خودتأییدگری مقابل خود انسجامی: مقایسه دو رقیب خود نظریه پدیده ناهماهنگی

Abstract High or low self-esteem individuals participated in a role-playing paradigm in which a friend stood them up for a dinner date. The participants received either a good explanation from the friend for the missed date (sufficient justification) or a poor explanation (insufficient justification). As predicted by self-affirmation theory (Steele, 1988), but not self-consistency theory (Aronson, 1999), low-esteem participants derogated the friend more than high-esteem participants under both insufficient and sufficient justification. Also supporting self-affirmation theory, sufficient/low-esteem participants reported more offense for being stood-up than sufficient/high-esteem participants. Discussion centers on the role of self-esteem in dissonance processes and on the need for more research that focuses on dissonance/self-threats that result from the behavior of other(s) rather than one’s own behavior.

. Introduction Few theories in social psychology have been as influential or enduring as Festinger’s (1957) cognitive dissonance theory. The 1990s and early 2000s, in fact, have seen a revival of interest in dissonance phenomena as evidenced by the publication of two books (Beauvois & Joule, 1996; Harmon-Jones & Mills, 1999) and numerous journal articles (e.g., Aronson, Blanton, & Cooper, 1995; Blanton, Cooper, Skurnik, & Aronson, 1997; Harmon-Jones, 2000; Nail et al., 2001; Simon, Greenberg, & Brehm, 1995; Stone & Cooper, 2001; Van Overwalle & Jordens, 2002). In the first published statement of the theory, Festinger, Riecken, and Schachter (1956) proposed that two cognitions are dissonant “if they do not fit together––that is, if they are inconsistent, or if, considering only the particular two items, one does not follow from the other” (p. 25). Dissonance results in a feeling that is psychologically aversive or unpleasant. Accordingly, it creates a drive-like state toward reducing this unpleasantness. Dissonance can be reduced in numerous ways, for example, by changing one of the dissonant cognitions, decreasing the importance of the dissonant cognitions, and/or obtaining social support (Festinger, 1957; Festinger et al., 1956). One popular method of testing the theory is known as the free-choice paradigm ( Brehm, 1956). Here, participants are asked to evaluate several similar items in terms of how desirable they are (i.e., music CDs). After the ratings, participants are given the opportunity to possess one of the items, but they must chose between two items that they initially rated almost equally. Because the items are similarly valued, participants presumably do not have sufficient justification for picking one item over the other. Being forced to make this choice creates dissonance because the chosen alternative invariably has some negative qualities, whereas the non-chosen alternative has some positive qualities (i.e., some bad songs, some good, respectively). After the choice, participants are asked to re-rate the items. Typically, they increase their rating of the chosen alternative but decrease that of the non-chosen alternative. Dissonance is measured by the absolute difference in the change of these ratings, the so-called spread of alternatives. Although early dissonance research generated a substantial body of evidence supporting the theory (e.g., Aronson & Carlsmith, 1963; Festinger & Carlsmith, 1959; Gerard & Mathewson, 1966), researchers soon discovered that inconsistency, in and of itself, is not enough to create the cognitive/behavioral changes postulated by the theory. Specifically, dissonance-arousing behavior must be perceived as: (a) having been freely chosen (Davis & Jones, 1960; Sherman, 1970), (b) having little external justification (Aronson & Carlsmith, 1963; Festinger & Carlsmith, 1959), and (c) entailing a commitment (Carlsmith, Collins, & Helmreich, 1966). These exceptions to the conditions that produce dissonance “strained the limits of the original inconsistency-based explanation” for dissonance phenomena (Scher & Cooper, 1989, p. 899). Consequently, several theorists proposed modifications in dissonance theory’s tenets in an effort to offer a more parsimonious account of dissonance effects (e.g., Aronson, 1968 and Aronson, 1999; Brehm & Cohen, 1962; Wicklund & Brehm, 1976). Other theorists offered alternative theories (e.g., Collins & Hoyt, 1972; Cooper & Fazio, 1984; Steele, 1988; Van Overwalle & Jordens, 2002).1 1.1. Self-consistency theory One prominent modification is Aronson, 1968 and Aronson, 1999 self-consistency theory. The theory proposes that it is not just any two inconsistent cognitions that produce dissonance; rather, dissonance is dependent upon how the cognitions reflect upon a person’s self-concept. According to Aronson, difficult choices produce dissonance and attitude change not fundamentally because of cognitive inconsistency but because of how the choice reflects on the self. The spread of alternatives occurs in the free-choice paradigm because the specter of making a bad choice is inconsistent with the positive image most