More Than a Lack of Control

Abstract

People’s explanations for social events powerfully affect their socioemotional responses. We examine why explanations affect emotions, with a specific focus on how external explanations for negative aspects of an outgroup can create compassion for the outgroup. The dominant model of these processes suggests that external explanations can reduce perceived control and that compassion is evoked when negative aspects of an outgroup are perceived as beyond their control. We agree that perceived control is important, but we propose a model in which explanations also affect perceived suffering of an outgroup, and that perceived suffering is an additional mechanism connecting external explanations to compassion. Studies are presented that support our integrative dual-mediation model and that pinpoint factors—depth of cognitive processing, expansive sense of identity—that modulate the extent to which the external explanation/perceived suffering mechanism evokes compassion.

Keywords

social explanation perceived suffering compassion perceived control moral judgment attitudes toward outgroups

Explanation is a core activity of the mind, energized by the need to understand. It plays a central role in social cognition and social interaction. Here, we examine what happens when people explain negative aspects of an outgroup, such as their socioeconomic disadvantage or militant violence. Much existing work reveals how such explanations can be biased by desires for positive ingroup identity (Tajfel & Turner, 1986) or system justification (Jost & Banaji, 1994), with the consequence that explanations focus on character deficiencies of the outgroup. Less work examines whether certain explanations can foster compassion for an outgroup. Of relevance here, Weiner (2006) presented a model describing why certain explanations might evoke compassion. His model emphasizes the controllability implications of explanations. Specifically, targets seen as lacking control over their negative aspects (e.g., Alzheimer’s patients) evoke compassion.

We argue for an additional mechanism connecting explanations to compassion. Our focus is on external explanations—which point to enduring causal forces that surround a group and shape its trajectory “from the outside” (e.g., unfair treatment by dominant groups)—and we propose that such explanations can evoke compassion because of their implications for outgroup suffering. A major contribution of our proposal is to extend attributional models of moral emotions. In so doing, we will challenge a core implication of existing models: That compassion will be withheld from targets seen as having control over their negative aspects. Furthermore, our studies will outline conditions under which our proposed external explanations/suffering mechanism will be especially evocative of compassion.

The Explanatory Default: Outgroup Derogation

Heider (1958) first elucidated the “naïve psychology” of social explanation. Specifically, he described the common-sense view that other people “are affected by their personal and impersonal environment, [and] they cause changes in the environment” (p. 17, italics in original). This distinction is seen in the attribution theory concepts of external causes, which involve personal and impersonal forces in the target’s environment that influence her acts or outcomes, and internal causes, which involve traits and other features that are internal to the target. These concepts have underlain hundreds of studies of self-, interpersonal, and intergroup perception (for reviews, see Bradbury & Fincham, 1990; Buchanan & Seligman, 1995; Gilbert & Malone, 1995; Hewstone, 1990).

In the realm of intergroup perception, research has focused on how explanation selection is biased when people think about outgroups. For example, Allport (1954) noted people’s “insatiable hunger for explanations” and their preference for explanation in terms of “human agency” (p. 170). These dual tendencies, Allport suggested, foster scapegoating, with negative events being “explained” in terms of despicable inner traits of the outgroup (Glick, 2005). Allport’s analysis was prescient, as there now exists abundant support for the notion that people’s default tendency is to “blame the outgroup” for negative events via internal, stereotypic explanations (Hoffman & Hurst, 1990; Jost & Banaji, 1994; Kluegel & Bobo, 1993; Sidanius & Pratto, 1999; Yzerbyt & Rogier, 2001).

This might leave one wondering if external explanations are ever applied to outgroups. They are. But, like internal explanations, they are often strategically deployed to diminish the outgroup (see Hewstone, 1990, for a review). This possibility was first suggested by Taylor and Jaggi (1974) and then incorporated into Pettigrew’s (1979) notion of the ultimate attribution error: People will attribute negative events to internal deficiencies of the outgroup (they failed because they are lazy), whereas positive events will be “explained away” via external explanations (they succeeded because they received abundant governmental assistance).

Beyond Outgroup Derogation: Explanations and Compassionate Responding

Although outgroup derogation is pervasive, people are capable of prosocial responding to outgroups (Doosje, Branscombe, Spears, & Manstead, 1998; Harth, Kessler, & Leach, 2008; Plant & Devine, 1998). Indeed, contrary to the notion of the ultimate attribution error, several studies suggest that people sometimes endorse external explanations for negative aspects of an outgroup (Andreychik & Gill, 2009, 2012; Gill & Andreychik, 2007, 2009; Guimond, Begin, & Palmer, 1989; Iyer, Leach, & Crosby, 2003; Lopez, Gurin, & Nagda, 1998; Zucker & Weiner, 1993). Furthermore, acceptance of such explanations can evoke compassionate responses (Iyer et al., 2003; Zucker & Weiner, 1993), even at the implicit level (Andreychik & Gill, 2012), albeit not always (Andreychik & Gill, 2009). Given this prosocial potential of external explanations, the key question of the present article is as follows: Why do external explanations for negative aspects of an outgroup evoke compassion?

As noted, the reigning answer comes from Weiner (2006). Weiner’s model places perceived control in the mediating role between explanations and emotion. To elaborate, Weiner highlights the fact that different explanations have different controllability implications, with some explanations (cognitive deficit) implying that a target has no control over his negative aspects (academic failures) and other explanations (laziness) implying that a target has ample control. In Weiner’s model, perceived control of negative aspects (he could have avoided this) inhibits compassion, whereas perceived lack of control (it was impossible for him to avoid this) evokes compassion.

Most tests of Weiner’s (2006) model examine responses to individual targets and focus on the distinction between internal-uncontrollable (cognitive deficit) and internal-controllable (laziness) explanations. Thus, that work is not directly relevant for understanding the psychology of external explanations and responses to groups. Studies by Zucker and Weiner (1993), however, are highly relevant. Zucker and Weiner had respondents indicate the extent to which the plight of the poverty-stricken was caused by external forces (e.g., lack of jobs). They also measured perceived control, compassion, and attitudes toward helping the poverty-stricken. Finally, they tested a path model in which external explanations decrease perceived control, perceived lack of control evokes compassion, and compassion increases helpfulness. Support for this model was found across two studies. Based on these studies and others, the importance of perceived control has become a core truth in the literature on social explanation and compassion (Goetz, Keltner, & Simon-Thomas, 2010; Weiner, 2006).

Crucial for present purposes—but not highlighted by Zucker and Weiner (1993)—is the fact that their analyses also revealed substantial direct paths from external explanations to compassion, paths not mediated by perceived control. This raises the possibility that there are additional mechanisms connecting external explanations to compassion. Our focus is on one such mechanism. Specifically, we highlight the fact that external explanations can imply suffering by the outgroup. Our proposed model is (a) external explanations for negative aspects of the outgroup increase the perception that the outgroup has suffered and (b) perceptions of suffering evoke compassion. Below, we elaborate on the three core constructs in our model: external explanations, perceived suffering, and compassion.

