When does the communication of group-based anger increase outgroup empathy in intergroup conflict? The role of perceived procedural unfairness and outgroup consensus

Abstract

Increasing outgroup empathy is an important first step toward reducing intergroup conflict. The communication of group-based anger has been found to increase outgroup empathy due to its presumed relational function (as it signals to the outgroup that they unfairly treat the ingroup, but also that the ingroup wants to maintain a positive intergroup relationship). Yet, little is known about when communicating group-based anger increases outgroup empathy. We therefore examine two antecedent conditions, namely perceived procedural unfairness (which makes the communication of anger more appropriate) and outgroup consensus (which makes the communication of anger more group-based). Three experiments suggest that the communication of group-based anger increases outgroup empathy only when the outgroup was treated unfairly (Experiment 1) and when there was high outgroup consensus (Experiment 2). Results from Experiment 3 revealed that either antecedent seems sufficient to facilitate the positive, empathy-inducing effects of the communication of group-based anger. We discuss the implications of these findings for the theory and practice of communicating anger in intergroup conflicts to increase outgroup empathy.

Keywords

anger consensus empathy intergroup conflict procedural fairness

Peace requires something far more difficult than revenge or merely turning the other cheek; it requires empathizing with the fears and unmet needs that provide the impetus for people to attack each other.

Marshall Rosenberg (2005a, p. 129)

As the previous quote suggests, empathizing with an outgroup is a first step toward more harmonious intergroup relations (e.g., Batson et al., 1997; Pettigrew & Tropp, 2008). Specifically, we view outgroup empathy as a combination of a concern for the outgroup’s welfare (i.e., empathic concern) and the ability to take their perspective (i.e., perspective taking), which leads to more positive attitudes and motivation to act toward them (Batson, 1990, 2009; Batson et al., 1997; Dovidio et al., 2004; Stephan & Finlay, 1999; Stephan & Renfro, 2002). As such, increasing outgroup empathy might be an important first step toward reducing intergroup conflict (Stephan, 2008). However, increasing outgroup empathy in intergroup conflicts is often easier said than done (Gill, Andreychik, & Getty, 2013), particularly when such conflicts include high stakes and a long history of contention (e.g., Halperin, Russell, Dweck, & Gross, 2011; Shnabel & Nadler, 2008) that involve all kinds of psychological reasons for not empathizing beforehand (e.g., structural blaming of the other group). In this article, we therefore focus on lower stake intergroup conflicts that enable us to more clearly and cleanly examine an essential antecedent of increasing outgroup empathy in intergroup conflict: The communication of group-based anger.

Indeed, the communication of group-based anger has been found to increase outgroup empathy because it serves a distinctly relational function (De Vos, van Zomeren, Gordijn, & Postmes, 2013). That is, among the broad spectrum of meanings and intentions potentially conveyed by group-based anger, a central element of meaning through its communication is a relational signal (a) of unfair treatment by the outgroup (van Zomeren, Spears, Fischer, & Leach, 2004; see also Frijda, 1986; Lazarus, 1991; Roseman, 2001; Scherer, 2001; Scherer, Schorr, & Johnstone, 2001), yet (b) emphasizes the importance of maintaining a long-term relationship with the outgroup (van Zomeren, 2016a, 2016b). Especially because of the latter aspect, the communication of group-based anger thus entails that individuals approach the other (Carver & Harmon-Jones, 2009) and implicitly ask the recipient of their anger to empathize with them as well as engage in reconciliatory behavior (and in this sense, group-based anger functions no differently than interpersonal anger; see Fischer & Roseman, 2007).

Although De Vos et al. (2013) showed that the communication of group-based anger, rather than contempt, increased outgroup empathy in recipients, there should of course be boundaries to this positive potential of the communication of group-based anger in intergroup conflict. The main aim of this article is to zoom in on two likely antecedent conditions of the empathy-inducing effects of the communication of group-based anger, which we derive from the emotion and intergroup conflict literatures. First, the communication of group-based anger signals to recipients that the outgroup has been treated unfairly (i.e., procedural unfairness, which is a key appraisal of group-based anger; van Zomeren et al., 2004), which makes anger an appropriate emotion to communicate (e.g., van Kleef, 2009; van Kleef & Côté, 2007). As such, we want to test whether the communication of group-based anger still increases outgroup empathy when the recipient perceives the group to be treated fairly (and thus seemingly lacks an appropriate basis for communicating anger).

Second, we examine whether the communicated anger needs to be perceived as group-based by the recipient (i.e., as consensually shared within the outgroup; see Smith, Seger, & Mackie, 2007), because the communication of group-based anger signals that one wants to maintain the long-term intergroup relationship (De Vos et al., 2013; van Zomeren, 2016a, 2016b). Put differently, we test whether the communication of anger still increases outgroup empathy when the anger is perceived as individual-based, rather than group-based (i.e., as unshared vs. shared by the outgroup as a whole). We report three experiments that put these two antecedent conditions of the positive, empathy-inducing effects of the communication of group-based anger to the test in the context of lower stake intergroup conflicts.

Perceived Procedural Unfairness as an Appropriate Basis for “Their” Anger

According to appraisal theories of emotions (for an overview, see e.g., Scherer et al., 2001), one of the key appraisals of the emotional experience of anger is unfairness (Miller, 2001; van Zomeren et al., 2004), which some refer to as illegitimacy (e.g., Roseman, Spindel, & Jose, 1990). The communication of anger, however, effectively signals the desire to address this sense of unfairness and demands from the other party some form of reconciliation aimed at restoring justice (De Vos et al., 2013). For this reason, we focus on how group members, through communicating their anger, signal to the other group that the treatment they are receiving is unfair (i.e., perceived procedural unfairness), which makes their communication of anger appropriate (van Kleef, 2009; van Kleef & Côté, 2007). Indeed, recipients need to perceive their treatment of the group as unfair in order to consider their anger appropriate (e.g., Shields, 2005).

