Abstract
Researchers who study perspective taking are generally optimistic about the potential for interventions to improve intergroup perceptions. The current research provides new insight into the conditions that frame the intergroup outcomes of perspective taking. The results show that the effects of perspective taking are not always positive but depend on perspective takers’ degree of identification with the in-group. In two experiments, we demonstrated that adopting the perspective of an out-group member can have damaging effects on intergroup perceptions among group members who are highly identified with the in-group. Specifically, compared with less committed members, those who identified highly with the in-group used a greater number of negative traits to describe the out-group following perspective taking. Such perspective taking also led participants with high in-group identification to judge the out-group less favorably. Understanding how social identity concerns frame the outcome of perspective taking is crucial to its effective employment in intergroup-relations programs.
A well-documented finding in the prejudice-reduction literature is that encouraging people to adopt the perspective of out-group members can lead to the development of more harmonious intergroup perceptions (e.g., Batson et al., 1997; Dovidio et al., 2004; Galinsky & Moskowitz, 2000; Vescio, Sechrist, & Paolucci, 2003). Reflecting its potential as a conflict-resolution tool, perspective taking has emerged as a regular feature of intergroup-relations programs such as diversity training (Pendry, Driscoll, & Field, 2007). However, such initiatives are rarely systematically evaluated for effectiveness (Paluck & Green, 2009). This observation, together with recent research showing that perspective taking can increase perspective takers’ self-serving behavior (Epley, Caruso, & Bazerman, 2006; Galinsky, Maddux, Gilin, & White, 2008; Trötschel, Hüffmeier, Loschelder, Schwartz, & Gollwitzer, 2011; see also Vorauer & Sasaki, 2009), indicates the risk of generalizing the effects of perspective taking. It also necessitates a more nuanced examination of when (or for whom) perspective taking might not be beneficial—and might even be damaging—to intergroup relations.
We sought unique insight into this important issue by testing the prediction that the outcome of perspective taking depends on perspective takers’ own relationship with their in-group. We examined the effects of perspective taking in group members who varied in their degree of psychological commitment to—or identification with—their in-group. We predicted that adopting the perspective of an out-group member would not be beneficial for intergroup perceptions of group members who were highly identified with their in-group. Rather, we predicted that such perspective taking would backfire and lead these group members to perceive the out-group more negatively.
In developing our predictions for the research, we drew on a rich literature on intergroup relations that has established social group membership as a lens through which people interact with the social world (Tajfel & Turner, 1986; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). A central tenet of the social identity approach is that group members are motivated to secure a positive and distinct image for their own group (see Turner, 1999). However, because of their strong emotional and cognitive ties to the in-group, group members who are highly identified with the in-group (high identifiers) are especially driven to behave in ways that serve this goal (Ellemers, Spears, & Doosje, 2002). Such group members do not automatically perceive out-groups in negative terms, but they do take strategic actions to defend their social identity when it is undermined: This can include derogating the out-group (Crisp, Stone, & Hall, 2006; Jetten, Spears, & Postmes, 2004).
Although previous research in this area has examined intergroup outcomes of perspective taking (e.g., out-group helping, prejudice reduction), that research has typically been conducted at the level of the individual and has not fully accounted for perspective takers’ own social identities (Iyer & Leach, 2008). The effects of perspective taking may not generalize to people who think about intergroup relations primarily as group members, rather than as individuals who just happen to have group memberships. Support for this suggestion comes from research showing that people are more likely to behave in ways that secure positive social identity when that identity is salient than when it is not salient (Dumont, Yzerbyt, Wigboldus, & Gordijn, 2003). Indeed, when a salient social identity is directly threatened (e.g., by being implicated in the harming of out-group members), perspective taking can reduce support for the out-group among in-group members most concerned about protecting the in-group (Mallett, Huntsinger, Sinclair, & Swim, 2008; Zebel, Doosje, & Spears, 2009).
