The Communication of “Pure” Group-Based Anger Reduces Tendencies Toward Intergroup Conflict Because It Increases Out-Group Empathy

Abstract

The communication of group-based anger in intergroup conflict is often associated with destructive conflict behavior. However, we show that communicating group-based anger toward the out-group can evoke empathy and thus reduce intergroup conflict. This is because it stresses the value of maintaining a positive long-term intergroup relationship, thereby increasing understanding for the situation (in contrast to the communication of the closely related emotion of contempt). Three experiments demonstrate that the communication of group-based anger indeed reduces destructive conflict intentions compared with (a) a control condition (Experiments 1-2), (b) the communication of group-based contempt (Experiment 2), and (c) the communication of a combination of group-based anger and contempt (Experiments 2-3). Moreover, results from all three experiments reveal that empathy mediated the positive effect of communicating “pure” group-based anger. We discuss the implications of these findings for the theory and practice of communicating emotions in intergroup conflicts.

Keywords

anger contempt communication intergroup conflict empathy

Extensive literature on anger management suggests that anger is a potentially destructive emotion that when expressed actively harms relationships (e.g., DiGuiseppe & Tafrate, 2003). The entrenched view is thus that conflicts are best resolved when anger is kept under wraps. Indeed, the advice you would least likely get in a conflict situation is to communicate anger toward the other party. Several recent studies, however, seem to suggest that it can actually have positive effects to do just that (e.g., Van Kleef & Côté, 2007; see also Halperin, Russel, Dweck, & Gross, 2011). An important distinction here is between experienced anger that motivates behavior and the communication of anger as an act in itself. Whereas the former may likely motivate destructive behavior in the angered person (but see Halperin et al., 2011), we argue that the latter can potentially de-escalate conflicts by motivating constructive conflict behavior in the recipient of the communicated anger.

We take a novel approach by focusing on the relational function of communicating anger and by applying this notion to the context of intergroup conflict. We argue that the communication of group-based anger (e.g., Smith, 1993; Smith, Seger, & Mackie, 2007) reduces tendencies toward intergroup conflict because it stresses the value of maintaining a positive long-term relationship (e.g., Fischer & Roseman, 2007). It is this positive relational signal that increases empathy in the recipients of the communicated anger for those who communicated it. This can be contrasted to the communication of an emotion such as contempt (e.g., Rozin, Lowery, Imada, & Haidt, 1999) that, although it is closely related to anger, should not have this positive effect because it signals a strong devaluation of the relationship (e.g., Fischer & Roseman, 2007). The communication of group-based contempt should therefore undermine any positive effect of group-based anger on intergroup conflict reduction through reducing empathy with the other group. In fact, we believe that negative effects of communicating group-based anger in intergroup conflict are likely to stem from communicating group-based contempt or a mixture of these emotions. In three experiments, we therefore examined the idea that only communicating “pure” group-based anger has positive effects in terms of reducing tendencies toward intergroup conflict because of its relational function. For this reason, we also predicted that this positive effect of communicating “pure” group-based anger is mediated by increased empathy for the out-group.

Discovering the Relational Function of Communicating Anger

Emotions have an important social function because they communicate how individuals feel about others and how they are likely to behave toward them (Averill, 1982; Fischer & Roseman, 2007). Group-based emotions are emotions felt and expressed or communicated on behalf of one’s group membership (Smith, 1993; Smith et al., 2007). The most prototypical emotion in intergroup conflict is group-based anger (Allred, 1999), which stems from a group-based appraisal of out-group members’ intergroup behavior as unjust (e.g., Frijda, 1986; Lazarus, 1991; Roseman, 2001; Scherer, 2001; Scherer, Schorr, & Johnstone, 2001). An important social function of anger is to ameliorate one’s own situation by causing a change in the behavioral pattern of the other (Fischer & Roseman, 2007). Indeed, anger is a unique emotion in the sense that it has both a negative valence, necessary to signal injustice, while being approach-oriented, vital to engage in dialogue about or confrontation of the perceived injustice. This sets anger apart from other negatively valenced emotions such as fear, sadness, and contempt, all of which communicate more avoidance-oriented motivation (for an overview, see Carver & Harmon-Jones, 2009). Whereas anger is typically associated with destructive responses toward the other (e.g., Berkowitz, 1989; but see Averill, 1982), its expression might in fact be beneficial for maintaining and improving relationships (Fischer & Roseman, 2007).

We therefore propose a positive relational function of group-based anger in intergroup conflict. This means that communicating group-based anger stresses the injustice of the situation while asking the other group to take steps toward reconciliation. Communication of group-based anger will therefore be perceived by the other group as a signal that one cares about maintaining a positive intergroup relationship. We argue that this in turn should encourage the recipient to take the other party’s perspective and feel empathy for them. Past research has extensively documented positive effects of empathy in reducing intergroup conflict (e.g., Batson et al., 1997; Pettigrew & Tropp, 2008; Stephan & Finlay, 1999), yet empathy has not been linked to communication of group-based anger. Because of its relational function, the communication of anger should reduce conflict through increasing empathy.

