The Politics of Affirmation Theory

Abstract

It has been well established in the literature that affirming the individual self reduces the tendency to exhibit group-favoring biases. The limited research examining group-affirmation and bias, however, is inconclusive. We argue that group-affirmation can exacerbate group-serving biases in certain contexts, and in the current set of studies, we document this phenomenon directly. Unlike self-affirmation, group-affirmation led to greater ingroup-favoring evaluative judgments among political partisans (Experiment 1). This increase in evaluative bias following group-affirmation was moderated by political party identification and was not found among those who affirmed a non-political ingroup (Experiment 2). In addition, the mechanism underlying these findings is explored and interpreted within the theoretical frameworks of self-categorization theory and the multiple self-aspects model (Experiments 2 and 3). The broader implications of our findings for the understanding of social identity and affirmation theory are discussed.

Keywords

group-affirmation self-affirmation intergroup processes ingroup bias

People tend to evaluate information and make judgments in a manner that best serves the interests of groups to which they belong (Dovidio & Gaertner, 2010; Hewstone, Rubin, & Willis, 2002). One way to mitigate the expression of ingroup bias is through self-affirmation (Steele, 1988). By bolstering the self-concept in one domain, one becomes less defensive in the face of threats to other self-relevant domains, including group memberships (Sherman & Cohen, 2002). Some researchers have argued that, if affirming an aspect of one’s individual identity can lead to a reduction in group-level biases, then affirming an aspect of one’s social identity should produce a similar effect. The limited research that has addressed this question, however, has been inconclusive. In fact, we argue that group-affirmation can have the opposite effect—accentuating certain biases. In a series of experiments, we demonstrate that affirming one’s group identity can lead to greater group-serving biases, both in terms of intergroup evaluative ratings and correspondence between personal beliefs and perceived ingroup beliefs. We also present some preliminary evidence elucidating the mechanism behind this bias-enhancing effect and lend some clarity to the contradictory findings in the literature.

Social Identity and Self-Affirmation

Social identity theorists argue that people possess various social identities derived from their group memberships (Tajfel & Turner, 1986). Social identity theory is based on the notion that people continually strive to enhance or maintain positive self-images and, further, that group memberships often comprise important components of the self-concept (Tropp & Wright, 2001). From this perspective, ingroup biases result from a need to see one’s group (and thus oneself) in a positive light. It follows that when people define themselves in terms of a social identity, the personal self-image will overlap with the image of the ingroup. Thus, phenomena that reflect favorably on the group will also reflect favorably on the individual. Likewise, threats to the group will be perceived as threats to the individual’s self-concept. People can counter threats to the individual or to the group and maintain a positive view of the self by exhibiting biases that make the ingroup seem more favorable than other groups (Fein & Spencer, 1997; Riek, Mania, & Gaertner, 2006).

Self-affirmation theory (Steele, 1988) emphasizes an alternative means by which individuals can maintain positive views of the self and the ingroup. In the face of threatening information, people can repair their damaged self-images and maintain a sense of integrity by affirming a valued domain of the self unrelated to the threat (Spencer, Fein, & Lomore, 2001). This can be accomplished, for example, by reflecting on a positive aspect of one’s self-concept or engaging in an activity that makes such an aspect salient, such as writing about an important self-relevant value (Crocker, Niiya, & Mischkowski, 2008). While doing this, individuals experience a boost to their self-worth and become more confident and secure, enabling them to evaluate potentially self-threatening information in a less defensive manner. For example, people tend to interpret ambiguous self-relevant information in a way that is consistent with their preexisting beliefs and desires (Nisbett & Ross, 1980). After affirming the self, however, people are more likely to consider evidence that contradicts their own beliefs or view an issue from perspectives other than their own (Sherman & Cohen, 2002). This self-affirmation effect extends to group-based threats and biases. If the motivation to protect the self is satisfied through self-affirmation, one can respond to information and threats relevant to one’s group memberships (other domains of the self) in a more objective way, thus making the group-serving biases normally exhibited following a threat less necessary. This bias-reducing pattern has been found in a number of group-based contexts, including group-serving attributions made by intramural sports teams (Sherman & Kim, 2005), perceptions of racism exhibited by ingroup members (Adams, Tormala, & O’Brien, 2006), and self-reported ethnic prejudice (Zarate & Garza, 2002).

Group-Affirmation

Although a number of studies have shown that affirming an aspect of the individual self can reduce group-serving judgments, there is some ambiguity with respect to the effect of group-affirmation on ingroup biases. Derks, van Laar, and Ellemers (2009) argue that self-affirmation enhances personal identity whereas group-affirmation enhances social identity. To the extent that these identities overlap and are highly connected, it could be reasoned that group-affirmation should be just as effective at countering group-related threats as is self-affirmation. Thus, it might be expected that affirming a group identity will bolster one’s sense of self-integrity, making one more open to possible group-threatening information and mitigating the need to make group-serving judgments. Some empirical findings seem to support this logic. For example, Sherman, Kinias, Major, Kim, and Prenovost (2007) found that affirming the group reduced the tendency of intramural athletes to attribute a positive game outcome to internal, team-related factors (just as a self-affirmation exercise had done). Additionally, Gunn and Wilson (2011) found that men (Experiment 1) and Canadians (Experiment 2) who performed a group-affirmation task were more willing to express feelings of collective guilt and shame after reading about historical ingroup transgressions against women and Canadian Aboriginals, respectively.

The bias-reducing effect of group-affirmation has not been reproduced consistently, however. Cehajic-Clancy, Effron, Halperin, Liberman, and Ross (2011) found that Israeli students who performed a self-affirmation task were more willing to express group-based guilt, acknowledge responsibility for wrongdoing, and support reparative measures with respect to the Israeli–Palestinian conflict. Students who performed a group-affirmation task, however, did not show this reparatory shift and continued to demonstrate a significant bias favoring the ingroup. Similarly, Glasford, Dovidio, and Pratto (2009) found that students who were given the opportunity to affirm their American identities showed no increase in collective guilt after reading a report describing a bombing run made by American soldiers in which many Iraqi civilians were killed. In addition, participants who affirmed their nationality exhibited less psychological discomfort after reading the report than did those who did not group-affirm, and this pattern was especially strong for those highly identified with their nationality. The authors explain this decrease in psychological discomfort following group-affirmation as a reduction in “intragroup dissonance.” This finding also suggests an increased willingness to tolerate ingroup transgressions (and thus a greater tendency to exhibit biases favoring the ingroup) among the group-affirmed. There seems, then, to be some contradictory hypotheses and findings concerning the effect of group-affirmation on ingroup biases. In some cases, affirming one’s group identity appears to have a bias-reducing effect similar to that of self-affirmation, but in other cases group-affirmation seems to be ineffective at reducing group-favoring biases and might, in fact, exacerbate them.