people have of themselves as competent and effective decision-makers. Changing one’s ratings of the items allows one to rationalize self-relevant negative thoughts that typically accompany making a tough choice. Aronson, 1968 and Aronson, 1999 conceptualization of dissonance is in some ways more parsimonious than Festinger’s (1957). Specifically, if dissonance is defined as inconsistency with one’s self-concept, all of the exceptions to inconsistency-based dissonance listed above neatly align themselves as corollaries of dissonance (Aronson) rather than as qualifiers of dissonance (Festinger). For example, under Aronson’s account, one would not expect behavior that is not freely chosen to cause much dissonance because such coerced behavior could be easily disconnected or disassociated from the self. At the same time, Aronson’s conceptualization represents only a revision of Festinger’s theory because inconsistency remains the driving force in dissonance-reducing behaviors. However, the inconsistency is not a lower-order inconsistency between just any two cognitions. Rather, dissonance is determined by how the inconsistency interfaces with one’s self-concept, hence the moniker, self-consistency theory. 1.2. Self-affirmation theory Steele’s (1988) self-affirmation theory can explain much of the dissonance literature, yet it stands as an alternative and competing theory of dissonance phenomena (Thibodeau & Aronson, 1992; see Aronson, Cohen, & Nail, 1999, for a review). The theory is similar to self-consistency theory in two ways. First, both theories point to the critical role of the self in producing dissonance; consequently, both theories conceptualize dissonance as being grounded in the ego and not in a free standing need for cognitive consistency (see Greenwald & Ronis, 1978, p. 55). Second, like self-consistency theory and for the same reasons, self-affirmation theory is relatively parsimonious. Unlike E. Aronson, however, Steele maintains that consistency–inconsistency is NOT at the heart of dissonance motivation. According to Steele, dissonance is not the aversive tension of logic-like inconsistency (Festinger, 1957), or the tension of self-inconsistency (Aronson, 1968 and Aronson, 1999), but rather the tension of a “threatened sense of self-integrity” (Steele, Spencer, & Lynch, 1993, p. 893). Steele postulates the existence of a “self system for maintaining a perception of global integrity, that is, of overall moral and adaptive adequacy” (Steele et al., 1993, p. 885). Dissonance occurs when one engages in behavior or learns something about one’s self that threatens this overall perception of self-integrity. Thus, the spread of alternatives in the free-choice paradigm occurs NOT to reduce inconsistency, but rather to restore one’s overall sense of self. 1.3. Self-consistency theory versus self-affirmation theory On first reflection, the theories seem almost identical. Specifically, it appears that any inconsistency with the self would also always represent a threat to the self. Indeed, in most situations the theories do make identical predictions, and they can explain much of the same data (see Thibodeau & Aronson, 1992, pp. 597–598). However, the theories make conflicting predictions when considering high versus low self-esteem. According to self-consistency theory, high-esteem individuals should experience more conflict in dissonance-arousing situations than low-esteem individuals. Assuming that self-inconsistency is the driving force in dissonance motivation, it follows that the possibility of a bad choice, for example, would be more inconsistent with the self-concept of high-esteem individuals than with that of low-esteem individuals. Thus, high-esteem individuals should display a greater spread of alternatives than lows. This possibility is generally referred to as a self-discrepancy effect ( Aronson et al., 1995). In contrast, self-affirmation theory predicts that high-esteem individuals will experience less dissonance than individuals with low-esteem. High-esteem individuals have positive self-concepts precisely because they have large reservoirs of positive information and thoughts regarding the self. Assuming that a threatened sense of self-integrity is the driving force in dissonance, it follows that high-esteem individuals would have greater self-resources to draw on than low-esteem individuals. Thus, high-esteem individuals should display a smaller spread of alternatives following a difficult choice. This possibility is generally referred to as a self-affirmation effect. Steele et al. (1993, Study 2) put these ideas to the test in a 2 × 2 design that crossed chronic high/low self-esteem × focus/no-focus on one’s self-concept. Chronic high/low self-esteem was determined based on a median split of scores on the Rosenberg (1965) Self-esteem scale. Self-focus/no-self-focus was manipulated by having randomly selected participants re-take the Rosenberg scale at the start of the experimental session or not. Dissonance was evoked employing a choice between music CDs in the free-choice paradigm. The results showed that in the no-self-focus