Prior research has neither considered this relational function, nor has it studied empathy as a mediator of anger’s positive effects. Instead, prior research has focused on the equally important but different strategic function of anger expressions (for an overview, see Van Kleef, 2009). Focusing mostly on interpersonal negotiations, this research showed that communication of anger about the other’s behavior increases this other’s approach behavior, enforcing cooperation within a competitive setting (e.g., Van Kleef, De Dreu, & Manstead, 2004). In these cases, the communication of anger serves the strategic function of informing the recipient of the sender’s intentions with a focus on the personal gain of the communicator (e.g., your last offer makes me angry). By contrast, the relational function of communicating anger focuses more on shared gains, which fits with our emphasis on empathy. Thus, like prior research, we expect positive effects of communication of anger, but we explain them through a relational function. Moreover, we apply this idea to intergroup rather than interpersonal settings through the notion of group-based emotion (Smith, 1993; Smith et al., 2007).

Communicating “Pure” Anger Versus Contempt and Mixed Anger/Contempt

If it is indeed the relational function of communicating anger that results in a reduction of conflict, then a comparable negative emotion with an opposite relational function should result in quite adverse effects. An excellent candidate for this is contempt because anger and contempt represent responses to violations of particular norms (e.g., Rozin et al., 1999). At the same time, however, research shows that contempt communicates the desired exclusion of the other person (Fischer & Roseman, 2007), which should, if anything, undermine empathy. Indeed, people perceive communication of contempt as a signal that the other desires to end the relationship. Communicating anger should therefore have more positive consequences than communicating contempt.

In the context of communicating anger or contempt, it is important to note that these two emotions do not necessarily exclude one another. In conflict situations, communication of anger is often combined with contempt (e.g., Cottrell & Neuberg, 2005). In fact, it has been argued that when anger persists over a longer period of time, it might start to coexist or completely develop into contempt (Fischer & Roseman, 2007). If communication of contempt is mixed with anger, this should, we believe, transform the interpretation of the communication: Because contempt signals the desired exclusion of the other, communication of anger in conjunction with it should resemble the adverse effects of contempt. Thus, a perhaps even more interesting test of the relational function of communicating anger would be to pit communication of “pure” anger against communication of “mixed” anger and contempt. The relational function of communicating anger should only function effectively when it is communicated free of contempt (given its function to strongly devalue the relationship). Applied to intergroup conflict, this means that the communication of “pure” group-based anger should have more positive effects than the communication of group-based contempt, or the combination of both group-based emotions.

We thus predict that only the communication of “pure” group-based anger decreases destructive intergroup conflict intentions because it increases empathy for the out-group. We sought support for our novel line of thought in three experiments. Each of these experiments used an intergroup conflict situation within which we assessed the positive effects of the communication of “pure” group-based anger in terms of reduced destructive conflict intentions. Experiment 1 manipulated the communication of group-based anger versus the absence of it to show the positive effect of using anger’s relational function. Experiment 2 manipulated the communication of “pure” group-based anger versus its absence, the communication of “pure” group-based contempt, and versus the communication of “mixed” group-based anger/contempt. Experiment 3 specifically focused on the contrast between communicating “pure” group-based anger and “mixed” group-based anger/contempt, while crossing it with a manipulation of the proposed mediator.

Experiment 1

The main aim of Study 1 was to find support for the relational function of the communication of group-based anger in intergroup conflict. According to this function, the communication of group-based anger should reaffirm the value of the relationship while also signaling that harm is done and as such reconciliation is needed. Consequently, we investigated whether the communication of group-based anger would reduce destructive conflict intentions compared with a condition in which the communication of group-based anger was absent, and whether this effect would be explained by increased empathy for the out-group. Furthermore, to link this positive effect and the role of empathy in it specifically to the relational function of the communication of group-based anger, we investigated whether a similar pattern of results would be obtained on measures of recipients’ perception of whether the communicator values their relationship.

Method

Participants were 44 Dutch psychology students (37 women; M_age = 21.80; SD = 1.91) who participated for course credit. Participants were randomly assigned to one of two conditions (out-group communicating anger vs. a control condition).

In the experiment, participants read a fictitious newspaper article and subsequently answered questions about it. The article outlined a realistic situation in which German students were discriminated against by Dutch students: A majority (78%) indicated having negative attitudes toward German students, arguing German students should remain in Germany. Dutch students justified their opinion, by referring to common stereotypes of Germans as being “badly integrated,” “loners,” “humorless strivers.”

In the article, the German students reacted to what they considered to be unjust discrimination. In this reaction, we manipulated whether they communicated anger or not. The communicated emotion was manipulated in the title of the article, the introductory paragraph, and in the final paragraph in which a German student (Lena) responded on behalf of her group (“We are viewed negatively based on unjust stereotypes and prejudices!” [control] / “I am very angry that we are viewed negatively based on unjust stereotypes and prejudices!” [anger]). We thus manipulated only the presence or absence of the anger label in the communication by the out-group.

Every variable in each of the three experiments, unless otherwise specified, was measured on 7-point Likert-type scales (1 = absolutely not, 7 = very much). An anger manipulation check measured to what extent they perceived the German students to be angry. Destructive conflict intentions were measured with 12 items (α = .86). Originally, three subscales consisting of four items each were devised, two of which measured tendencies to move against and move away (cf. Horney, 1945; for example, “To what extent would you, as a Dutch student, want to seek a direct confrontation with German students such as Lena to release some frustration?” and “To what extent would you, as a Dutch student, exclude German students such as Lena from your social network?”), and one measured inaction (“To what extent would you, as a Dutch student, do nothing to accommodate German students such as Lena?”). As results on all separate subscales revealed a highly consistent pattern across all experiments, we decided to collapse all items into one scale measuring destructive conflict intentions for all subsequent analyses.