An alternative way to consider the potential effects of group-affirmation on group-level biases may help resolve these apparent contradictions. Perhaps there are distinct processes occurring when one affirms a group compared with when one affirms the personal self. In some cases, these processes may result in similar outcomes (such as reducing ingroup biases) and sometimes they may lead to contrasting outcomes. Derks et al. (2009), for example, found that both self-affirmation and group-affirmation led to feelings of self-improvement and achievement motivation among highly identified members of stigmatized groups. They argue, however, that these similar outcomes resulted from different processes. Self-affirmation increased achievement motivation by inducing participants to focus on their individuality and sparking their desire for personal success and mobility. Group-affirmation, by contrast, did so by inducing highly identified group members to focus on their collective identity and by sparking their desire for group-level success. Thus, these authors contend that affirming a group identity encourages group members to think and behave in ways that improve the position of the group as a whole, whereas affirming an individual identity directs group members’ attention away from the group and allows them to focus on personal outcomes.

A similar discontinuity between self- and group-affirmation may exist in the context of group-serving biases. Two models help provide a theoretical basis for this distinction: self-categorization theory (SCT; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) and the multiple self-aspects framework (McConnell, 2011). According to SCT, one’s self-concept can be categorized at different levels of abstraction, ranging from the self as an individual to the self as the embodiment of an ingroup. Whether one categorizes the self at the group level depends, in part, on the contextual salience of that group identity. The SCT framework also posits that, at the individual level, one’s personal goals and motivations are most salient. At the group level, the goals and achievements of the collective are most salient and become merged with those of the individual (Dovidio & Gaertner, 2010). In addition, McConnell’s multiple self-aspects framework posits that one’s overall self-concept can be construed as a collection of various context-dependent selves. Different contexts activate different aspects of the self and make the affective traits and emotional states associated with those aspects more salient. Greater salience can lead to a greater tendency to exhibit specific behaviors and attitudes corresponding to those traits and emotional states.

A self-affirmation manipulation may provide a boost to the self by inducing a focus on positive aspects of one’s global (individual-level) sense of self, or “overall self-concept of worth” (Steele, 1988, p. 266), thus making aspects of the self that are related to certain group memberships less salient and mitigating the need to defend those identities. In fact, Critcher and Dunning (2015) demonstrated that affirming the self reduces defensiveness in the face of threat precisely because it expands the working self-concept and directs attention away from the threatened domain. We posit that a group-affirmation manipulation, in contrast, may directly activate that group’s self-aspect or social identity, making any constructs associated with that group psychologically predominant. Presumably, these constructs would include biases toward salient outgroups and the desire to enhance and defend the worth and integrity of the ingroup. If so, affirming the group may exacerbate, rather than reduce, ingroup-favoring biases in the face of threat. In addition, it is likely that the link between group-affirmation and self-aspect activation would be especially strong for groups with which one is highly identified. Such groups, which would comprise a significant part of an individual’s self-concept (Tropp & Wright, 2001), may be especially sensitive to threat and prone to bias if their identities are made salient through a group-affirmation procedure.

There is some tangential evidence for our proposed process in research that has looked at self-affirmations that are domain relevant versus domain irrelevant (i.e., whether an individual is affirmed in the same or different domain in which he or she is threatened). Blanton, Cooper, Skurnik, and Aronson (1997), for instance, found that, in the context of attitude–behavior symmetry, individuals who were affirmed in a domain relevant to their dissonant behavior actually showed exacerbated levels of dissonance as compared with those who were affirmed in an irrelevant domain. Similarly, in a series of studies exploring escalation of commitment, Sivanathan, Molden, Galinsky, and Ku (2008) found that participants who affirmed themselves on traits with low relevance to decision-making ability demonstrated a reduced tendency to justify their decisions and escalate their commitment, whereas those who affirmed themselves on traits relevant to this ability continued to engage heightened commitment. Thus, affirming the self-concept in the same domain in which one is threatened may actually backfire and produce a different outcome than affirming the self-concept in an irrelevant domain. Similarly, it is possible that responding to a group-level threat by affirming that very same group (as opposed to an irrelevant group or an aspect of the individual self) would have a similar, bias-enhancing outcome. In a series of experiments, we document this phenomenon directly.

The Current Research

In three experiments, we investigated the relationship between group-affirmation and ingroup bias. Specifically, we tested whether, in certain contexts, affirming the group (in contrast to affirming the self) not only fails to reduce defensiveness and biases but can also exacerbate them. In Experiment 1, we examined the tendency of American political partisans to engage in evaluative biases following either a self-affirmation or group-affirmation. We hypothesized that affirming the self and affirming the group in this threatening context would have contrasting effects on the strength of the ingroup–outgroup evaluative bias. In Experiment 2, we sought to lend some clarity to the mechanism behind the bias-enhancing effect of group-affirmation observed in Experiment 1. Specifically, we examined the effects of two different group-affirmations (one related to politics and one unrelated to politics) on a broader array of political attitudes. In addition, we tested whether group-affirmation effects were stronger for persons highly identified with their party. In Experiment 3, we sought to further delineate the processes underlying group-affirmation. Specifically, we tested whether group-affirmation (in contrast to self-affirmation) makes ideas associated with that group more salient.

Experiment 1

In Experiment 1, we provide a basic test of whether self-affirmation and group-affirmation can have contrasting effects on ingroup biases. To determine what group-level contexts members of our participant pool found particularly important (and thus were likely to identify strongly with), we conducted a pilot study in which undergraduates at our university were asked to indicate the most important groups to which they belonged. Results from the pilot study revealed that political party membership was the social identity selected most frequently behind only athletic team affiliation. Given the accessibility of political partisans on a college campus and the applicability of political affiliation to a broad range of people (as opposed to identification with a particular sports team), that was the social identity context we selected. Given our research question, we felt that political party affiliation would be an especially informative identity on which to focus for additional reasons. In the United States, there are two major political parties (Democrats and Republicans) providing a built-in intergroup comparison. This intergroup context is especially contentious and bias ridden. By definition and in practice, the Democratic and Republican parties in the United States are highly threatening to one another (Binning, Sherman, Cohen, & Heitland, 2010; Chambers & Melnyk, 2006; Ware, 2011), and it is likely that hostile interparty attitudes and biases would be a core component of partisans’ political party self-aspects. Therefore, we thought that this would be an ideal context within which to examine our hypotheses at the intergroup level.

Specifically, we examined the effects of both self-affirmation and group-affirmation on evaluations of one’s own political party (ingroup) and the opposing party (outgroup). Participants in this study performed a self-affirmation, group-affirmation, or control task and then rated members of their political party and the opposing party on positive and negative traits. Because of the contentious nature of the political domain¹ and the inherent threat that results from rating outgroup members in such a domain (McQueen & Klein, 2006), we expected that, by default, participants would demonstrate an evaluative bias favoring their own political party. That is, we predicted that participants in a control condition would evaluate the ingroup more positively than the outgroup. However, consistent with previous research, we predicted that participants who affirmed a personal self-aspect would express this bias to a lesser degree. Most importantly, we hypothesized that group-affirmation would not lessen intergroup bias in this context. Indeed, we predicted that affirming one’s political party would exacerbate the propensity to evaluate the ingroup more favorably than the outgroup.

