A Cognitive Dissonance Theory Perspective on Persuasion

Eddie Harmon-Jones. The Persuasion Handbook: Developments in Theory and Practice. Editor: James Price Dillard & Michael Pfau. Sage Publications. 2002.

Research and conceptual development on the theory of cognitive dissonance continues in full force more than 40 years after the theory’s conception. Part of the reason for the theory’s longevity may be that it was stated abstractly; could be applied to a wide array of issues; dealt with the interaction of cognition, motivation, and affect; and generated research that suggested ways of inducing lasting attitude, belief, and behavior change. For example, dissonance research has demonstrated that dissonance processes can reduce prejudice (Leippe & Eisenstadt, 1994), increase water conservation (Dickerson, Thibodeau, Aronson, & Miller, 1992), increase the purchasing of condoms (Stone, Aronson, Crain, Winslow, & Fried, 1994), reduce hunger and thirst (Brehm, 1962), and reduce pain (Zimbardo, Cohen, Weisenberg, Dworkin, & Firestone, 1969). In addition, dissonance processes can lead to changes in attitudes toward a variety of objects and issues, such as boring tasks (Festinger & Carlsmith, 1959), boring reading passages (Harmon-Jones, Brehm, Greenberg, Simon, & Nelson, 1996), delicious chocolate (Harmon-Jones, 2000a), eating grasshopPers (Zimbardo, Weisenberg, Firestone, & Levy, 1965), sour beverages made with vinegar (Harmon-Jones et al., 1996), increasing tuition at one’s university (Elliot & Devine, 1994), and mandatory comprehensive final exams (Simon, Greenberg, & Brehm, 1995).

Currently, researchers are approaching issues related to cognitive dissonance processes with a level of intensity that typically occurs at the birth of a theory. Several generative revisions to the original theory have been proposed, and there is much controversy regarding the processes by which the cognitive and behavioral changes occur. In this chapter, I provide a brief review of original theory of cognitive dissonance and revisions of the original theory. Then, I critically evaluate the revisions of the theory by discussing recent research that has challenged each of these revisions. Finally, I present a recently proposed revision and preliminary tests of this revision.

Original Version of Cognitive Dissonance Theory

According to the original version of the theory (Festinger, 1957), the presence of a cognitive inconsistency of sufficient magnitude will evoke an aversive motivational state—dissonance—that drives cognitive work aimed at reducing the cognitive inconsistency. The magnitude of dissonance aroused in regard to a particular cognition is a function of the number and psychological importance of cognitions dissonant (inconsistent) and consonant (consistent) with this cognition. Thus, dissonance will increase as the number and importance of dissonant cognitions relative to consonant cognitions increase. From the ratio used to compute the magnitude of dissonance (dissonant cognitions/dissonant + consonant cognitions, with each weighted by its importance), the routes of dissonance reduction are as follows: subtracting dissonant cognitions, adding consonant cognitions, decreasing the importance of dissonant cognitions, and increasing the importance of consonant cognitions. These modes of dissonance reduction may manifest themselves in attitude, belief, value, or behavior maintenance or change. Dissonance reduction will be aimed at altering the cognition least resistant to change.

The term dissonance has been used to refer to both the cognitive inconsistency and the aversive motivational state the inconsistency produces. However, it is important to distinguish the affective-motivational state of dissonance from the cognitive inconsistency and the cognitive and behavioral changes that result from the affective-motivational state of dissonance. I refer to the affective-motivational state as dissonance, to the cognitive inconsistency as cognitive discrepancy, and to cognitive and behavioral changes as cognitive discrepancy reduction.

Major Paradigms Used in Dissonance Research

Four research paradigms have been used extensively in the investigation of dissonance processes. To facilitate the presentation of experiments within this chapter, I briefly review these paradigms.

Free Choice Paradigm. In the free choice paradigm, developed by Brehm (1956), it is assumed that once a decision is made, dissonance may be aroused. After the Person makes a decision, each of the negative aspects of the chosen alternative and the positive aspects of the rejected alternative is dissonant with the decision. By contrast, each of the positive aspects of the chosen alternative and the negative aspects of the rejected alternative is consonant with the decision. Difficult decisions arouse more dissonance than do easy decisions because there is a greater proportion of dissonant cognitions after a difficult decision than after an easy one. Postdecision dissonance can be reduced by subtracting negative aspects of the chosen alternative or positive aspects of the rejected alternative, or it can be reduced by adding positive aspects to the chosen alternative or negative aspects to the rejected alternative. Thus, dissonance may be reduced by viewing the chosen alternative as more desirable and the rejected alternative as less desirable. This effect has been termed spreading of the alternatives.

