Cognitive Dissonance and Consumer Behavior: a Review of the Evidence

William H. Cummings, University of Iowa
M. Venkatesan, University of Iowa
[ to cite ]:
William H. Cummings and M. Venkatesan (1975) ,"Cognitive Dissonance and Consumer Behavior: a Review of the Evidence", in NA - Advances in Consumer Research Volume 02, eds. Mary Jane Schlinger, Ann Abor, MI : Association for Consumer Research, Pages: 21-32.

Advances in Consumer Research Volume 2, 1975      Pages 21-32

COGNITIVE DISSONANCE AND CONSUMER BEHAVIOR: A REVIEW OF THE EVIDENCE

William H. Cummings, University of Iowa

M. Venkatesan, University of Iowa

[William H. Cummings is a doctoral student in social psychology and M. Venkatesan is Professor of Business Administration at the College of Business Administration. The University of Iowa.]

The theory of cognitive dissonance has generated a good deal of research in consumer behavior, however, this research has not yet been fully reviewed and critiqued. The relevant evidence from three phases of consumer behavior is reviewed here: pre-decisional determinants of product preference, post-decisional determinants of product preference, and information seeking behavior. The evidence from the first two areas generally supports the dissonance-based predictions, while the evidence from the third area generally fails to support the predictions.

There have been a number of nasty rumors lately that "cognitive dissonance theory is dead" and that "cognitive dissonance theory began to be out of date a long time ago." This may well be so. It would seem, however, that before we bury dissonance theory, we should make sure that all the phenomena and events which it explains (particularly post-decisional attitude change) are also dead or out of date. We should make sure that the theory is entirely incorporated within new and more general theories. And we should be sure that dissonance theory is incapable of generating further predictions of interest to social scientists.

A few years ago, dissonance theory generated considerable interest and enthusiasm. The generalizability of the theory from the social psychology laboratory to consumer behavior appeared quite certain. Twenty-three studies have since examined the applicability of dissonance theory to consumer behavior. [We have attempted to include all relevant studies from the Consumer Behavior literature while including as few studies as possible from the social psychological literature. Only three studies from the social psychological literature have been included in this review. Three studies from the consumer behavior literature have been excluded. Oshikawa's study (1970) is not directly relevant to consumer behavior. Reizenstein's study (1971) hints at relevance to dissonance theory but fails to provide necessary choice-making conditions for producing dissonance among the subjects. Bell's study (1967) is questionable on basic conceptual grounds, viz., "direct" indication of magnitude of dissonance arousal.] Although a number of reviews have examined the theoretical relevance of dissonance theory to the consumer behavior area, no review has apparently yet attempted to integrate and critique the empirical evidence on this matter. This will be the major purpose of this review. Before we lay dissonance theory to rest, it may be worthwhile to examine the studies and answer a few questions:

(a) What does the available evidence tell us about the applicability of dissonance theory (and related approaches) to consumer behavior?

(b) Does the evidence (both positive and negative) stand up to methodological scrutiny? Can the methodological problems be avoided in the future?

(c) What sort of research is needed for a better understanding of the applicability--if any--of dissonance theory (and related approaches) to consumer behavior?

(d) What interesting predictions--if any--can still be derived from dissonance theory and related approaches?

The initial presentation of the theory of cognitive dissonance (Festinger, 1957) clearly demonstrated the relevance of dissonance theory to the purchase situation. Festinger derived his theory from two basic principles: (a) Dissonance is uncomfortable and will motivate the person to reduce it, and (b) the dissonant individual will avoid situations which produce further dissonance. Dissonance is a post-decisional phenomenon and should therefore be a postpurchase phenomenon. Dissonance theory works according to a reality principle in that the cognition least anchored in reality will be the one most likely to change; since the purchase will be a strongly anchored behavior, the cognitions most likely to change will be those concerning satisfaction with the purchased product relative to rejected products. Dissonance should increase with the importance of the cognitions, and to the extent that the individual commits some time and some money in the purchase, many purchase decisions should be important ones. So the potential application of dissonance theory to consumer behavior is considerable. In answering the question of whether the empirical findings have supported this application, we have found it convenient to divide the research into three areas: (a) p{e-decisional determinants of product preference/satisfaction; i.e., those variables which increase magnitude of dissonance; (b) post-decisional determinants of product preference/satisfaction; i.e., those variables which decrease magnitude of dissonance; and (c) effects of dissonance arousal on information seeking behavior. We will now examine the evidence within each area.