Method

Participants and design

A total of 170 undergraduates in an introductory psychology course received partial course credit for participating in this study. The sample consisted of 109 women and 61 men aged 18 to 29 years (M = 18.79, SD = 1.21). Participants were randomly assigned to the self-affirmation, group-affirmation, or no-affirmation (control) condition.

Procedure and measures

Participants completed an online questionnaire containing three sections. The first section consisted of demographic questions and a set of political party affiliation questions. The second section consisted of the affirmation manipulation, and the third section consisted of the political attitude measures.

Political party affiliation

Participants first indicated whether they were a Democrat, a Republican, or an Independent. If participants indicated that they were Independents, they were asked if they saw themselves as closer to the Republican or Democratic Party. This selection of political party affiliation determined which party was the ingroup and which party was the outgroup for each participant. Overall, there were a total of 43 Republicans, 82 Democrats, and 45 Independents. When Independents were collapsed across the party in which they felt closest to, the final sample resulted in 58 Republicans and 112 Democrats.²

Affirmation manipulation

Participants in the self- and group-affirmation conditions completed the standard values-affirmation manipulation adapted from previous research (Cohen, Aronson, & Steele, 2000; Fein & Spencer, 1997; Gunn & Wilson, 2011). Participants were presented with 15 values (e.g., “Art/music,” “Social life,” and “Concern for others”). Participants in the self-affirmation condition were asked to select the value that they thought was most important to them as individuals and to write a paragraph explaining how that value pertained to them and why they thought it was so important. Participants in the group-affirmation condition were asked to select the value that they thought was most important to their political party as a group. They were then instructed to write a paragraph explaining how that value pertained to their political party and why members of that party thought it was important. Participants in the no-affirmation (control) condition were asked to list everything they had eaten or drank in the previous 48 hours. The no-affirmation instructions were taken from Cohen et al. (2000), who argue that participants can turn any values-based task (such as writing about an unimportant value) into a self-affirming task and, thus, something completely irrelevant and arbitrary should be used as a control.

Dependent measures

Following the affirmation manipulation, participants were presented with a scale on which they indicated how well they thought various positive traits (e.g., “Honest,” “Ethical”) and negative traits (e.g., “Ignorant,” “Immoral”) characterized members of both their own political party and members of the opposing political party. These traits and rating procedures were selected primarily from Chambers and Melnyk (2006) and the order in which the groups were evaluated was counterbalanced. Ratings were made on a scale ranging from 1 (not at all characteristic) to 7 (extremely characteristic). Participants’ responses for the positive traits and negative traits were averaged for each target group producing four trait-related outcome variables: positive ingroup (α = .83), negative ingroup (α = .69), positive outgroup (α = .83), and negative outgroup (α = .67).

Results

We conducted a 2 (group: ingroup, outgroup) × 2 (valence: positive, negative) × 3 (affirmation: self, group, control) mixed ANOVA on trait ratings, with group and valence as within-subjects factors. The three-way interaction was significant, F(2, 167) = 3.62, p = .029, $η_{p}^{2}$ = .04. We examined the patterns contributing to this interaction separately by trait valence.

For the positive traits, there was a significant group main effect, F(1, 167) = 103.17, p < .001, $η_{p}^{2}$ = .38. Overall, participants rated members of their own political party (M = 5.00, SD = 0.80) more positively than members of the opposing party (M = 4.08 SD = 0.91). There was not a significant main effect for affirmation condition, F(2, 167) < 1, nor was there a Group × Affirmation interaction, F(2, 167) = 2.10, p = .126.

For the negative traits (Figure 1), the group main effect was significant, F(1, 167) = 45.03, p < .001, $η_{p}^{2}$ = .21. Overall, participants rated members of their own political party (M = 3.57, SD = 0.95) less negatively than members of the opposing party (M = 4.30, SD = 1.03). There was not a significant main effect for affirmation condition, F(2, 167) = 2.24, p = .110. However, there was a significant Group × Affirmation interaction, F(2, 167) = 3.20, p = .043, $η_{p}^{2}$ = .04.

Figure 1.

Effect of affirmation condition on negative attitudes toward members of participants’ own political party (ingroup) and opposing political party (outgroup; Experiment 1).

Contrast analyses revealed that this interactive pattern for negative traits was driven by a linear contrast across affirmation condition for outgroup evaluations, F(1, 167) = 6.87, p = .01. Participants in the group-affirmation condition (M = 4.57, SD = 1.00) rated the outgroup most negatively, followed by the participants in the control condition (M = 4.25, SD = 0.95), and then those in the self-affirmation condition (M = 4.07, SD = 1.10).

In addition, simple effects tests indicated that the difference between ingroup and outgroup ratings on the negative traits was greatest for those who affirmed their political party. Specifically, the discrepancy between ingroup and outgroup ratings was significantly greater for those who affirmed the group (M_diff = 1.11, SD_diff = 1.35) compared with those who affirmed the self (M_diff = 0.58, SD_diff = 1.23), t(112) = 2.17, p = .032, 95% confidence interval (CI) = [0.05, 1.00], or for those in the control condition (M_diff = 0.49, SD_diff = 1.61), t(111) = 2.21, p = .029, 95% CI = [0.06, 1.17].

Discussion

As hypothesized, affirming the self and affirming the group had contrasting effects on the tendency to exhibit group-based evaluative biases. These findings provide support for the notion that, unlike affirming a personal aspect of the self, affirming a group identity in the face of threat can actually increase biased judgments relevant to that group. Importantly, the focus on political parties provided a naturally threatening context and facilitated the examination of bias at the intergroup level. It is noteworthy that, in this context, group-affirmation elevated intergroup bias specifically when examining negative trait evaluations. That is, the outgroup was evaluated more negatively than the ingroup to a significantly greater degree in the group-affirmation condition than in either the self-affirmation or control conditions (although ingroup bias was observed to some extent in all conditions).

Experiment 2

The findings from the previous experiment can readily be interpreted based on the multiple self-aspects framework (McConnell, 2011) and SCT (Turner et al., 1987). The group-affirmation task presumably activated the self-aspect associated with a specific social identity (political party), thus making associated constructs (e.g., political issues) and motivations (e.g., negative attitudes toward the opposing party) more salient. Experiment 2 was designed to test three additional predictions that emerge from this interpretation. First, we predicted that the effect of group-affirmation on bias would be most pronounced when that group is a particularly important aspect of the self. Consistent with this prediction, a number of studies have shown that group identification predicts greater levels of ingroup and intergroup bias (e.g., Aberson, Healy, & Romero, 2000). The procedure for determining political affiliation used in Experiment 1 did not assess the strength of party identification. Therefore, we included a continuous measure of party identification in Experiment 2. We hypothesized that the effect of political party-affirmation on bias would be strongest for those highly identified with their party.

