Jump to ContentJump to Main Navigation
Altruism in Humans$

C. Daniel Batson

Print publication date: 2010

Print ISBN-13: 9780195341065

Published to Oxford Scholarship Online: May 2011

DOI: 10.1093/acprof:oso/9780195341065.001.0001

Benefits of Empathy-Induced Altruism

(p.163) 7 Benefits of Empathy-Induced Altruism
Altruism in Humans

C. Daniel Batson (Contributor Webpage)

Oxford University Press

Abstract and Keywords

Empathy-induced altruism offers a number of benefits. Most obvious are the benefits that accrue to individuals whose needs elicit empathic concern, but research suggests that empathy-induced altruistic motivation can also benefit groups in need. It may even benefit the person experiencing this motivation. The potential benefits discussed in this chapter are: More, more sensitive, and less fickle help; less aggression and derogation of the victims of injustice; increased cooperation and care in conflict situations, including conflict in bargaining and negotiations, ethnic, religious, and political conflicts, and racial conflicts in educational settings; improved attitudes toward, and action on behalf of, stigmatized groups; more positive close relationships, including friendships and romantic relationships; and better mental and physical health for the altruistic helper.

Keywords:   aggression, attitudes toward stigmatized groups, bargaining, close relationships, cooperation, derogation, empathy-induced altruism, mental health, physical health

The most obvious benefits of empathy-induced altruism are those for the individuals whose needs elicit empathic concern, but there are other benefits as well. Research suggests that empathy-induced altruistic motivation can also benefit groups in need. It may even benefit the person experiencing this motivation. Let us consider these different benefits—and the relevant research—in turn.

More, More Sensitive, and Less Fickle Help

Clearly, empathy-induced altruism is not the only motive for helping. One can help to gain rewards, avoid punishments, or reduce one’s own distress caused by witnessing another’s distress (see Chapter 3). But empathy-induced altruistic motivation can produce more, more sensitive, and less fickle help than these egoistic motives.

More Help

The research reviewed in Chapter 5 reveals a number of specific circumstances in which empathy-induced altruism can increase the likelihood of help being offered: when escape from the need situation is easy (see Appendix B), when helping is anonymous (Appendix C), when failure to help is justified (Appendix D), when there will be no feedback about the effectiveness of one’s helping effort (Appendix F), and when one anticipates a mood-enhancing experience even if one does not help (Appendix G). In each of these situations, empathy-induced altruism has been found to produce more help than egoistic motives alone.

More Sensitive Help

Not only can empathy-induced altruistic motivation produce more help across a range of circumstances; it can produce more sensitive help as well. Because altruistic motivation (p.164) is directed toward enhancing the welfare of the person in need, the behavior it evokes is likely to be responsive to the need. Egoistic goals of gaining rewards and avoiding punishments can often be reached even if the help offered does not effectively address the needy person’s suffering. To satisfy these motives, the thought counts. When the motivation is altruistic, it’s the result that counts, not the thought. Failure to address the need, even when the failure is in no way one’s fault, will be disappointing. Research reviewed in Chapter 5 again supports this reasoning. Unlike those feeling little empathic concern, empathically aroused individuals feel bad if their own or another person’s helping effort does not succeed, even when they can in no way be blamed for the ineffectiveness (Batson, Dyck et al., 1988; Batson & Weeks, 1996). Capitalizing on this distinction, economists have used concern for the effectiveness of one’s help to differentiate egoistic and altruistic motives for contributing to charities (Ribar & Wilhelm, 2002). Empathy-induced altruism is directed toward what is good for the target(s) of empathy, not toward a display of one’s own goodness.

An experiment by Sibicky, Schroeder, and Dovidio (1995) provided a nice demonstration of the sensitivity that characterizes help evoked by empathic concern. Participants in their experiment either were or were not induced to feel empathy for a person in need; then they were given a chance to help this person. In addition to the typical condition in which helping would provide benefit, there was also a condition in which helping beyond a minimal level would provide short-term benefit but could harm the person in the long-term. Based on the empathy-altruism hypothesis, Sibicky et al. expected that participants induced to feel empathic concern would help at a lower level in the new condition. Results supported this expectation. In contrast, those not induced to feel empathy did not lower their level of help in the new condition. Sibicky et al. concluded that empathic concern enhances sensitivity to the real need of the person for whom empathy is felt, prompting consideration of the long-term as well as short-term consequences of one’s help.

Even more dramatic are the findings of Penner, Cline, Albrecht, Harper, Peterson, Taub, and Ruckdeschel (2008). These researchers assessed the level of empathic concern felt by parents for their child when he or she was about to undergo an invasive and stressful treatment for pediatric cancer. They found a significant negative correlation between the parent’s level of empathic concern and the level of pain and distress the child experienced during the treatment (as assessed by the child, by nurses, and by trained observers).