condition, there was no difference in the spread of alternatives between high (M=0.62) and low-esteem (M=0.56) participants. In the self-focus condition, however, high-esteem participants (M=−0.29) displayed significantly less spread of alternatives than low-esteem participants (M=1.14). These results support self-affirmation theory over self-consistency theory. Presumably, the self-focus manipulation caused participants to reflect on their self-concepts and self-resources. Self-reflection resulted in low-esteem participants rationalizing their choice more than high-esteem participants (i.e., a self-affirmation effect). Apparently, with their self-systems restored, high-esteem participants simply had no need to rationalize their choice. Similar supporting evidence for self-affirmation theory over self-consistency theory regarding self-esteem has been obtained in a replication of Steele et al. (1993, Study 2) by Stone (1999) and in a related study by Steele et al. (1993, Study 1).2 Given that all known published research to date examining chronic self-esteem and self-consistency theory versus self-affirmation theory has been based on the free-choice paradigm, we wondered how the theories might fare employing a different dissonance paradigm. Specifically, our line of research has led to situations in which dissonance is caused by another’s behavior rather than one’s own behavior ( Nail, Bedell, & Little, in press; Nail et al., 2001; see also Tesser, 1988). Consistent with both E. Aronson’s and Steele’s definitions of dissonance, a self-inconsistency or a self-threat, respectively, could be caused as easily by another person’s actions as it could be by one’s own actions. For example, being stood-up for a date by a friend in the absence of a good reason (i.e., sufficient justification) should cause dissonance in most people. Note that the dissonance in this case would be the result of the friend’s behavior rather than one’s own. Yet, who will experience the greatest dissonance in this situation, those with high self-esteem, consistent with Aronson, or those with low self-esteem, as Steele would predict? To address this issue, we employed the role-playing/dissonance paradigm developed by Nail et al. (2001). Participants were asked to imagine a friend, Chris, standing them up for a dinner date and later having an adequate explanation for missing the date (sufficient justification) or no adequate explanation (insufficient justification). Participants were then allowed the chance to derogate Chris as a friend, the primary measure of dissonance reduction. Consistent with the Nail et al. (2001) findings and both theories, in Hypothesis 1 we predicted that insufficient participants would derogate Chris more than sufficient participants. Derogating Chris allows insufficient participants the chance (a) to decrease the inconsistency between how they were treated and how they view themselves (à la Aronson and self-consistency theory) or (b) to restore their overall sense of self-integrity (à la Steele and self-affirmation theory). The biggest consideration, however, concerns how high and low self-esteem participants might differ. If dissonance is caused by inconsistency with one’s self-concept as self-consistency theory proposes, Hypothesis 2 states that insufficient justification participants with high self-esteem should derogate Chris more strongly than insufficient participants with low-esteem, a self-discrepancy effect. The logic is that being stood-up is more inconsistent with high- than with low-esteem, and derogating Chris provides a means of reducing this inconsistency. Interestingly, Hypothesis 2 is consistent with related theories of the self in social settings, Swann’s (1990) self-verification theory and the work of Baumeister and colleagues on self-esteem and interpersonal rejection (e.g., Baumeister, Smart, & Boden, 1996). On the other hand, if dissonance is caused by a threat to one’s self-system as self-affirmation theory maintains, Hypothesis 3 states that insufficient participants with high-esteem should derogate Chris less strongly than insufficient participants with low-esteem, a self-affirmation effect. High-esteem individuals have greater resources to counteract the negativity of being stood-up than do low-esteem individuals, and thus, they should have less of a need to derogate Chris. Finally, if the Steele et al. (1993) results with free-choice dissonance paradigm generalize to the Nail et al. (2001) role-playing dissonance paradigm, Hypothesis 4 states that we should find differences between high- and low-esteem individuals only under the condition that self-esteem is primed at the beginning of the experimental session. Only when participants are self-focused should self-esteem have an influence on the evaluations of Chris as a friend. We investigated these issues in the context of a 2 × 2 × 2 design. One independent variable was a chronic individual difference variable: self-esteem (high versus low). The other two were manipulated, situational variables: justification (sufficient versus insufficient) and self-focus (self-focus versus no-self-focus).