In the current research, we examined the effects of perspective taking on intergroup perceptions in everyday group contexts (university group, nationality). We predicted that when social identity was salient, the effects of an instruction to adopt an out-group member’s perspective would be moderated by in-group identification. Group members who have low identification with their in-group (low identifiers) tend not to be particularly motivated by the need to secure a positive image for the in-group (Ellemers & Van Rijswijk, 1997). Thus, perspective taking might be expected to have generally positive consequences for these members’ perceptions of the out-group, that is, to reflect the pattern found in previous research (e.g., Batson et al., 1997; Galinsky & Moskowitz, 2000).
By contrast, we predicted that perspective taking would elicit a defensive reaction among high identifiers. Recent tests of prejudice-reduction models, such as the common in-group identity model (Gaertner, Dovidio, Anastasio, Bachman, & Rust, 1993), have shown that attempts to change group members’ intergroup perceptions are experienced by high identifiers as threatening to social identity and can lead them to form more negative attitudes toward out-groups (Crisp, Stone, & Hall, 2006; Hornsey & Hogg, 2000).We predicted that a similar reaction would be shown by high identifiers when they took the perspective of an out-group member. Specifically, we predicted that perspective taking would lead high identifiers to express a more negative evaluation of the out-group.
Experiment 1
Method
Before being admitted to the experiment, university students (N = 127; 41 males, 86 females; mean age = 20.45 years, SD = 4.00 years) were asked to confirm their university group membership (Keele University). The salience of this identity was further raised by informing participants that the study was about their university group’s perceptions of other social groups. Participants then reported their identification with this in-group using three items adapted from Ellemers, Kortekaas, and Ouwerkerk (1999; e.g., “I feel good about this university”). Participants rated the items on scales ranging from 1 (strongly disagree) to 7 (strongly agree; α = .81). Using the procedure of Galinsky and Moskowitz (2000), we then presented participants with a photograph of a student from an out-group university (University of Sussex) and gave them 2 min to write a short paragraph about a “day in the life” of this individual.
Participants in the perspective-taking condition were instructed to actively take the perspective of the target while writing their paragraph—“to go through a typical day in his shoes, as if you were him.” Participants in the control condition did not receive this instruction. The proportion of first-person pronouns (relative to the total number of words) in participants’ paragraphs enabled a check on the manipulation of perspective taking: We expected participants to employ more first-person pronouns in the perspective-taking condition than in the control condition (cf. Galinsky & Ku, 2004). A single item assessing the degree to which participants tried to take the perspective of the out-group target (Dovidio et al., 2004) served as a second manipulation check. Participants rated the item on a scale ranging from 1 (not at all) to 7 (extremely).
Participants were then presented with four positive traits (“organized,” “tolerant,” “wise,” and “practical”) and four negative traits (“irritable,” “boastful,” “indecisive,” and “moody”) and were asked to indicate which of those traits they believed could be applied to the out-group (see Crisp, Walsh, & Hewstone, 2006). Trait-attribution scores were calculated by separately summing the numbers of positive and negative traits that participants assigned to the out-group.
Results
One participant was removed from the analysis because he or she did not write anything during the perspective-taking task. Confirming the manipulation’s effectiveness, we found that, compared with participants in the control condition, participants in the perspective-taking condition used a higher proportion of first-person pronouns in their paragraphs (M = .08, SD = .06, vs. M = .01, SD = .02), F(1, 122) = 74.62, p < .01, η p 2 = .38, and reported a greater degree of perspective taking (M = 4.65, SD = 1.38, vs. M = 3.76, SD = 1.34), F(1, 121) = 12.64, p < .01, η p 2 = .10. In-group identification did not moderate these effects (ps > .51).