In the intergroup literature, external explanations generally refer to causal forces that have surrounded and impinged on the outgroup over an extended period (e.g., enduring exclusion from important societal domains—educational, economic, political—by more dominant groups; Guimond et al., 1989; Iyer et al., 2003; Lopez et al., 1998). This is different from the typical case studied in interpersonal perception, in which external causes are brief events that exist at the time of the target’s action (e.g., she was nervous because he mentioned an uncomfortable topic). Relatedly, in the case that concerns us, what is being explained is not a single act (as is typical in the interpersonal literature) but rather a general pattern associated with the outgroup: socioeconomic disadvantage, high rates of militant violence, and so on. In sum, in the present context, external explanations build a “story”—a coherent narrative of interconnected events—that explains a current negative aspect of the outgroup (socioeconomic disadvantage) in terms of what was done to the outgroup/what the outgroup has experienced (external causes) over an extended period (see Bruner, 2002; Costabile & Klein, 2008; Wyer, Adaval, & Colcombe, 2002; on the role of narratives as a basis of social understanding). The core meaning of an external explanation is that the outgroup per se is not the sole cause of its negative aspects, and thus, the causal role of internal outgroup features should be discounted (Kelley, 1973).

External explanations do not explicitly focus on what a target is feeling. Thus, we conceive of perceived suffering as a potential implication of external explanations that involves “going beyond the information given.” Because perceived suffering is not a necessary consequence of processing external explanations, we expect that inferences about outgroup suffering will be made to varying degrees depending on factors we will outline below. Suffering can involve experiences that are physical (e.g., excessive hunger, illness) or psychological (e.g., hopelessness, powerlessness), and which would be labeled as subjectively unpleasant or painful. We assume that when perceivers consider the types of external causes sufficient to produce salient negative aspects of an outgroup, they can recognize that such causes would likely cause suffering. Indeed, groups do not become mired in poverty or violence because their experiences have been primarily joyful (or even neutral).

Finally, because compassion is an emotion “elicited by the perception of suffering” (Haidt, 2003, p. 862), to the extent that one recognizes the implications for suffering in external explanations, one should experience compassion. Compassion is arguably a distinct emotion (Goetz et al., 2010) that creates tendencies “to help, comfort, or otherwise alleviate the suffering of the other” (Haidt, 2003, p. 862). Unsurprisingly, then, compassion for outgroups has been linked to support for policies aimed at improving their well-being (Iyer et al., 2003; Smith, 2009; Zucker & Weiner, 1993), willingness to volunteer to help the outgroup (e.g., Karaçanta & Fitness, 2006; Omoto, Malsch, & Barraza, 2009), and even an impulse to avoid harming an “enemy” outgroup (Glover, 2000).

The External Explanations/Perceived Suffering Mechanism: Moderators of Its Strength

Beyond establishing the existence of the external explanations/perceived suffering mechanism, we will explore factors that moderate its strength. First, as noted, we conceive of perceived suffering as an implication of external explanations that involves going beyond the information explicitly in the explanation. This implies that superficial acknowledgment of external explanations (“They’ve often been the targets of discrimination”) need not evoke a compassionate response. It is only when one cognitively engages with the explanation that one “sees” the implications for suffering (“Being treated like that must have created a feeling of powerlessness”). Because of this, we predict that implications for suffering will be recognized to a greater degree as a function of individual differences and situational manipulations related to depth of cognitive processing. We will test this prediction using individual difference measures (Openness, Need for Cognitive Closure [NFCC]) and situational manipulations (cognitive load).

Second, we note that existing literature suggests that perceptions of suffering do not necessarily evoke compassion (Batson, 1991; Cameron & Payne, 2011; Eisenberg, 2010). Conceptually, perceived suffering and compassion are distinguished by the fact that perceived suffering is about events “inside the head” of the target (e.g., emotional pain), whereas compassion is a feeling and action tendency toward the target (e.g., support for the target’s well-being). Empirically, the link between perceived suffering and compassion is moderated by a variety of factors. For example, among poor emotion regulators, perceived suffering can lead to overarousal and a consequent egoistic focus on the self (see Eisenberg, 2010, for a review). In addition, perceived suffering can foster defensive apathy if one feels powerless to remedy the suffering (Cameron & Payne, 2011).

An additional crucial factor that facilitates the transformation of perceived suffering into compassion is a sense of identification with the sufferer (Cwir, Carr, Walton, & Spencer, 2011; Gaertner & Dovidio, 2000; Oliner & Oliner, 1988). According to Schopenhauer (1840/1998), identification happens when the “entire difference between me and [the other], which is the very basis of my egoism, is eliminated, to a certain extent at least” (p. 144, italics in original). This notion of identification as involving a blurring of the boundary between self and other is evident in much work in social psychology (see Ashmore, Deaux, & McLaughlin-Volpe, 2004, for a review). When a perceiver identifies with a target, the perceiver is affected by the positive and negative experiences of the target almost as if those experiences are happening to the self (see Ashmore et al., 2004). Below, we will test whether the relation between perceived suffering and compassion is moderated by the possession of an expansive sense of identification that includes outgroups. Consistent with prior work (Gaertner & Dovidio, 2000), we will measure identification in terms of the perceiver’s tendency to view the self and the outgroup as constituents of a more inclusive “we.”

The Present Studies

Below, we present four studies. Study 1 uses correlational methods to test our predictions in the context of Whites’ attitudes toward African Americans. Specifically, we will test a dual-mediation path model in which external explanations simultaneously and independently increase perceived suffering and decrease perceived control, and thereby evoke compassion. In addition to mediation, Study 1 will also examine moderation: Can the external explanations/perceived suffering mechanism evoke compassion even when perceived control is high?¹ In general, the literature suggests that compassion is inhibited when perceived control is high (Weiner, 2006). This literature, however, did not vary perceived suffering. Our model posits that perceived control and perceived suffering are independent bases of compassion, and thus, we predict that perceived suffering can evoke compassion even when perceived control is high. Our rationale is that, based on moral intuitions about harm (Haidt & Joseph, 2007), it intuitively “feels wrong” to add more suffering to those who have already suffered, regardless of perceived control. In Study 1, we will examine this issue by using moderated mediation procedures (Preacher, Rucker, & Hayes, 2007) to test the strength of the relations among external explanations, perceived suffering, and compassion across different levels of perceived control.

In Study 2, we will use an experimental approach and a different target group. Specifically, we will manipulate external explanations regarding a group that has engaged in unambiguously controllable acts of violence: Chechen militants. We will test whether this manipulation increases perceived suffering and evokes compassion despite uniformly high levels of perceived control across explanation conditions. Furthermore, Study 2 will test our hypothesis that depth of processing moderates the link between external explanations and perceived suffering. Specifically, we will test whether the relation between manipulated external explanations and perceived suffering is stronger among those high as compared with low in the trait of Openness.

Studies 3 and 4 will further explore the proposed moderators of the external explanations/perceived suffering mechanism. Study 3 is a replication of Study 1, but with added measures of depth of cognitive processing and identification with outgroups. Our indicator of cognitive processing depth will be the NFCC scale (Webster & Kruglanski, 1994), and our indicator of identification will be a modified version of the Relationship Profiles scale (Haslam, Reichert, & Fiske, 2002). Moderated mediation procedures will be used to test whether the relation between external explanations and perceived suffering is stronger among those low rather than high in NFCC, and whether the relation between perceived suffering and compassion is stronger among those high rather than low in expansive identification.

Finally, Study 4 will use experimental methods to examine the depth of processing hypothesis. Participants will learn about the negative behaviors of a fictitious group, and we will vary the presence versus absence of external explanations. Furthermore, participants will be under either a low or high cognitive load as they learn about the group. We will test whether high cognitive load hinders the tendency for external explanations to heighten perceptions of suffering, despite not hindering participants’ capacity to grasp the causal implications of the explanation (i.e., to engage in causal discounting). Such a result would support the notion that external explanations contain implications regarding causality and suffering, but that the suffering implications in particular require deeper processing to extract.