Second, we predicted that, if a group-affirmation indeed activates a specific group-level aspect of the self, such a manipulation would result in an increased correspondence between perceptions of the self and of the group. One manifestation of this phenomenon would be greater overlap between attitudes attributed to the self and attitudes attributed to the ingroup. To explore this prediction, additional dependent measures were added in Experiment 2. Specifically, participants indicated how strongly they personally supported various controversial political issues and how strongly they thought the issues were supported by each political party. We hypothesized that participants who affirmed their political party (relative to those who did not) would show more overlap between the degree to which they personally support the issues and the degree to which they perceive their ingroup supports the issues.

Third, we predicted that these group-affirmation effects would be specific to the affirmed group. If one interprets the act of affirming the same group that has just been threatened as a domain-relevant affirmation (Sivanathan et al., 2008), then it should be the case that affirming one particular group-level identity would not activate self-aspects associated with other, unrelated groups. Therefore, an affirmation of one group identity should not influence biases related to a different group. In Experiment 2, participants affirmed their political party, affirmed their university identity, or completed a control task. We hypothesized that affirming one’s political party would increase evaluative bias favoring the ingroup party (replicating Experiment 1) and lead to greater self–ingroup overlap in perceived issue support. In contrast, affirming one’s university should not enhance group-related biases in the political domain.

Method

Participants and design

Participants were 178 undergraduates in an introductory psychology course who received partial course credit for participating. The sample consisted of 89 women and 89 men aged 18 to 30 years (M = 18.93, SD = 1.33). Participants were randomly assigned to the political party-affirmation, university-affirmation, or no-affirmation condition.

Procedure and measures

Participants completed an online questionnaire containing three sections. The first section contained demographic questions and a set of political party identification questions. The second section consisted of the affirmation manipulation, and the third section consisted of the political attitude measures.

Political party identification

As in Experiment 1, participants indicated whether they were a Republican, Democrat, or Independent. If they indicated that they were Independent, they were asked whether they saw themselves as closer to the Republican or Democratic Party. Overall, the sample consisted of 42 Republicans, 69 Democrats, and 67 Independents. When Independents were collapsed across the party with which they felt closest to, the final sample resulted in 65 Republicans and 113 Democrats.

In addition, as a measure of party identification, participants were asked to indicate how strongly they identified with both the Republican and Democratic parties on a 7-point scale ranging from 1 (not at all) to 7 (extremely).

Affirmation manipulations

The procedure was identical to Experiment 1 for the political-party-affirmation and control conditions. Participants in the university-affirmation condition were asked to select the value that they thought was most important to students at their university and to write a paragraph explaining how this value pertained to those students and why it was important to them.

Dependent measures

Participants next completed a series of attitudinal measures. They evaluated each political party on the same traits used in Experiment 1. Participants’ responses for the positive traits and the negative traits were averaged for each group producing four outcome variables: positive ingroup (α = .87), negative ingroup (α = .80), positive outgroup (α = .89), and negative outgroup (α = .80).

Participants also were presented with a list of 12 politically sensitive issues/policies (e.g., “Same-sex marriage,” “Obamacare,” and “Strict gun control laws”) and asked to indicate separately how strongly they personally, the Republican Party, and the Democratic Party supported each policy. Responses were recorded on a scale from 1 (totally oppose) to 7 (totally support). The order of all dependent measures was counterbalanced.

Results

Trait ratings

Using a general linear model (GLM) procedure, we examined the effects of affirmation condition, political party identification, and the Affirmation × Identification interaction on participants’ attitudes toward members of their political party (the ingroup) and the opposing political party (the outgroup). We treated target group (ingroup, outgroup) as a within-subjects factor, and affirmation condition (no-affirmation, university-affirmation, party-affirmation) and party identification (centered continuous) as between-subjects predictors.

For the positive traits, only the group main effect was significant, F(1, 175) = 103.75, p < .001, $η_{p}^{2}$ = .37. Participants rated members of their party more positively (M = 4.78, SD = 0.92) than the opposing party (M = 3.78, SD = 1.06). Importantly, this ingroup preference did not differ as a function of affirmation condition and party identification: three-way interaction, F(2, 172) < 1.

For the negative traits, the group main effect was significant, F(1, 175) = 72.50, p < .001, $η_{p}^{2}$ = .29, such that participants rated members of the opposing party more negatively (M = 4.25, SD = 1.03) than their own party (M = 3.41, SD = 0.96). Importantly, there also was a significant Group × Affirmation × Identification interaction, F(2, 172) = 3.76, p = .025, $η_{p}^{2}$ = .04.

To specify the precise patterns contributing to the overall interaction (Figure 2), we used a series of regression analyses to examine the effects of party identification (a continuous variable) and target group (the within-subjects factor) on negative trait ratings within each affirmation condition. In the no-affirmation condition (Figure 2a), only the main effect for group was significant, such that the outgroup was rated more negatively than the ingroup, b = .765, SE = .163, t(172) = 4.70, p < .001, 95% CI = [0.44, 1.09]. This effect was not moderated by party identification, b = .115, SE = .122, t(172) = .95, p = .34.

Figure 2.

Party identification predicting negative trait ratings of the ingroup and the outgroup in the (a) control condition, (b) university-affirmation condition, and (c) political party-affirmation condition (Experiment 2).

Likewise, in the university-affirmation condition (Figure 2b), only the main effect for group was significant, such that the outgroup was rated more negatively than the ingroup, b = .743, SE = .166, t(172) = 4.48, p < .001, 95% CI = [0.42, 1.07]. This effect was not moderated by party identification, b = .144, SE = .107, t(172) = 1.34, p = .18.

In the party-affirmation condition (Figure 2c), again there was a significant main effect for group, such that the outgroup was rated more negatively than the ingroup, b = .969, SE = .173, t(172) = 5.59, p < .001 95% CI = [0.63, 1.31]. This effect, however, was moderated by party identification, b = .612, SE = .156, t(172) = 3.92, p < .001, 95% CI = [0.30, 0.92]. As shown in Figure 2c, party identification was a significant predictor of negative trait ratings for both the ingroup, b = −.272, SE = .117, t(172) = 2.33, p = .021, 95% CI = [−0.50, −0.04], and outgroup, b = .340, SE = .122, t(172) = 2.78, p = .006, 95% CI = [0.10, 0.58]. Thus, only for participants who affirmed their political party, the more they identified with that party the less negatively they rated members of their party and the more negatively they rated members of the opposing party.

