Can Cognitive Dissonance TheoryExplain Consumer Behavior?

SADAOMI OSHIKAWA

Cognitive dissonance theoryis applicable to very limitedareas of consumer behavioraccording to the author. Pub-lished findings in support ofthe theory are equivocal;they fail to show that cogni-tive dissonance is the onlypossible cause of observed"dissonance-reducing" behav-ior.

Experimental evidences areexamined and their weak-nesses pointed out by theauthor to justify his position.He also provides suggestionsregarding the circumstancesunder which dissonance re-duction may be useful inincreasing the repurchaseprobability of a purchasedbrand.

Journal of Marketing, Vol. 33 (October,1969). pp. 44-49.

l tTORE than a decade has passed since the publication of Festing-•L'-'- er's original book on the theory of cognitive dissonance.'Marketing researchers and psychologists have conducted numerousexperiments to test the theory. Whether dissonance theory can beapplied to marketing is a question which has raised considerableinterest among marketing writers.^ The theory asserts that aperson has certain cognitive elements which are "knowledges" abouthimself, his environment, his attitudes, his opinions, and his pastbehavior. If one cognitive element follows logically from another,they are said to be consonant to each other. They are dissonant toeach other if one does not follow logically from the other.

Dissonance can be aroused in three ways and can motivate theperson to reduce this tension in a variety of ways. Dissonance maybe aroused 1) after making an important and difficult decision, 2)after being coerced to say or do something which is contrary toprivate attitudes, opinions, or beliefs, and 3) after being exposedto discrepant information.

The theory does not specify the mode of dissonance reduction butindicates that there are many possible ways to reduce dissonance.Attitude change, opinion change, seeking and recall of consonantinformation, avoidance of dissonant information, perceptual dis-

' Leon Festinger, A Theory of Cognitive Dissonance (Stanford, Cali-fornia: Stanford University Press, 1957).

2 James F. Engel, "Are Automobile Purchasers Dissonant Consumers?"JOURNAL OF MARKETING, Vol. 27 (April, 1963), pp. 55-58; James F.Engel, "Further Pursuit of the Dissonant Consumer: A Comment,"JOURNAL OF MARKETING, Vol. 29 (April, 1965), pp. 33-34; Robert J.Holloway, "An Experiment on Consumer Dissonance." JOURNAL OFMARKETING, Vol. 31 (January, 1967), pp. 39-43; Harold H. Kassarjianand Joel B. Cohen, "Cog^nitive Dissonance and Consumer Behavior,"California Management Revietv. Vol. 8 (Fall, 1965), pp. 55-64; GeraldD. Bell, "The Automobile Buyer After the Purchase," JOURNAL OFMARKETING, Vol. 31 (July, 1967), pp. 12-16; Donald Auster, "Atti-tude Change and Cognitive Dissonance," Journal of Marketing Re-search, Vol. 2 (November, 1965), pp. 401-405; Gerald D. Bell, "Self-Confidence and Persuasion in Car Buying," Journal of MarketingResearch, Vol. 4 (February, 1967), pp. 46-52; and James F. Engeland M. Lawrence Light, "The Role of Psychological Commitment inConsumer Behavior: An Evaluation of the Theory of Cog^nitive Dis-sonance," in Applications of the Sciences in Marketing Management,Frank M. Bass, Charles W. King, and Edgar A. Pessemier (eds.),(New York: John Wiley and Sons, Inc., 1968), pp. 179-206.

44

Can Cognitive Dissonance Theory Explain Consumer Behavior? 45

tortion, and behavioral change are some of the com-mon modes of dissonance reduction.^

Since the theory does not designate the expectedmode of reducing dissonance, most researchers haveadopted the experimental method in which subjectscould reduce dissonance in only one predetermined way.When the subjects responded to the experimentalmanipulation in the manner predicted by dissonancetheory, the dissonance researcher took the results asevidence for the support of the theory. However,some psychologists have suggested that many ofthe findings are the results of built-in artifacts (orbiases) or can be explained by other competingtheories, and that the affirmative result is not neces-sarily unequivocal evidence for the theory.* Thisarticle will attempt to assess this possibility.

This article will also attempt to examine theo-retical issues and experimental findings for each ofthe three dissonance arousal conditions mentionedabove to determine the relevance of the theory tothe study of consumer behavior. In general it ap-pears that the findings are contradictory and are notalways supportive of the theory; however, an at-tempt will be made to sort out and evaluate theevidence.