By contrast, when the treatment of one person by another is perceived to be fair, the communication of anger is perceived as inappropriate and therefore as not requiring any reconciliation (van Kleef, 2009; van Kleef & Côté, 2007). If anything, such an inappropriate display of anger may have detrimental consequences because it can be perceived as a sign of aggression (Lickel, Miller, Stenstrom, Denson, & Schmader, 2006; van Kleef & Côté, 2007). Therefore, for any empathy-inducing effects of the communication of group-based anger to occur, recipients of anger should view it as appropriate and thus based in perceived procedural unfairness (i.e., about unfair treatment). Put differently, in conflicts where group members cannot imagine the other party’s anger to be appropriate because they perceive their treatment of the outgroup as fair, communicating group-based anger will not have the positive, empathy-inducing effects it would otherwise have.

Perceived Outgroup Consensus as “Their” Basis for Group-Based Anger

Just as perceived procedural unfairness is a key antecedent of the empathy-inducing effects of the communication of group-based anger, a second antecedent is the perceived consensus in the outgroup about the anger communicated (Smith et al., 2007). Indeed, the very notion of group-based anger implies a sharedness of anger within the group, such that recipients of that anger will realize it is not just a single individual in the group that feels angry. Thus, when people believe the outgroup as a whole to communicate anger, it implies perceived outgroup consensus on this dimension (Smith et al., 2007).

As a consequence, and in line with the presumed relational function of anger (De Vos et al., 2013; Fischer & Roseman, 2007), the communication of group-based anger should be perceived by recipients as a signal that the outgroup as a whole cares about maintaining a positive intergroup relationship and therefore encourage recipients to empathize with them. By contrast, if the communicated anger is not perceived as shared within the group (Smith et al., 2007), then one is faced with a lone angry voice that can be easily ignored or avoided. We therefore test whether the communication of anger still increases outgroup empathy in recipients if there is low perceived outgroup consensus.

The Current Research

We report the three experiments we ran to test whether perceived procedural unfairness and outgroup consensus reflect antecedent conditions of the positive, empathy-inducing effects of the communication of group-based anger in lower stakes intergroup conflicts. In Experiments 1 and 2 we therefore test whether an absence of perceived procedural unfairness or outgroup consensus still leads to these effects. However, Experiments 1 and 2 do not include both factors in one design and thus cannot tell us much about the interrelationship between these two antecedent conditions. For this reason, Experiment 3 manipulates both variables in an explorative fashion to test whether each antecedent is sufficient to produce any positive, empathy-inducing effects of the communication of group-based anger.

Across the three studies, we employed an experimental approach to allow causal inferences from the findings. We thereby made deliberate choices in the experimental designs to increase experimental control, while keeping the contexts as real and believable for our participants as possible. All three experiments made use of actual lower stakes intergroup conflicts, which included a clear intergroup differentiation (and thus the potential for a manipulation of perceived outgroup consensus), and a clear dimension of anger communication (and thus the potential for a manipulation of procedural unfairness). Intergroup differentiation was clearly visible in Experiment 1 where we used an intergroup conflict between university students and higher vocational education students; in Experiment 2 where we used an intergroup conflict between university students and local citizens in a university town; and in Experiment 3 where we used an intergroup conflict between dog owners and non-dog owners. Furthermore, these intergroup conflicts entailed a mistreatment of one group over the other, suggesting that perceived procedural unfairness would be relevant with respect to whether the communication of group-based anger would be appropriate.

Finally, we note that across the three studies, we also varied the medium through which group-based anger was communicated, either through mass media (Experiment 1), through statements of outgroup members (Experiment 2), or an imagined direct encounter with the communicator (Experiment 3). We purposefully chose indirect ways of communicating group-based anger because group-based anger is often not encountered in direct contact with an outgroup member, but through indirect channels such as mass media (Halualani, Chitgopekar, Morrison, & Dodge, 2004; Hargrave & Livingstone, 2009; Marsden, 1987). For all of these reasons, the experiments were designed to have both adequate internal and external validity.

Experiment 1: Perceived Procedural Unfairness

Method

Participants were 51¹ first-year psychology students from the University of Groningen (38 women; M_age = 21.10; SD = 1.47) who participated for course credit. Participants were randomly assigned to one of four conditions in a 2 (Communication of Group-Based Anger: Present/Absent) x 2 (Perceived Treatment: Unfair/Fair) between-subjects design.

Upon entering the lab, participants were given a paper questionnaire consisting of a short, ostensibly real newspaper article, and a series of questions related to the article. The article described the plans of the Dutch government to make education cuts and reduce student funding. University students reacted to this by presenting a petition with 5,000 signatures to the state secretary in which they pleaded for a specific approach to applying these cuts to reduce study financing. In this petition they argued that students at universities were entitled to higher funding than students at higher vocational education (VE) institutions, because university tuition fees are higher, and because people with a university diploma are scarcer and therefore more valuable in the job market. In the fair treatment condition, university students argued that VE students should in the future get 95% of the funding of university students (i.e., almost the same as the ingroup), whereas in the unfair treatment condition the percentage was only 50% (i.e., much less).

Subsequently, the manipulation of the communication of group-based anger took place. A VE student, Michel de Vries, reacted to this treatment by either specifically communicating anger (“This treatment by university students makes me really angry”) or by communicating a general disapproval (but no specific emotion; “I disapprove of this unfair treatment by university students”), which was specifically mentioned to be on behalf of the group (and thus was group-based). Additionally, the communicated anger, or lack thereof, was manipulated in the title of the article (e.g., “VE students angry over/disapprove of treatment by university students”).