Moderated hierarchical regression analysis was used to test the study’s predictions. Perspective taking (dummy coded) and in-group identification (mean centered) were entered as predictors in the first step, and their interaction was entered in the second step (see Table 1). For attribution of negative traits to the out-group, a significant Perspective Taking × Identification interaction was observed. Simple-slopes analysis indicated that in the control condition, there was no relationship between in-group identification and trait attribution, β = −0.03, t(122) = −0.23, p = .82. However, as predicted, participants in the perspective-taking condition attributed significantly more negative traits to the out-group as in-group identification increased, β = 0.33, t(122) = 2.53, p < .05 (Fig. 1, upper panel). For positive-trait attribution, none of the effects achieved significance (ps > .32).
Effects of Perspective Taking and In-Group Identification on Out-Group Trait Attributions in Experiment 1
p < .05.
Mean number of negative traits attributed to the out-group as a function of in-group identification (1 SD below and 1 SD above the mean) and perspective-taking condition, in Experiment 1 (upper panel) and Experiment 2 (lower panel). Error bars show standard errors.
Experiment 2
Experiment 1 demonstrated that in a relatively innocuous, everyday intergroup context, perspective taking elicits a negative reaction among some group members. In-group identification did not predict participants’ trait attributions in the control condition. However, when they adopted the perspective of the out-group member, participants attributed more negative trait attributions to the out-group as in-group identification increased. In Experiment 2, we tested the robustness of this effect in a different intergroup context (nationality) and manipulated participants’ in-group identification in order to test the causal effect of this variable. We also included a broader set of out-group ratings.
Method
One hundred British university students participated in the experiment (23 males, 77 females; mean age = 18.81 years, SD = 1.74 years). As in Experiment 1, social identity was made salient by asking participants to confirm their in-group membership (in this case, nationality) and by informing them that the study concerned how British people perceive other nations. National identification was manipulated using Jetten, Spears, and Manstead’s (1997) procedure, which required participants to indicate whether they agreed or disagreed with statements about their in-group that were designed to make salient some positive or negative aspects of their identity. High identification was induced by presenting participants with moderately positive statements (e.g., “In general, I like being British”) and very negative statements (e.g., “British people have nothing at all in common with each other”). Conversely, low identification was induced by presenting moderately negative statements (e.g., “British weather is amongst the wettest in Europe”) and very positive statements (e.g., “In terms of how I define myself, being British is crucial to my identity”). Participants were expected to more readily endorse the moderate statements in each condition, which would lead them to infer that they had high or low identification, respectively, with the in-group (Jetten et al., 1997). The relative numbers of positive and negative statements participants agreed with in each condition constituted the manipulation check.
Perspective taking was manipulated by the method in Experiment 1, which was adapted for the national group context (the out-group target was German). In addition to completing the trait-attribution task from Experiment 1, participants reported how favorably or unfavorably they judged the out-group, on a scale ranging from 1 (not at all) to 10 (extremely; cf. Haddock, Zanna, & Esses, 1993).
Results
One participant did not write anything during the perspective-taking task and was therefore removed from the analysis. The manipulation check confirmed that the manipulation of group identification was successful: Participants in the high-identification condition agreed with more positive statements about the in-group (M = 4.34, SD = 0.82) than negative statements (M = 0.28, SD = 0.61), F(1, 97) = 396.00, p < .01, η p 2 = .80, and participants in the low-identification condition agreed with more negative statements (M = 4.27, SD = 1.25) than positive statements (M = 1.02, SD = 0.90), F(1, 97) = 247.90, p < .01, η p 2 = .72. The perspective-taking manipulation was also successful. Compared with participants in the control condition, participants in the perspective-taking condition used a higher proportion of first-person pronouns in their paragraphs (M = .10, SD = .06, vs. M = .01, SD = .03), F(1, 95) = 102.67, p < .01, η p 2 = .52, and self-reported a greater degree of perspective taking (M = 4.64, SD = 1.54, vs. M = 3.35, SD = 1.67), F(1, 95) = 15.83, p < .01, η p 2 = .14. These effects were not moderated by in-group identification (ps > .40).