Self-party issue overlap

To test the prediction that group-affirmation would increase overlap between personal attitudes and ingroup attitudes, we examined patterns of personal, perceived ingroup, and perceived outgroup support for the controversial political issues. For each participant, we computed two within-person correlations: one for the correlation between ratings of self-support and perceived ingroup support across the 12 issues (self–ingroup overlap) and a second for the correlation between ratings of self-support and perceived outgroup support (self–outgroup overlap).³

Using a GLM procedure, we examined the effects of affirmation condition, party identification, and the Affirmation × identification interaction on each of the two within-person correlations (self–ingroup overlap and self–outgroup overlap). We treated group (ingroup, outgroup) as a within-subjects factor and affirmation condition (no-affirmation, university-affirmation, party-affirmation) and party identification (centered continuous) as between-subjects factors.

Overall, there was a significant group main effect, F(1, 175) = 59.61, p < .001, $η_{p}^{2}$ = .25, such that participants exhibited greater self–ingroup overlap (M = .33, SD = .38) than self–outgroup overlap (M = −.043, SD = .419). In addition, a significant Group × Affirmation × Identification interaction was found, F(2, 172) = 3.95, p = .021, $η_{p}^{2}$ = .05.

To specify the precise patterns contributing to the three-way interaction (Figure 3), we examined the group and identification effects within each affirmation condition. In the no-affirmation condition (Figure 3a), there was a significant group main effect, such that participants exhibited more perceived self–ingroup overlap than self–outgroup overlap, b = .385, SE = .077, t(172) = 4.99, p < .001, 95% CI = [0.23, 0.54]. This effect was not moderated by party identification, b = .046, SE = .058, t(172) = .80, p = .43.

Figure 3.

Party identification predicting self-party issue overlap for the ingroup and the outgroup in the (a) control condition, (b) university-affirmation condition, and (c) political party-affirmation condition (Experiment 2).

In the university-affirmation condition (Figure 3b), there was also a significant group main effect, such that participants showed more perceived self–ingroup overlap than self–outgroup overlap, b = .350, SE = .079, t(172) = 4.45, p < .001, 95% CI = [0.20, 0.51]. Surprisingly, this effect was moderated by party identification, b = .156, SE = .051, t(172) = 3.07, p = .003, 95% CI = [0.06, 0.26]. As shown in Figure 3b, identification was a significant predictor of both self–ingroup overlap, b = .064, SE = .030, t(172) = 2.11, p = .04, 95% CI = [0.004, 0.12] and self–outgroup overlap, b = −.092, SE = .034, t(172) = −2.67, p = .008, 95% CI = [−0.16, −0.02].

A similar, but significantly stronger, pattern was found in the party-affirmation condition (Figure 3c). There was a significant group main effect, such that participants showed more perceived self–ingroup overlap than self–outgroup overlap, b = .377, SE = .082, t(172) = 4.59, p < .001, 95% CI = [0.22, 0.54]. As predicted, this effect was moderated by party identification, b = .309, SE = .074, t(172) = 4.18, p < .001, 95% CI = [0.16, 0.46]. As shown in Figure 3c, identification was a significant predictor of both self–ingroup overlap, b = .189, SE = .044, t(172) = 4.29, p < .001, 95% CI = [0.10, 0.28], and self–outgroup overlap, b = −.120, SE = .050, t(172) = −2.39, p = .02, 95% CI = [−0.22, −0.02].

Although a similar pattern was found in both the university-affirmation and party-affirmation conditions (Figures 3b and 3c, respectively), the positive relationship between party identification and self–ingroup overlap was significantly stronger in the party-affirmation condition (Figure 3c, dashed line) than in the university-affirmation condition (Figure 3b, dashed line), b = .125, SE = .053, t(172) = −2.35, p = .018. The strength of the relationship between party identification and self–outgroup overlap in these two conditions (solid lines), however, was not significantly different, b = .028, SE = .061, t(172) = .46, p = .65.

Discussion

Experiment 2 provides further support for the notion that affirming a group identity can lead to greater group-serving bias, while extending the findings from the first experiment in three main ways. First, higher levels of group identification were associated with greater levels of group-serving evaluative bias in this experiment. Specifically, following a political-party affirmation, the more identified a participant was with that ingroup, the less negatively he or she evaluated the ingroup party and the more negatively he or she evaluated the outgroup. This finding is consistent with the prediction that the effect of group-affirmation on bias should be stronger when that group is a particularly important aspect of the self. Second, group-affirmation also predicted greater overlap between a participant’s personal attitudes toward controversial political issues and the attitudes he or she attributed to the ingroup party. Likewise, group-affirmation predicted less overlap between personal attitudes and the attitudes attributed to the outgroup. As with evaluative intergroup bias, the tendency to perceive greater self–ingroup overlap (and less self–outgroup overlap) was most pronounced for participants who were highly identified with their political party. This finding is consistent with the notion that a given group-affirmation activates that particular group-level aspect of the self, yielding a greater psychological correspondence between the personal self and that group identity. Third, affirming one’s political party led to greater bias specific to that group identity, whereas affirming one’s university identity did not (evaluative intergroup bias) or did so to a significantly lesser degree (self–party overlap). This pattern is consistent with the prediction that the bias-enhancing effects of group-affirmation are largely restricted to the affirmed group.

Experiment 3

Experiment 3 was designed to further address the processes underlying group-affirmation. Specifically, we sought to test directly our assumption that affirming a group identity would activate concepts associated with that particular group membership. The self–party issue overlap findings from Experiment 2 provided initial evidence for this process. However, we sought to demonstrate at the most direct and basic level that affirming a group identity makes ideas associated with that group more salient. In comparison, we sought to show that affirming the self activates different concepts than does affirming the group. Based on the notion that self-affirmation directs attention to positive aspects of one’s global (individual-level) self-concept (Critcher & Dunning, 2015) and in accordance with research conducted by Crocker et al. (2008) on the positive feelings associated with self-affirmation, we anticipated that affirming the self (as opposed to affirming the group) would make positive constructs/feelings more accessible. We tested these hypotheses in Experiment 3 using a word accessibility task. Specifically, we predicted that affirming one’s political party would lead to greater accessibility of politically relevant words/constructs, whereas affirming the self would lead to greater accessibility of positive constructs not necessarily related to politics.

Method

Participants and design

A total of 263 U.S. citizens were recruited using Amazon’s Mechanical Turk, and they received US$0.20 for participating in this study. The sample consisted of 175 women and 88 men aged 18 to 79 years (M = 35.76, SD = 13.24). Participants were randomly assigned to the self-affirmation, group-affirmation, or no-affirmation (control) condition.

Procedure and measures

Participants completed an online questionnaire containing two sections. The first section consisted of the affirmation manipulation and the second section consisted of the word accessibility task.

Affirmation manipulation

The affirmation manipulation was identical to that used in Experiment 1.