Each item in the questionnaires used in the three experiments was, unless mentioned otherwise, measured on 7-point Likert-type scales (1 = absolutely not, 7 = very much). Two manipulation check items verified the effectiveness of the anger manipulation (“To what extent did VE students express anger?”), and the procedural fairness manipulation (“To what extent did you perceive VE students to be treated fairly?”). Outgroup empathy was measured with eight items, four of which tapped into empathic concern (i.e., “I empathize with/feel sorry for/feel compassion for/could not care less for VE Students such as Michel de Vries,” with the last item reverse-coded), and four items into perspective taking (i.e., “I am able to take the perspective of/I can easily place myself in the shoes of/I don’t find it difficult in this case to take the perspective of/I completely understand the reaction of VE students such as Michel”). A principal axis factor analysis with oblique rotation extracted only one main factor that explained 52.55% of the variance with an eigenvalue of 4.57. A small second factor explained only 9.59%, with an eigenvalue of 1.04, with inspection of the scree plot suggesting one main factor. Furthermore, inspection of the factor loadings did not support a differentiation between empathic concern and perspective taking. We therefore combined all eight items into a highly reliable outgroup empathy scale (α = .88).²

Results and Discussion

Manipulation checks

A 2 x 2 analysis of variance (ANOVA) on the manipulation check of perceived anger revealed only the expected main effect of anger, F(1, 47) = 13.49, p = .001, η_p² = .22. VE students were indeed perceived to be angrier in the anger conditions (M = 5.35, SD = 1.09) than in the no-anger conditions (M = 3.88, SD = 1.74). The main effect of procedural unfairness, F(1, 47) = 2.48, p = .122, and the interaction effect, F(1, 47) = 0.06, p = .813, were not significant.

The manipulation check of procedural unfairness revealed only the expected main effect of procedural unfairness, F(1, 47) = 13.90, p < .001, η_p² = .23: Participants perceived the treatment to be more fair in the fair treatment conditions (M = 4.20, SD = 1.56) than in the unfair treatment conditions (M = 2.65, SD = 1.38). The main effect of anger, F(1, 47) = 0.91, p = .344, and the interaction effect, F(1, 47) = 0.11, p = .736, were not significant. Thus, both manipulations were effective.

Outgroup empathy

A 2 x 2 ANOVA on outgroup empathy showed a significant main effect of procedural unfairness, F(1, 47) = 16.61, p < .001, η_p² = .26. Participants empathized more with the outgroup when they perceived their treatment to be unfair (M = 5.73, SD = 0.82) rather than fair (M = 4.92, SD = 0.64). Importantly, this main effect was qualified by the predicted significant two-way interaction effect, F(1, 47) = 7.67, p = .008, η_p² = .14 (see Figure 1).

Figure 1.

Outgroup empathy towards vocational education students as a function of the communication of group-based anger and perceived procedural unfairness manipulations, Experiment 1.

As expected, simple main effect analyses showed that when the treatment of the outgroup was perceived as unfair, significantly more outgroup empathy was felt after the communication of anger (M = 6.01, SD = 0.63) compared to no anger (M = 5.44, SD = 0.91), F(1, 47) = 4.31, p = .043. Yet, when treatment of the outgroup was perceived as fair, a marginally significant effect was found so that even somewhat less outgroup empathy was felt when anger was communicated (M = 4.67, SD = 0.59) compared to no anger (M = 5.19, SD = 0.61), F(1, 47) = 3.40, p = .071. Further simple main effects analyses showed that when the outgroup communicated anger, participants felt more outgroup empathy when the treatment was perceived as unfair compared to fair, F(1, 47) = 23.92, p < .001. Importantly, when they did not communicate anger, there was no significant difference in outgroup empathy between the fairness and unfairness conditions, F(1, 47) = 0.84, p = .37.

Discussion

The Experiment 1 results supported our hypothesis that the communication of group-based anger increased outgroup empathy only under conditions of perceived unfair treatment of the outgroup. This suggests that perceived procedural unfairness may be a sufficient antecedent condition for the positive, empathy-inducing effects of the communication of group-based anger (De Vos et al., 2013), at least within the lower stakes intergroup conflict that we studied. Moreover, our findings are in line with van Kleef and Côté (2007) who showed that appropriate displays of anger lead the other to give in more easily in negotiation settings. Our results suggest that appropriate displays of group-based anger, such as those based in procedural unfairness, can increase outgroup empathy in intergroup conflicts as well. Finally, and in line with the relational function of the communication of group-based anger (De Vos et al., 2013; van Zomeren, 2016a, 2016b), in case of perceived procedural fairness, the communication of anger appeared ineffective, and, if anything, to backfire.

However, conceptually there is certainly more to the communication of group-based anger than just signaling perceived procedural unfairness. Experiment 2 therefore tested a second antecedent condition for the positive, empathy-inducing effects of communicating group-based anger, which is perceived outgroup consensus. Indeed, because an important function of the communication of group-based anger is to emphasize the importance of maintaining a positive intergroup relationship (De Vos et al., 2013; van Zomeren, 2016a, 2016b), we hypothesized that only when there was a perceived high outgroup consensus regarding group-based anger, its communication would increase outgroup empathy.