Post-decision Dissonance

A review of literature on the psychological studyof decision making led Festinger to hypothesize thatdecision making almost always provokes dissonancebecause, after a decision is made to choose one al-ternative, a person has to cope with the cognitiveelements concerning the attractive attributes of therejected alternatives.^

Since decision making entails the rejection ofalternative (s), the theory asserts that post-decisiondissonance is an inevitable consequence of decisionmaking. The magnitude of dissonance dependsupon the importance of the decision and the relativeattractiveness of the rejected alternative (s). There-fore, the more important the decision and/or the

'Jack W. Brehm and Arthur R. Cohen, Explorationsin Cognitive Dissonance (New York: John Wiley &Sons, Inc., 1962), pp. 306-308.

*S. E. Asch, "Review of L. Festinger, A Theory ofCognitive Dissonance," Contemporary Psychology,Vol. 3 (July, 1958), pp. 194-195; Milton J. Rosen-berg, "\Vhen Dissonance Fails: On Eliminating Eval-uation Apprehension from Attitude Measurement,"Journal of Personality and Social Psychology, Vol. 1(January, 1965), pp. 28-42; Alan C. Elms, "RolePlaying, Incentive, and Dissonance," PsychologicalBulletin, Vol. 68 (August, 1967), pp. 132-148; KarlE. Weick, "When Prophecy Pales: The Fate of Dis-sonance Theory," Psychological Reports, Vol. 16(June, 1965), pp. 1261-1275; Howard L. Fromkin,"Reinforcement and Effort Expenditure: Predictionsof 'Reinforcement Theory' Versus Predictions of Dis-sonance Theory," Journal of Personality and SocialPsychology. Vol. 9 (August, 1968), pp. 347-352.

6 Same reference as footnote 1.

more attractive the rejected alternative(s), thegreater the dissonance.

One derivation of the theory is that the greaterthe number of alternatives a consumer considersbefore his purchase decision and/or the more equalthe positive and negative attributes of the alterna-tives, the greater the post-purchase dissonance.

Equivocal Evidence for Post-purchase Dissonance

The experimental evidence frequently quoted tosupport the existence of post-purchase dissonancewas reported by Ehrlich et al. They found that thelarger number of alternative automobiles the con-sumer considered before his purchase, the greaterthe frequency of reading the automobile advertise-ments of the make he bought. This finding sup-ported dissonance theory. They also found, however,that both recent and not-recent purchasers noticedand read more advertisements of considered-but-re-jected makes of automobiles than those of not-con-sidered makes. This evidence cast doubt on thehypothesis that purchasers experienced dissonance.According to dissonance theory they should haveavoided the advertisements of the rejected makes.'

It may be that, soon after the purchase of a newautomobile, Ehrlich et al's consumers read automo-mobile advertisements not because they experienceddissonance but because automobile buying was aninfrequent undertaking and the topic of automobileswas relevant and useful to them. An experiment byBerkowitz and Cottingham supports the view thatpeople tend to be interested in the topics which arerelevant to themselves.^ They found that safety-belt users were more interested in communication onsafety-belts than were nonusers because the topicwas relevant to them, their interest having beenaroused previously.

Danuta Ehrlich, Isaiah Guttman, Peter Schonbach,and Judson Mills, "Post-Decision Exposure To Rele-vant Information," Journal of Abnormal and SocialPsychology, Vol. 54 (January, 1957), pp. 98-102;Table I.L. Berkowitz and D. Cottingham, "The Interest Valueand Relevance of Fear Arousing Communications,"Journal of Abnormal and Social Psychology, Vol. 60(January, 1960), pp. 37-43.

• ABOUT THE AUTHOR. SadaomiOshikawa is assistant professor oimarketing at the School of BusinessAdministration of the University ofWashington. A graduate of WasedaUniversity. Tokyo and the Universityof Colorado, he studied marketing andsocial psychology and received aDBA degree from the University ofWashington. Professor Oshikawa haspublished articles on cognitive disson-ance in the Journal of Maiketing Research and BehavioraiScience.

46 Journal of Marketing, October, 1969

Canon, Freedman, and Lowe and Steiner foundthat when information is dissonant but also useful,utility outweighs dissonance and the information willnot be avoided.** However, Berkowitz and Cotting-ham's relevance hypothesis and Canon's utility hy-pothesis explain the finding equally well and the evi-dence is not unequivocally supportive of dissonancetheory.