In addition, we measured the proposed mediator, empathy, with two items (“To what extent do you empathize with German students such as Lena?” and “To what extent do you feel indifferent toward German students such as Lena?” latter item recoded; r = .48, p = .01). Finally, communicated relationship importance was measured with five items (α = .94; for example, “To what extent do you think German students such as Lena value a good relationship with Dutch students?” and “To what extent do you think German students such as Lena want to increase positive contact with Dutch students?”).

Results and Discussion

Manipulation Checks

As intended, the results of an analysis of variance (ANOVA) showed significant differences in the level of perceived anger between conditions, F(1, 42) = 6.36, p = .004, η²_p = .23. Participants perceived more anger in German students in the anger condition (M = 5.47, SD = 0.83) than in the control condition (M = 4.41, SD = 0.87), t(43) = 3.31, p = .002. This result indicates that the manipulation was a success.

Destructive Conflict Intentions

In line with the predicted positive effect of communicating group-based anger, an ANOVA showed significant differences on the Destructive Conflict Intentions Scale, F(1, 42) = 5.21, p = .028, η²_p = .11. Participants indeed reported less destructive conflict intentions when anger had been communicated (M = 2.22, SD = 0.57) than in the control condition (M = 2.62, SD = 0.58; see Table 1 for an overview of all results).

Table 1.

Means and Standard Deviations of the Dependent Variables as a Function of the Presence or Absence of Anger, Experiment 1.

	Anger		Control
Variable	M	SD	M	SD
Destructive conflict intentions	2.22	0.57	2.62	0.58
Empathy	4.91	1.20	3.95	1.27
Communicated relationship importance	4.76	1.26	3.72	1.12

Empathy

In line with anger’s relational function, an ANOVA showed significant differences on the Empathy Scale, F(1, 42) = 6.55, p = .014, η²_p = .14. Participants reported to feel more out-group empathy after anger had been communicated (M = 4.91, SD = 1.20) than in the control condition (M = 3.95, SD = 1.27).

Communicated Relationship Importance

Corroborating anger’s relational function, an ANOVA showed significant differences on the Relationship Importance Scale, F(1, 42) = 8.42, p = .006, η²_p = .17. Participants indicated that they thought German students valued the relationship more after they had communicated anger (M = 4.76, SD = 1.26) than in the control condition (M = 3.72, SD = 1.12).

Mediation Analyses

To test the idea that the positive effect of the communication of group-based anger is explained by increased empathy, we performed a series of regression analyses. Results showed, first, that the communication of anger (compared with the control condition) reduced destructive conflict intentions (β = −.33, p = .028) and increased empathy (β = .37, p = .014). Second, empathy significantly predicted destructive conflict intentions (β = −.55, p < .001) above and beyond the effect of the manipulation (which turned nonsignificant: β = −.13, p = .331). A Sobel z test confirmed that the indirect effect was significant, Z = −2.21, p = .027. In line with our ideas, empathy fully mediated the positive effect of “pure” group-based anger on destructive conflict intentions.

Furthermore, we reasoned that the communication of anger (compared with the control condition) signals that the out-group assigns a higher value to the relationship, and thus links the communication of anger to empathy and decreased destructive conflict intentions. We therefore ran two sets of regression analyses. In line with our ideas, whereas the communication of anger (compared with the control condition) decreased destructive conflict intentions (β = –.33, p = .028) and increased communicated relationship importance (β = .41, p = .006), the latter indeed significantly predicted the former (β = −.46, p = .003) above and beyond the effect of the manipulation (which turned nonsignificant: β = −.14, p = .324). A Sobel z test confirmed that the indirect effect was significant, Z = 2.14, p = .032. Thus, in line with anger’s relational function, communicated relationship importance fully mediated the positive effect of “pure” group-based anger on destructive conflict intentions.

In the second set of analyses, we wanted to test which of the two variables was more closely related to decreased destructive action tendencies. According to our reasoning, increased empathy should follow from the perception of communicated anger because it communicated relational importance. Indeed, at the bivariate level, the variables were positively correlated (r = .47, p = .001). Moreover, entering both constructs in the mediation analyses at the same time results in a stronger predictive value for empathy (β = −.44, p = .003) than for relationship importance (β = −.29, p = .048) with respect to destructive conflict intentions (above and beyond the effect of the manipulation, which turned nonsignificant: β = −.05, p = .70). A bootstrapping macro for multiple mediators (Preacher & Hayes, 2008) was used to provide a more powerful test of the significance of both indirect paths. Confirming the above results, bootstrapping 5,000 samples at a 95% confidence interval (CI) showed that the indirect effect of empathy, 95% CI = [−0.472, −0.037], was more clear than the indirect effect of relationship importance, 95% CI = [−0.374, −0.0001]. These results suggest, in line with our reasoning, that empathy for the out-group is the more proximal predictor of decreased destructive action tendencies as a function of the presence or absence of the communication of group-based anger. At the same time, these results corroborate the workings of anger’s relational function by linking its communication directly to increased communicated relationship importance, increased empathy for the out-group, and decreased destructive action tendencies toward them.