Experiment 2: Perceived Outgroup Consensus

Method

Participants were 51 first-year psychology students from the University of Groningen (32 women; M_age = 21.69; SD = 2.43) who participated for course credit. Participants were randomly assigned to one of four conditions in a 2 (Communication of Anger: Present/Absent) x 2 (Perceived Outgroup Consensus: High/Low) between-subjects design. Participants were told that the experiment was about the conflict between students and “Stadjers” (a term referring specifically to city-born inhabitants of Groningen, or “locals”). Groningen is a medium-sized city with a university founded in 1614, and similar to many inner-city universities where students have historically formed a sizable and visible minority among the local population, it has a pedigree of “town and gown” tensions between students and locals (which in centuries past occasionally erupted into physical violence and even homicide). Although contemporary conflicts, fortunately, tend to be mild by comparison, there are still pervasive tensions that are extensively covered in the media and well known to all involved.

In the experiment, participants were first informed about the ongoing conflict between students and locals in Groningen, and were subsequently told that they would be reading ostensibly real statements of two locals (Alfred and Tanja) talking about their experience of Groningen students causing excessive noise, littering, and other forms of antisocial behavior in their neighborhood (which signal a mistreatment of other people in the city). The statements either contained a specific communication of anger (e.g., “I am very angry that students in Groningen behave in such an antisocial way”), or a general disapproval (but no specific emotion; “Students in Groningen behave in an antisocial way”). Perceived outgroup consensus was manipulated by including a final statement that research conducted by the university showed that 85% (high outgroup consensus) or 15% (low outgroup consensus) of locals share this emotion (whereby “emotion” was replaced by “anger” or “disapproval,” depending on the specific condition).

We used the same manipulation check item as in Experiment 1 for the anger manipulation. Additionally, the manipulation of perceived outgroup consensus was checked with a dichotomous item asking participants to indicate the extent to which they perceived low (0) or high (1) consensus concerning the message that the locals had communicated. Outgroup empathy was measured with eight items similar to those in Experiment 1 (yet with both empathic concern and perspective-taking items referring to Stadjers), from which a principal axis factor analysis with oblique rotation again extracted one main factor that explained 39.69% of the variance with an eigenvalue of 3.61, and a smaller second factor that explained an additional 9.40% with an eigenvalue of 1.38, which a scree plot suggested to be potentially meaningful. However, as in Experiment 1, inspection of the factor loadings did not show support for two different empathic concern and perspective-taking subfactors. Therefore, as in Experiment 1, we decided to collapse these eight items into a highly reliable outgroup empathy scale (α = .89).

Results and Discussion

Manipulation checks

A 2 x 2 ANOVA on the manipulation check of perceived anger revealed the expected main effect of anger, F(1, 47) = 12.34, p < .001, η_p² = .21. Stadjers were indeed perceived to be angrier in the anger conditions (M = 5.50, SD = 1.98) than in the no-anger conditions (M = 3.68, SD = 1.68). The main effect of consensus, F(1, 47) = 1.01, p = .321, and the interaction effect, F(1, 47) = 0.001, p = .976, were not significant. Furthermore, a logistic regression on the manipulation check of consensus revealed the expected main effect of consensus, χ²(1) = 19.45, p < .001: Participants perceived the group of Stadjers to have more consensus about the message in the high-consensus conditions (M = 0.92, SD = 0.27) than in the low-consensus conditions (M = 0.36, SD = 0.49). The main effect of anger, χ²(1) = 0.23, p = .629, and the interaction effect, χ²(1) = 1.96, p = .161, were not significant. Thus, both manipulations were effective.

Outgroup empathy

A 2 x 2 ANOVA on outgroup empathy showed a significant two-way interaction effect, F(1, 47) = 5.43, p = .024, η_p² = .10 (see Figure 2). As expected, simple main effect analyses showed that when there was high outgroup consensus, participants felt significantly more outgroup empathy after the communication of anger (M = 5.03, SD = 0.65) compared to when no anger was communicated (M = 4.09, SD = 0.67), F(1, 47) = 9.54, p = .003. Yet, when there was low outgroup consensus, there were no differences in outgroup empathy between the anger condition (M = 4.21, SD = 0.61) and the no-anger condition (M = 4.28, SD = 1.10), F(1, 47) = 0.06, p = .82. Further comparisons of conditions showed that when the outgroup communicated anger, participants felt more outgroup empathy when perceived outgroup consensus was high versus when it was low, F(1, 47) = 7.33, p = .009. When they did not communicate anger, however, there was no difference in felt outgroup empathy between the two perceived outgroup consensus conditions, F(1, 47) = 0.37, p = .55. Thus, participants empathized with the outgroup most when they communicated anger and when there was high perceived outgroup consensus—a pattern confirmed by testing the appropriate 3-versus-1 contrast, F(3, 47) = 3.92, p =.014.

Figure 2.

Outgroup empathy towards “Stadjers” as a function of the communication of anger and perceived outgroup consensus manipulations, Experiment 2.

Discussion

The Experiment 2 findings provided support for our hypothesis that the communication of group-based anger has positive, empathy-inducing effects only when there is high perceived outgroup consensus concerning its communication. This fits with the idea that perceived outgroup consensus emphasizes the group-based nature of the anger communicated. That is, high perceived outgroup consensus turns the communication of anger from a lone angry voice into an intergroup signal that functions as a relational “push” for the other group to empathize with the outgroup. Indeed, it signals that the group feels wronged, rather than a lone individual voice within that group that can easily be dismissed or ignored.

Taken together, Experiment 1 and 2 supported the idea that perceived procedural unfairness and outgroup consensus reflect two antecedent conditions of the positive, empathy-inducing effects of the communication of group-based anger in lower stakes intergroup conflicts. We therefore designed Experiment 3 to explore whether each antecedent condition is a sufficient explanation of when anger has empathy-inducing effects. Because of this aim and because we wanted to retain a parsimonious design of the experiment, we chose to drop the conditions used in Experiments 1 and 2 in which no anger was communicated, thus holding constant the communication of group-based anger. We further chose a third lower stake intergroup conflict context (i.e., dog owners vs. non-dog owners) in order to increase external validity.