PRE-DECISIONAL DETERMINANTS OF PRODUCT PREFERENCE

Most applications of dissonance theory in consumer behavior fall into this category--attempts to modify product satisfaction via the control of events prior to the purchase decision. Brehm provided the landmark study in 1956 on the effect of magnitude of dissonance arousal on attitude change. Brehm observed--as predicted by dissonance theory--that when subjects had to choose between two products which were similar in desirability, there was substantial attitude change in favor of the chosen product relative to the unchosen product. When subjects chose between two products that were disparate in desirability, there was much less attitude change. In all conditions, the chosen product was favorably reevaluated and the unchosen product was derogated or unchanged in desirability.

Later studies have replicated and extended Brehm's findings. LoSciuto & Perloff (1967) observed significantly more favorable reevaluation of the chosen product (relative to the unchosen product) when subjects had to choose between two record albums close in desirability than when subjects chose between two albums disparate in desirability. Thus, the more similar the products which are considered, the greater is the magnitude of dissonance, and the greater the "spread" in ultimate preference rankings of the two products. Sheth (1968, 1970) also demonstrated the predicted positive reevaluation of chosen products and derogation of unchosen products for various product classes. for both students and housewives.

Likewise, Anderson, et al. (1966), have demonstrated the predicted post-decisional spread in desirability of the alternatives. Anderson, et al., also demonstrated that the reevaluation tendency increased with the number of alternatives. However, this effect was significant only when the initial difference in desirability between the alternatives was large. The number of alternatives seems to have little effect on magnitude of dissonance if dissonance is already quite high.

Mittelstaedt (1969) used a modified free choice paradigm to demonstrate that "buyers" who became committed to one particular "brand" would persist in their preference for that "brand," even when offered the choice of a "brand" they originally preferred more. Subjects first rank-ordered nine swimsuits and were then given a choice either between their third-ranked and fourth-ranked choices (high dissonance condition) or between their third-ranked and fifth-ranked choices (low dissonance condition). Later, most of the high dissonance "purchasers" preferred the chosen (third-ranked) swimsuit over a swimsuit they originally ranked higher. Most of the low dissonance subjects did not.

A simulated purchase study by Holloway (1967) examined the effects of inducement to purchase, cognitive overlap, and presence/absence of supportive information. These factors had the predicted -but a nonsignificant--effect on product reevaluations. Holloway also manipulated subjects' anticipated dissonance and found (contrary to expectations) greater product reevaluation under low anticipated dissonance, but, again, nonsignificantly so.

Dissonance theory also predicts that product satisfaction should increase when the amount of effort expended in obtaining the product is increased. Cardozo (1965) found that subjects who expended a good deal of effort to obtain a product (a ball-point pen) rated this product significantly higher than did subjects who expended little effort. However, this effect was significant only when subjects were expecting a higher quality product than they actually received.

A common feature of all the above studies of consumer behavior with one exception (Sheth, 1970) is that they were conducted in the laboratory. Doob, et al., (1969) have employed a field study and a basic prediction from the forced compliance paradigm, e.g., "the less the pressure...put upon the person to perform the act, the greater the dissonance (Kiesler et al, 1969, p. 206)." Doob, et al., observed that introducing a product brand at a special, low, discount price would hurt later sales of that product. Thus, the higher the initial price paid for the product, the greater the magnitude of dissonance aroused, the greater the ultimate satisfaction should be, and thus, the greater the tendency to repurchase that brand. Doob, et al., matched this prediction for five different household product classes.