Induced Compliance Paradigm. In the induced compliance paradigm, which was used by Festinger and Carlsmith (1959), it is assumed that dissonance is aroused when a Person does or says something that is contrary to a prior belief or attitude. From the cognition of the prior belief or attitude, it follows that one would not engage in such behavior. On the other hand, inducements to engage in such behavior—promises of reward or threats of punishment—provide cognitions that are consonant with the behavior. That is, these cognitions provide external justifications for the behavior. The greater the number and importance of these cognitions (justifications), the less the dissonance aroused. The amount of justification for engaging in the action has been manipulated using different amounts of money or different amounts of social influence that give Persons the Perception of having little choice or much choice over engaging in the action. When Persons engage in the counterattitudinal action and believe they have little justification for doing so because they were paid a relatively small amount or because they were left with the Perception that they chose to engage in the action, they will experience dissonance and reduce it by changing the belief or attitude to correspond more closely to what was said or done (see reviews by Beauvois & Joule, 1996; Brehm & Cohen, 1962; Harmon-Jones & Mills, 1999a; Wicklund & Brehm, 1976).

Belief Disconfirmation Paradigm. A third paradigm used in the investigation of dissonance theory is the belief disconfirmation paradigm, first used by Festinger, Riecken, and Schachter (1956). In this paradigm, it is assumed that dissonance is aroused when Persons are exposed to information inconsistent with their beliefs. If the dissonance is not reduced by changing one’s belief, the dissonance can lead to misperception or misinterpretation of the information, rejection or refutation of the information, seeking support from those who agree with the belief, and attempts to Persuade others to accept the belief.

Hypocrisy Paradigm. Aronson, Fried, Stone, and colleagues developed another paradigm to test the theory—the hypocrisy paradigm (Aronson, Fried, & Stone, 1991). In this paradigm, Persons are induced to make a public attitudinally consistent statement, and then they are reminded of times when they did not act in accord with their statement. In other words, participants are induced to say one thing and then are reminded of times when they did not practice what they “preached.” Research using this paradigm has demonstrated that individuals will reduce the dissonance by being more likely to act in accord with their proattitudinal statement (Stone et al., 1994) or by changing their attitudes to be more consistent with their past behavior, depending on whether the recent speech or past behavior was more resistant to change (Fried, 1998).

Revisions of the Original Theory

What is the psychological mechanism that is responsible for these Persuasion effects? Festinger (1957) proposed that cognitive inconsistency evoked motivation to reduce inconsistency. In an effort to increase the predictive power of the theory, other theorists have suggested that different motivations underlie the Persuasion effects observed in dissonance settings. Currently, this is the source of serious inquiry by dissonance theorists (see Harmon-Jones & Mills, 1999a). In what follows, I review these alternative motivational explanations.


One of the first motivational alternative explanations for dissonance theory was proposed as a refinement of Festinger’s original statement. In this self-consistency theory, Aronson (1968) proposed that dissonance theory made its most precise predictions when “a firm expectancy was involved as one of the cognitions in question” (p. 23). He further proposed that the theory is

clearer still when that firm expectancy involves the individual’s self-concept, for—almost by definition—our expectancies about our own behavior are firmer than our expectancies about the behavior of another Person. Thus, at the very heart of dissonance theory, where it makes its clearest and neatest prediction, we are not dealing with just any two cognitions; rather, we are usually dealing with the self-concept and cognitions about some behavior. If dissonance exists, it is because the individual’s behavior is inconsistent with his self-concept. (p. 23)

Aronson (1968, 1969, 1992, 1999) thus proposed that a violation of a self-concept was necessary to create dissonance. According to this theory, the self-concepts of morality, competence, and consistency are the self-concepts that serve as the standards against which behavior is compared. These self-concepts are standards for behavior that are often in “accord with the conventional morals and prevailing values of society” (Thibodeau & Aronson, 1992, p. 592). However, Aronson noted that Persons with negative self-concepts would expect to behave immorally and incompetently and hence would experience dissonance when they behave in contrary ways, that is, morally or competently.

Evaluation of Self-Consistency Explanation

The self-consistency revision retains inconsistency as the motivating force behind dissonance and its effects. Although the self-consistency revision is quite similar to the original theory, it limits dissonance theory to a self theory and thus restricts dissonance processes to organisms with self-concepts, thereby excluding most nonhuman animals and humans under 3 years of age. However, experiments demonstrating that dissonance effects occur in white rats, which presumably lack self-concepts (Lawrence & Festinger, 1962), contradict this revision. In addition, the self-consistency revision restricts dissonance processes to organisms whose self-concepts become accessible as a standard for comparison whenever cognitive discrepancies are encountered. This restriction would probably omit several situations because it does not seem plausible that a self-concept would become accessible every time a cognitive discrepancy is encountered (for further criticisms of this theory, see Brehm, 1992, and the discussion below).