Experiment 3: Perceived Procedural Unfairness and Outgroup Consensus

Method

Participants were 56 dog owners from Groningen (36 women; M_age = 42.16; SD = 13.12). Participants were randomly assigned to one of four conditions in a 2 (Perceived Treatment: Fair/Unfair) x 2 (Perceived Outgroup Consensus: Low/High) between-subjects design. Participants were approached at a dog training school and asked to read a scenario and fill out a short questionnaire.

The introduction to the scenario read that the local government had passed legislation that dogs were allowed to walk off-leash in designated areas only. Participants were asked to imagine themselves in a scenario in which their dog was roaming free in the park. Perceived fair treatment was manipulated by stating that this occurred in either an off-leash area or one of the newly designated on-leash areas. We chose this particular manipulation because it signals that this dog owner either conforms or does not conform to a rule that presumably protects others from any misbehavior by the owner’s unleashed dog. We reasoned that not conforming to the rule will be more strongly perceived as procedurally unfair, that is, a perception of unfairness based in how people mistreat each other, as compared to conforming to the rule.

Indeed, participants subsequently read that their dog spotted another person, enthusiastically bounded towards the other and jumped up on its hind legs in a nonaggressive way. In response to the dog jumping up, the person responded in anger by saying: “Keep your dog leashed! It really makes me so angry when this happens!” In the fair treatment condition, the dog is walking in an off-leash area and is therefore fairly roaming free, which was intended to make the communicated anger seem less appropriate. In the unfair condition, the dog is running around in an on-leash area and is therefore clearly in the wrong, which was intended to make the communicated anger seem more appropriate.

Furthermore, participants read a statement saying that these situations occur more often and that, given this exact situation, recent studies have shown 25% (low outgroup consensus) or 75% (high outgroup consensus) of non-dog owners to agree with the anger communicated by the person in the scenario. We used the same manipulation check items for perceived procedural unfairness and outgroup consensus as in Experiments 1 and 2. Outgroup empathy was again measured with eight³ items, which were again factor analyzed through principal axis factoring with oblique rotation, resulting again in one main factor, explaining 56.19% of the variance with an eigenvalue of 4.80. As in Experiments 1 and 2, there was also a smaller second factor, explaining 9.69% of the variance with an eigenvalue of 1.09, but unlike Experiments 1 and 2, inspection of the factor loadings this time supported different empathic concern (four items) and perspective-taking (four items) subscales. For comparison purposes with Experiments 1 and 2, we nevertheless created one highly reliable outgroup empathy scale (α = .90), while reporting findings for the two reliable subscales (α = .89 and .84 for empathic concern and perspective taking, respectively) in Endnote 4.

Results and Discussion

Manipulation checks

A 2 x 2 ANOVA on the manipulation check of perceived procedural unfairness revealed only the expected main effect of procedural unfairness, F(1, 52) = 5.13, p = .028, η_p² = .09. Participants perceived the non-dog owner to be treated more unfairly (because of the unleashed dog’s behavior) in the on-leash conditions (M = 3.39, SD = 1.07) than in the off-leash conditions (M = 2.61, SD = 1.42). The main effect of perceived outgroup consensus, F(1, 52) = 0.18, p = .68, and the interaction effect, F(1, 52) = 0.22, p = .64, were not significant.

Furthermore, a logistic regression on the dichotomous manipulation check of perceived outgroup consensus revealed only the expected main effect of consensus, χ²(1) = 12.34, p < .001. Participants perceived more outgroup consensus in the high-consensus conditions (M = 0.85, SD = 0.37) than in the low-consensus conditions (M = 0.40, SD = 0.50). The main effect of procedural fairness, χ²(1) < 0.00, p = 1.00, and the interaction effect, χ²(1) = 0.55, p = .456, were not significant. Thus, both manipulations were effective.

Outgroup empathy

A 2 x 2 ANOVA on the outgroup empathy scale showed a significant two-way interaction effect, F(1, 52) = 4.68, p = .044, η_p² = .08 (see Figure 3). Simple main effect analyses further showed that when the outgroup was treated fairly, dog owners felt more outgroup empathy when there was high perceived outgroup consensus (M = 4.61, SD = 0.77) than when there was low perceived outgroup consensus (M = 3.58, SD = 0.83), F(1, 52) = 6.72, p = .026. Moreover, when the treatment of non-dog owners was unfair, no differences in outgroup empathy were found; low perceived consensus, M = 4.67, SD = 1.56; high perceived consensus, M = 4.41, SD = 1.07; F(1, 52) = 0.349, p =.557. Another way of summarizing this pattern of results is that only in the case of fair treatment and low perceived outgroup consensus did the communication of anger lead to the lowest level of outgroup empathy, as tested by the appropriate 3-versus-1 contrast: F(3, 52) = 7.48, p = .009, which suggests that each antecedent condition seems sufficient.⁴

Figure 3.

Outgroup empathy towards non-dog-owners after the communication of anger, as a function of the perceived procedural unfairness and perceived outgroup consensus manipulations, Experiment 3.

Discussion

Whereas Experiments 1 and 2 supported the idea that perceived procedural unfairness and outgroup consensus are two antecedent conditions for the positive, empathy-inducing effect of the communication of group-based anger, the results from Experiment 3 explored their interrelationship. Its findings suggested that both antecedent conditions seem sufficient, rather than necessary, conditions for increasing outgroup empathy as a consequence of the communication of group-based anger in lower stakes intergroup conflicts.