The findings presented above are summarized in Table 1. This summary indicates that 11 findings have supported predictions from dissonance theory, while 2 findings have failed to support the predictions, and there are 4 findings which could be considered as "mixed." Before drawing any conclusions from this simple count, we must examine the findings in view of the methodological problems in this area. The methodological problems of each study are presented in the third column of Table 1.

One criticism that has been generally made of investigations in consumer behavior which have employed the free choice paradigm is the criticism of "regression artifact" or "ceiling effect." Specifically, Oshikawa (1968) has demonstrated that free choice-type reevaluations made in the absence of any choice may show a dissonance-type effect. But when corrected for ceiling effect, this dissonance-type effect disappears. To be sure, however, two of the five free choice studies reported above have included a correction or control for this sort of regression artifact (Brehm, 1956; LoSciuto & Perloff, 1967). In one other study (Sheth, 1970), findings were reported that directly contradict the ceiling effect (see Sheth, 1971). Thus, the appropriate control or correction for regression should be included in any study where regression artifact is a possibility. However, these studies have, in general, demonstrated the predicted reranking of the chosen and unchosen products, despite the possibility of regression artifact.

TABLE 1

PRE-DECISIONAL DETERMINANTS OF PRODUCT PREFERENCE

Another problem of studies in this area is a possible failure to meet the prerequisite conditions for producing dissonance. Accumulated evidence in social psychology has suggested that predictions from dissonance theory cannot clearly be made unless the choice situation meets three conditions: (a) volition concerning the choice, (b) irrevocable commitment to the decision (product choice), and (c) importance of the choice to the individual's self-concept. The prerequisite of irrevocable commitment has generally been met, while the prerequisite of volition has not always been met. The condition of importance of choice is most questionable: Has it always been upheld? Certainly the role-playing situations, which have been employed in some studies, would necessarily reduce the importance of the decision for the individual and his future behavior. Although some studies which have failed to clearly meet the prerequisite conditions may have supported the theory, these effects may be due to causes other than dissonance arousal.

Some studies which have cited dissonance theory as a source of hypotheses do not, in fact, appear to be relevant to the main theory. For example, three studies have examined the effect of prior expectations on subsequent product evaluation (Cardozo, 1965; Olshavsky and Miller, 1972; Anderson, 1973). While this is certainly an important question, it is a question distantly relevant to dissonance theory. The effects of such manipulations cannot be taken as either support or failure to support the theory.

Two studies cited in Table 1 have taken correlational findings as causal support or disconfirmation of the theory. For example, one study found no effects of subjects' post-decisional ratings of pre-decisional conflict on subsequent measures of brand preference, intention to repurchase, or brand loyalty. These findings were interpreted as a disconfirmation of dissonance theory. What the results do demonstrate is that there is little if any correlation between the initial measure of conflict and the subsequent measures. Moreover, the technique of measuring dissonance by asking the subject about his level of conflict or worry has been questioned (Oshikawa, 1972; Hawkins, 1972). These investigations have demonstrated that such measures are correlated with measures of general confidence and anxiety.

A final problem in interpreting these studies is the compelling alternative explanations generated by studies which appear to be otherwise methodologically sound. For example, Doob et al. (1969) demonstrated that introducing a product at a special low price decreases later sales, as predicted by the theory. But there was no safeguard against buyers stockpiling goods at the initial purchase, during the sale. Also, there was no indication that a sale was in progress, and some buyers may have been "frustrated" after the sale when they were suddenly faced with a more expensive product.

In summary, the evidence has generally supported the predicted relationships between pre-decisional factors and subsequent product preference. No finding from this area has conclusively disconfirmed any prediction from the theory. On the other hand the positive evidence is importantly qualified by these major methodological concerns. But it should be noted that the evidence in favor of the applicability of dissonance theory to this area is more voluminous and somewhat more substantial than the evidence against. There is no single methodological or conceptual limitation that runs throughout all the positive instances of the dissonance effect. There is no other single explanation -- other than that of cognitive dissonance theory -- that can fully account for the results of these studies. These findings, then, have substantiated the predicted relation between pre-decisional factors which would tend to affect magnitude of dissonance and subsequent product preference.