In sum, all four studies test our idea that external explanations evoke compassion by increasing perceived suffering. We think that our model might prove particularly relevant in the realm of intergroup relations, as we expect that many outgroups are seen as having ample control over their trajectory. Our model illuminates how and when outgroups might receive compassion in spite of such high levels of perceived control.

Study 1

Method

Participants

Participants were 105 White undergraduates (54 female) who participated for course credit.

Procedure

Participants attended testing sessions in small groups. On arrival, they signed a consent form and received a survey packet. The survey packet contained four measures of relevance to the present study and several irrelevant measures. Surveys were arranged in one of four random orders. All measures for the present study utilized the same 6-point response scale with endpoints labeled strongly disagree (1) and strongly agree (6).

Of relevance here, participants reported their endorsement of external explanations regarding current social problems of African Americans. They did this using the Social Explanations Questionnaire (Gill & Andreychik, 2007). Consistent with our conceptualization, the items referred to longstanding external historical forces molding the current social reality of African Americans (e.g., The history of slavery, segregation, and discrimination faced by African Americans has surely contributed to any current economic and social problems they are facing; M = 3.63, SD = 0.92, α = .82).

Participants also completed measures of our two proposed mediators. Perceived suffering was assessed by four items that we created (e.g., African Americans have experienced suffering and/or frustration as result of their experiences in the United States; The social conditions of African Americans in the U.S. have created emotional pain for them; M = 4.0, SD = 0.78, α = .76). Perceived control was assessed by three items that we created (e.g., African Americans have always had some control over their own social and economic outcomes; African Americans could have done things differently in the past that would have resulted in their group being better off today; M = 3.34, SD = 0.98, α = .71).

Finally, participants completed a six-item measure of compassion, which was based on items commonly used in the literature and which was highly similar to the measure used by Andreychik and Gill (2009; for example, I feel compassion for African Americans; I feel sympathetic toward African Americans; M = 4.1, SD = 1.0, α = .91).²

Results and Discussion

Our primary analysis was the path model presented in Figure 1, which represents our proposed dual-mediation model. As can be seen there, our predictions were supported. Specifically, endorsement of external explanations was associated with increased perceptions of suffering and with decreased perceptions of control. Furthermore, as expected, perceptions of suffering were associated with increased compassion and perceptions of control were associated with reduced compassion. The mediated pathway from external explanations through perceived suffering to compassion was significant: Sobel z = 2.65, p < .008; bootstrap 95% confidence interval (CI) = [0.03-0.39] (N = 5,000). The mediated pathway from external explanations through perceived control to compassion was marginal in one test and significant in another: Sobel z = 1.79, p = .07; bootstrap 95% CI = [0.003-0.15] (N = 5,000). A variety of indicators suggested that the overall fit of the model was good: χ²(2) = 5.3, p > .07; goodness of fit index (GFI) = 0.98; adjusted goodness of fit index (AGFI) = 0.88; comparative fit index (CFI) = 0.96; root mean square residual (RMR) = 0.04. This suggests that, taken together, perceived suffering and perceived control do an adequate job explaining why external explanations evoke compassion. In addition, consistent with the idea that they are independent mediators, the zero-order relation between perceived suffering and perceived control was modest, r(103) = −.21, p = .03.

Figure 1.

Both perceived suffering and perceived control connect external explanations to compassion (Study 1)

Next, we examined moderation: Can external explanations increase perceived suffering and evoke compassion even when perceived control is high? We used moderated mediation procedures (Preacher et al., 2007) to examine this issue. See Figure 2 for the two models we tested. We computed each model using SPSS macros available at http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html. Results offered no evidence for moderation by perceived control (ts < 1.25, ps > .21). These results suggest that (a) the relation between external explanations and perceived suffering is equally strong across all levels of perceived control and (b) the relation between perceived suffering and compassion is equally strong across all levels of perceived control. Thus, even when an outgroup is perceived as having substantial control over its negative aspects, external explanations can foster compassion by increasing the perception that the outgroup has suffered.

Figure 2.

Moderated mediation models examined in Study 1

Study 2

Study 1 was correlational and thus cannot support causal conclusions. Thus, Study 2 is experimental. In addition, to establish generalizability, we focus on a very different target group: Chechen militants. By focusing on a group that has committed planned acts of violence, Study 2 provides an alternative means for disentangling the external explanations/perceived suffering mechanism from perceived control. That is, perceived control will be uniformly high, and we will show that external explanations nevertheless can evoke compassion by increasing perceived suffering. Study 2 will also test our assertion that deep cognitive processing strengthens the tendency for external explanations to increase perceived suffering.

Method

Participants

Participants were 84 (41 female) undergraduates who participated for course credit.

Procedure

At a pretest, participants completed the Big Five Inventory (BFI; John & Srivastava, 1999). We were particularly interested in the trait of Openness. Those high in Openness tend to process information deeply and to pursue thorough understanding (John & Srivastava, 1999). The BFI taps Openness using 10 descriptions that the respondent rates in terms of how well they describe the self (e.g., Is ingenious, a deep thinker; Likes to reflect, play with ideas). After appropriate reverse scoring, the items were averaged (M = 3.47 of 5, SD = .61, α = .77).

Weeks later, participants returned individually for an ostensibly unrelated experiment. They signed a consent form and were randomly assigned to either the no explanation or external explanation condition. All participants read background information about the Russia/Chechnya conflict and then learned about the Beslan incident in which Chechen militants took schoolchildren hostage, rigged their school with explosives, and issued demands for Chechen independence from Russia. During an ensuing shootout with Russian forces, nearly 200 children were killed.

Next, we introduced our explanation manipulation. Those in the no explanation condition learned nothing else,³ whereas those in the external explanation condition learned that the militants were driven by the treatment their people have received from the Russians over decades, including large-scale relocations, lack of freedom, and ultimately invasion and strict control of Chechnya by Russian forces.⁴ In a pilot study, participants were exposed to this manipulation and then completed two items tapping perceived control: Instead of taking hostages at the school, the Chechens could have chosen a different approach in their conflict with the Russians and The hostage-taking incident was a deliberate, intentional act taken by the Chechens. These two items were correlated—r(102) = .45, p = .000001—and thus were averaged to form an index of perceived control (M = 3.64, SD = 0.58). Notably, the mean score was close to the maximum possible score of 4. Thus, participants overwhelmingly viewed the Chechens’ violence as controllable, and this perception was unaffected by our explanation manipulation (F < 1).

In the present study, after reading the relevant information, participants completed a three-item measure of perceived suffering (e.g., I got a strong sense of the pain and suffering of the Chechens as I learned about this conflict). Responses were made on a 6-point scale with endpoints labeled strongly disagree (1) and strongly agree (6; M = 3.63, SD = 1.14, α = .84). Participants also completed a six-item measure of compassion (e.g., I wished I could alleviate the suffering of the Chechens; I had feelings of care and concern toward the Chechens). Responses were made on a 6-point scale with endpoints labeled strongly disagree (1) and strongly agree (6; M = 3.67, SD = 1.35, α = .93).