In an attempt to reconcile the Ehrlich et al findingwith dissonance theory. Mills hypothesized that auto-mobile purchasers liked considered-but-rejectedmakes better than not-considered makes. He rea-soned further that, if they preferred to read ads ofa chosen (and liked) product to those of a con-sidered-but-rejected (less liked) product, then theEhrlich et al finding that considered-but-rejectedmakes were noticed more frequently than not-con-sidered makes could be explained by their liking forthe former makes.*

In an experiment Mills proceeded to show thatconsumers preferred to read advertisements of thechosen product to those of the unchosen product.However, in this experiment the consumers had beenpromised that they would receive the chosen productas a free gift, but had not received it when theyexpressed their ad preferences. Consequently, theirpreference may have been infiuenced by their curiosityabout the free gift. His findings would have beenless ambiguous if the consumers had, in fact, re-ceived the gift. Even granting that Mills provedhis hypothesis, the Ehrlich et al experiment showedthat their consumer subjects did not experiencestrong enough dissonance to overcome the interest-in-the-liked product tendency.

Ehrlich and others ascribed their unexpectedfinding to the possibility that some recent pur-chasers sought the advertisements of unchosenmakes in order to find faults and reduce dissonance.This, however, is not a satisfactory explanation re-garding the behavior of the purchasers in their at-tempts to reduce dissonance. If the experimenter isallowed to do this, the findings will always supportpredictions and there is no room for the rejection ofthe theory. A better research approach would havebeen to clearly specify the predicted mode of dis-

*L. K. Canon, "Self-Confidence and Selective Expo-sure to Information," in L. Festinger (ed.). Conflict,Decision, and Dissonance (Stanford: Stanford Uni-versity Press, 1964), pp. 83-95; Jonathan L. Freed-man, "Confidence, Utility, and Selective Exposure:A Partial Replication," Journal of Personality andSocial Psychology, Vol. 2 (November, 1965), pp. 778-780; Rosemary H. Lowe and Ivan D. Steiner, "SomeEffects of the Reversibility and Consequences of De-cisions on Postdecision Information Preferences,"Journal of Personality and Social Psychology, Vol.8 (February, 1968), pp. 172-179.

9 Judson Mills, "Avoidance of Dissonant Information,"Journal of Personality and Social Psychology, Vol. 2(October, 1965), pp. 589-593.

sonance reduction and to block other possible modesbefore the execution of the experiment.

Problems in Experimental DesignOne of the criticisms raised by Chapanis and

Chapanis and Janis and Gilmore was that some ofthe experimental findings in support of the theorywere the results of built-in bias. They argued thatsome experiments were designed and manipulated insuch a way as to produce supporting results. Conse-quently, the findings could not be accepted as evi-dence for dissonance theory.̂ **

An experiment reported by LoSciuto and Perloffillustrates this problem. Dissonance theory postu-lates that if a person, given a choice between twoequally desirable products, chooses one and rejectsthe other, he will experience dissonance. Such dis-sonance will lead the person to evaluate the chosenproduct more favorably and the rejected productless favorably. In their experiment, LoSciuto andPerloff had their subjects rank nine phonographrecords according to desirability. To arouse strongdissonance, one group of subjects was given achoice between the third- and fourth-ranked albums;the other group was given a choice between the third-and eighth-ranked albums.*'

The experimenters found that the first group ofhigh-dissonance subjects tended to rerank the chosenalbums as more desirable and the rejected albumsas less desirable. The low-dissonance group did notshow this tendency as strongly. According to dis-sonance theory, a greater proportion of high-disson-ance subjects would show divergent changes in rank-ing (that is, reranking of the chosen albums as moredesirable and the rejected ones as less desirable),while a greater proportion of low-dissonance sub-jects would show convergent changes. A chi-squaretest supported this prediction at the .001 level ofsignificance.

Analysis of the design showed that the experi-ment was set up in such a way as to make it easierfor high-dissonance subjects to show divergentchanges and for low-dissonance subjects to showconvergent changes. Since high-dissonance sub-jects chose between the third-ranked and fourth-ranked albums, the third-ranked albums had twoplaces to move up, and the fourth-ranked albums hadfive places to move down, totaling seven places tomove divergently. Seven (44%) of the total 16

Natalia P. Chapanis and Alphonse Chapanis, "Cogni-tive Dissonance: Five Years Later," PsychologicalBulletin, Vol. 61 (January, 1964), pp. 1-22; I. L.Janis and J. B. Gilmore, "The Influence of IncentiveConditions on the Success of Role Playing in Modify-ing Attitudes," Journal of Personality and SocialPsychology, Vol. 1 (January, 1965), pp. 17-27.Leonard LoSciuto and Robert Perloff, "Influence ofProduct Preference on Dissonance Reduction," Jour-nal of Marketing Research, Vol. 4 (August, 1967),pp. 286-290.