POST-DECISIONAL DETERMINANTS OF PRODUCT PREFERENCE

Three studies have employed a postpurchase reinforcement technique to increase purchasers' satisfaction with the obtained product. Two studies found greater product satisfaction with the technique, as predicted, while the third study obtained mixed results. Table 2 summarizes these findings. VanDyke (19662 demonstrated that a supportive letter from the dealer increased consonance, [VanDyke's estimate of consonance is theoretically appropriate and should be of some interest. The relative weighting of each of the five most important attributes involved in a car purchase (as rated by each subject) was multiplied by the difference between the subject's rating of the chosen brand and the unchosen brand on that attribute. This product was summed over all five attributes and constituted the consonance estimate for each subject.] following the purchase of a new car. Purchasers who did not receive the supportive letter showed no such increase in consonance. Donnelly and Ivancevich (1970) found that two supportive phone calls within two weeks following the purchase of a new car would decrease the back-out rate from about six percent to about two and one-half percent. Hunt (1970) likewise found that a supportive letter following the purchase of a refrigerator reduced post-transaction anxiety and increased disposition to repurchase. However, a supportive telephone contact -- which was similar to the letter in content -had the opposite effect. (Both effects were statistically nonsignificant.)

TABLE 2

POST-DECISIONAL DETERMINANTS OF PRODUCT PREFERENCE

How shall we interpret these results? First, it is obvious that the positive effects (postpurchase contact increasing satisfaction) are all interpretable within common reinforcement notions. Postpurchase reinforcement provided in the form of social support should directly increase produce satisfaction and tendency to repurchase. This is a much simpler explanation than is dissonance theory. The interpretation of these studies via dissonance theory, however, is beneficial for two reasons. (@ Dissonance theory explains why these social contacts should be particularly reinforcing immediately following the purchase, i.e., because the purchaser of a major product is experiencing considerable dissonance and is in need of social support. (b) The dissonance interpretation is consistent with the interpretation given the studies in the first section (predecisional determinants of product satisfaction). Thus, while dissonance theory can explain both sets of findings (pre and post), reinforcement notions work best only with the latter set of findings.

Thus, despite the presence of one negative finding and some compelling alternative explanations, it appears that dissonance theory offers the most parsimonious account of the studies reviewed in this section.

EFFECTS OF DISSONANCE AROUSAL ON INFORMATION SEEKING BEHAVIOR

Festinger's second basic hypothesis regarding dissonance theory was that the dissonant individual should actively avoid information which would tend to increase the dissonance and seek information which would tend to support his decision. However, the evidence within social psychology -- obtained under optimal laboratory conditions -- has failed to support this hypothesis. Freedman and Sears (196$) and Sears (1968) have concluded that the available evidence does not support either (a) a general preference for supportive over nonsupportive information or (b) a greater information seeking/avoidance tendency by high dissonance subjects.

Is there evidence within the area of consumer behavior to modify this conclusion? The relevant findings are summarized in Table 3, and the answer is a clear "no," there is no evidence here which strongly supports a link between magnitude of dissonance and information seeking behavior. All the findings in this area are, in turn, qualified by some compelling methodological problems. Ehrlich et al. (1957) and Engel (1963) found the predicted effects of magnitude of dissonance on some measures but failed to obtain the predicted effects on a number of additional measures. Kassarjian and Cohen (1965) found the predicted inverse relation between the amount a person smoked and how believable that person found the smoking-cancer link to be. However, this is correlational evidence. Perhaps strongly anchored smoking behavior reduced belief in the cancer evidence. Or perhaps people who didn't believe the link became the heaviest smokers. Or maybe a certain type of person tended to be both a nonbeliever and a heavy smoker. LoSciuto and Perloff (1967) found no effect of magnitude of dissonance on a recognition task. However, this recognition task followed a reranking task on which differential reranking of the products was obtained. If dissonance was reduced by reranking the products, it is unlikely that there will be any effect of magnitude of dissonance on subsequent trials.

