1990) ,"Attitude Function: Is It Related to Attitude Structure?", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 98-100.

Advances in Consumer Research Volume 17, 1990      Pages 98-100


Mark P. Zanna, University of Waterloo

Historically, attitudes have been conceptualized as either one component, consisting of an individual's feelings or affective responses to an attitude object, or three components, consisting of, in addition, an individual's knowledge or cognitions about and behavior towards an attitude object. With the widespread awareness that affective and evaluative responses should no longer be considered synonymous (cf. Abelson et al., 1982), a consensus seems to be emerging around a resolution to this conceptual controversy. Following Zanna and Rempel (1984,1988) among others (e. g., Cacioppo et al., 1989; Eagly & Chaiken, in preparation), attitudes are now viewed as evaluations based on,-or developed from, three general classes of information: (1) affective or emotional information, (2) cognitive information, and/or (3) information concerning past behaviors or behavioral intentions. Further, evaluations or attitudes are viewed as influencing three modes of response, including affective, cognitive and behavioral responses.

When evaluations are based primarily on (utilitarian) beliefs about the attitude object, this present view can be reduced to something like the formulation proposed by Fishbein and Ajzen (1975). When evaluations are based primarily on affects produced by or associated with the attitude object, this view can resemble the formulation proposed by Zajonc (1980). Finally, when evaluations are based on inferences from past behavior, this view can be similar to Bern's theory of self-perception (1972). Thus, this general framework or formulation not only seems to resolve the one vs. three component controversy, it suggests a way of integrating several of the most prominent theories of attitude formation. More important, perhaps, it suggests exciting new avenues of research.

Because one's overall evaluation is assumed to be a function of what evaluative-relevant information is chronically or, even, acutely salient, one of the interesting implications of this perspective is that an individual can hold more than one attitude toward a given attitude object. This notion is nicely captured by Tim Wilson's "slide rule" metaphor. Rather than having a single attitude toward the object, individuals are presumed to have a range of positions that they view as acceptable. This interesting possibility was exploited in creative ways by each of the participants in the symposium.

Wilson and his colleagues conducted a series of intriguing studies in which subjects were asked to introspect about their attitudes, explaining why they held particular attitudes (Wilson, Lisle, & Kraft, this symposium). To the extent that introspecting about reasons for attitudes makes salient otherwise nonsalient beliefs, attitudes temporarily change. As a consequence, individuals come to make unwise (consumer) decisions in the short run and, in the long run, behave inconsistently with these unstable evaluations. These studies are extremely fascinating. Not only are the results generally consistent with this emerging perspective, I believe the experiments themselves make little sense without such a perspective.

Abe Tesser and Murray Millar's research is also fascinating. This research follows as well from the notion that because attitudes can be based on different sources of information, we can hold more than one attitude toward the attitude object depending upon, in Tesser and Millar's case, whether we are led to focus on our feelings or our beliefs (Tesser & Millar, this symposium). In their research, Tesser and Mill-ar go further by suggesting, and demonstrating, that attitudes based on affect seem to influence consummatory behavior while attitudes based on cognition seem to influence instrumental behavior. Again, it would appear that this research is generally consistent with this emerging perspective and would have made little sense without it.

Finally, the research of Sharon Shavitt and Russell Fazio (Shavitt & Fazio, this symposium) capitalizes on the notion that the salience of information dimensions (e. g., taste vs. social impression) can temporarily influence one's overall attitudes, and, thus, behavior, toward consumer products, especially for individuals with high self-monitoring tendencies (i. e., individuals who tend to be more attentive to situational cues, especially cues concerning social image, and who tend to hold less accessible attitudes). Again, this research fits nicely with this emerging perspective. What is particularly impressive to me, at least, is the ease (and creativity) with which Shavitt and Fazio (as well as Wilson, Tesser and their respective colleagues) have been able to change people's attitudes, at least temporarily, by manipulating the salience of I different dimensions of evaluative information.