Word accessibility task

Following the affirmation manipulation, participants engaged in a word accessibility task. Participants were told that a list of 55 words would be flashed on the screen. The list contained five categories of words: positive traits (e.g., “Honest,” “Trustworthy,” “Considerate”), positive general words (e.g., “Love,” “Laughter,” “Joy”), negative traits (e.g., “Ignorant,” “Immoral,” “Stubborn”), negative general words (e.g., “Rape,” “Cancer,” “Pain”), and politically relevant words (e.g., “Obama,” “Abortion,” “Congress”). The positive and negative traits were identical to the trait ratings used in Experiments 1 and 2. The politically relevant words were taken from research demonstrating that people tend to associate those words with politics and political identities (Burdein, Lodge, & Taber, 2006; Lodge & Taber, 2005). The words were organized into seven columns, and their placement in those columns was randomly assigned to each participant. Participants were instructed to “Please scan the following list of words and, as quickly as possible, select the five words that stand out to you the most.”

Results

A GLM procedure was used to test the effect of affirmation condition on the accessibility of various word categories. It was found that for negative general words, there was no significant difference in accessibility across affirmation conditions, F(2, 260) = 1.85, p = .16. For negative traits, however, there was a significant difference across conditions, F(2, 260) = 4.61, p = .011, η² = .034. Simple effects tests revealed that participants in the control condition (M = 0.35, SD = 0.60) selected significantly more negative traits than those in both the self-affirmation condition, M = 0.13, SD = 0.37; t(260) = 2.95, p = .003, 95% CI = [0.07, 0.37], and in the group-affirmation condition, M = 0.20, SD = 0.46; t(260) = 2.00, p = .046, 95% CI = [0.01, 0.30].

Most critical to our analyses, however, was that there were significant differences in accessibility across conditions for both positive and politically relevant words. For positive words, this held for positive words in general, F(2, 260) = 8.13, p < .001, η² = .06, and for positive traits, F(2, 260) = 4.10, p = .018, η² = .03). Participants in the self-affirmation condition selected significantly more positive general words (M = 1.98, SD = 1.20) than those in the group-affirmation condition, M = 1.33, SD = 1.19; t(260) = 3.60, p < .001, 95% CI = [0.30, 1.03], and the control condition, M = 1.36, SD = 1.18; t(260) = 3.47, p = .001, 95% CI = [0.27, 0.99]. Similarly, self-affirmed participants selected significantly more positive traits (M = 1.66, SD = 1.02) than those who affirmed the group, M = 1.22, SD = 1.05; t(260) = 2.57, p = .011, 95% CI = [0.12, 0.75]; or did not affirm at all, M = 1.25, SD = 1.18; t(260) = 2.44, p = .015, 95% CI = [0.07, 0.76].

In contrast, the significant overall effect for politically relevant words, F(2, 260) = 16.52, p < .001, η² = .11, was driven by higher values in the group-affirmation condition. Participants in the group-affirmation condition selected significantly more politically relevant words (M = 1.91, SD = 1.33) than those in the self-affirmation condition, M = 0.86, SD = 0.93; t(260) = 5.71, p < .001, 95% CI = [0.70, 1.40], and control condition, M = 1.50, SD = 1.23; t(260) = 2.31, p = .022, 95% CI = [0.03, 0.78].

Discussion

Experiment 3 provides some additional insight into the processes underlying group-affirmation. As we had anticipated, participants who affirmed their political party demonstrated greater accessibility of politically relevant concepts than those who performed self-affirmation or control tasks. Also, consistent with our reasoning, self-affirmation increased the accessibility of positive concepts and traits. It should be noted, however, that this experiment does not provide a direct test of group-relevant self-aspect activation following affirmation. Although these data demonstrate that political party-affirmation increases the accessibility of words generally associated with political identities (Burdein et al., 2006; Lodge & Taber, 2005), it cannot be determined whether affirming the group also made salient the idiosyncratic constructs and motivations that made up participants’ individual political party self-aspects.

General Discussion

It is well established that self-affirmation reduces a host of self-serving and group-serving biases. Typically, self-affirmation tasks make reference to personal aspects of the self (for a review, see Sherman & Cohen, 2002). An emerging question is whether similar psychological processes occur when affirming a group-level aspect of the self. Although some intriguing findings relevant to this topic have been reported (e.g., Cehajic-Clancy et al., 2011; Gunn & Wilson, 2011; Sherman et al., 2007), not all of these studies specifically were designed to address this issue. Moreover, the findings from these studies could lead to varying conclusions. The present research was designed to address systematically the effects of group-affirmation on group-serving bias to bring some initial clarity to this issue.

The central question addressed in the current research is whether affirming one’s group identity has similar effects on group-serving biases as does affirming one’s personal identity. We proposed that group-affirmation activates the self-aspect associated with a given group membership and, by doing so, may actually enhance or exacerbate biases associated with that group identity. Support for this guiding hypothesis was found across three experiments.

In Experiment 1, we found that participants tended to evaluate members of the opposing political party more negatively than they evaluated members of their own party. This group-serving evaluative bias was magnified when participants first affirmed a value endorsed by the group (i.e., group-affirmation) compared with when they endorsed a personally held value (i.e., self-affirmation). Experiments 2 and 3 followed up on these findings and explored more thoroughly the mechanisms involved in group-affirmation. Experiment 2 provided evidence for the proposed link between group-affirmation and self-aspect activation. For participants who affirmed their political party, the more they identified with that party, the more negatively they rated members of the opposing party and the more overlap there was between their personal and perceived ingroup support for controversial political policies. For participants who affirmed a group unrelated to politics (university affiliation), however, stronger party identification produced no increase in negative outgroup attitudes and was only a modest predictor of self–party overlap. Finally, in Experiment 3 we found that, whereas self-affirmation induced a focus on positive words and concepts in general, group-affirmation induced a focus on concepts specific to that affirmed group identity.

Theoretical Implications

Findings from the present research can be used to consider conceptual ideas stemming from SCT (Turner et al., 1987; Turner, Oakes, Haslam, & McGarty, 1994) and the multiple self-aspects framework (McConnell, 2011) in the context of psychological processes proposed by affirmation theory (Steele, 1988). The current and existing research establishes that, when focusing on the personal self, an affirmation task reduces group-serving biases. Such a self-affirmation effect is a fundamental tenet of affirmation theory (Sherman & Cohen, 2002). We consider this pattern also to be consistent with SCT and the multiple self-aspects framework because it implies that focusing on one level of the self renders judgments and biases related to other specific self-aspects or identities less psychologically relevant.