Discussion

The results of Experiment 1 provide first support for our hypothesis that the communication of group-based anger (versus its absence) results in less destructive intergroup conflict intentions and that this decrease is explained by an increase in empathy for the out-group. Further substantiating the relational function of the communication of group-based anger in intergroup conflict, Experiment 1 showed that communicated anger increased the importance recipients perceive the out-group to assign to their relationship. In line with our argument, communicated relationship importance was positively related to empathy, but the latter explained more variance in destructive action tendencies.

Experiment 2 was designed to replicate the positive effect of communicating group-based anger (versus its absence), but moved beyond Experiment 1 by testing whether its predicted positive effects would still hold when compared with the communication of group-based contempt or of a mixture of group-based anger/contempt. Indeed, because the proposed relational function of communicating group-based anger is distinct and different from the function of communicating group-based contempt (which is to devaluate the intergroup relationship and terminate it), it follows that only the communication of “pure” anger should reduce destructive conflict intentions.

Experiment 2

Method

Participants were 51 Dutch psychology students (41 women; M_age = 20.21, SD = 1.59) who participated for course credit and were randomly assigned to one of four conditions within a 2 (Group-Based Anger: Absent/Present) × 2 (Group-Based Contempt: Absent/Present) between-subjects design. As in Experiment 1, in the experiment, participants read a (fictitious) newspaper article and subsequently answered questions about it. The article described the current shortage on the Groningen student dorm market. A majority (74%) of Dutch students indicated rather having a Dutch than a German roommate, which was justified by referring to typical stereotypes of Germans as being “badly integrated,” “loners,” and “humorless strivers.”

The article continued by noting how the German students reacted to what they considered to be unjust discrimination. In this reaction, we manipulated whether they communicated “pure” anger, “pure” contempt, “mixed” anger and contempt (in which participants expressed both emotions), or no emotion. The communicated emotion was manipulated in the title of the article (“Discrimination of German Students” / “Anger Among German students” / “Contempt Among German students” / “Anger and Contempt Among German students”), the introductory paragraph (e.g., “Dutch Discrimination of Germans Makes German Students Very Angry”), and in the final paragraph in which a German student (Kati) responded on behalf of her group (“I am very angry that we are discriminated against based on unjust stereotypes and prejudices!” [anger] / “I really look down on Dutch students for discriminating us based on unjust stereotypes and prejudices!” [contempt] / “I am very angry that we are discriminated against based on unjust stereotypes and prejudices and I really look down on Dutch students for doing that!” [anger & contempt] / “We are discriminated based on unjust stereotypes and prejudices.” [no emotion]).

Two manipulation checks verified whether participants had perceived the correct emotion. Participants indicated to what extent they perceived German students to be angry and contemptuous. We measured the proposed mediator, empathy, with the same two items as in Experiment 1 (r = .34, p = .01). Similarly, destructive conflict intentions were measured with 16 items, 12 of which were identical to the ones used in Experiment 1, while the last 4 items were specifically related to the conflict situation at hand (e.g., “To what extent would you, as a Dutch student, choose Dutch students over German students such as Kati as a future roommate?”; α = .87).

Results and Discussion

Manipulation Checks

A 2 × 2 ANOVA on the manipulation check of perceived group-based anger revealed only the intended main effect of group-based anger, F(1, 47) = 5.46, p = .024, η²_p = .10. German students were indeed perceived to be more angry in the anger present condition (M = 6.27, SD = 0.60) than in the anger absent condition (M = 5.72, SD = 1.06). The main effect of contempt and the interaction effect were not significant, both Fs < 2.86. Another 2 × 2 ANOVA on the manipulation check of perceived contempt revealed only the expected main effect of contempt, F(1, 47) = 12.36, p = .009, ηp2 = .21: Participants perceived more contempt in the contempt-present conditions (M = 5.48, SD = 0.82) than in the contempt-absent conditions (M = 4.42, SD = 1.28). The main effect of anger and the interaction effect were not significant, both Fs < 1.3. Thus, both manipulations were effective.

Destructive Conflict Intentions

A 2×2 ANOVA on destructive conflict intentions showed a significant main effect of group-based contempt, F(1, 47) = 4.80, p = .033, η²_p = .09. Participants reported more destructive conflict intentions after the German students communicated contempt (M = 2.45, SD = 0.72) than when they did not communicate contempt (M = 2.02, SD = 0.72). More importantly, this main effect was qualified by the predicted significant two-way interaction effect, F(1, 47) = 6.75, p = .012, η²_p = .17 (see Figure 1). Replicating the Experiment 1 findings, simple main effect analyses showed that when devoid of contempt, communication of (thus “pure”) group-based anger resulted in less destructive conflict intentions (M = 1.68, SD = 0.48) than when no anger was communicated (M = 2.39, SD = 0.76), F(1, 47) = 6.65, p = .013. On the other hand, when group-based contempt was communicated, there was no difference between anger (thus “mixed” anger/contempt; M = 2.60, SD = 0.62) and no anger (thus “pure” contempt; M = 2.31, SD = 0.81), F(1, 47) = 1.17, p = .285. Moreover, in line with our hypothesis, a one-degree of freedom contrast between “pure” anger and “pure” contempt also proved significant, t(47) = −2.34, p = .024, d = −0.63, 95% CI = [−1.16, −0.09]. Indeed, as expected, only the communication of “pure” group-based anger (versus the other three conditions) resulted in the least destructive conflict intentions, t(47) = 3.42, p = .001.