Can Cognitive Dissonance Theory Explain Consumer Behavior? 47

movements would be considered divergent and the re-maining nine movements (56%) would be converg-ent. For the low-dissonance subjects, 13 of the pos-sible 16 movements (81%) contributed to convergentchanges and only three (19%) to divergent changes.

If all subjects reranked the albums randomly, agreater proportion of low-dissonance subjects wouldshow convergent changes and a greater proportionof high-dissonance subjects divergent changes. Al-though dissonance theory indicates that cognitivedissonance produces the above pattern of changes,an altemative explanation is that the observed pat-tern of changes is the result of the experimentaldesign.

A Replication of the ExperimentThis writer conducted an experiment to test this

possibility.!^ One hundred fifty-four undergraduatestudents were told that he was conducting a surveyamong college students on the popularity of ninerecord albums. The students were asked to rankeach of the nine albums. One week later they weregiven the impression that the first preferencequestionnaires had been misplaced and were askedto rank the same albums again. Because the sub-jects were not asked to choose between any twoalbums, dissonance was not provoked.

The pattern of changes of the third-, fourth-, andeighth-ranked albums from the first to the secondsurvey was studied. The chi-square test rejectedthe null hypothesis of independence between the pat-tern of changes and the initial location of albums atthe .0005 level of significance (x^ = 25.58). Thisfinding showed that the experimental design ensuredthe statistically significant outcome even when sub-jects did not experience cognitive dissonance. Sheth'sexperimental findings in support of dissonancetheory can be largely explained as the result of thesame built-in bias of using changes in rank positionsas the dependent variable.^^

Summary: Post-decision DissonanceThe Ehrlich et al study did show that new car

owners sought out dissonance-reducing information,and supported dissonance theory. The difficulty,however, is that alternative explanations (ortheories) predict the same results. Consequently,the findings were not unequivocal in support ofdissonance theory. Furthermore, dissonance theoryalso postulates avoidance of dissonance-increasing in-formation. Although some experiments supportedthis postulate, many others failed to do so. Thus,

12 Sadaomi Oshikawa, "The Theory of Cognitive Dis-sonance and Experimental Research," Journal of Mar-keting Research, Vol. 5 (November, 1968), pp. 429-430.

i3Jagdish N. Sheth, "Cognitive Dissonance, BrandPreference, and Product Familiarity," in Johan Arndt(ed.). Insights into Consumer Behavior (Boston:Allyn and Bacon, Inc., 1968), pp. 41-53.

the theory has not fared too well in the area of in-formation seeking and avoidance. In addition, prob-lems associated with the experimental design mayhave produced a "built-in" bias for many post-de-cision dissonance studies.

Forced Compliance

Another way to create dissonance is to have aperson verbalize or behave in a manner which iscontrary to his original attitude, belief, opinion orconviction. In most experiments, subjects wereforced to comply with the request of the experi-menter to create dissonance. Hence, this process iscalled "forced compliance."" The theory has somesupport in forced compliance experiments.

As applied to automobile-purchasing behavior,forced compliance resulted in dissonance when aconsumer clearly knew that a particular make wassuperior to other makes in relevant attributes butwas induced, on his own volition, to buy an inferiormake. The knowledge of superior attributes of re-jected makes is dissonant with the knowledge thathe bought an inferior make. The less the amountof inducement to buy the inferior make and thegreater the freedom he had in rejecting superiormakes, the greater his post-purchase dissonance.

The importance of volition cannot be overempha-sized. Without it, the person will not experiencestrong enough dissonance to motivate a dissonance-reducing behavior. Its importance was well illus-trated in the Festinger and Carlsmith experiment.^'

Forced Compliance in Consumer Behavior

When the forced compliance paradigm is appliedto consumer behavior, the consumer has to be in-duced to buy the make he knows is inferior if hisdissonance is to be aroused. If he believes that onealternative is not a good one but is forced to choosethat altemative, he will not experience dissonancebecause he can explain the poor choice as forcedupon him. If, on the other hand, he has completefreedom in making the decision and chooses thewrong alternative, then he will experience dissonanceas he cannot ascribe his poor choice to the force im-posed upon him. The problem, however, is that in arealistic market situation it is impossible to force acustomer to buy a product which he knows is inferior.