Thus, the evidence on this issue is most equivocal. In view of repeated failures to demonstrate any consistent effect of magnitude of dissonance on information seeking, avoidance, and perceptual processing, we cannot conclude at this time that dissonance arousal factors are relevant to postpurchase information seeking variables in consumer behavior.

SUMMARY, CONCLUSIONS, AND EXTENSIONS

We feel that the relevant evidence on predecisional and postdecisional determinants of product preference (satisfaction) has given generous support to dissonance theory. Each study reviewed here has been interpreted in terms of its methodological and conceptual limitations. In view of these limitations, the negative evidence has not conclusively disconfirmed any aspect of the theory. Moreover, dissonance theory appears to offer a unique and parsimonious explanation for the positive findings. We conclude, then, that the generalizability of cognitive dissonance theory to consumer behavior -- excluding information seeking behavior -- has been substantiated.

The questions which would now appear to be of primary interest in this area are the interaction questions; specifically, what are the precise conditions under which we would expect the dissonance effects to occur. The studies in the consumer behavior area have not been primarily oriented toward this question. From the social psychological literature we know that predictions from the theory can be most accurately applied when there is high perceived volition concerning the choice, when there is an irrevocable commitment to the choice, and when the choice is an important one (ego-involving) for the individual. These factors raise a number of interesting questions in application: What sorts of products are sufficiently important to arouse dissonance? Are certain product classes more susceptible to dissonance effects than others, even when cost is held constant? What is the optimal balance between volition and positive inducement in maximizing both initial purchase tendency and postpurchase satisfaction? What modes of dissonance reduction are normally used by the consumer in everyday life? What are some of the long-term effects of differential dissonance arousal and reduction; i.e., is the increased product satisfaction that results from dissonance reduction a lasting effect?

TABLE 3

EFFECTS OF DISSONANCE AROUSAL ON INFORMATION SEEKING BEHAVIOR

Experimental research in this area has concentrated primarily on the free choice paradigm. However, there are other paradigms which may be used for research and application. The following appear promising: (a) forced compliance paradigm, (b) fait accompli paradigm, (c) attribution theory, and (d) commitment theory.

In conclusion, we feel that the application of dissonance theory to consumer behavior -- at least within the areas of predecisional and postdecisional determinants of product satisfaction -- has been substantiated. There have been a number of methodological problems in this line of research, but most of them are of a sort that can be avoided by means of a simple inventory of possible pitfalls or the addition of the appropriate control conditions. It does not appear that dissonance theory can be adequately incorporated within any other framework and retain its power to generate new and intriguing hypotheses. And that power -- to generate hypotheses D should continue to be the major selling point of dissonance theory.

REFERENCES

Anderson, L. Y., J. R. Taylor, and R. J. Holloway. The consumer and his alternatives: An experimental approach. Journal of Marketing Research, 1966, 3, 62-67.

Anderson, R. E. Consumer dissatisfaction: The effect of disconfirmed expectancy on perceived product performance. Journal of Marketing Research, 1973, 10, 38-44.

Bell, G. D. The automobile buyer after the purchase, Journal of Marketing, 1967, 31 (3), 12-16.

Brehm, J. W. Post-decision changes in the desirability of alternatives. Journal of Abnormal and Social Psychology, 1956, 52, 384-389.

Cardozo, R. N. An experimental study of customer effort, expectations and satisfaction, Journal of Marketing Research, 1965, 2, 244-249.

Chapanis, N., and A. Chapanis. Cognitive dissonance five years later. Psychological Bulletin, 1964, 61, 1-22.

Cohen, J. B., and M. E. Goldberg. The dissonance model in post-decision product evaluation, Journal of Marketing Research, 1970, 7, 315-321.

Cohen J. B., and M. J. Houston. Cognitive consequences of brand loyalty. Journal of Marketing Research, 1972, 9, 97-99.