At the University of Waterloo my students and I have also begun to explore the idea that attitudes can be based on different sources of information which may not have entirely consistent implications for an overall evaluation. For example, Megan Thompson, in her Masters thesis research (Thompson, 1989), wondered what would happen if, and when, two or more sources of information have contradictory implications for the overall evaluation. This is a state of affairs we might call attitude ambivalence, as in the case when your "mind" (i. e., cognition) tells you one thing (e.g., that exercise is good for your health) but your "heart" (i. e., affect) tells you another (e. g., that exercise is painful). To investigate the possibility that attitudes are often ambivalent, Megan induced her subjects to provide separate positive and negative evaluations of several social attitude topics (such as AIDS, euthanasia, reinstating capital punishment, and abortion) by having subjects consider separately, first, their positive beliefs and feelings and, then, their negative beliefs and feelings for each social issue. Following Kalman Kaplan (1972), Megan defined ambivalence as "the total amount of evaluation less the degree to which the summary evaluation is polarized." For example, using four-point scales created by splitting seven-point bipolar scales (e. g., favorable-unfavorable) with the neutral point retained in each, + 2 units of positive evaluation indicated at the same time as - 3 units of negative evaluation produces 5 units of total evaluation (I+ 21 + I- 31), which can be decomposed into 1 unit of absolute polarization (l[+ 2] + 1- 3]1), in this case a slight overall negative evaluation, and 4 units of ambivalence. Ambivalence is thus determined by the amount of the lesser evaluation, all of which overlaps with an equivalent amount of the stronger competing evaluation.

Using this procedure, Megan discovered first that there were strong individual differences in the ambivalence individuals seem to experience in their social attitudes. Some individuals seemed to hold ambivalent social attitudes while others did not. Next she determined that theoretically relevant dispositional variables, such as the Need for Cognition (NFC), i. e., stable differences in individuals' tendencies to engage in and enjoy effortful cognitive endeavours (Cacioppo & Petty, 1982) and the Fear of Invalidity (FOI), i. e., stable differences in individuals' concern with error in decision-making (Thompson et al., 1989), were related to this ambivalence. As expected, those high in NFC were less likely to hold ambivalent social attitudes while those high in FOI were more likely to do so. Next she determined that experts, i. e., those who possessed a great deal of knowledge concerning social issues, were less likely to be ambivalent, but, interestingly, the effect for knowledge or expertise was moderated by both personality variables. Expertise was associated with less ambivalence, but only for those high in NPC and low in FOI. In fact, there was a hint that experts high in FOI actually experienced more ambivalence. In any event, it is clear that ambivalent attitudes exist and that some people are more likely to hold ambivalent attitudes than others. The consequences of holding such attitudes will be an important topic for research in the future.

Recently some of my graduate students and I have taken the notion of attitudes being based on different sources of information one step further. Specifically, we began to wonder what it might mean if there were to be a multidimensional structure to the beliefs upon which attitudes were based. For example, in the realm of social perception it seems clear that attributes along the Social Desirability dimension (e. g., warm vs. cold) seem to influence our judgments of liking vs. disliking, while attributes along the Intellectual Desirability dimension (e. g., intelligent vs. unintelligent) seem to influence our judgments of respect vs. disrespect (cf. Lydon, Jamieson, & Zanna, 1988). If so, which judgment, liking or respect, determines our overall evaluation of or attitude toward other persons? Does this depend on the context (e. g., task-related vs. social-emotional-related situations) in which we are asked to form (or retrieve) such attitudes? Do we experience ambivalence when we like but do not respect another (or, conversely, respect but do not like another), or, for the most part, do we compartmentalize these seemingly contradictory judgments? These and other questions will be important topics for future research in the area of impression formation. From the point of view of the consumer researcher, it might be interesting to determine whether two dimensions, similar to the Social and Intellectual Desirability dimensions of implicit personality theory, exist for consumer products, and, if so, whether two dimensions of evaluative judgment, similar to liking and respect, exist as well.

