When bad things happen to good feedback: exacerbating the need for self-justification with self-affirmations.

In numerous self-affirmation studies, Claude Steele and colleagues have demonstrated that self-affirmations reduce the need to justify dissonant behavior even when the affirmation is unrelated to the dissonance-evoking action. However, research has not sufficiently examined the impact of reaffirming self-aspects that are related to the dissonance. The authors argue that relevant affirmations of this sort can make salient the standards that are violated in the course of dissonant behavior, thereby increasing dissonance and the need for self-justification. In a laboratory study using the induced-compliance paradigm, it was demonstrated that dissonance can be exacerbated by reaffirming standards that are violated in the course of the dissonant behavior.

In 1957, Festinger proposed that a person holding two inconsistent cognitions would experience the psychological state of cognitive dissonance, which motivates efforts to reduce dissonance and achieve consonance. Numerous theorists have since proposed revisions to Festinger's original theory of dissonance. Typically using the induced-compliance paradigm, they have sought to redefine dissonance by finding the limiting conditions under which counterattitudinal behavior leads to attitude change and when it does not. For instance, Aronson (1968) proposed that dissonance occurs only when a cherished self-concept is threatened. Thus, Nel, Helmreich, and Aronson (1969) had participants advocate a counterattitudinal position to legalize marijuana and demonstrated that it lead to attitude change only in conditions designed to challenge their sense of moral decency. More recently, Cooper and Fazio (1984) proposed that dissonance is evoked only when an individual feels personally responsible for bringing about an aversive event. In this tradition, Cooper and colleagues have shown that counterattitudinal behavior leads to dissonance only when one freely and knowingly chooses to create an unwanted outcome (e.g., Cooper, 1971; Cooper & Brehm, 1971; Cooper & Worchel, 1970).

Throughout such revisions, however, theorists have agreed with Festinger's (1957) original thesis that dissonance arousal motivates dissonance reduction. Moreover, it has been assumed that such efforts must address the action that originally evoked the dissonance if they are to be effective. Thus, when participants in induced-compliance studies alter their prior attitudes, it has been interpreted as evidence that they are attempting to undo their previous action--as a way of either restoring cognitive consistency (Festinger & Carlsmith, 1959), reinstating a cherished self-belief (Aronson, 1968), or retaining a belief that they have not brought about unwanted consequences (Cooper & Fazio, 1984). Once the evoking action is disarmed through self-justifying attitude change, dissonance motivation subsides.

A well-reasoned alternative to these traditional approaches has been articulated by Steele and his colleagues (e.g., Steele, 1988). Like others, they have advanced a definition of dissonance that differs from that proposed in Festinger's (1957) original theory. They argue that dissonance is created when one's global self-evaluation is threatened. This self-affirmation theory of dissonance deviates from prior revisions, however, by suggesting that dissonance can be reduced or eliminated through actions that do not directly address the dissonance-evoking action. As Steele and Liu (1983) stated, "Because the disturbing thing about dissonant behavior is its ego threat, any self-affirming activity may reduce dissonance even when it does not resolve or dismiss the particular provoking inconsistency" (p. 18). Research using the induced-compliance paradigm, they argue, has uncovered only self-justifying reactions to dissonance because dissonance researchers have failed to give their participants opportunities to make other responses. Had they provided opportunities to reaffirm, participants would not have to respond by changing their attitudes.

Steele and Liu (1983) demonstrated the ameliorating effect of affirmations with a group of students who had advocated tuition increases. In high-dissonance conditions, participants did not change their attitudes toward the cost of tuition if they were first given an opportunity to reaffirm an important, but unrelated, political and economic value orientation. Steele, Spencer, and Lynch (1993) took this a step further by suggesting that some people possess a high number of affirmational resources that they can use to buffer the self from the threat of future dissonant actions. Affirmational resources are alternative positive self-conceptions that one can have at the ready when the self is threatened in specific situations. In one study (Steele et al., 1993, Study 1), participants were provided with affirmational resources in the form of positive feedback from a bogus personality test. Later, when forced to choose between two attractive record albums, participants did not exhibit the postdecisional spread of alternatives indicative of dissonance arousal.

These results provide interesting evidence consistent with self-affirmation theory because the actions that eliminate the attitude change are completely unrelated to those that evoke the dissonance. It seems remarkable that an individual who has acted in a way that creates unwanted and aversive consequences would lose the motivation to justify such behavior simply by reaffirming important but distinct aspects of the self-concept. As Prentice and Miller (1992) have pointed out, results of this kind provide strong intuitive support for a theory because the independent variables (e.g., political/economic orientation and personality feedback) seem unlikely to influence the dependent variables (e.g., attitudes toward tuition and preference for record albums) in any meaningful manner were it not for the theoretical backdrop. In this regard, Steele and colleagues eliminate dissonance in what seem to be the most unlikely of ways. Steele, Hopp, and Gonzales (1986, reported in Steele, 1988), for instance, eliminated the postdecisional spread of alternatives among a group of students who valued a scientific orientation simply by allowing them to wear a scientist's lab coat. Such results demonstrate that dissonance motivation can be reduced in ways not originally envisioned by Festinger (1957) or other traditional dissonance theorists.

Dissonance and the Relevant Affirmation

Our research begins by asking what would happen if one were to reaffirm the aspect of the self that is threatened by the dissonant behavior, rather than an aspect that is unrelated to the evoking action. The appeal of self-affirmation theory has been its ability to predict a nonobvious relationship between global affirmations and seemingly unrelated attitudes. However, it does not appear to be necessary that the affirmation and …