A second motivational explanation of dissonance effects is self-affirmation theory (Steele, 1988; Steele & Liu, 1983; Steele, Spencer, & Lynch, 1993). This revision proposes that the effects observed in dissonance situations are not the result of cognitive inconsistency or self inconsistency. In fact, Steele et al. (1993) suggested that “not only is a motive for self-consistency absent from dissonance processes, but it is absent from mental life altogether” (p. 894). The self-affirmation revision posits that situations that create dissonance exert their effects because of the threat to the individual’s need to Perceive oneself as having global integrity, that is, being morally and adaptive adequate in general.

Research on self-affirmation theory has suggested that providing Persons with an opportunity to affirm their self-integrity reduces the attitude change that typically occurs in the induced compliance paradigm (Steele & Liu, 1983). Steele has argued that these effects result because acting counterattitudinally threatens self-integrity, and the attitude change that is typically observed occurs to reduce this threat to self-integrity. Having Persons affirm their self-worth will thus reduce the need to change attitudes.

In contrast to E. Aronson’s self-consistency model, the self-affirmation model asserts that the motivating force in producing dissonance effects is not inconsistency but rather the need for self-integrity. Thus, for self-affirmation theory, the need for a positive self-concept is a condition necessary for dissonance.

Evaluation of Self-Affirmation Explanation

Relevance of the Affirmation. Recent research suggests that the self-affirmation manipulations prevent attitude change only when the affirmation is irrelevant to the counterattitudinal action. When the affirmation is relevant to the counterattitudinal action, attitude change may actually increase because it enhances the accessibility of a Personal standard. Blanton, CooPer, Skurnik, and Aronson (1997) conducted an experiment in which Persons reflected on a positive aspect of the self that was either relevant or irrelevant to the counterattitudinal action. They found that when Persons had freely chosen to write essays arguing against funding for handicapped services at their university and then had received Personality feedback indicating that they were high in the trait of compassion (relevant affirmation), they changed their attitudes even more in the direction of their behavior. By contrast, when Persons had freely chosen to write essays on the same issue and then received Personality feedback indicating that they were high in the trait of creativity (irrelevant affirmation), they did not change their attitudes. Blanton et al. interpreted these effects to indicate that dissonance reduction is not the result of a motive for self-affirmation.

These effects could be reinterpreted in original dissonance theory terms. Accordingly, the counterattitudinal behavior is the cognition most resistant to change, the conception of oneself as compassionate is dissonant with the counterattitudinal behavior, and thus the magnitude of dissonance is increased relative to a situation in which the conception of oneself as compassionate is not made more accessible. Thus, this increased dissonance caused more attitude change. By contrast, the conception of oneself as creative is a cognition that is not relevant to the cognitions involved in the dissonant relationship. Hence, it should have no effect on the magnitude of dissonance. However, affirming aspects of the self that are not relevant to the discrepancy may alter the amount of dissonance aroused because this cognition distracts one from the dissonance (McGregor, Newby-Clark, & Zanna, 1999), reduces the importance of the relevant cognitions (Simon et al., 1995), or decreases the negative affect associated with the dissonance (Harmon-Jones, 2000b).

Importance of Cognitions. Research by Simon et al. (1995) has shown that a typical self-affirmation manipulation (i.e., participants complete a scale on which they report their important values) causes Persons to Perceive their counterattitudinal behavior and preexisting attitudes as less important, which reduces the likelihood of observing dissonance-related attitude change. In one experiment, Simon et al. found that making an important but non-self-affirming issue salient (i.e., world hunger) caused reductions in the Perceived importance of the counterattitudinal behavior and preexisting attitude rather than attitude change following freely choosing to write a counterattitudinal statement. Taken together, these results show that self-affirmation may exert its effects on the dissonance process by reducing Persons’ Perception of the importance of the cognitions associated with the preexisting attitude and counterattitudinal behavior. Thus, the original theory of dissonance can explain the effects generated by self-affirmation theory.