Results and Discussion

We computed the path model shown in Figure 3. Again, results supported our predictions: Manipulated external explanations significantly increased perceived suffering, and perceived suffering contributed to feelings of compassion. The mediated pathway from external explanations through perceived suffering to compassion was significant: Sobel z = 4.71, p = .00001; 99% CI = [0.17-0.59] using bootstrap methods (N = 5,000). Conceptually replicating the moderation analyses from Study 1, these results suggest that even when an outgroup is perceived as having control over its negative aspects, it is still possible to generate compassion for them by using external explanations to increase perceptions of suffering.⁵

Figure 3.

Perceived suffering carries the effect of external explanations to compassion (Study 2)

Next, we tested whether Openness moderated the extent to which external explanations increase perceived suffering. We tested a model with the same structure as Model 1 in Figure 2 but with Openness substituted for perceived control. Thus, the model tests whether the effect of the external explanations manipulation on perceived suffering varied as a function of Openness. It did: The interaction between external explanations and Openness for predicting perceived suffering was significant, t(78) = −2.33, p = .02. Figure 4 shows mean levels of perceived suffering in the no explanation and external explanation conditions at −1 SD/+1 SD of Openness. These are consistent with our predictions: Deep thinkers (+1 SD of Openness) show a stronger tendency to perceive suffering as they process external explanations than do shallow thinkers (−1 SD of Openness). So, external explanations can evoke compassion by increasing perceived suffering, but this mechanism is more effective among those prone to deep cognitive processing.⁶

Figure 4.

The effect of external explanations on perceived suffering is stronger among deep thinkers (+1 SD of Openness) as compared with shallow thinkers (−1 SD of Openness)

Study 3

In Study 3, we return to African Americans as the target outgroup. Study 3 will further explore our hypotheses regarding moderation. Specifically, as in Study 2, we will use moderated mediation procedures to test whether the strength of the external explanations/perceived suffering link varies depending on one’s tendency to think deeply versus shallowly. We will use NFCC—rather than Openness—as our indicator of cognitive processing habits. We will also use moderated mediation procedures to test whether the strength of the perceived suffering/compassion link varies depending on one’s tendency toward expansive identification with outgroups. To measure such identification, we will use a scale based on the Relational Models framework of Fiske and his colleagues (Fiske, 1992).

Method

Participants

Participants were 98 White undergraduates (54 female) who participated for course credit.

Procedure

Participants attended pretests in small groups. On arrival, they signed a consent form and received a survey packet. Questionnaires were presented in a variety of random orders. The packet contained the measure of external explanations used in Study 1 (M = 4.12, SD = 0.85, α = .77) and a five-item measure of compassion that was highly similar to that in Study 1 (M = 4.1, SD = 0.93, α = .81). To assess identification with outgroups, participants completed a measure of intergroup communal sharing orientation, which we developed based on the Relationship Profile Scale (RPS; Haslam et al., 2002). Communal sharing orientation fundamentally involves a sense of “oneness” with others (i.e., identification). Here is a slightly shortened version of what participants read:

Every social group takes a “one for all and all for one” approach toward the other social groups. Each group feels that “what’s ours is yours” and that what happens to other groups is as important as what happens to one’s own group. . . . Groups willingly share resources such as food, wealth, knowledge, and so on.

Participants then responded to four items tapping whether the description sounded consistent with their preferred form of intergroup identification (e.g., Would you like social groups to relate in this manner?). Responses were made using 6-point scales with endpoints labeled no (1) and yes (6). Responses were averaged to create an identification score (M = 4.2, SD = 1.24, α = .91).

Participants returned a couple of weeks later. Again, they received a variety of surveys in one of several random orders. They completed a six-item measure of perceived suffering that was similar to the one in Study 1 (e.g., I definitely get a sense of the suffering of African Americans when I think about their experiences in the United States; M = 3.8, SD = 0.88, α = .80). Finally, to measure cognitive processing habits, participants completed the NFCC scale (Webster & Kruglanski, 1994). Those high in NFCC prefer quick, certain answers in all domains, shunning extended contemplation (e.g., I usually make important decisions quickly and confidently; I do not usually consult many different opinions before forming my own view; M = 3.8 on a 6-point scale, SD = 0.53, α = .86).

Results and Discussion

We computed the path model shown in Figure 5. It is a moderated mediation model that simultaneously examined relations among the core constructs of our model (external explanations, perceived suffering, compassion) and the proposed moderators of those relations. The presented path coefficients reveal that Study 3 replicated the results of Study 1: Endorsement of external explanations predicts increased perceptions of suffering, and perceived suffering is associated with increased compassion. The mediated pathway from external explanations through perceived suffering to compassion was significant: Sobel z = 3.69, p = .0002; bootstrap 99% CI = [0.07-0.39] (N = 5,000).

Figure 5.

Moderated mediation model examined in Study 3

Next, we examined moderation. Both our hypotheses regarding moderation were supported. Specifically, the interaction between external explanations and NFCC when predicting perceived suffering was significant, t(91) = −2.31, p = .019. Figure 6 shows the simple slopes of the relation between external explanations and perceived suffering at −1 SD/+1 SD of NFCC. The pattern replicates what was found in Study 2: Deep thinkers (−1 SD NFCC) show a stronger tendency to perceive suffering from external explanations than do shallow thinkers (+1 SD NFCC). Notably, the zero-order relation between NFCC and external explanations was not significant (r = −.07, ns). Thus, NFCC was not associated with endorsement of external explanations but rather was specifically associated with the tendency of external explanations to give rise to perceptions of suffering.

Figure 6.

The relation between external explanations and perceived suffering is stronger among deep thinkers (−1 SD of NFCC) as compared with shallow thinkers (+1 SD of NFCC; Study 3)

The expected interaction between perceived suffering and identification when predicting compassion was marginal, t(91) = 1.84, p = .068. Figure 7 shows the simple slopes of the relation between perceived suffering and compassion at −1 SD/+1 SD of identification. The pattern supported our predictions: Those who prefer to think of social groups as “members of one team” (+1 SD on identification) show a stronger tendency to feel compassion in response to perceptions of outgroup suffering than do those who lack this oneness mind-set (−1 SD on identification). Notably, the zero-order relation between identification and perceived suffering was not significant (r = .11, ns). Thus, identification was not associated with greater perceptions of suffering but rather was specifically associated with transformation of perceived suffering into compassion, or caring about the suffering one perceived in the outgroup.

Figure 7.

The relation between perceived suffering and compassion is stronger among those with an expansive sense of identification that includes outgroups (Study 3)

To sum up, external explanations can evoke compassion by increasing perceived suffering, but this mechanism will be more effective among those prone to deep cognitive processing and among those with a sense of expansive identification or oneness with outgroups.

Study 4

Our conceptual approach assumes that external explanations and perceived suffering are two different constructs. Another way of stating our position is to say that external explanations carry multiple implications. Their primary implication is for causality, in particular the notion that the outgroup per se is not the sole cause of its negative aspects. Implications for suffering are separate from this. Our moderation analyses above are consistent with this view: The amount of suffering inferred from an external explanation varies depending on depth of cognitive processing. However, our moderation analyses do not definitively show that causal implications and suffering implications are two different facets of external explanations. The reason is that we presented no evidence regarding the extent to which perceivers were drawing appropriate causal conclusions from their external explanations. Study 4 will provide evidence relevant to this issue. Specifically, it will examine whether a manipulation of cognitive processing capacity selectively obstructs recognition of the suffering implications of external explanations while leaving the recognition of causal implications intact.