In another recent Masters thesis at the University of Waterloo, Richard Ennis explored the multidimensionality of beliefs upon which evaluations, in this case attitudes toward consumer products, were based. Rich proposed that one way to conceptualize (and operationalize) a functional approach to attitudes would be to suggest that overall evaluations based on different dimensions of beliefs might fulfill different functions (Ennis, 1989). Specifically,_Ennis suggested that the Fishbein and Ajzen (1975) model of attitudes was incomplete because their belief elicitation procedure essentially "pulled" for utilitarian or instrumental beliefs about the attitude object. Following the lead of Sharon Shavitt (1989), Rich demonstrated that it was possible to elicit beliefs concerning consumer items that were more symbolic in nature, and that this was, in fact, easier for products that Shavitt suggested served more symbolic functions, such as a graduation ring or a gold neck-chain, than for products that she suggested served a more instrumental function, such as an air conditioner or coffee. Interestingly, Rich was also able to demonstrate that the overall evaluations of symbolic products, such as graduation rings, were based to a greater extent on (or, at least, correlated to a greater extent with) the symbolic belief dimension (e. g., "a graduation ring is a memento of one's university years"; "a graduation ring is an item to show off your status to others") than on the utilitarian or instrumental belief dimension (e. g., "a graduation ring usually is expensive"; "a graduation ring is not terribly useful"). In contrast, overall evaluations of instrumental products, such as air conditioners, were based to a greater extent on instrumental beliefs (e. g., "an air conditioner enables you to relax comfortably") than on symbolic beliefs (e. g., "in our climate an air conditioner symbolizes extravagance"). Attitudes toward products thought to fulfill both instrumental and symbolic functions (e. g., a University of Waterloo jacket) seemed to be based on both instrumental (e. g., "a U of W jacket looks good") and symbolic (e. g., "a U of W jacket provides a feeling of status because U of W is well respected") beliefs. Future research might determine whether either individual difference variables, such as Self-Monitoring tendencies (Snyder & DeBono, 1989), or situational variables (such as social context "primes"), for that matter, moderate the belief dimension-overall evaluation relation. Further, future research on persuasion might create communications designed to change either utilitarian or symbolic beliefs in order to test the functional hypothesis that in order to change an attitude one must change the function-relevant bases (in this formulation, informational dimensions) from which the attitude is derived.

Finally, Ennis' perspective on how to think about and operationalize a functional approach to attitudes may provide a way to conceptualize symbolic attitudes, including symbolic racism (Sears, 1988). Here the notion is simply that symbolic attitudes (e. g., symbolic racism) (in addition, perhaps, to being based to a greater extent on affect) are based upon symbolic beliefs (e. g., "Blacks are getting more than they deserve," "the streets are unsafe today," and so on) rather (or to a greater extent) than on traditional stereotypic beliefs (e. g., "Blacks are unintelligent," "lazy," and so on).

In any event, by not equating affective response with evaluation and by focusing upon the various, possibly multidimensional, sources of information upon which attitudes are based, attitude researchers (as attested to by the participants of the present symposium) will have more than enough interesting questions to pursue in the future. Most exciting, at least for me, is the possibility that, by considering attitude structure, specifically the information upon which an attitude is based, we can learn more about attitude function.


Abelson, R. P., Kinder, D. R., Peters, M. D., & Fiske, S. T. (1982). Affective and semantic components in political person perception. Journal of Personality and Social Psychology, 42, 619-630.

Bern, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology, Vol 6. New York: Academic Press.

Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42, 116-131.

Cacioppo, J. T., Petty, R. E., & Geen, T. R. (1989). Attitude structure and function: From tripartite to the homeostasis model of attitudes. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function Hillsdale, NJ: Erlbaum.

Eagly, A. H., & Chaiken, S. (in preparation). The psychology of attitudes

Ennis, P. R. (1989). Attitude function: A procedure for its identification and measurement. Unpublished Masters Thesis. University of Waterloo, Waterloo, Ontario.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research Reading, MA: Addison-Wesley.

Lydon, J. E., Jamieson, D. W., & Zanna, M. P. (1988). Interpersonal similarity and the social and intellectual dimensions of first impressions. Social Cognition, 6, 269-286.

Sears, D. O. (1988). Symbolic racism. In P. A. Katz & D. A. Taylor (Eds.), Eliminating racism: Profiles in controversy New York: Plenum.

Shavitt, S. (1989). Operationalizing functional theories of attitude. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function Hillsdale, NJ: Erlbaum.

Snyder, M., & De Bono, K. G. (1989). Understanding the functions of attitudes: Lessons from personality and social behavior. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function Hillsdale, NJ: Erlbaum.

Thompson, M. M. (1989). When a little knowledge is a dangerous thing: Personality based and domain specific antecedents of ambivalent social attitudes. Unpublished Masters Thesis. University of Waterloo, Waterloo, Ontario.

Thompson, M. M., Naccarato, M. E., & Parker, K. C. H. (June 1989). Measuring cognitive needs: The need for structure and fear of invalidity scales. Paper presented at the Annual Meeting of the Canadian Psychological Association, Halifax.

Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35, 151-175.

Zanna, M. P., & Rempel, J. K. (June 1984). Attitudes: A new look at an old concept. Paper presented at the Conference on the Social Psychology of Knowledge, Tel Aviv, Israel.

Zanna, M. P., & Rempel, J. K. (1988). Attitudes: A new look at an old concept. In D. Bar-Tal & A. Kruglanski (Eds.), The social psychology of knowledge New York: Cambridge University Press.