One methodological problem of the forced com-pliance experiment is that the subjects were inducedto comply and those who did, in fact, comply reduceddissonance in a variety of ways as predicted by theexperimenters. Even in the artificial experimentalsituations, however, experimenters have been plaguedby a loss of subjects who refused to comply. In the

Same reference as footnote 3, pp. 84-91.Leon Festinger and James M. Carlsmith, "CognitiveConsequences of Forced Compliance," Journal of Ab-normal and Social Psychology, Vol. 58 (March, 1959),pp. 203-210.

48 Journal of Marketing, October, 1969

Festinger-Carlsmith experiment, 11 out of 71 sub-jects had to be discarded because of their refusal tocomply. In another experiment, only 72 of theoriginal sample of 203 subjects could be used.i"

Experimenters must strike a balance betweenexercising too much force and not exercising enough.The implication of this methodological problem tomarketing is that, in a natural setting, consumersare not likely to experience post-purchase dissonancevia forced compliance because they will not behavein ways which they know will later arouse dissonance.

Exposure to Discrepant InformationAnother set of circumstances under which dis-

sonance may occur is when the consumer is exposedto new information not available to him at the timeof decision making and which is obverse to the in-formation he already has. This condition is calledcognitive intrusion because new dissonant cognitions"intrude" upon one's cognitive structure.'^

For example, suppose a consumer studied exten-sively and carefully the attributes of different makesof automobiles and purchased a particular makewhich he judged to be the best. Will he experiencedissonance when he is later exposed to new infor-mation describing unfavorable attributes of thechosen make and/or favorable attributes of unchosenmakes ?

Whether exposure to discrepant information willarouse dissonance depends upon a variety of factors.The most important of these is the degree of commit-ment and ego-involvement.

The findings of several experiments suggested thatwhen the discrepant information is not salient andthe degree of public ego-involvement is small, dis-sonance will not occur. For example, Rosen foundthat when students made decisions individuallywithout announcing publicly, more (67%) soughtdissonance-producing information regarding the de-cisions made and less (33%) sought dissonance-re-ducing information.!*

On the other hand, dissonance may be provokedbecause the consumer has publicly committed him-self to the position that the choice he made is a goodone. Public commitment results in ego-involvementwhich in turn increases the importance of thatcognitive element on which one has committed him-

» "The magnitude of dissonance is a function

18 Chapanis and Chapanis, same reference as foot-note 10.

17 Bruce C. Straits, "The Pursuit of the Dissonant Con-sumer," JOURNAL OF MARKETING, Vol. 28 (July, 1964),pp. 62-66.

18 Sidney Rosen, "Postdecision Affinity for IncompatibleInformation," Journal of Abnormal and Social Psy-chology, Vol. 63 (July, 1961), pp. 188-190.

19 A. R. Cohen, J. W. Brehm, and B. Latane, "Choiceof Strategy and Voluntary Exposure to Informationunder Public and Private Conditions," Journal ofPersonality, Vol. 27 (March, 1959), pp. 63-73; samereference as footnote 3.

of the ratio of dissonant to consonant cognitions,where each cognitive element is weighted for itsimportance to the person."-" Consequently, publiccommitment tends to increase the magnitude ofdissonance by increasing the relative weight of thedissonant cognitions.

Relating the above discussion to the Ehrlich et alstudy, it should be noted that most automobile pur-chasers are not put in a position to publicly defendthe adequacy of their purchase decision, and dis-crepant information which they read in the news-paper will not arouse strong enough dissonance tomake them resort to a dissonance-reducing behavior.

Alternative Theory SupportedTo test if dissonance theory can be applied in a

more natural situation where individuals do notcommit themselves publicly, this writer conducted anexperiment and examined whether, after being ex-posed to dissonant information, strong dissonanceleads individuals to convince themselves that theoriginal decision was correct.-i Students were giventhe choice of essay type, objective type or any combi-nation of both tests for midterm and final examina-tions. After indicating their preference, some wereexposed to consonant information which supportedtheir original choice while others were exposed todissonant information.