We believe that this same process can account for the contrasting effect that group-affirmation had on group-serving biases in the current set of experiments. By inducing a focus on a specific group-level aspect of the self, a group-affirmation presumably activates the psychological attributes (e.g., attitudes, stereotypes, norms, behavioral tendencies, intergroup comparisons) pertinent to that group membership. Once activated, this group-specific mind-set guides subsequent thoughts, judgments, motivations, and behaviors. A by-product of this mind-set would be an increase in biases that serve to establish, protect, and enhance the worth and integrity of that group. In short, a self-affirmation allows for a more objective orientation toward a particular group membership by leading people to categorize themselves at the individual level and, thus, highlighting aspects of the broader self unrelated to the group. A group-affirmation, in contrast, increases group-relevant subjectivity by leading people to recategorize themselves specifically at that group level and, thus, reinforces any biases inherent to that group membership. This logic is also consistent with research exploring the contrasting effects of affirmations in domains relevant and irrelevant to threat (Blanton et al., 1997; Sivanathan et al., 2008). Indeed, we found in Experiment 2 that affirming a group identity unrelated to politics did not have the same bias-enhancing effect as did affirming one’s political party.

We believe that our findings and theoretical reasoning help clear up some of the confusion in the literature regarding group-affirmation and bias. For instance, the Israeli students in Cehajic-Clancy et al.’s (2011) study who affirmed their nationality may have shown no increase in collective guilt or responsibility because the biases and motives associated with their self-aspects as “Israelis” had become more salient (presuming that, given the historical and ongoing intergroup tension, Israelis and Palestinians are not “supposed” to feel guilty or responsible for the suffering of the other group). On the contrary, Gunn and Wilson (2011) may have found an increase in feelings of collective guilt and shame for an ingroup’s historical mistreatment of a particular outgroup (men’s mistreatment of women and Canadians’ mistreatment of aboriginal tribes) after group-affirmation because that type of behavior is no longer considered socially acceptable. Thus, one’s “male” and “Canadian” self-aspects may contain the sentiment that mistreatment of women and aboriginals, respectively, is wrong and that guilt and shame are the proper emotions to exhibit toward such transgressions. Affirming these group identities presumably would make such feelings particularly salient, intense, and influential. Thus, this interpretation could also be used to predict decreases in bias in some group contexts, provided those unbiased feelings are associated with that particular social identity.

There were also a couple of unanticipated findings in these studies that are worth further consideration because of their theoretical relevance. The first is the observation from Experiment 2 that participants who affirmed their university showed a similar pattern with respect to self–ingroup overlap as the participants who affirmed their political party (although this pattern was significantly stronger for those who affirmed their university). Thus, a university-affirmation induced participants who were highly identified with their political party to exhibit greater issue overlap with their party, but it did not affect their intergroup trait evaluations. This result may be attributable to the notion that one’s political identity and university identity are not mutually exclusive. In addition to being a component of one’s political self-aspect, for example, one’s stance on such issues as abortion or gay rights may be a part of one’s university self-aspect as well. So, although there may not be an explicit, direct connection between one’s political and university identities, such that making one’s university affiliation salient would make one view the opposing political party more negatively, there may be an indirect relationship between the two self-aspects, such that both are associated with certain politically relevant issues.

Another result that warrants further discussion is the lack of an affirmation effect on positive trait ratings in Experiments 1 and 2. In Experiment 2, for example, although there was a significant three-way interaction between group type, affirmation condition, and party identification for negative trait ratings, there was only a significant group type main effect for the positive ratings (such that participants rated members of their own party more positively than members of the opposing party). This finding is somewhat surprising given that much of the intergroup bias literature emphasizes tendencies toward ingroup positivity and favoritism (Brewer, 1979). Nevertheless, people have also been shown to explicitly hold and express negative attitudes toward members of outgroups (Tajfel, 1982). It is possible that the current findings were restricted to negative ratings due to the nature of the groups being studied. The American political arena is characterized by disagreements between two opposing parties over fundamental values, and this frequently leads to the expression of negative and vitriolic feelings toward members of the opposite party (Ware, 2011). Thus, negative attitudes may become especially salient if one’s political identity is activated following a group-affirmation task.

Future Directions

Our experiments address a relatively untouched and inconsistent area of research, and the current findings open the door to many interesting questions and directions for future investigation. For instance, future research should explicitly assess the link between group-affirmation, self-aspect activation, and bias. Our results lend support to the notion that affirming the group activates that group’s self-aspect, thus leading to the biased tendencies associated with that self-aspect. However, our evidence for this process is indirect in nature (increased overlap between individual and ingroup issue support and greater salience of constructs relevant to that identity). It would be useful to test whether group-affirmation directly activates the idiosyncratic attitudes and biases that individuals associate with their group identities and whether it is this activation that leads to greater bias.

Future research should also assess the boundary conditions of the affirmation effects observed in our experiments. In the context of political affiliation, we found that affirming the group can lead to a greater tendency to engage in biases favoring that group in the face of threat. The political domain is unique, however, in that it is a highly competitive, and sometimes even hostile, intergroup context. Is the bias-enhancing effect of group-affirmation limited to contexts in which there are clearly defined and opposing groups? Also, would a similar pattern of results be found if these studies were conducted in the context of a group where members do not have a choice in membership, such as ethnicity? Would the results hold in a minimal-groups setting, or does group-affirmation only have an effect on bias if the groups are well established?

In addition, are there contexts in which a group-affirmation manipulation would actually “work”? In other words, are there situations in which the collective self would overlap with the individual self to the extent that reflecting on one’s collective identity would provide a boost to one’s overall sense of self-worth and result in less of a need to exhibit intergroup bias? Perhaps group-affirmation would be more effective if a different technique was used. Although we employed the traditional and commonly used values-affirmation manipulation, a manipulation in which participants are provided false positive feedback about their group, for example, may provide a self-integrity boost without necessarily making attitudes and biases associated with that group psychologically predominant. In particular, such a manipulation might increase positive feelings about the ingroup without necessitating increased intergroup bias.

Finally, future research should explore the effects of group-affirmation on people’s tendencies to exhibit biases at other levels of identity. It has been demonstrated that self-affirmation can reduce biases at both the individual level (Cohen et al., 2000) and the collective level (Sherman & Kim, 2005). In the current research, we show that group-affirmation does not reduce biases at the collective level (and, in fact, can even exacerbate them), but does this effect translate to the individual level? In other words, would affirming a collective level self-aspect reduce personal self-relevant biases?

Conclusion

The current research suggests that affirming the group does not have the same psychological consequences as does affirming the self. Whereas self-affirmation buffers against threats to one’s social identity and reduces group-favoring biases, affirmations of the groups themselves appear to maintain or magnify these biases. The current findings document that there are contexts wherein affirming one’s group identity can lead to a greater willingness to defend the ingroup and exhibit biases toward outgroups. These findings reinforce the fundamental distinction between the personal, individual-level self and the collective, group-level self.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

The online supplemental material is available at .

Notes

References

Aberson

C. L.

Healy

Romero

(2000). Ingroup bias and self-esteem: A meta-analysis. Personality and Social Psychology Review, 4, 157-173.

Adams

Tormala

T. T.

O’Brien

L. T.

(2006). The effect of self-affirmation on perceptions of racism. Journal of Experimental Social Psychology, 42, 616-626.