Pitting Consistency Against Self-Affirmation. Stone, Wiegand, Cooper, and Aronson (1997) provided further evidence suggesting that the primary motive underlying discrepancy reduction is not the restoration of the global moral and adaptive integrity of the entire self-system, as self-affirmation theory predicts; rather, the primary motive is the resolution of the specific discrepancy, as the original theory and self-consistency theory predict. Using a hypocrisy paradigm in which participants delivered a public speech advocating using condoms to prevent AIDS and were then reminded of their failures to use condoms, Stone et al. (1997) found that when given a choice between purchasing condoms (which would reduce the specific discrepancy) and donating to a homeless project (which would restore global self-worth), more participants chose to purchase condoms. In a second experiment using a hypocrisy paradigm in which participants delivered a public speech advocating the importance of volunteering to help the homeless and were then reminded of their failures to volunteer, Stone et al. found that more participants chose to donate to the homeless than to purchase condoms, even when participants had rated using condoms to prevent AIDS as more important to their global self-worth than donating to feed the homeless. Results from the experiments also indicated that Persons were more likely to choose an option that would restore global self-worth in conditions where a sufficient discrepancy was aroused (hypocrisy) than in comparison conditions. These results suggest that Persons will choose to restore global self-worth following a discrepant action if it is the only available option. However, if given a choice, Persons opt for direct discrepancy reduction, suggesting that avoiding discrepancy rather than restoring global self-worth is a more prominent concern of Persons in dissonance situations. Thus, the results from these experiments cast doubt on the self-affirmation explanation of dissonance effects.

In sum, recent research from several different Perspectives has challenged the self-affirmation interpretation of dissonance effects. However, the research on self-affirmation has identified important ways of reducing dissonance, has contributed to a more thorough understanding of dissonance processes, and has been useful in other arenas examining self-protective motivations (Aronson, Cohen, & Nail, 1999).

Aversive Consequences

In another motivational explanation for dissonance theory, Cooper and Fazio (1984) posited that cognitive inconsistency was neither necessary nor sufficient for dissonance to occur. They proposed that for dissonance to occur, individuals must engage in behavior that has the Perceived potential to cause an irrevocable unwanted consequence. That is, CooPer and Fazio asserted that a sense of Personal responsibility for producing foreseeable aversive consequences is necessary for dissonance to be aroused.

CooPer and Fazio (1984) reviewed evidence suggesting that when Persons engaged in a counterattitudinal action but did not produce aversive consequences, they did not change their attitudes to be consistent with their behavior. For instance, Cooper and Worchel (1970) suggested that participants in the low-justification condition of Festinger and Carlsmith (1959) changed their attitudes not because of cognitive inconsistency but rather because they felt Personally responsible for producing the aversive consequence of convincing another Person to believe that boring tasks were interesting. To test this explanation, Cooper and Worchel replicated and extended the design of the Festinger and Carlsmith (1959) experiment. They found that when low-justification participants were led to believe that they did not convince another Person that a boring task was interesting, they subsequently did not rate the task as more interesting. Other experiments have replicated these results by finding that when participants believe that their counterattitudinal statements do not Persuade others, they do not change their attitudes (e.g., Cooper, Zanna, & Goethals, 1974; Goethals & CooPer, 1972; Hoyt, Henley, & Collins, 1972; Nel, Helmreich, & Aronson, 1969).

Evaluation of Aversive Consequences Explanation

It is important to note that the evidentiary basis for the aversive consequences model relies solely on the production of no attitude change in conditions where aversive consequences are not produced. Because this is a null effect, several alternative explanations can be offered. For example, in these past experiments, participants were encouraged to produce lengthy counterattitudinal statements. These statements may have reduced the likelihood of detecting dissonance-related attitude change, as research has demonstrated that length of statement is inversely related to amount of dissonance-related attitude change (Beauvois & Joule, 1996, 1999; Rabbie, Brehm, & Cohen, 1959). This inverse relation may result because longer statements allow for more consonant cognitions that support the counterattitudinal behavior and hence reduce the dissonance. Thus, the overall level of dissonance in these experiments may have been rather low, and the addition of the production of aversive consequences was necessary to produce sufficient dissonance to cause attitude change. Another explanation for the past failures to find attitude change in nonaversive consequences experiments is that dissonance was aroused but reduced in a manner other than attitude change in the no aversive consequences conditions. Other alternative explanations have been presented (Harmon-Jones, 1999).