Method

Participants

Participants were 85 (53 female) undergraduates who participated for course credit.

Procedure

Participants arrived individually, signed a consent form, and were introduced to a study of “the effect of distraction on social judgment.” They were randomly assigned to the no explanation or external explanation condition and to the low load or high load condition. All participants read about a fictitious group, the Niffites, whose behaviors included dishonesty, unfairness, lack of generosity, and criminality (materials from Andreychik & Gill, 2012). In the no explanation condition, participants learned only about the negative behaviors. In the external explanation condition, participants learned about the negative behaviors and also learned that those behaviors were caused by a history of negative experiences with other groups (e.g., Because other groups don’t like them, Niffites receive little kindness from others. As a result, Niffites have grown to distrust others and to believe that people should “only look out for themselves.”).

As participants read about the Niffites, they experienced either low or high levels of cognitive load from a simultaneous “tone-detection task” (Gilbert, Gill, & Wilson, 2002). Specifically, participants listened to a recording of high-, medium-, and low-frequency tones played in quasirandom patterns with varying amounts of time elapsing between tones. In the low load condition, participants were told to press a clicker each time they detected the high-frequency tone. In the high load condition, participants were told to press a clicker each time they detected the high tone, followed by the medium tone, and then followed by the low tone. Although most participants made some errors (e.g., confusing the medium tone with the high tone), all participants correctly responded to their target tone(s) in the majority of instances. The average ratio of hits to misses/false alarms was 4:1.

Finally, after reading the relevant vignette and while remaining under low or high load, participants responded to our primary dependent variables. These consisted of several ratings of the Niffites on a 5-point scale with endpoints labeled no (1) and yes (5). Two items tapped perceived suffering: The Niffites have experienced a lot of emotional pain and suffering and The experience of the Niffites includes a lot of frustration, unhappiness, and other painful feelings. These two items were correlated—r(83) = .54, p < .001—and thus were averaged to form an index of perceived suffering (M = 3.1, SD = 1.09). Next, to assess causal discounting, participants responded to an item about the inherent badness of the Niffites: The Niffites’ behavior is caused by the fact that they are just inherently bad people (M = 2.73, SD = 1.29). We also included one item tapping perceived control: The Niffites have control over whether they behave negatively (M = 3.78, SD = 1.19). Finally, participants responded to a compassion item: I feel sympathy and compassion for the Niffites (M = 2.39, SD = 1.23).

Results and Discussion

Did cognitive load undermine the capacity to see suffering in external explanations? Yes, it did. We computed a 2 (explanation: none, external) × 2 (cognitive load: low, high) ANOVA with perceived suffering as the dependent variable. This revealed a main effect of explanation, F(1, 81) = 77.1, p = .0000000000002 (M_{no explanation} = 2.39, M_{external explanation} = 3.87). The predicted interaction was also significant, F(1, 81) = 4.66, p = .03. The pattern can be seen in Figure 8. As expected, within the external explanation condition, perceived suffering was significantly reduced in the high as compared with the low load condition, t(81) = 2.48, p = .015. In contrast, also as expected, the load manipulation had no effect on perceived suffering within the internal explanation condition, t < 1. Thus, cognitive load impaired participants’ capacity to perceive suffering in the external explanation condition.

Figure 8.

Cognitive load impairs the capacity to perceive suffering when processing external explanations (Study 4)

Next, to test our hypothesis regarding causal implications, we computed the same ANOVA but with inherent badness as the dependent variable. This revealed only a main effect of explanation, F(1, 81) = 24.46, p = .000004 (M_{no explanation} = 3.3, M_{external explanation} = 2.1). The interaction with cognitive load was not significant (p > .22). Indeed, the means within the external explanation condition were virtually identical, t < 1 (M_{external explanation/low load} = 2.15, M_{external explanation/high load} = 2.05), and the difference between the no explanation and external explanation conditions was significant within both load conditions (both ts > 2.62, ps < .02). This suggests that participants in the external explanation condition were able to engage in causal discounting—recognizing that external explanations imply a lack of inherent badness—even under a high level of cognitive load. Under cognitive load, causal implications were fully recognized even though recognition of suffering implications was impaired.

Next, we examined perceived control using the same ANOVA. The analysis revealed only a main effect of explanation, F(1, 81) = 4.43, p = .038 (M_{no explanation} = 4.04, M_{external explanation} = 3.51). The interaction with cognitive load was not significant (p > .26). Indeed, as with inherent badness, the means across load conditions within the external explanation condition were quite close, t < 1 (M_{external explanation/low load} = 3.45, M_{external explanation/high load} = 3.57). This suggests that participants were able to glean the (un)controllability implications from external explanations even under a high level of cognitive load. Intriguingly (and unexpectedly), the difference between the no explanation and external explanation conditions was actually larger in the high load condition (M_{no explanation} = 4.38 vs. M_{external explanation} = 3.57; t(81) = 2.27, p = .026) than in the low load condition (M_{no explanation} = 3.69 vs. M_{external explanation} = 3.45; t < 1). Of course, because this pattern was not expected and because the interaction did not approach significance, we will not give it much attention. Suffice it to say that although cognitive load seems to impair recognition of suffering implications, it does not seem to impair recognition of controllability implications.

Finally, we examined responses to the compassion item using the same ANOVA. This revealed only a main effect of explanation, F(1, 81) = 22.86, p = .000008 (M_{no explanation} = 1.84, M_{external explanation} = 2.98). Although the lack of an interaction parallel to that involving perceived suffering might seem to challenge our theorizing, we believe that it does not. Our model suggests that compassion is evoked by multiple mediators, including perceived control. It could be the case that, within the external explanation condition, reduced perceptions of control are contributing to compassion thereby hampering our ability to find results that perfectly parallel those for perceived suffering. Of clearer relevance to our account, we computed correlations between perceived suffering and compassion within the external explanation/low load and external explanation/high load conditions. These suggested that, as we would predict, perceived suffering contributed to compassion in the low load condition, r(18) = .44, p = .05, but not in the high load condition, r(19) = .12, p = .61.

General Discussion

Explanations shape emotional responses to social events. In the present article, we explore this fundamental truth in the context of trying to understand compassionate responses to outgroups. We challenge the notion that perceived control is the sole mediator of the effects of external explanations on compassion. We propose a novel mechanism involving perceived suffering. Specifically, we suggest (a) external explanations for negative aspects of the outgroup increase the perception that the outgroup has suffered and (b) perceptions of suffering evoke compassion. We further suggest that the link between external explanations and perceived suffering is moderated by the depth of cognitive processing, and that the link between perceived suffering and compassion is moderated by expansive identification with outgroups. In contrast to prior models of explanations and moral emotions, our moderator approach makes clear that compassion is not simply a matter of how one chooses to explain negative aspects of an outgroup. Rather, external explanations must join forces with deep thinking and an expansive sense of identification for compassion to result.