Perspective taking and in-group identification were dummy coded and mean centered prior to entry into a moderated hierarchical regression model (see Table 2). Exactly replicating Experiment 1, Experiment 2 showed a significant interaction effect of perspective taking and identification on negative out-group trait attributions. In the control condition, participants in the high- and low-identification conditions did not differ in the number of negative traits they assigned to the out-group, β = −0.22, t(95) = −0.45, p = .55. However, as predicted, after perspective taking, participants in the high-identification condition assigned significantly more negative traits to the out-group than did participants in the low-identification condition, β = 0.93, t(95) = 2.55, p = .01 (Fig. 1, lower panel). None of the effects for the positive-trait attributions achieved significance (ps > .19).
Effects of Perspective Taking and In-Group Identification on Out-Group Trait Attributions and Out-Group Favorability in Experiment 2
p < .05.
Perspective taking and identification also had an interaction effect on out-group favorability ratings (Table 2): Participants in the high-identification condition expressed a significantly less favorable out-group judgment in the perspective-taking condition compared with the control condition, β = −1.70, t(95) = −2.67, p < .01. For participants in the low-identification condition, out-group favorability judgments did not differ between the perspective-taking and control conditions, β = 0.39, t(95) = 0.62, p = .54.
General Discussion
People’s responses to intergroup phenomena are shaped by social identity concerns (Tajfel & Turner, 1986), and group members who identify highly with their in-group are especially motivated to secure a positive social identity (Crisp, Stone, & Hall, 2006; Jetten et al., 2004). Although group members often employ identity-protection strategies in contexts in which their social identity is directly threatened (Miron, Branscombe, & Biernat, 2010; Tarrant, Branscombe, Warner, & Weston, 2012), our findings demonstrate that social identity structures the effects of perspective taking even in relatively innocuous, everyday intergroup contexts. Highly identified in-group members in the current research expressed more negative out-group perceptions after adopting the perspective of an out-group member, as predicted. Specifically, compared with low identifiers, high identifiers used a greater number of negative traits to describe the out-group following perspective taking. Moreover, in Experiment 2, perspective taking led high identifiers to judge the out-group less favorably.
Initiatives such as diversity training often encourage participants to adopt the perspective of out-group members. However, such initiatives themselves may appeal most to people who are already positively oriented toward changing their intergroup representations (Paluck & Green, 2009). Forcing people who are not so oriented to participate in such initiatives may undermine their effectiveness. Perhaps because they threaten social identity distinctiveness (Crisp, Stone, & Hall, 2006), interventions targeted at high group identifiers, who are especially sensitive to such threats, may actually backfire and lead to negative out-group perceptions. An implication of our research, then, is that perspective-taking initiatives might work best for group members who are not highly identified with the in-group and whose intergroup attitudes actually are least likely in need of modification.
Encouraging in-group members to reflect on an out-group’s perspective is an intuitively appealing way of promoting better intergroup relations, and various studies have provided empirical support for the effectiveness of this technique. The current research provides valuable insight into the conditions that frame the effects of perspective taking and highlights the importance of considering perspective takers’ own relationship with their in-group. For some group members, perspective taking has negative effects, and when it backfires in this way, there may be serious and damaging consequences for intergroup relations. A challenge for future research, therefore, is to develop ways in which the beneficial effects of perspective taking can be harnessed for all group members. Understanding the meaning of perspective taking for social identity is crucial to realizing this technique’s full potential.
Footnotes
Acknowledgements
We thank Jessica Salvatore for her helpful comments on an earlier version of this manuscript.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.
Funding
This research was supported by an Economic and Social Research Council (ESRC) grant (RES-000-22-2441) and by the National Institute for Health Research (NIHR). The views expressed in this publication are those of the authors and not necessarily those of the ESRC, NIHR, National Health Service, or Department of Health.