. Results With the present sample, the Rosenberg scale yielded M=3.28, Mdn=3.3, SD=0.42, and α=0.81, indicating good overall reliability. Consistent with our pilot work, there were no effects on Rosenberg scores as a function of justification or self-focus (F’s < 1). Further, the results indicated no main effects or interactions for self-focus or gender on either of the dependent variables. Thus, these variables are excluded from the analyses reported below. The lack of self-focus effects is inconsistent with Hypothesis 4––that self-esteem would have to be primed to produce differences between high- and low-esteem individuals in dissonance reduction. Participants were classified as having high or low self-esteem based on a tercile split of the Rosenberg scale. One participant was excluded from the analyses because of an extreme Offended score (2.62 standard deviations from the mean of his experimental cell). Participants falling on the tercile break points were included to more closely approach equal cell sizes. Mean post ratings regarding Friend are presented in Table 1. A 2 (self-esteem: high versus low) × 2 (justification: sufficient versus insufficient) ANOVA of these data yielded a significant main effect for justification, F(1,79)=86.19, p<0.0001, η2=0.51. Consistent with Hypothesis 1, after being stood-up insufficient participants overall rated Chris as less of a friend (M=5.13) than sufficient participants (M=9.11), thus supporting the basic dissonance theory/insufficient justification effect. Table 1. Frienda means and standard deviations for justification condition by self-esteem Justification condition Self-esteem Low High n M SD n M SD Sufficient 20 8.5 1.64 25 9.6 1.78 Insufficient 18 4.56 2.09 20 5.65 2.21 a Friend ratings were made on an 11-point scale (1=not a very good friend; 11=a good friend). Table options The self-esteem × justification analysis regarding Friend also produced a significant main effect for self-esteem. Consistent with Hypothesis 3 but contrary to Hypothesis 2, high self-esteem participants overall derogated Chris less (M=7.84) than low-esteem participants (M=6.63), F(1,79)=6.66, p<0.02, η2=0.05. The interaction term, however, did not approach significance (F<1). 3 The data regarding Offended are presented in Table 2. A 2 × 2 ANOVA of these data produced significant main effects for justification, F(1,79)=47.85, p<0.0001, η2=0.34, and for self-esteem, F(1,79)=4.99, p<0.03, η2=0.04. Insufficient participants overall reported that they were more offended for being stood-up (M=7.0) than sufficient participants (M=3.42), and low-esteem participants reported more offense (M=5.74) than high-esteem participants (M=4.49). However, both main effects are qualified by a significant justification × self-esteem interaction, F(1,79)=9.01, p<0.003, η2=0.10. Internal examination of the simple main effects revealed three significant contrasts. Insufficient/high self-esteem participants reported that they were more offended (M=7.25) than sufficient/high self-esteem participants (M=2.28), F(1,79)=80.39, p<0.0001, η2=0.50. Likewise, insufficient/low-esteem participants reported more offense (M=6.72) than sufficient/low-esteem participants (M=4.85), F(1,79)=4.27, p<0.05, η2=0.05. Further, sufficient/low-esteem participants reported greater offense (M=4.85) than sufficient/high-esteem participants (M=2.28), F(1,79)=13.06, p<0.0001, η2=0.14. There was no significant difference in Offended ratings, however, between insufficient/low-esteem (M=6.72) and insufficient/high-esteem participants (M=7.25), F<1. 4 Table 2. Offendeda meansb and standard deviations for justification condition by self-esteem Justification condition Self-esteem Low High n M SD n M SD Sufficient 20 4.85b 2.89 25 2.28a 1.49 Insufficient 18 6.72c 2.67 20 7.25c 2.22 a Offended ratings were made on an 11-point scale (1=not at all offended; 11=very offended). b Means with different subscripts within rows and columns differ beyond the 0.05 level.