Studies 1 and 2 focused on demonstrating that our proposed mechanism is dissociable from perceived control. Study 1 examined Whites’ attitudes toward African Americans and provided evidence that perceived suffering and perceived control are dual mediators that independently connect external explanations to compassion. Study 1 also revealed that relations among external explanations, perceived suffering, and compassion are equally strong across all levels of perceived control. Thus, heightened perceptions of control do not obstruct the operation of the external explanations/ perceived suffering mechanism. In Study 2, we took an experimental approach. The target group was Chechen militants, a group that committed planned acts of violence. Thus, we created a context in which perceived control would be uniformly high. Study 2 revealed that manipulated external explanations increase perceived suffering and thereby evoke compassion even when perceived control is uniformly high. In sum, the evocation of compassion via explanatory information depends on more than controllability implications. Heightened perceptions of suffering can also link explanations to compassion, and this happens independently of controllability perceptions.

A major implication of our model is that—contra Weiner—it is possible via attributional interventions to evoke compassion even toward those perceived as exerting control over their negative aspects. Thus, the lazy student (Weiner & Kukla, 1970), the drug abuser (Weiner, Perry, & Magnusson, 1988), and even the terrorist (see Study 2) can receive compassion. Future work is needed to understand how compassion influences downstream consequences such as punishment decisions regarding those who have committed controllable transgressions. One possibility is that compassion— generated via external explanations/perceived suffering—might encourage a tendency to respond in a way that is less focused on causing pain to the transgressor and more focused on restorative or rehabilitative considerations. In other words, the retributive impulse for “just desserts” (Carlsmith, Darley, & Robinson, 2002) might be weakened by compassion. This possibility could prove important in situations of internecine intergroup conflict, weakening the sense that an outgroup’s controllable negative acts mean that they should be “taught a lesson” through harsh punishment.

Studies 2, 3, and 4 examined proposed moderators of the external explanations/perceived suffering mechanism. All three studies supported our prediction that deep cognitive processing would increase the extent to which external explanations increase perceived suffering. The studies included individual difference measures and situational manipulations of cognitive processing depth. In all cases, we found that the relation between external explanations and perceived suffering was stronger when depth of processing was high. Going beyond depth of processing, Study 3 supported our prediction that expansive identification with outgroups would increase the extent to which perceived suffering gives rise to compassion. In sum, these studies suggest that the external explanations/perceived suffering mechanism will have a particularly powerful effect on compassion among those who think deeply and among those who have an expansive sense of identity that includes outgroups.

Additional meaningful contrasts can be drawn between Weiner’s (2006) model and our own. For example, Weiner situates his model within the metaphor of person as judge who must “rationally interpret evidence” (p. 4) to determine guilt or responsibility (see Fincham & Jaspars, 1980). Our focus on perceived suffering suggests that people are more than rational judges assessing target agency. An apt metaphor for our approach is that people are humanitarians, sensitive to the suffering of others and motivated to support or to refrain from further harming those who have suffered. One way of conceptualizing our differences from Weiner is to suggest that Weiner’s work emphasizes the fairness/justice dimension of morality (i.e., do not blame people for things they cannot control), whereas our work emphasizes the care/harm dimension (i.e., respond with care to those who suffer; see Haidt & Joseph, 2007).

Our focus on sensitivity to suffering means that our model, unlike Weiner’s, builds bridges between work on social explanation and work on empathy. Our approach implies that explanation—despite being the epitome of abstract, uniquely human cognition—can be a stimulus to the empathy system, which is the epitome of a perceptually driven, primitive emotional system. We will attempt to spell out how our work relates to existing work on empathy. First, unlike our work, much work on empathy involves direct perception of targets who are experiencing physical pain (e.g., Avenanti, Sirigu, & Aglioti, 2010; Batson, Duncan, Ackerman, Buckley, & Birch, 1981; Stotland, 1969) or expressing emotional pain (e.g., Coke, Batson, & McDavis, 1978). In this work, “empathy” refers to a process in which the direct perception of distress automatically activates a neural “matching” system in the perceiver, thereby eliciting congruent distress within the perceiver (Keysers, 2011; Preston & de Waal, 2002). In our work, given our focus on abstract causal cognition, there is no physically present target and hence no perceptual input to be “matched.” What is happening in our studies, then, must be different in some ways from the typical situation studied in the empathy literature. In our case, the empathy system is being stimulated by the imagination of events that one has not directly perceived. We are unaware of research that systematically explores similarities and differences between perceptually versus imaginatively activated empathic responses. Are the same neural areas involved? Is the emotional arousal equally strong? Are they equally likely to lead to prosocial responding? These are important questions with practical implications for how to design interventions that will effectively elicit compassion and prosocial behavior.

Another major concept from the empathy literature is perspective taking. How is our external explanations/perceived suffering mechanism related to perspective taking? In a sense, all empathic activity involves responding to the inner state of another and thus could be said to involve perspective taking. However, thinking of empathy in such an overly broad manner prevents one from seeing important distinctions (Batson, 2009). For example, perceivers frequently have automatically activated empathic reactions without trying to take the other’s perspective (Keysers, 2011). Yet, empathic reactions can be intensified when a perceiver deliberately tries to adopt the perspective of the other (Batson et al., 1997; Toi & Batson, 1982). Presumably because of this latter phenomenon, perspective taking has come to be defined as involving an active, intentional effort to “step into the shoes” of another (Batson, 1991). Our phenomenon does not depend on inducing such efforts and thus is not identical to what is called perspective taking in the literature.

In fact, more tellingly, when people are asked to take the perspective of a stigmatized outgroup, they do not experience compassion. For example, Finlay and Stephan (2000) and Dovidio et al. (2004) asked White participants to take the perspective of African Americans. In both studies, perspective taking had no effect on compassion. Instead, perspective taking increased feelings of anger and moral outrage regarding injustice against African Americans. So, perspective taking and external explanations seem to have quite different effects on emotion. Future research is needed to explore the reasons for these divergent effects. One possibility is that compassion depends on having a third-person perspective (which seems likely when one is explaining a target), whereas outrage/anger depends on a first-person perspective (which should be present if one truly is engaged in perspective taking). Another important question flows out of this analysis: What are the differential effects of compassion versus moral outrage/anger on intergroup helping (Thomas, McGarty, & Mavor, 2009)?

Our approach highlights perceived suffering and perceived control as crucial, independent contributors to compassion. An important research agenda grows out of this conceptualization: What are the personal and situational factors that affect whether people are moved primarily by perceived suffering (responding as humanitarians) or primarily by perceived control (responding as judges)? Above we showed that the external explanations/perceived suffering mechanism can evoke compassion even when perceived control is high. Yet, we are certain that perceived suffering will not “trump” perceived control for all perceivers in all contexts. For example, those in a prosecutorial mind-set (Tetlock et al., 2007) might be prone to minimize suffering and to respond punitively whenever perceived control is high. In our lab, we are currently exploring the possibility that there is also a humanitarian mind-set that makes perceivers prone to be moved primarily by perceived suffering and to focus less on perceived control. Research exploring moderators of the weight given to perceived suffering versus perceived control would have major implications for settings in which a “judge” and a “humanitarian” might reach different conclusions, such as in legal settings that involve perpetrators with histories of suffering.

Finally, the present work has practical implications for impression management by devalued groups. Work on “moral typecasting” (Gray & Wegner, 2009) suggests that targets tend to be viewed as either moral patients (who have suffered) or moral agents (who are capable of enacting good or bad deeds). The extent to which a target is typecast into one category reduces the extent to which they will be seen as belonging to the other. This pattern has negative implications for disadvantaged groups, as it implies that telling the story of their suffering will reduce perceptions of group agency. Fortunately, our work suggests that it is quite possible to view a group as having endured suffering and also as possessing agency. Indeed, links between perceived suffering and perceived control in our data were consistently weak (albeit always in the direction suggested by Gray and Wegner’s work). A key task for future research, then, is to illuminate precisely how it is possible to communicate information about group suffering without undermining perceptions of group agency.