General Discussion

Three experiments employing different lower stakes intergroup conflicts showed support for our two hypotheses that perceived procedural unfairness and outgroup consensus are antecedent conditions of the positive, empathy-inducing effects of the communication of group-based anger. Specifically, the results of Experiment 1 showed that the communication of group-based anger was only effective in increasing outgroup empathy when treatment of the outgroup was perceived as unfair. In fact, consistent with research on the effects of inappropriate anger displays (van Kleef & Côté, 2007), when treatment of the outgroup was perceived as fair, the communication of group-based anger was ineffective or, if anything, even back-fired. Furthermore, the results of Experiment 2 showed that the communication of group-based anger was only effective in increasing outgroup empathy when participants perceived high outgroup consensus. In line with the relational function of group-based anger (De Vos et al., 2013; van Zomeren, 2016a, 2016b), only perceived outgroup consensus regarding group-based anger (compared to a more general disapproval) increased outgroup empathy. And finally, results from Experiment 3 indicated that either antecedent condition seemed sufficient, but not necessary, for the communication of group-based anger to increase outgroup empathy. This suggests that, at least in lower stake intergroup conflicts, communicating group-based anger may be a promising way to increase outgroup empathy.

Theoretical and Practical Implications

The current findings replicate, qualify, and nuance previous work on the empathy-inducing effects of the communication of group-bases anger (De Vos et al., 2013). Specifically, they replicate the positive, empathy-inducing effect of communicating group-based anger across different lower stakes intergroup conflicts, which increases both the internal and external validity of our findings. Moreover, it importantly qualifies De Vos et al.’s findings by outlining two antecedent conditions that derive from a relational perspective on the communicative function of group-based anger (De Vos et al., 2013; van Zomeren, 2016a, 2016b). Specifically, the main reason for why the communication of group-based anger should increase outgroup empathy in the first place is that it (a) stresses perceived procedural unfairness, and (b) emphasizes the importance of maintaining a positive intergroup relationship (De Vos et al., 2013; van Zomeren, 2016a, 2016b). Thus, if the treatment of the outgroup is perceived as fair, and if there is low outgroup consensus, the communication of group-based anger should be ineffective in increasing outgroup empathy. The current experiments provide support for this line of reasoning by showing that both antecedent conditions facilitate the empathy-inducing effects of the communication of group-based anger, at least within the lower stakes intergroup conflicts we studied.

Conceptually, the communication of group-based anger reflects a social call for attention to an intergroup relationship at risk, which communicates both a perceived unfair treatment by the outgroup as well as a motivation to maintain a long-term positive relationship (De Vos et al., 2013; Fischer & Roseman, 2007; van Zomeren, 2016a, 2016b; see also Rosenberg, 2005b). Indeed, as Bowlby (1973, p. 278) already put it decades ago: “Anger acts to promote, and not to disrupt, the bond.” As such, the communication of anger implicitly invites the recipient to empathize with the outgroup, which is often viewed as an important first step toward intergroup conflict reduction (Pettigrew & Tropp, 2008). This also fits with an increased recognition of anger as an approach emotion (Carver & Harmon-Jones, 2009), and with Lazarus’s (1991) notion of the relational meaning of anger (i.e., a demeaning offense against me or mine). Our perspective thus uniquely combines relational and approach-oriented aspects of group-based anger to examine its communicative effects.

As such, the current work is different from theory and research about the instrumental function of communicating anger (van Kleef, 2009) because that line of work focuses on the strategic use of anger in conflict settings (i.e., anger as an indicator of higher negotiation limits with the aim of forcing the recipient to cooperate). A strategic function of anger entails utilizing anger (felt or not) for personal gain, whereas the relational function of anger entails tailoring the communication of group-based anger for shared gain. Future research should consider integrating these different aspects of the communication of the very same emotion in interpersonal and intergroup conflicts.

We note that Experiment 3 was more explorative in nature and its findings therefore may warrant further discussion and nuance. Specifically, the pattern of results indicates how the two antecedents interact with each other, but there may be several explanations for why, for instance, perceived procedural unfairness did not seem to require perceived outgroup consensus with respective to the positive, empathy-inducing effects of the communication of anger. First, in line with approaches like moral foundation theory (Graham et al., 2011; Haidt & Graham, 2007), it is possible that the communicated messages might have influenced the recipients’ perception that they were somehow responsible for the harm done to the outgroup communicating the anger. In this sense, regardless of the presence or absence of its group-based nature, the communication of anger may induce outgroup empathy because individuals perceive a moral violation by the ingroup (independent of whether the anger comes from a group or a lone voice). However, not all perceptions of unfairness have moral underpinnings (van Zomeren, 2013) and one can wonder whether the current lower stakes intergroup conflicts touched upon violations of core moral principles. Future research should test this explanation in more detail.

A second explanation revolves around the interpretation of perceived outgroup consensus. Experiment 3 showed that the communication of anger increased outgroup empathy when there was perceived outgroup consensus, even when treatment of the outgroup was perceived as fair. In our view, this could be because even though participants may have believed that the outgroup was treated fairly, a collective plea for reconciliation obligates one to feel outgroup empathy in order to maintain a positive intergroup relationship (i.e., it is a relational “push”; van Zomeren, 2016a). In fact, the absence of an interaction effect of the manipulations on the unfairness manipulation check may actually point to this explanation, but clearly future research is needed to substantiate this explanation.

A final implication of our findings refers to the type of intergroup conflicts we studied in the three experiments. Although we made use of different groups and issues, we intentionally focused on relatively lower stakes intergroup conflicts (for instance, as compared to high-stakes conflicts such as the Israeli–Palestinian conflict; see Halperin et al., 2011). It is doubtful, but as far as we know still an open empirical question, whether the current findings will generalize to high-stakes intergroup conflicts. Indeed, in such contexts, there may simply be too much defensiveness associated with the very notion of empathizing with the outgroup. Nevertheless, even in those contexts research has shown that, for example, the emotional experience of anger toward the outgroup has more positive implications for resolving the conflict than, for instance, the emotional experience of hatred (Halperin et al., 2011). Future research should thus examine whether the communication of group-based anger will help or hinder outgroup empathy in high-stakes intergroup conflicts.