My colleagues and I were not convinced by the past evidence regarding the necessity of aversive consequences. To test whether cognitive discrepancy in the absence of aversive consequences can produce dissonance, we conducted induced compliance experiments in which participants behaved counterattitudinally but did not produce aversive consequences (Harmon-Jones, 1999, 2000a; Harmon-Jones et al., 1996). We designed these experiments so that factors that may have inhibited attitude change in the past experiments were omitted in our experiments. In these experiments, which were presented as studies on how writing different types of statements affects memory, participants were exposed to a simple stimulus about which they held a positive or negative attitude (e.g., boring passage, delicious chocolate) and then were given low or high choice to write sentences in privacy and with anonymity that were inconsistent with their attitudes. Participants were told to discard their counterattitudinal statements once they finished writing them. Then, they expressed their attitudes toward the stimulus in privacy and with anonymity. Results have indicated that participants given high choice to write the counterattitudinal statements shifted their attitudes to align them with their behavior, whereas participants given low choice did not. Moreover, results have indicated that participants given high choice, as compared to low choice, to write the statements evidence increased electrodermal activity (Harmon-Jones et al., 1996, Experiment 3) and report more discomfort (Harmon-Jones, 2000a, Experiments 1 and 2) during the Period of time between the writing of the statement and the assessment of the attitude, suggesting that this manipulation produced dissonance. In addition, following the attitude change opportunity, Persons report less discomfort, suggesting that the attitude change reduced the dissonance (Harmon-Jones, 2000a, Experiment 2).

Evidence from belief disconfirmation studies also suggests that the production of aversive consequences is not necessary to create dissonance. In these studies (e.g., Batson, 1975; Brock & Balloun, 1967; Burris, Harmon-Jones, & Tarpley, 1997; Russell & Jones, 1980), Persons are exposed to information inconsistent with a highly important and highly resistant to change belief, and they show dissonance effects (e.g., negative affect, belief intensification, selective avoidance of inconsistent information, transcendence) as a result of this exposure. In the belief disconfirmation paradigm, Persons have not produced aversive consequences and thus cannot feel responsible for having done so. Persons are simply exposed to information from an external source; they have not done anything for which to feel responsible.

Aronson (1992, 1999; see also Thibodeau & Aronson, 1992) and Stone and colleagues (e.g., Stone et al., 1994) also questioned whether the production of aversive consequences is necessary to create dissonance effects. In their program of research on hypocrisy, they have found that when Persons give a public speech advocating a prosocial policy that will produce positive consequences and then are reminded of times when they have not behaved in accord with the position advocated in their speech, they experience dissonance and engage in cognitive and behavioral changes to reduce the dissonance.

McGregor et al. (1999) further pointed out that recent attitudinal ambivalence research has provided evidence of dissonance-related negative affect in the absence of feeling Personally responsible for producing negative consequences.

Assessment of the Current Status of the Theory

Recent research has strongly supported the idea that dissonance is a motivational theory and that dissonance produces genuine and lasting attitude, belief, and behavior changes. Festinger proposed that cognitive discrepancy was inherently aversive. Other theorists have proposed alternate motivations that underlie the motivating effects putatively due to cognitive discrepancy (Aronson, 1968; Cooper & Fazio, 1984; Steele, 1988). Much debate centers around determining the underlying motivation (Harmon-Jones & Mills, 1999a). Most of the revisions concur with the original theory and postulate that the situations that create a discrepancy evoke negative affect and that this negative affect motivates the cognitive and behavioral adjustments found. However, these revisions differ in their explanations of why these situations evoke negative affect and why individuals engage in the cognitive and behavioral adjustments. That is, each of the revisions implicitly views negative affect as a proximal motivation for the cognitive changes observed in dissonance settings, and each revision proposes a different distal motivation underlying the arousal of the negative affect. Although each of these revisions has yielded important data, recent research has been presented that challenges each of these revisions and supports the original version of the theory (e.g., Beauvois & Joule, 1996, 1999; Harmon-Jones, 1999; Harmon-Jones et al., 1996; McGregor et al., 1999; Simon et al., 1995).

An Action-Based Model of Cognitive Dissonance

Thus, the evidence supports the idea that a cognitive inconsistency of sufficient magnitude evokes an aversive motivational state that causes cognitive and behavioral changes, in accord with the original version of the theory. However, the understanding of dissonance processes could be improved and extended with an explanation of why cognitive inconsistency arouses an aversive motivational state and why this state causes the cognitive and behavioral adjustments.

The action-based model of cognitive dissonance has recently been proposed to address these questions. The model begins with the assumption that cognitions, broadly defined, can serve as action tendencies, an idea endorsed by several scholars (Cacioppo & Berntson, 1994; James, 1890/1950). For the model, the cognitions that are of primary concern are those that provide useful information, and usefulness of information is defined by its relevance to actions and goals. When information inconsistent with cognitions that guide action is encountered, an aversive motivational state—dissonance—is aroused because the dissonant information has the potential to interfere with effective and unconflicted action.