Figure 1.

Means and standard errors of destructive conflict intentions towards German students as a function of the presence or absence of anger and contempt, Experiment 2.

Empathy

A 2 × 2 ANOVA with empathy as the dependent variable showed the predicted significant two-way interaction effect, F(1, 47) = 8.48, p = .005, η²_p = .17 (see Figure 2). Replicating the Experiment 1 findings, and in line with anger’s relational function, simple main effect analyses showed that, when no group-based contempt was communicated, more empathy was felt after communication of (thus “pure”) group-based anger (M = 5.54, SD = 1.25) than when no group-based anger was communicated (M = 4.04, SD = 1.15), F(1, 47) = 10.10, p = .003. However, when group-based contempt was communicated, no differences were found between the conditions in which group-based anger was (M = 4.15, SD = 1.11) or was not (M = 4.58, SD = 1.19) communicated, F(1, 47) = 0.84, p = .364 (thus no differences between “pure” contempt and “mixed” anger/contempt). Moreover, in line with our hypothesis, a one-degree of freedom contrast between “pure” anger and “pure” contempt also proved significant, t(47) = 2.08, p = .043, d = 0.96, 95% CI = [0.03, 1.89]. Finally, as expected, only the communication of “pure” group-based anger (compared with the other three conditions) resulted in the most empathy for the German student, t(47) = 3.39, p = .001.

Figure 2.

Means and standard errors of empathy for German students as a function of the presence or absence of anger and contempt, Experiment 2.

Mediation Analyses

We hypothesized that the positive effect of “pure” group-based anger (versus all other conditions) on destructive conflict intentions would be explained by increased empathy. We therefore ran a series of regression analyses. Results showed, first, that the two-way interaction predicted destructive conflict intentions (β = .34, p = .012) and empathy (β = −.38, p = .005). Second, empathy significantly predicted destructive conflict intentions (β = −.40, p = .005) above and beyond the interaction effect (which turned nonsignificant: β = .19, p = .161). A Sobel z test confirmed that the indirect effect was significant, Z = 2.08, p = .038. As in Experiment 1, empathy thus fully mediated the positive effect of “pure” group-based anger on destructive conflict intentions.¹

Discussion

The results of Experiment 2 replicate those of Experiment 1 by showing that the communication of “pure” group-based anger results in less destructive conflict intentions than its absence. Moreover, Experiment 2 replicated the mediational role of empathy in this effect. Furthermore, Experiment 2 showed no such positive effects for the communication of group-based contempt, and of “mixed” anger/contempt. This increases confidence in the proposed relational function of communicating group-based anger in the context of intergroup conflict. However, because empathy was measured in Experiments 1 and 2, we are not yet in the position to make claims about causality with respect to its mediational role. Experiment 3 thus manipulated empathy (through an often-used perspective taking [PT] manipulation; see Batson et al., 1997). For reasons of efficiency, we wanted to restrict Experiment 3 to a focused examination of empathy’s effects as a function of the main contrast between “pure” and “mixed” group-based anger. Because we already replicated support for the prediction that the communication of group-based anger by and of itself already has positive effects, we reasoned that the inclusion of such a control condition was less relevant in Experiment 3. Furthermore, our choice between contrasting a “pure” group-based anger condition with a “mixed” anger condition rather than a “pure” contempt condition was based on the idea that the former presents a more conservative test of our ideas regarding the relational function of communicating group-based anger.

Against this backdrop, in Experiment 3 participants were instructed to take the perspective of an in-group or an out-group member, or participants received no PT instructions. We hypothesized that in the absence of PT instructions, the pattern should resemble that of Experiment 1, constituting a direct replication of the positive effect of “pure” compared with “mixed” anger. Furthermore, when instructed to take the perspective of the out-group, we expected a similar pattern because taking the out-group’s perspective matches the relational function of anger but not of contempt. Indeed, in line with the social function of contempt (e.g., Rozin et al., 1999), we reasoned that its communication would effectively undermine empathizing with the out-group because it signals a devaluation of the relationship. Finally, we hypothesized that being instructed to take the perspective of one’s own in-group would clash with any tendencies toward taking the perspective of the out-group member, effectively canceling out “pure” anger’s positive effects.

Experiment 3

Method

Participants were 101 Dutch students (74 women; M_age = 20.98; SD = 1.34) who participated in exchange for €2.50 and were randomly assigned to one of six conditions within a 2 (Communicated Emotion: “Pure” anger vs. “Mixed” anger) × 3 (PT: Dutch students vs. German students vs. control) between-subjects design. As in the first experiment, participants read a fictitious newspaper article about Dutch students discriminating German students. The context was that of a future increase in college tuition fees. We chose this particular context because we wanted to show that the communication of “pure” anger can be beneficial even in intergroup conflicts with a clear instrumental basis (which was not the case in Experiments 1 and 2). In the article, a majority (79%) of Dutch students expressed the opinion that tuition fees should increase more for German students than for Dutch students, justifying their opinion by referring to common German stereotypes.