Anderson

C. A.

Deuser

W. E.

DeNeve

K. M.

(1995). Hot temperatures, hostile affect, hostile cognition, and arousal: Tests of a general model of affective aggression. Personality and Social Psychology Bulletin, 21, 434-448.

Binning

K. R.

Sherman

D. K.

Cohen

G. L.

Heitland

(2010). Seeing the other side: Reducing political partisanship via self-affirmation in the 2008 presidential election. Analyses of Social Issues and Public Policy, 10, 276-292.

Blanton

Cooper

Skurnik

Aronson

(1997). When bad things happen to good feedback: Exacerbating the need for self-justification with self-affirmation. Personality and Social Psychology Bulletin, 23, 684-692.

Blascovich

Mendes

W. B.

Hunter

S. B.

Lickel

Kowai-Bell

(2001). Perceiver threat in social interactions with stigmatized others. Journal of Personality and Social Psychology, 80, 253-267.

Brewer

M. B.

(1979). Ingroup bias in the minimal intergroup situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307-324.

Burdein

Lodge

Taber

(2006). Experiments on the automaticity of political beliefs and attitudes. Political Psychology, 27, 359-371.

Cehajic-Clancy

Effron

D. A.

Halperin

Liberman

Ross

L. D.

(2011). Affirmation, acknowledgment of in-group responsibility, group-based guilt, and support for reparative measures. Journal of Personality and Social Psychology, 101, 256-270.

10.

Chambers

J. R.

Melnyk

(2006). Why do I hate thee? Conflict misperceptions and intergroup mistrust. Personality and Social Psychology Bulletin, 32, 1295-1311.

11.

Cohen

G. L.

Aronson

Steele

C. M.

(2000). When beliefs yield to evidence: Reducing biased evaluation by affirming the self. Personality and Social Psychology Bulletin, 26, 1151-1164.

12.

Critcher

C. R.

Dunning

(2015). Self-affirmations provide a broader perspective on self-threat. Personality and Social Psychology Bulletin, 41, 3-18.

13.

Crocker

Niiya

Mischkowski

(2008). Why does writing about important values reduce defensiveness? Self-affirmation and the role of positive other-directed feelings. Psychological Science, 19, 740-747.

14.

Derks

van Laar

Ellemers

(2009). Working for the self or working for the group: How self versus group affirmation affects collective behavior in low-status groups. Journal of Personality and Social Psychology, 96, 183-202.

15.

Dovidio

J. F.

Gaertner

S. L.

(2010). Intergroup bias. In Fisk

S. T.

Gilbert

D. T.

Lindzey

(Eds.), Handbook of social psychology (5th ed., pp. 1084-1121). Hoboken, NJ: John Wiley.

16.

Fein

Spencer

S. J.

(1997). Prejudice as self-image maintenance: Affirming the self through derogating others. Journal of Personality and Social Psychology, 73, 31-44.

17.

Glasford

Dovidio

Pratto

(2009). I continue to feel so good about us: In-group identification and the use of social identity-enhancing strategies to reduce intragroup dissonance. Personality and Social Psychology Bulletin, 35, 415-427.

18.

Gunn

G. R.

Wilson

A. E.

(2011). Acknowledging the skeletons in our closet: The effect of group affirmation on collective guilt, collective shame, and reparatory attitudes. Personality and Social Psychology Bulletin, 37, 1474-1487.

19.

Hewstone

Rubin

Willis

(2002). Intergroup bias. Annual Review of Psychology, 53, 575-604.

20.

Lodge

Taber

(2005). The automaticity of affect for political leaders, groups, and issues: An experimental test of the hot cognition hypothesis. Political Psychology, 26, 455-482.

21.

McConnell

A. R.

(2011). The multiple self-aspects framework: Self-concept representations and its implications. Personality and Social Psychology Review, 15, 3-27.

22.

McQueen

Klein

W. M. P.

(2006). Experimental manipulations of self-affirmation: A systematic review. Self and Identity, 5, 289-354.

23.

Nisbett

R. E.

Ross

L. D.

(1980). Human inference: Strategies and shortcomings of social judgment. Englewood Cliffs, NJ: Prentice Hall.

24.

Riek

B. M.

Mania

E. W.

Gaertner

S. L.

(2006). Intergroup threat and outgroup attitudes: A meta-analytic review. Personality and Social Psychology Review, 10, 336-353.

25.

Sherman

Cohen

G. L.

(2002). Accepting threatening information: Self-affirmation and the reduction of defensive biases. Current Directions in Psychological Science, 11, 119-123.

26.

Sherman

Kim

(2005). Is there an “I” in “team”? The role of the self in group-serving judgments. Journal of Personality and Social Psychology, 88, 108-120.

27.

Sherman

Kinias

Major

Kim

Prenovost

(2007). The group as a resource: Reducing biased attributions for group success and failure via group affirmation. Personality and Social Psychology Bulletin, 33, 1100-1112.

28.

Sivanathan

Molden

D. C.

Galinsky

A. D.

(2008). The promise and peril of self-affirmation in de-escalation of commitment. Organizational Behavior and Human Decision Processes, 107, 1-14.

29.

Spencer

S. J.

Fein

Lomore

C. D.

(2001). Maintaining one’s self-image vis-à-vis others: The role of self-affirmation in the social evaluation of the self. Motivation and Emotion, 25, 41-65.

30.

Steele

(1988). The psychology of self-affirmation: Sustaining the integrity of the self. In Berkowitz

(Ed.), Advances in experimental social psychology, Vol. 21: Social psychological studies of the self: Perspectives and programs (pp. 261-302). San Diego, CA: Academic Press.

31.

Tajfel

(1982). Social psychology of intergroup relations. Annual Review of Psychology, 33, 1-39.

32.

Tajfel

Turner

J. C.

(1986). The social identity theory of intergroup behavior. In Worchel

Austin

W. G.

(Eds.), Psychology of intergroup relations (pp. 7-24). Chicago, IL: Nelson-Hall.

33.

Tropp

L. R.

Wright

S. C.

(2001). Ingroup identification as inclusion of ingroup in the self. Personality and Social Psychology Bulletin, 27, 585-600.

34.

Turner

J. C.

Hogg

M. A.

Oakes

P. J.

Reicher

S. D.

Wetherell

M. S.

(1987). Rediscovering the social group: A self-categorization theory. New York, NY: Basil Blackwell.

35.

Turner

J. C.

Oakes

P. J.

Haslam

S. A.

McGarty

(1994). Self and collective: Cognition and social context. Personality and Social Psychology Bulletin, 20, 453-463.

36.

Ware

(2011). Political conflict in America. New York, NY: Palgrave MacMillan.

37.

Zarate

M. A.

Garza

A. A.

(2002). In-group distinctiveness and self-affirmation as dual components of prejudice reduction. Self and Identity, 1, 235-249.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.18 MB

0.19 MB