Limitations and Directions for Future Research

The current set of studies has at least three important limitations. First, all three experiments relied on experimental manipulations based in vignettes. Although results have been fairly consistent across several contexts and populations (see also De Vos et al., 2013), it is unclear whether these findings would generalize to more natural interactions (e.g., face-to-face interactions or computer-mediated communication involving anger). As mentioned in the introduction, however, this is not necessarily a weak point of the current experiments, as the type of contact people most often have with outgroup members is through second-hand information such as mass media outlining intergroup attitudes and behavior (Hargrave & Livingstone, 2009). Future research is nevertheless needed to test whether the current findings could also be translated to real-time interactions in which anger is communicated.

Second, all three experiments admittedly had relatively low sample sizes, which can potentially pose problems for the robustness of our key findings. Future research should certainly seek to replicate our findings with larger samples. We examined the severity of this limitation by specifically comparing the effect sizes of the manipulation of the communication of group-based anger in Experiment 1 (under conditions of perceived procedural unfairness; Cohen’s d = 0.78) and Experiment 2 (under conditions of perceived outgroup consensus; Cohen’s d = 1.25) to those reported in Experiments 1 and 2 in De Vos et al. (2013; with ds of 0.73 and 1.42). This comparison suggests that the effects obtained in the current experiments seem equal in size as those in De Vos et al. (2013). This conclusion was further corroborated by a power analysis in which we used the average effect size (1.075) in the two studies in De Vos et al. (2013) as input, together with a β/α ratio of 4, and assuming 12 participants per condition (which in total, for the full design of each study, is 48—note that we had 51 participants in each of those two studies). A priori power to detect such an effect for each study is then .81. Taken together, this pattern of results suggests that larger samples would have been ideal, but perhaps not necessary.⁵

A third limitation is that the relatively small sample sizes make factor analyses less reliable; that is, a smaller sample size makes it more difficult to empirically differentiate between factors. This may be one reason for why the outcomes of our factor analyses across the studies were inconsistent, both in terms of the number of meaningful factors extracted and whether the pattern of factor loadings supported a conceptual distinction between perspective taking and empathic concern. This inconsistency led us to feel on safer ground by interpreting our findings in terms of the more general notion of outgroup empathy, rather than for arguing for a differentiation between empathic concern and perspective-taking aspects of outgroup empathy (that we could empirically differentiate only in Experiment 3). Nevertheless, future research can focus more specifically on these different aspects of outgroup empathy to examine which one(s) may be driving the effects.

Finally, it is important to note that, as indicated in Endnote 2, we did not find any effects of the communication of group-based anger in the current set of studies on a variety of conflict intention items, which runs counter to the findings reported in De Vos et al. (2013). This means that although the communication of group-based anger has positive, empathy-inducing effects, this does not always translate into intentions related to the conflict. Future research should examine these differential psychological consequences of outgroup empathy, and identify potential moderators of the relationship between outgroup empathy and conflict intentions.

We further welcome research on other antecedent conditions under which the communication of group-based anger increases outgroup empathy. Indeed, we certainly do not expect the communication of group-based anger to always have positive, empathy-inducing effects and therefore it is important to identify other relevant boundary conditions. For instance, it is possible that the communication of group-based anger could result in a host of other emotional responses such as pride and guilt (e.g., Harth, Kessler, & Leach, 2008) or decrease the perception of suffering of the outgroup (Gill et al., 2013). To us, such more in-group defensive responses seem particularly likely when contempt, hatred, or a mixture of those emotions is communicated, rather than “pure” anger. Indeed, only the latter was found to increase outgroup empathy and de-escalate intergroup conflict in previous work (De Vos et al., 2013). In higher stakes intergroup conflicts, we suspect that it may be much harder to communicate group-based anger in such a “pure” way and that it is likely to be perceived as infused with other, more detrimental emotions such as contempt, hatred, or fear (Cottrell & Neuberg, 2005; Halperin et al., 2011), leading to more defensive responses that are likely to increase, rather than decrease, intergroup conflict. However, this interesting line of thought is obviously speculative at present and thus invites future empirical tests.

Conclusion

The findings of three experiments point to the importance of communicating group-based anger to increase outgroup empathy in intergroup conflict, which may be a first step toward conflict reduction (Pettigrew & Tropp, 2008). Importantly, however, we identified two antecedent conditions for this effect: perceived procedural unfairness (i.e., mistreatment of the outgroup) and perceived outgroup consensus (i.e., whether the outgroup as a whole is angry). More exploratively, we further found that the presence of either may be sufficient to facilitate the empathy-inducing effects of the communication of group-based anger. Although we certainly do not want to overinterpret our data given the limitations of our studies, we do think this is quite a hopeful message. If peace indeed requires empathy, as Marshall Rosenberg (2005a) was quoted as suggesting at the start of this article, then our research adds that the communication of group-based anger can sometimes be an effective way to increase outgroup empathy.

Footnotes

Funding

This research was financially supported by a Netherlands Organization for Scientific Research (NWO) Grant 432-08-026, awarded to Ernestine H. Gordijn, Martijn van Zomeren, and Tom Postmes. All materials and data related to this article are deposited in a secure server at the University of Groningen in line with the university’s data storage protocol and are available upon request.