They were also told either that they were com-mitted to their original preference or that they couldchange their preference after reading the informa-tion. Dissonance theory predicts that those whowere committed to the original preference and wereexposed to discrepant information would try to re-duce dissonance by becoming more convinced of thewisdom of their original decision. Kurt Lewin'sfield theory asserts, on the other hand, that discrep-ant information reduces the desirability of thechosen test and increases that of the rejected testand predicts the opposite outcome from dissonancetheory.22

Experimental evidence supported Lewin's theory,showing that the students were positively infiuencedby both the discrepant and the consonant informationregardless of their commitment. It appears that,under a natural circumstance, individuals do not re-spond to discrepant information in the way disson-ance theory predicts.

Evaluation of the Theory andApplicable Circumstances

An attempt has been made to examine the experi-mental findings on the theory of cognitive dissonance.

2" Same reference as footnote 6.-' Sadaomi Oshikawa, "Consumer Pre-decision Conflict

and Post-decision Dissonance," Behavioral Science,Vol. 15 (March, 1970).

-2 Kurt Lewin, Field Theory in Social Science (NewYork: Harper & Brother, 1951), p. 274; same refer-ence as footnote 3, p. 234.

Can Cognitive Dissonance Theory Explain Consumer Behavior? 49

Many findings concerning exposure to discrepantinformation and post-decision dissonance arousalhave been shown to be equivocal. In the forcedcompliance experiments, the artificial conditionsunder which compliance was obtained and some sub-jects' refusal to comply have reduced the usefulnessof the experimental findings to the analysis of con-sumer behavior. It was shown, however, that if thesubjects did comply, they attempted to reduce dis-sonance in a predicted manner.

Analysis of the theory and experimental findingssuggested that the necessary condition for provok-ing dissonance strong enough to motivate disso-nance-reducing behavior is that one be committedon his own volition to an undesirable product,position, or behavior and be unable to retract thiscommitment. However, consumers are unlikely toexperience strong dissonance since they will notknowingly commit themselves to undesirable or in-ferior products in a natural market setting.

Role of Advertising as Dissonance ReducerDissonance theory sheds new light on the role of

advertising of increasing the repurchase probabilityof the advertised product. A seller's product ad-vertisement reassures the consumer as to the wisdomof the purchase by emphasizing its desirable featuresand therefore helps to reduce post-purchase disso-nance. Dissonance reduction, in turn, reinforceshis purchase. It may increase the probability ofhis purchasing the same brand.

Dissonance reduction may not operate as a strongreinforcer in the case of frequently-purchased mer-chandise. The more frequently the product is pur-chased, the less important becomes the question ofwhich brand is purchased at any one time, and theless the post-purchase dissonance. The consumerwho has purchased a convenience good usually wouldnot experience strong dissonance because he knowsthat he is not irrevocably tied to that particularchoice, but can easily switch brands. Since hisdissonance is not strong, advertising's role in rein-forcing the purchase is diminished.

On the other hand, the consumer who has justpurchased an expensive specialty good is likely toexperience strong dissonance if his purchase is irrev-ocable and if it is important in some psychologicalsense. For example, if a substantial financial out-lay is involved or if his taste and intelligence arejudged by the purchase, strong dissonance may bepresent. Under these circumstances, an advertise-ment which emphasized the desirable features of thechosen brand can reduce the dissonance which maylead the consumer to form a more favorable attitudetoward the brand.

Wearing-Out of Reinforcing Effect

However, the longer the time lapse before productreplacement, the less reinforcing will be the effects ofthe advertising. In the meantime, the seller's advertis-ing must compete with that of his competitors, andit may not operate effectively as a reinforcing agentlong after the purchase. By the time he is ready toreplace the product, the effect of the firm's originaladvertising may have worn off and the attitude andpreference of the purchaser may have been infiuencedby the more recent advertising efforts of the firm'scompetitors. Since dissonance is reduced overtime,2-'' it is reasonable to expect that the greaterthe post-purchase dissonance, the longer the periodduring which the seller's advertisement operates asa reinforcer.

Thus, for the consumer who purchased an expen-sive product, advertising can act as a reinforcementfor some period of time following the purchase. Thisreinforcing effect, however, does not necessarily in-sure a repeat purchase because of the counteractingeffects of competitive advertisements.

In summary, the theory of cognitive dissonance isdesigned to explain and predict post-decisional be-havior, but in most instances it is not adequate toexplain consumer behavior before a purchase deci-sion.

23 Same reference as footnote 6, p. 99.