Conclusion

Compassion for outgroups depends on how perceivers explain negative aspects of the outgroup. When perceivers endorse external explanations for the outgroup’s negative aspects, this can increase the perception that the outgroup has suffered. Perceived suffering can evoke compassion. The external explanations/perceived suffering mechanism will be more facilitative of compassion when perceivers are thinking deeply and when they have an expansive sense of identification with others. Finally, the external explanations/perceived suffering mechanism operates independently of perceived control, and thus illuminates how compassion might be generated even toward those perceived as having control over their negative aspects. We imagine that Henry Wadsworth Longfellow had in mind something like what we are describing here when he wrote, “If we could read the secret history of our enemies, we should find in each man’s life sorrow and suffering enough to disarm all hostility.”

Footnotes

Acknowledgements

Sincere thanks are extended to Maggie Barr, Samantha Berman, Kelly Chu, Anna Fishbon, Rachel Kobren, Catherine Maryon, Kirsten Peterson, Dheera Reddy, Erin Schwartz, and Ben Welkie for their assistance with data collection and management.

Authors’ Note

Study 3 data come from an honors thesis completed by Catherine Maryon.

Declaration of Conflicting Interests

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

Funding

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

Notes

References

Allport

G. W.

(1954). The nature of prejudice. Reading, MA: Addison-Wesley.

Andreychik

M. R.

Gill

M. J.

(2009). Ingroup identity moderates the impact of social explanations on intergroup attitudes: External explanations are not inherently prosocial. Personality and Social Psychology Bulletin, 35, 1632–1645.

Andreychik

M. R.

Gill

M. J.

(2012). Do negative implicit associations indicate negative attitudes?: Social explanations moderate whether ostensible “negative” associations are prejudice-based or empathy-based. Journal of Experimental Social Psychology, 48, 1082–1093.

Ashmore

R. D.

Deaux

McLaughlin-Volpe

(2004). An organization framework for collective identity: Articulation and significance of multidimensionality. Psychological Bulletin, 130, 80–114.

Avenanti

Sirigu

Aglioti

S. M.

(2010). Racial bias reduces empathic sensorimotor resonance with other-race pain. Current Biology, 20, 1018–1022.

Batson

C. D.

(1991). The altruism question: Toward a social-psychological answer. Hillsdale, NJ: Lawrence Erlbaum.

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.

Duncan

B. D.

Ackerman

Buckley

Birch

(1981). Is empathic emotion a source of altruistic motivation? Journal of Personality and Social Psychology, 40, 290–302.

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.

10.

Bradbury

T. N.

Fincham

F. D.

(1990). Attributions in marriage: Review and critique. Psychological Bulletin, 107, 3–33.

11.

Brigham

J. C.

(1993). College students’ racial attitudes. Journal of Applied Social Psychology, 23, 1933–1967.

12.

Bruner

J. S.

(2002). Making stories: Law, literature, life. Cambridge, MA: Harvard University Press.

13.

Buchanan

G. M.

Seligman

M. E. P.

(1995). Explanatory style. Hillsdale, NJ: Erlbaum.

14.

Cameron

C. D.

Payne

B. K.

(2011). Escaping affect: How motivated emotion regulation creates insensitivity to mass suffering. Journal of Personality and Social Psychology, 100, 1–15.

15.

Carlsmith

K. M.

Darley

J. M.

Robinson

P. H.

(2002). Why do we punish? Deterrence and just deserts as motives for punishment. Journal of Personality and Social Psychology, 83, 284–299.

16.

Coke

J. S.

Batson

C. D.

McDavis

(1978). Empathic mediation of helping: A two-stage model. Journal of Personality and Social Psychology, 36, 752–766.

17.

Costabile

K. A.

Klein

S. B.

(2008). Understanding and predicting social events: The effects of narrative construction on inference generation. Social Cognition, 26, 420–437.

18.

Cwir

Carr

P. B.

Walton

G. M.

Spencer

S. J.

(2011). Your heart makes my heart move: Cues of social connectedness cause shared emotions and physiological states among strangers. Journal of Experimental Social Psychology, 47, 661–664.

19.

Davis

M. H.

(1996). Empathy: A social psychological approach. Boulder, CO: Westview Press.

20.

Doosje

Branscombe

N. R.

Spears

Manstead

A. S. R.

(1998). Guilty by association: When one’s group has a negative history. Journal of Personality and Social Psychology, 75, 872–886.

21.

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.

22.

Eisenberg

(2010). Empathy-related responding: Links with self-regulation, moral judgment, and moral behavior. In Mikulincer

Shaver

P. R.

(Eds.), Prosocial motives, emotions, and behavior: The better angels of our nature (pp. 129–148). Washington, DC: American Psychological Association.

23.

Fincham

F. D.

Jaspars

J. M.

(1980). Attribution of responsibility: From man the scientist to man as lawyer. In Berkowitz

(Ed.), Advances in experimental social psychology (pp. 81–138). New York, NY: Academic Press.

24.

Finlay

K. A.

Stephan

W. G.

(2000). Improving intergroup relations: The effects of empathy on racial attitudes. Journal of Applied Social Psychology, 30, 1720–1737.

25.

Fiske

A. P.

(1992). The four elementary forms of sociality: Framework for a unified theory of social relations. Psychological Review, 99, 689–723.

26.

Gaertner

S. L.

Dovidio

J. F.

(2000). Reducing intergroup bias: The common ingroup identity model. Ann Arbor, MI: Sheridan Books.

27.

Gilbert

D. T.

Gill

M. J.

Wilson

T. D.

(2002). The future is now: Temporal correction in affective forecasting. Organizational Behavior and Human Decision Processes, 88, 430–444.

28.

Gilbert

D. T.

Malone

P. S.

(1995). The correspondence bias. Psychological Bulletin, 117, 21–38.

29.

Gill

M. J.

Andreychik

M. A.

(2009). Getting emotional about explanations: Social explanations and social explanatory styles as bases of prosocial emotions and intergroup attitudes. Social & Personality Psychology Compass, 3, 1038–1054.

30.

Gill

M. J.

Andreychik

M. R.

(2007). Explanation and intergroup emotion: Social explanations as a foundation of prejudice-related compunction [Special Issue on Intergroup Emotion]. Group Processes & Intergroup Relations, 10, 87–106.

31.

Glick

(2005). Choice of scapegoats. In Dovidio

J. F.

Glick

Rudman

L. A.

(Eds.), On the nature of prejudice: 50 years after Allport (pp. 244–261). Malden, MA: Blackwell.

32.

Glover

(2000). Humanity: A moral history of the twentieth century. New Haven, CT: Yale University Press.

33.

Goetz

J. L.

Keltner

Simon-Thomas

(2010). Compassion: An evolutionary analysis and empirical review. Psychological Bulletin, 136, 351–374.

34.

Gray

Wegner

D. M.

(2009). Moral typecasting: Divergent perceptions of moral agents and moral patients. Journal of Personality and Social Psychology, 96, 505–520.

35.

Guimond

Begin

Palmer

D. L.

(1989). Education and causal attributions: The development of “person-blame” and “system-blame” ideology. Social Psychology Quarterly, 52, 126–140.