As in the previous experiments, the article was illustrated with a citation of a German student. Communication of “pure” and “mixed” anger was manipulated in a similar way as in Experiments 1 and 2. The manipulation of PT took place prior to reading the article. In the no instructions conditions, participants read that they were going to read an article “about a conflict situation between German and Dutch students in Groningen.” Participants in the other PT conditions subsequently read:

While reading the article, try to put yourself in the shoes of the Dutch student, Michiel Dijkstra / German student, Jonas Kruger. Try to identify as best as you can with his feelings and reaction to the situation and try to get a sense of how he really feels.

We decided to use this manipulation after a pilot study (N = 48) confirmed that this manipulation was effective in affecting actual PT and subsequently felt empathy.² We used the same manipulation checks for perceived anger and contempt as in Experiment 2. Destructive conflict intentions were again measured with 16 items, 12 of which measured the move away, move against, and inaction tendencies as in Experiments 1 and 2, while 4 new conflict-specific items were related to the situation in the article (α = .82; for example, “As a Dutch student, to what extent would you want to sign a petition to increase tuition fees for German students more than those of Dutch students?”). As in the previous experiments, we collapsed all items into one Destructive Conflict Intention Scale.

Results

Manipulation Checks

As expected, a 2 × 3 ANOVA on perceived contempt showed only a significant main effect of group-based contempt, F(1, 95) = 16.48, p < .01, ηp2 = .15: Participants perceived more contempt in German students in the “mixed” group-based anger/contempt conditions (M = 4.10, SD = 1.50) than in the “pure” anger conditions (M = 2.86, SD = 1.47). However, note that participants should perceive an equal amount of anger across conditions, which was indeed the case. The 2 × 3 ANOVA on perceived anger showed no significant differences between the “pure” group-based anger conditions (M = 5.66, SD = 1.08) and the “mixed” group-based/contempt anger conditions (M = 5.80, SD = 1.27), F(1, 95) = 0.35, p = .556. In both tests, the main effects of PT and the interaction effects were not significant, all Fs < 1.06, ps > .352. Thus, the manipulations again were successful.

Destructive Conflict Intentions

A 2 × 3 ANOVA on destructive conflict intentions yielded a significant main effect for communicated emotion, F(1, 95) = 11.51, p < .01, η²_p = .11. Participants reported weaker destructive conflict intentions when German students had communicated “pure” group-based anger (M = 2.67, SD = 0.68) compared with when they had communicated “mixed” group-based anger/contempt (M = 3.17, SD = 0.80). This main effect was qualified by a marginally significant interaction effect, F(2, 95) = 2.41, p = .096, ηp2 = .05 (see Figure 3). However, the more focused test of our predictions is at the level of the simple main effects. Replicating Experiment 2, in the no-instruction condition, participants reported less destructive conflict intentions after German students communicated “pure” group-based anger (M = 2.53, SD = 0.68) than when they communicated “mixed” group-based anger/contempt (M = 3.29, SD = 1.05), F(1, 95) = 8.73, p = .004. As expected, exactly the same pattern arose in the German students’ perspective condition: Communication of “pure” group-based anger (M = 2.48, SD = 0.53) led to less destructive conflict intentions than communication of “mixed” group-based anger/contempt (M = 3.16, SD = 0.71), F(1, 95) = 7.72, p = .007. By contrast, in the Dutch students’ perspective conditions, no significant differences in reported destructive conflict intentions were found between the communicated “pure” group-based anger (M = 3.00, SD = 0.72) and “mixed” group-based anger/contempt (M = 3.04, SD = 0.56) conditions, F < 1. Thus, as expected, taking the perspective of Dutch students undermined the positive effect of communicating “pure” group-based anger versus “mixed” group-based anger/contempt. All these findings substantiate the hypothesized mediational role of empathy as found in Experiments 1 and 2 through experimental manipulation of this mediator. As such, the results of Experiment 3 support the proposed relational function of group-based anger in the context of intergroup conflict and its communicative potential toward improving conflictuous intergroup relations.

Figure 3.

Means and standard errors of destructive conflict intentions towards German students as a function of the communicated emotion and perspective taking, Experiment 3.

General Discussion

Three experiments provide convergent evidence for the novel hypothesis that the communication of “pure” group-based anger in intergroup conflicts has positive consequences. Across the board, our results showed that communicating “pure” group-based anger reduced destructive conflict intentions among recipients of the communication by increased empathy for the communicator (in comparison with communicating no group-based anger, “pure” group-based contempt, and “mixed” group-based anger/contempt). These findings map onto our argument that the communication of group-based anger can have such positive effects in intergroup conflicts because it stresses not only the perceived injustice of the intergroup situation (i.e., the undesirable treatment by the other) but at the same time the desire to maintain a positive long-term intergroup relationship (Fischer & Roseman, 2007). Indeed, Experiment 1 results directly support this claim by showing that the communication of “pure” group-based anger increased the perception that the out-group valued the intergroup relationship, and connected this to increased empathy for the out-group. Furthermore, Experiments 1 to 3 all showed evidence for the mediational role of empathy in explaining the positive effects of communicating “pure” group-based anger, with Experiment 3 even showing evidence for its causal role in the mediation sequence.