Thus, cognitive discrepancy may create dissonance because discrepancy among cognitions undermines the potential for effective and unconflicted action (Beckmann & Irle, 1985; Harmon-Jones, 1999; Jones & Gerard, 1967). Previous dissonance scientists viewed commitment as occurring when an individual freely chose to engage in a behavior of which they knew or could know the consequences (Beauvois & Joule, 1996; Brehm & Cohen, 1962; Festinger, 1964). However, it is also possible that commitment or attachment to a cognition or Perception can occur without an obvious behavioral commitment and that individuals would experience dissonance if information contradicted this commitment. Consider the law of gravity. Most Persons would probably experience considerable dissonance if, while they were walking through a calm forest, a tree suddenly flew into the atmosphere. Thus, regarding a cognition or Perception as true or as reality established a commitment in the Perceivers’ minds. This commitment to a reality then serves the function of guiding information processing, which then serves the ultimate function of activating and directing behavior.

When dissonant information is encountered, dissonance may result in motivating the individual to engage in information processing that would support the commitment. With further increases in dissonant information, the dissonance may increase to the point where the individual is motivated to accept the information and reject the commitment. Which cognition is supported or accepted depends on the resistance to change of each cognition.

As an example, consider a situation where a hiker is lost deep in the woods. Based on her calculations, she is in the middle of the woods, has just enough time and resources to cover 15 miles, and knows that an exit (and thus safety) is 15 miles to the east or west. She Perceives that she has two options: She can begin walking either to the west or to the east. Each option has advantages (she is very familiar with the east route, while she has heard that the west route requires less effort) and disadvantages (the east route will require crossing a river, while the west route will require climbing a mountain). Once the hiker makes a decision to walk east, the disadvantages of walking east and the advantages of walking west become discrepant cognitions. If she continues to weigh the relative merits of the options, vacillating between them, beginning to walk one way and then turning around and walking the other, action will be impeded and she might never arrive to safety. She must reduce the discrepancy and follow through with the option she has chosen. Even if the route she chooses is not the best, she will be better off Persisting in one course of action than going halfway only to turn back.

From the current perspective, the proximal motivation to reduce cognitive discrepancy stems from the need to reduce the aversive state of dissonance, while the distal motivation to reduce cognitive discrepancy stems from the requirement for effective action. When the potential for effective action is threatened by information that is sufficiently discrepant from the psychological commitment, dissonance results, which prompts attempts at the restoration of cognitions supportive of the commitment (i.e., discrepancy reduction). It is important to note that while the proximal and distal motivations are often linearly related—that is, an increase in the need to produce effective behavior will be related to an increase in negative affect—the two motivations may operate independently in some circumstances.

The current model is consistent with views that have been presented previously but have not been given due consideration (e.g., Beckmann & Irle, 1985; Gollwitzer, 1990; Heckhausen, 1986; Jones & Gerard, 1967; Kuhl, 1984; Lewin, 1951). One reason for the lack of consideration of the these conceptual ideas may be due to the difficulty of understanding how the effects observed in the laboratory experiments would be produced by a concern over effective action. For example, why would women in Brehm’s (1956) experiment rate a toaster more positively after choosing it? At least two arguments can be advanced to address this question. First, it would be beneficial to value the objects that one owns to maintain their quality and get use out of them. Second, the current model can be applied to results obtained in dissonance experiments by assuming that a mechanism that survived because of its adaptive value is able to produce effects in conceptually similar situations that do not have obvious adaptive significance.

Experimental Results Consistent with the Action-Based Model

We have recently completed experiments that test predictions derived from the action-based model. In two experiments, we have tested the hypothesis that processes that facilitate an action-orientation will increase cognitive discrepancy reduction (Harmon-Jones & Harmon-Jones, in press). The action-oriented mindset is posited to prime one toward being more able to enact decisions and hence bring cognitions in line with the decision. In other words, the action-oriented mindset will facilitate the ease with which one can reduce cognitive discrepancy without actually increasing the amount of negative affect. Such a finding would support the theoretically derived hypothesis that the distal motivation (effective behavior) can be activated independently of the proximal motivation (negative affect).

Action-Oriented Mindset Experiment

In the first experiment, university students evaluated how desirable they found seven physical exercises that were described in writing to them. They then decided to participate in one of two of the exercises. Either the two used in the decision were rated equally (difficult decision) or one was rated highly and the other lowly (easy decision). After the decision, the participants completed one of two questionnaires; they either provided demographic information (control condition) or described their implemental intentions for the upcoming exercise (action-oriented mindset). Then, the experimenter asked the participants to re-rate the desirability of the seven exercises. The change in rating of the chosen exercise minus the change in rating of the rejected exercise was used as the dependent variable. As predicted, a significant 2 (Decision Difficulty) × 2 (Mindset) interaction emerged. Follow-up tests revealed that spreading of alternatives was greatest in the action-oriented/difficult decision condition.