36.

Haidt

(2003). The moral emotions. In Davidson

R. J.

Scherer

Goldsmith

H. H.

(Eds.), Handbook of affective sciences (pp. 852–870). New York, NY: Oxford University Press.

37.

Haidt

Joseph

(2007). The moral mind: How 5 sets of innate moral intuitions guide the development of many culture-specific virtues, and perhaps even modules. In Carruthers

Laurence

Stich

(Eds.), The innate mind, Vol. 3 (pp. 367–391). New York, NY: Oxford University Press.

38.

Harth

N. S.

Kessler

Leach

C. W.

(2008). Advantaged groups’ emotional reactions to intergroup inequality: The dynamics of pride, guilt, and sympathy. Personality and Social Psychology Bulletin, 34, 115–129.

39.

Haslam

Reichert

Fiske

A. P.

(2002). Aberrant social relations in the personality disorders. Psychology and Psychotherapy: Theory, research and practice, 75, 19–31.

40.

Heider

(1958). The psychology of interpersonal relations. New York, NY: John Wiley & Sons.

41.

Hewstone

(1990). The “ultimate attribution error”? A review of the literature on intergroup causal attribution. European Journal of Social Psychology, 20, 311–335.

42.

Hoffman

Hurst

(1990). Gender stereotypes: Perception or rationalization? Journal of Personality and Social Psychology, 58, 197–208.

43.

Iyer

Leach

C. W.

Crosby

F. J.

(2003). White guilt and racial compensation: The benefits and limits of self-focus. Personality and Social Psychology Bulletin, 29, 117–129.

44.

John

O. P.

Srivastava

(1999). The big five trait taxonomy: History, measurement, and theoretical perspectives. In Pervin

L. A.

John

O. P.

(Eds.), Handbook of personality: Theory and research (2nd ed., pp. 102–138). New York, NY: Guilford.

45.

Jost

J. T.

Banaji

M. R.

(1994). The role of stereotyping in system-justification and the production of false consciousness. British Journal of Social Psychology, 33, 1–27.

46.

Karaçanta

Fitness

(2006). Majority support for minority out-groups: The roles of compassion and guilt. Journal of Applied Social Psychology, 36, 2730–2749.

47.

Kelley

H. H.

(1973). The processes of causal attribution. American Psychologist, 28, 107–128.

48.

Keysers

(2011). The empathic brain: How the discovery of mirror neurons changes our understanding of human nature. Charleston, SC: CreateSpace.

49.

Kluegel

J. R.

Bobo

(1993). Dimensions of whites’ beliefs about the black-white socioeconomic gap. In Sniderman

P. M.

Tetlock

Carmines

E. G.

(Eds.), Prejudice, politics, and the American dilemma (pp. 127–147). Stanford, CA: Stanford University Press.

50.

Lopez

G. E.

Gurin

Nagda

B. A.

(1998). Education and understanding structural causes for group inequalities. Political Psychology, 19, 305–329.

51.

Oliner

S. P.

Oliner

P. M.

(1988). The altruistic personality: Rescuers of Jews in Nazi Germany. New York, NY: Free Press.

52.

Omoto

A. M.

Malsch

A. M.

Barraza

J. A.

(2009). Compassionate acts: Motivations for and correlates of volunteerism among older adults. In Fehr

Sprecher

Underwood

L. G.

(Eds.), The science of compassionate love (pp. 257–282). West Sussex, UK: Wiley-Blackwell.

53.

Pettigrew

T. F.

(1979). The ultimate attribution error: Extending Allport’s cognitive analysis of prejudice. Personality and Social Psychology Bulletin, 5, 461–476.

54.

Plant

E. A.

Devine

P. G.

(1998). Internal and external motivation to respond without prejudice. Journal of Personality and Social Psychology, 75, 811–832.

55.

Preacher

K. J.

Rucker

D. D.

Hayes

A. F.

(2007). Addressing moderated mediation hypotheses: Theory, methods, and prescriptions. Multivariate Behavioral Research, 42, 185–227.

56.

Preston

S. D.

de Waal

F. B. M.

(2002). Empathy: Its ultimate and proximate bases. Behavioral and Brain Sciences, 25, 1–20.

57.

Schopenhauer

(1998). On the basis of morality ( Payne

E. F. J.

Trans.). Indianapolis, IN: Hackett. (Original work published 1840)

58.

Sidanius

Pratto

(1999). Social dominance: An intergroup theory of social hierarchy and oppression. New York, NY: Cambridge University Press.

59.

Skitka

L. J.

(1999). Ideological and attributional boundaries on public compassion: Reactions to individuals and communities affected by a natural disaster. Personality and Social Psychology Bulletin, 25, 793–808.

60.

Smith

T. W.

(2009). Loving and caring in the United States: Trends and correlates of empathy, altruism, and related constructs. In Fehr

Sprecher

Underwood

L. G.

(Eds.), The science of compassionate love (pp. 79–119). West Sussex, UK: Wiley-Blackwell.

61.

Stotland

(1969). Exploratory investigations of empathy. In Berkowitz

(Ed.), Advances in experimental social psychology (Vol. 4., pp. 271–314). New York, NY: Academic Press.

62.

Tajfel

Turner

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

Austin

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

63.

Taylor

D. M.

Jaggi

(1974). Ethnocentrism and causal attribution in a South Indian context. Journal of Cross-Cultural Psychology, 5, 162–171.

64.

Tetlock

P. E.

Visser

P. S.

Singh

Polifroni

Scott

Elson

S. B.

Rescober

(2007). People as intuitive prosecutors: The impact of social-control goals on attributions of responsibility. Journal of Experimental Social Psychology, 43, 195–209.

65.

Thomas

E. F.

McGarty

Mavor

K. I.

(2009). Transforming “Apathy into movement”: The role of prosocial emotions in motivating action for social change. Personality and Social Psychology Review, 13, 310–333.

66.

Toi

Batson

C. D.

(1982). More evidence that empathy is a source of altruistic motivation. Journal of Personality and Social Psychology, 43, 281–292.

67.

Webster

D. M.

Kruglanski

A. W.

(1994). Individual differences in Need for Cognitive Closure. Journal of Personality and Social Psychology, 67, 1049–1062.

68.

Weiner

(2006). Social motivation, justice, and the moral emotions: An attributional approach. Mahwah, NJ: Lawrence Erlbaum.

69.

Weiner

Kukla

(1970). An attributional analysis of achievement motivation. Journal of Personality and Social Psychology, 15, 1–20.

70.

Weiner

Perry

R. P.

Magnusson

(1988). An attributional analysis of reactions to stigmas. Journal of Personality and Social Psychology, 55, 738–748.

71.

Wyer

R. S.

Adaval

Colcombe

S. J.

(2002). Narrative-based representations of social knowledge: Their construction and use in comprehension, memory and judgment. In Zanna

M. P.

(Ed.), Advances in experimental social psychology (Vol. 34, pp. 131–197). San Diego, CA: Academic Press.

72.

Yzerbyt

Rogier

(2001). Blame it on the group: Entitativity, subjective essentialism, and social attribution. In Jost

J. T.

Major

(Eds.), The psychology of legitimacy (pp. 103–134). Cambridge, UK: Cambridge University Press.

73.

Zucker

G. S.

Weiner

(1993). Conservatism and perceptions of poverty: An attributional analysis. Journal of Applied Social Psychology, 23, 925–943.