Our studies extend prior theorizing and research in several ways. First, we move beyond Fischer and Roseman (2007) by extending their interpersonal analysis of anger and contempt to group-based anger and contempt in the intergroup domain (Smith, 1993; Smith et al., 2007). Moreover, we move beyond previous work on group-based emotions by showing that group-based emotions have the potential to be used as communicative devices that stress the value of maintaining an intergroup relationship. By contrast, Fischer and Roseman showed that the perception of the level of intimacy of a relationship with another person predicts whether anger or contempt will be expressed. Thus, our data show that emotions are not merely expressions that follow from a certain relationship (cf. Fischer & Roseman, 2007), but that individuals can communicate emotions that signal how they view the importance of that relationship. This can lead the other to either want to improve (in the case of anger) or discontinue (for contempt) the relationship.

Our results further help explain the paradox that the communication of anger can have positive (Halperin et al., 2011; Van Kleef et al., 2004) and negative effects (Berkowitz, 1989; Lickel, Miller, Stenstrom, Denson, & Schmader, 2006) in conflict situations. By differentiating between “pure” and “mixed” anger communications, it becomes clear that only communication of “pure” anger accentuates the value of maintaining a positive relationship, whereas the “mixing” of anger with contempt (and communication of “pure” contempt) transforms the relational intent that is communicated—signaling instead a desire to terminate the relationship. Our results offer evidence for the distinct social and communicative functions of anger and contempt (see also Fischer & Roseman, 2007), and also for the validity of the distinction between these emotions (Rozin et al., 1999). Thus, if one wants to reduce the potential negative effects of emotional communication in intergroup conflict, our analysis suggests that the communicator should take care to communicate “pure” anger and that the recipient should also perceive it as such.

Our analysis shares an emphasis on potentially positive consequences of communicating anger with the Emotions As Social Information (EASI) model (e.g., Van Kleef, 2009), and in fact we believe that our findings can make a valuable contribution to the EASI model by focusing attention to the relational function of the communication of anger (i.e., anger as a signal of, as it were, the personal pain that flows from bad treatment of the other, which contains within it a tacit expression of the desire to improve that relationship). Indeed, the EASI model focuses on the equally important but different strategic function of the communication of anger (i.e., anger as a signal of thwarted personal gain). Van Kleef (2009) argues that in competitive settings, such as conflicts, the expression of anger would affect the other person’s strategic inferences of the cause of this emotion, leading one to perceive the other as, for instance, having high negotiation limits. This would then result in giving in to the angry person’s demands (i.e., cooperation). Future research should focus on the role of empathy in this process to explore how the relational and strategic functions of the communication of anger relate to each other.

Of course, the positive effects of communicating “pure” group-based anger should have boundary conditions. For instance, the current studies point toward the importance of the “purity” of the anger communicated as well as the possibility to take the out-group’s perspective. We believe that other boundary conditions include the perception of fair treatment or legitimacy (e.g., Frijda, 1986; Lazarus, 1991; Roseman, 2001; Scherer, 2001; Scherer et al., 2001), and the extent to which the anger is expressed by all group members (and thus on behalf of the group; for example, De Vos, Van Zomeren, Gordijn, & Postmes, 2013). Indeed, when confronted with an out-group member’s anger, individuals tend to assess the extent to which the anger is just given the fairness of the treatment they received. If treatment is fair, anger is likely to be perceived as inappropriate and perhaps even as a sign of aggression, which might lead to conflict escalation rather than de-escalation. Furthermore, recipients should view the out-group communicator as a representative for the entire out-group if anger is to have positive effects. Otherwise, the anger might be perceived as inauthentic or “just a lone voice” and could be disregarded.

Furthermore, from the EASI model it can be derived that power differentials between the out-group communicating the anger and the target in-group may be important (e.g., Van Kleef & Côté, 2007) in the sense that groups with low power are less effective in bringing about change when expressing anger. In the current experiments, the communicator was part of a minority, but there existed no measurable power differences between the groups so this needs to be explored further. Moreover, also contextual differences between competitive and cooperative situations in which anger is expressed might influence its success: Anger is seen as more appropriate in situations where a competition or conflict of interests is present, whereas anger in more cooperative settings (like a reconciliation process) would be quite detrimental (Van Kleef, 2009). Indeed, whereas the current studies investigated how group-based anger might de-escalate intergroup conflict, the next step for future research is to uncover when it does and does not occur.

In sum, even though anger is often perceived to be an unpleasant and negative subjective state, the current experiments show that its communication has remarkably positive consequences in the context of intergroup conflict. Communicating “pure” group-based anger explicitly stresses the injustice of another party’s actions, while underlining the value of maintaining a positive intergroup relationship. The expression of group-based anger may thus have positive consequences for intergroup relations. Clearly, when the communication of “pure” group-based anger leads to conflict de-escalation, this expression might be an important first step toward reducing intergroup conflict. Anger, in this sense, should not always be kept under wraps.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was financially supported by a Netherlands Organisation for Scientific Research (NWO) Grant 432-08-026, awarded to Ernestine H. Gordijn, Martijn van Zomeren, and Tom Postmes.

Notes

References

Allred

K. G.

(1999). Anger and retaliation: Toward an understanding of impassioned conflict in organizations. In Bies

R. J.

Lewicki

R. J.

Sheppard

B. H.

(Eds.), Research in negotiation in organizations Vol. 7 (pp. 27-58). Greenwich, CT: JAI Press.