Notes

References

Batson

C. D.

(1990). How social an animal: The human capacity for caring. American Psychologist, 45, 336–346. doi:10.1037/0003-066x.45.3.336

Batson

C. D.

(2009). These things called empathy: Eight related but distinct phenomena. In Decety

Ickes

(Eds.), The social neuroscience of empathy (pp. 3–15). Cambridge, MA: MIT Press.

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

Bowlby

(1973). Attachment and loss: Volume II. Separation. New York, NY: Basic Books.

Carver

C. S.

Harmon-Jones

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

Cohen

(1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.

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). The communication of “pure” group-based anger reduces tendencies toward intergroup conflict because it increases outgroup empathy. Personality and Social Psychology Bulletin, 39, 1043–1052. doi:10.1177/0146167213489140

Dovidio

J. F.

Ten Vergert

Stewart

T. L.

Gaertner

S. L.

Johnson

J. D.

Esses

V. M.

. . . Pearson

A. R.

(2004). Perspective and prejudice: Antecedents and mediating mechanisms. Personality and Social Psychology Bulletin, 30, 1537–1549. doi:10.1177/0146167204271177

10.

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

11.

Frijda

N. H.

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

12.

Gill

M. J.

Andreychik

M. R.

Getty

P. D.

(2013). More than a lack of control: External explanations can evoke compassion for outgroups by increasing perceptions of suffering (independent of perceived control). Personality and Social Psychology Bulletin, 39, 73–87. doi:10.1177/0146167212460281

13.

Graham

Nosek

B. A.

Haidt

Iyer

Koleva

Ditto

P. H.

(2011). Mapping the moral domain. Journal of Personality and Social Psychology, 101, 366–385. doi:10.1037/a0021847

14.

Haidt

Graham

(2007). Beyond beliefs: Religions bind individuals into moral communities. Personality and Social Psychology Review, 14, 140–150. doi:10.1177/1088868309353415

15.

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

16.

Halualani

R. T.

Chitgopekar

A. S.

Morrison

J. H. T. A.

Dodge

P. S.-W.

(2004). Diverse in name only? Intercultural interaction at a multicultural university. Journal of Communication, 54, 270–286. doi:10.1093/joc/54.2.270

17.

Hargrave

A. M.

Livingstone

S. M.

(2009). Harm and offence in media content: A review of the evidence. Bristol, UK: Intellect.

18.

Harth

N. S.

Kessler

Leach

C. W.

(2008). Advantaged group’s emotional reactions to intergroup inequality: The dynamics of pride, guilt, and sympathy. Personality and Social Psychology Bulletin, 34, 115–129. doi:10.1016/j.jenvp.2012.12.005

19.

Lazarus

R. S.

(1991). Emotion and adaptation. New York, NY: Oxford University Press.

20.

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

21.

Marsden

P. V.

(1987). Core discussion networks of Americans. American Sociological Review, 52, 122–131. doi:10.2307/2095397

22.

Miller

D. T.

(2001). Disrespect and the experience of injustice. Annual Review of Psychology, 52, 527–553. doi:10.1146/annurev.psych.52.1.527

23.

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

24.

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). New York, NY: Oxford University Press.

25.

Roseman

I. J.

Spindel

M. S.

Jose

E. E.

(1990). Appraisal of emotion-eliciting events: Testing a theory of discrete emotions. Journal of Personality and Social Psychology, 59, 899–915. doi:10.1037/0022–3514.59.5.899

26.

Rosenberg

M. B.

(2005a). Speak peace in a world of conflict: What you say next will change your world. Encinitas, CA: PuddleDancer Press.

27.

Rosenberg

M. B.

(2005b). The surprising purpose of anger: Beyond anger management: Finding the gift. Encinitas, CA: PuddleDancer Press.

28.

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). New York, NY: Oxford University Press.

29.

Scherer

K. R.

Schorr

Johnstone

(2001). Appraisal processes in emotion. New York, NY: Oxford University Press.

30.

Shields

S. A.

(2005). The politics of emotion in everyday life: “Appropriate” emotion and claims on identity. Review of General Psychology, 9, 3–15. doi:10.1037/1089–2680.9.1.3

31.

Shnabel

Nadler

(2008). A needs-based model of reconciliation: Satisfying the differential emotional needs of victim and perpetrator as a key to promoting reconciliation. Journal of Personality and Social Psychology, 94, 116–132. doi:10.1037/0022–3514.94.1.116

32.

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. doi:10.1037/0022–3514.93.3.431

33.

Stephan

W. G.

(2008). Psychological and communication processes associated with intergroup conflict resolution. Small Group Research, 39, 28–41. doi:10.1177/1046496407313413

34.

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

35.

Stephan

W. G.

Renfro

C. L.

(2002). The role of threat in intergroup relations. In Mackie

D. M.

Smith

E. R.

(Eds.), From prejudice to intergroup emotions: Differentiated reactions to social groups (pp. 191–207). New York, NY: Psychology Press.

36.

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

37.

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

38.

Van Zomeren

. (2013). Four core motivations to undertake collective action. Social and Personality Psychology Compass, 7, 378–388. doi:10.1111/spc3.12031

39.

Van Zomeren

. (2016a). From self to social relationships: An essentially relational perspective on social motivation. Cambridge, UK: Cambridge University Press.

40.

Van Zomeren

. (2016b). Beyond individual emotion regulation and reconciliation: Toward an essentially relational perspective. Psychological Inquiry, 27, 133–135. doi:10.1080/1047840x.2016.1160761

41.

Van Zomeren

Spears

Fischer

A. H.

Leach

C. W

. (2004). Put your money where your mouth is!: Explaining collective action tendencies through group-based anger and group efficacy. Journal of Personality and Social Psychology, 87, 649–664. doi:10.1037/0022-3514.87.5.649