Action-Oriented Mindset and Prefrontal Cortical Activation Experiment

We conducted a second experiment to further test the hypothesis that an action-oriented mindset would facilitate the justification of a decision. Recent research (Gollwitzer & Kinney, 1989; Taylor & Gollwitzer, 1995) has found that after making a decision and thinking about steps needed to implement it (action-oriented mindset), Persons are more likely to have positive illusions and more of an illusion of control. In other words, thinking about implementing one decision has effects on actions and cognitions that are unrelated to the decision. That is, the effects of the action-oriented mindset can transfer to unrelated actions and cognitions. Therefore, in the second exPeriment, we tested whether having Persons think about implementing one decision would affect the spreading of alternatives for a different decision.

We also tested the effects of implemental intentions on prefrontal cortical activity. Specifically, we predicted that the activation of an approach-related, action-oriented mindset will increase left prefrontal cortical activity.

The left frontal cortical region has been described as an important center for intention, self-regulation, and planning (Kosslyn & Koenig, 1995; Petrides & Milner, 1982; Tomarken & Keener, 1998). These functions have often been described as properties of the will, a hypothetical construct important in guiding approach-related behavior. Damage to the left frontal region results in behavior and experience that may be described as involving a deficit in approach. Persons with damage to this region are apathetic, experience less interest and pleasure, and have difficulty in initiating actions (e.g., Robinson & Downhill, 1995).

Research using measures of electroenceph-alographic (EEG) activity has found that decreased left frontal activity relates to depression (Henriques & Davidson, 1990, 1991) and that increased left frontal cortical activation relates to trait and state measures of approach motivation (Harmon-Jones & Allen, 1997, 1998; Sobotka, Davidson, & Senulis, 1992; Sutton & Davidson, 1997) and trait repression (Tomarken & Davidson, 1994). It is interesting to note that repression has been linked to an increased likelihood of reducing dissonance via attitude change (Olson & Zanna, 1979; Zanna & Aziza, 1976). Thus, as approach-related, action-oriented thinking increases, greater left frontal activation may occur.

In the second experiment, participants evaluated how desirable they found eight experiments that were described to them in writing. Then, they decided to participate in one of two of them that would supposedly occur within a few minutes after the decision. In this experiment, the decision was always difficult, that is, between two equally and highly valued alternatives. After the decision, they completed “Personality questionnaires.” The Personality questionnaires contained the manipulation of mindset. Participants were randomly assigned to complete either (a) a questionnaire that asked them to write about a project they had decided to do and to describe the steps they would need to implement to accomplish the project (action-oriented mindset), (b) a questionnaire that asked them to write about a decision in their lives that they were currently deliberating and had not yet decided on how to act (deliberative mindset), or (c)a questionnaire that asked them to write about a typical day in their lives (control mindset). After completing the questionnaire, they were asked to think about the information they provided in the questionnaire for 2 minutes while their brain activity was recorded. Then, they re-rated the decision alternatives. As predicted, participants in the action-oriented mindset condition evidenced more spreading of alternatives than did participants in the other two conditions. Moreover, participants in the action-oriented mindset condition evidenced greater left prefrontal cortical activity after the mindset manipulation than did participants in the other two conditions. However, this latter effect resulted only for women and not for men. A similar Sex of Participant × Mindset condition interaction did not emerge for spreading of alternatives. In addition, relative left frontal activity mediated the effect of mindset on spreading of alternatives for women. The sex effect was not expected, and we are currently conducting an experiment to assess its replicability.

Emotion as Action Tendency Experiment

An experiment has also been conducted to test the hypothesis that dissonance should be increased as the salience of the action implications of cognitions that are involved in a dissonant relationship are increased. Several Perspectives consider emotion to involve action tendencies (Frijda, 1986; Lang, 1995). To the extent that an emotion generates an action tendency, as the intensity of one’s current emotion is increased and is involved in a dissonant relationship with other information, dissonance should be increased.

Much research has demonstrated that the emotion of sympathy (empathy) increases helping behavior because it evokes altruistic motivation, that is, motivation to relieve the distress of the Person in need of help (for a review, see Batson, 1991). We have conducted one experiment that tested whether an inconsistency between the emotion of sympathy and knowledge about past behavior evoked motivation to reduce this inconsistency (Harmon-Jones, Peterson, & Vaughn, in press).