Averill

J. R.

(1982). Anger and aggression: An essay on emotion. New York, NY: Springer.

Batson

C. D.

Polycarpou

M. P.

Harmon-Jones

Imhoff

H. J.

Mitchener

E. C.

Bednar

L. L.

Highberger

(1997). Empathy and attitudes: Can feeling for a member of a stigmatized group improve feelings toward the group? Journal of Personality and Social Psychology, 72, 105-118. doi:10.1037/0022-3514.72.1.105

Berkowitz

(1989). Frustration-aggression hypothesis: Examination and reformulation. Psychological Bulletin, 106, 59-73. doi:10.1037/0033-2909.106.1.59

Carver

C. S.

Harmon-Jones

(2009). Anger is an approach-related affect: Evidence and implications. Psychological Bulletin, 135, 183-204. doi:10.1177/0022002710383670

Cottrell

C. A.

Neuberg

S. L.

(2005). Different emotional reactions to different groups: A sociofunctional threat-based approach to “prejudice”. Journal of Personality and Social Psychology, 88, 770-789. doi:10.1037/0022-3514.88.5.770

De Vos

Van Zomeren

Gordijn

E. H.

Postmes

(2013). “We” have a “right” to be angry! Or when legitimacy and consensus increase our empathy for the angry outgroup Manuscript in preparation.

DiGuiseppe

Tafrate

R. C.

(2003). Anger treatment for adults: A meta-analytic review. Clinical Psychology: Science and Practice, 10, 70-84. doi:10.1093/clipsy/10.1.70

Fischer

A. H.

Roseman

I. J.

(2007). Beat them or ban them: The characteristics and social functions of anger and contempt. Journal of Personality and Social Psychology, 93, 103-115. doi:10.1037/0022-3514.93.1.103

10.

Frijda

N. H.

(1986). The emotions. Cambridge, UK: University Press.

11.

Halperin

Russell

A. G.

Dweck

C. S.

Gross

J. J.

(2011). Anger, hatred, and the quest for peace: Anger can be constructive in the absence of hatred. Journal of Conflict Resolution, 55, 274-291. doi:10.1177/0022002710383670

12.

Horney

(1945). Our inner conflicts. Oxford, UK: Norton.

13.

Lazarus

R. S.

(1991). Cognition and motivation in emotion. American Psychologist, 46, 352-367. doi:10.1037/0003-066X.46.4.352

14.

Lickel

Miller

Stenstrom

D. M.

Denson

T. F.

Schmader

(2006). Vicarious retribution: The role of collective blame in intergroup aggression. Personality and Social Psychology Review, 10, 372-390. doi:10.1207/s15327957pspr1004_6

15.

Pettigrew

T. F.

Tropp

(2008). How does intergroup contact reduce prejudice? Meta-analytic tests of three mediators. European Journal of Social Psychology, 38, 922-934. doi:10.1002/ejsp.504

16.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879-891. doi:10.3758/BRM.40.3.879

17.

Roseman

I. J.

(2001). A model of appraisal in the emotion system: Integrating theory, research, and applications. In Scherer

K. R.

Schorr

Johnstone

(Eds.), Appraisal processes in emotion (pp. 68-91). Oxford, UK: University Press.

18.

Rozin

Lowery

Imada

Haidt

(1999). The CAD-triad hypothesis: A mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). Journal of Personality and Social Psychology, 76, 574-586. doi:10.1037/0022-3514.76.4.574

19.

Scherer

K. R.

(2001). Appraisal considered as a process of multilevel sequential checking. In Scherer

K. R.

Schorr

Johnstone

(Eds.), Appraisal processes in emotion (pp. 92-120). Oxford, UK: University Press.

20.

Scherer

K. R.

Schorr

Johnstone

(2001). Appraisal processes in emotion. Oxford, UK: Oxford University Press.

21.

Smith

E. R.

(1993). Social identity and social emotions: Toward new conceptualisations of prejudice. In Mackie

D. M.

Hamilton

D. L.

(Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 297-315). San Diego, CA: Academic Press.

22.

Smith

E. R.

Seger

C. R.

Mackie

D. M.

(2007). Can emotions be truly group level? Evidence regarding four conceptual criteria. Journal of Personality and Social Psychology, 93, 431-446.

23.

Stephan

W. G.

Finlay

(1999). The role of empathy in improving intergroup relations. Journal of Social Issues, 55, 729-744. doi:10.1111/0022-4537.00144

24.

Van Kleef

G. A.

(2009). How emotions regulate social life: The emotions as social information (EASI) model. Current Directions in Psychological Science, 18, 184-188. doi:10.1111/j.1467-8721.2009.01633.x

25.

Van Kleef

G. A.

Côté

(2007). Expressing anger in conflict: When it helps and when it hurts. Journal of Applied Psychology, 92, 1557-1569. doi:10.1037/0021-9010.92.6.1557

26.

Van Kleef

G. A.

De Dreu

C. K. W.

Manstead

A. S. R.

(2004). The interpersonal effects of anger and happiness in negotiations. Journal of Personality and Social Psychology, 86, 57-76. doi:10.1037/0022-3514.86.1.57