Donnelly, J. H., and J. M. Ivancevich. PostPurchase reinforcement and back-out behavior, Journal of Marketing Research, 1970, 7, 399-400.

Doob, A. N., J. M. Carlsmith, J. Freedman, T. K. Landauer, and S. Tom. Effect of initial Selling price on subsequent sales. Journal of Personality and Social Psychology, 1969, 11, 345-350.

Ehrlich, D., I. Guttman, P. Schonbach, and J. Mills. Postdecision expOsure to relevant information. Journal of Abnormal and Social Psychology, 1957, 54, 98-102.

Engel, J. F. Are automobile purchasers dissonant consumers? Journal of Marketing, 1963, 27, 55-58.

Festinger, L. A theory of cognitive dissonance- Stanford: Stanford University Press, 1957.

Freedman, J. L., and D. O. Sears. Selective expoSure. In L. Berkowitz (Ed.), Advances in experimental social psychology. Vol. 2. New York: Academic Press. 1965.

Hawkins, D. I. Reported cognitive dissonance and anxiety: Some additional findings, Journal of Marketing, 1972, 36 ^3), &3-66.

Holloway, R. J. An experiment on consumer dissonance. Journal of Marketing, 1967, 31 (1). 39-43.

Hunt, S. D. Post-transaction communications and dissonance reduction, Journal of Marketing, 1970; 34 (3), 46-51,

Kassarjian, H. H., and J. B. Cohen. Cognitive dissonance and consumer behavior, California Management Review, 1965, 8, 55-64.

Kiesler, C. A., B. E. Collins, and N. Miller. Attitude change: A critical analysis of theoritical approaches. New York: Wiley, 1969.

LoSciuto, L., and R. Perloff. Influence of product preference on dissonance reduction, Journal of Marketing Research, 1967, 4, 286-290.

Mittelstaedt, R. A dissonance approach to repeat Purchasing behavior. Journal of Marketing Research, 1969, 6, 444-446.

Olshavsky, R. W., and J. A. Miller. Consumer expectations, product performance, and Perceived product quality, Journal of Marketing Research, 1972. 9. 19-21.

Oshikawa, S. The theory of cognitive dissonance and experimental research, Journal of Marketing Research, 1968, 5, 429-430.

Oshikawa, S. Can cognitive dissonance theory explain consumer behavior? Journal of Marketing, 1969, 33 (4), 44-49.

Oshikawa, S. Consumer pre-decision conflict and post-decision dissonance, Behavioral Science, 1970, 15, 132-140.

Oshikawa, S. Dissonance reduction or artifact? Journal of Marketing Research, 1971, 8, 514-515.

Oshikawa, S. The measurement of cognitive dissonance: Some experimental findings, Journal of Marketing, 1972, 36 54), 64-67.

Reizenstein, R. C. A dissonance approach to measuring the effectiveness of two personal selling techniques through decision reversal, Combined Proceedings, American Marketing Association, Spring and Fall, 1971, 176-180.

Sears, D. O. The paradox of de facto selective exposure without preference for supportive information. In R. P. Abelson, E. Aronson, W. J. McGuire, T. M. Newcomb, M. J. Rosenberg, and P. H. Tannenbaum (Eds.), Theories of cognitive consistency; A sourcebook. Chicago: Rand-McNally, 1968.

Sheth, J. N. Cognitive dissonance, brand preference, and product familiarity, In J. Arndt (Ed.), Insights into consumer behavior. Boston: Allyn and Bacon, 1968.

Sheth, J. N. Are there differences in dissonance reduction behavior between students and housewives? Journal of Marketing Research, 1970, 7, 243-245.

Sheth, J. N. Dissonance reduction or artifact? A reply, Journal of Marketing Research, 1971, 8, 516-517.

Straits, B. C. The pursuit of the dissonant consumer, Journal of Marketing, 1964, 28, 62-66.

VanDyke, J. E. G study of the arousal and reduction of cost-decision dissonance following the purchase of a new automobile, Unpublished master's thesis, The University of Iowa, 1966.

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