In the experiment, we tested the hypothesis that, after experiencing sympathy for a target Person in need of help, individuals will be more motivated to help that Person when they are reminded of times when they failed to help similar Persons (for evidence that feeling sympathy for one target Person can transfer to the target Person’s group and cause attitude change toward the group, see Batson et al., 1997). Participants were informed that they would be listening to a pilot broadcast for a local radio station and that the researchers would like students’ reactions to the tape. Participants then listened to a tape-recorded message that was supposedly from a Person in need of help (an adolescent with cancer). Before listening to the tape, participants were assigned to one of two conditions: one in which they tried to imagine how the Person must feel (high-empathy set) or one in which they tried to remain objective as they listened to the tape (low-empathy set). Then, they listened to the tape-recorded message and afterward completed questionnaires assessing self-reported emotional responses and evaluations of the tape-recorded message. Then, either participants were asked to list times when they failed to help other Persons who were in need of help or they completed a demographic survey. Finally, participants were given an opportunity to help by volunteering time to assist the Person with addressing letters that would request money from possible donors or by donating money to the Person’s family. The design was a 2 (Empathy: low vs. high) × 2 (Times When Did Not Help: reminded vs. not reminded) between-subjects factorial. Results indicated that more helping occurred in the high-empathy/reminded of past failures condition than in other conditions.

Future Directions

Serious challenges to each of the revisions of the original theory have been offered, thus questioning the necessity of the motivations for dissonance processes proposed by these revisions. The action-based model offers a new motivational explanation of the dissonance process, and its motivation—the need for effective and unconflicted action—may be super ordinate to the motivations offered by other models. However, the challenges to the revisions have not cogently demonstrated the insufficiency of these motivations for dissonance processes, and it is thus possible that these alternative motivations can influence dissonance processes, Perhaps depending on the context. Thus, each of the proposed motivations may independently influence dissonance processes. Future research will need to address whether these motivations are sufficient to independently alter dissonance processes or whether they are Perhaps subordinate to and part of another motivation such as the need for effective action.

To more thoroughly understand the motivation(s) underlying the dissonance process, future research will need to focus on the mediators of the process. In our recent research derived from the action-based model, we have examined neural mechanisms involved in the dissonance process, and we plan to continue this line of research. In addition to investigations of the neural mechanisms, future research will need to examine experiential mediators of the process, Perhaps in conjunction with the neural mechanisms.

Another suggestion for future research is the investigation of the role of individual differences in the dissonance process, an area that has been relatively neglected in past research. We have recently obtained evidence that trait action-orientation for decision-related contexts relates to spreading of alternatives. Not only will investigations of individual differences contribute to a more thorough understanding of dissonance motivation, but the examination of individual differences will provide suggestions for effectively implementing dissonance-related techniques for inducing Persuasion outside the laboratory with different groups of individuals. Investigations of the role of individual differences in the dissonance process, however, need to carefully consider the various places in the process at which the individual difference could play a role in affecting outcome variables.

The precise prediction of the manner in which dissonance will be reduced depends on a specification of the cognition that is most resistant to change. Thus, when a Person is exposed to a counterattitudinal communication, the Person will experience dissonance and the dissonance could lead to different outcomes, such as either rejection or acceptance of the communication. Which outcome will occur should depend on the resistance to change of the involved cognitions (Harmon-Jones, 2000b). In laboratory settings, the experimenter has control over which cognition is more resistant to change and thus can predict how dissonance will be reduced. But as dissonance research is used outside the laboratory and to test the moderating role of individual differences, the specification of the determinants of resistance to change and the measurement of resistance to change will become necessary. Future research is needed to address these issues.

An understanding of these basic mechanisms involved in the dissonance process will ultimately lead to more precise specifications of conditions necessary to produce Persuasion in naturally occurring circumstances.


The original theory proposed that cognitive inconsistency generated an aversive state that motivated cognitive and behavioral changes, which often result in Persuasion. Several revisions to the theory have proposed alternative motivations presumably responsible for the Persuasion effects observed in dissonance experiments. Challenges to each of these revisions have been offered, and recent research supports the original version over these revisions. However, the original theory never clearly specified why cognitive inconsistency generated an aversive motivational response. An action-based model of dissonance theory has been proposed that extends the original theory by specifying why cognitive inconsistency evokes an aversive motivational state. In addition to linking the study of dissonance to models of self-regulation, the action-based model suggests important lines of inquiry into understanding how affect and motivation influence the cognitive and behavioral changes that lead to Persuasion.