What produced this correlation? The parents faced a situation in which sensitive care did not involve freeing the child from pain. That was not possible. Instead, it involved being there with and for the child during the pain. Penner et al. observed that parents feeling high empathy differed from those feeling low empathy in both verbal and nonverbal behaviors when interacting with their child during the treatment. High-empathy parents were more likely to offer supportive and normalizing communication (e.g., comforting, reassuring, and engaging the child in everyday, non-medical activities such as reading and play) rather than invalidating communication (e.g., denying or minimizing the child’s pain). These findings are consistent with the possibility that empathy-induced altruistic motivation led to more sensitive care by the parent, which in turn enabled the child to endure the cancer treatment with less pain and distress.

(p.165) Less Fickle Help

In addition to producing more and more sensitive help, altruistic motivation is also likely to be less fickle than egoistic motivation. Research reviewed in Chapter 5 indicates that individuals experiencing relatively low empathic concern and, hence, a relative predominance of egoistic over altruistic motivation are far less likely to help when they can easily escape exposure to the need situation without helping, or when they can easily justify to themselves and others a failure to help (Batson et al., 1981; Batson et al., 1988; Toi & Batson, 1982—see Appendix B and Appendix D). The practical implications of these findings are more than a little troubling. Easy escape and high justification are common features of many helping situations we face in life. Amidst the blooming, buzzing confusion, we can almost always find a way to direct attention elsewhere, or to convince ourselves that inaction is justified. Given this fact, the practical potential of empathy-induced altruistic motivation looks promising indeed. Individuals experiencing relatively high empathic concern show no noticeable decrease in readiness to help under conditions of easy escape, high justification, or the two combined.

Less Aggression

A second possible benefit of empathy-induced altruism is inhibition of aggression. Altruistic motivation should inhibit any inclination to aggress against or harm the person for whom empathic concern is felt, even in the face of provocation. Empathy-induced altruism should not inhibit all aggressive impulses, only those directed toward the target of empathy. Indeed, it is easy to imagine altruistic aggression, in which empathic concern felt for Person A leads to empathic anger and, thereby, to increased aggression toward Person B if B is perceived to be a threat to A’s welfare (Hoffman, 2000; Vitaglione & Barnett, 2003).

Miller and Eisenberg’s Meta-Analysis

In apparent support of the idea that empathy-induced altruism can inhibit aggression, Miller and Eisenberg (1988) concluded from a meta-analysis of approximately fifty studies that “empathy is negatively related to aggression, externalizing [i.e., threatening, attacking, and fighting, as well as general disobedience], and antisocial behaviors” (p. 338). However, a close look at the studies reviewed by Miller and Eisenberg suggests the need for a more guarded conclusion.

First, in many of the reviewed studies, the negative relation between empathy and aggression was weak. Overall, there was “modest but not entirely consistent support for the notion that empathic responsiveness may be an inhibitor of aggression” (Miller & Eisenberg, 1988, p. 339). Second, the clearest evidence for inhibition was found in studies that assessed empathy using self-report questionnaire measures of a general disposition to experience empathic concern. As noted in Chapter 2, responses on questionnaire measures of dispositional empathy may reflect desire to present oneself as a nice, sensitive, (p.166) caring person rather than readiness to feel empathy (Batson et al., 1986). Thus, the lower aggression associated with these measures may not be the result of empathic concern but of desire to be—or to appear—nice. Such a desire could easily reflect egoistic motivation either to avoid social and self-censure or to gain social and self-rewards.

Third, virtually all of the reviewed studies that assessed empathy toward a specific person assessed empathy toward someone other than the target of aggression. To find that reporting empathic concern for one person is associated with displaying less aggression toward another could indicate a general disposition toward empathic concern that produces a general inhibition of aggression. But, once again, this association could also indicate a desire to be or to appear nice.

Miller and Eisenberg reported only four studies in which attempts were made to induce empathic concern for the target of potential aggression by experimental manipulation, either manipulation of perceived similarity or of perspective taking. The evidence from these four studies is inconclusive. Eliasz (1980) failed to find a negative relation between empathy and aggression. However, before the empathy manipulation in his study, participants received a harsh evaluation designed to provoke anger and retaliatory aggression. This ordering of events may have prevented empathic concern from ever developing. In each of the other three studies, the experimental induction of empathy significantly inhibited harming the person for whom empathy was felt. But these other three studies were all reported in unpublished dissertations.

Subsequent to the Miller and Eisenberg review, Richardson, Hammock, Smith, Gardner, and Signo (1994, Study 2) attempted to induce male undergraduates to feel empathic concern for a target before the target aggressed against them. Richardson et al. found that these undergraduates aggressed no less in return than did male undergraduates not induced to feel empathy. But no measure of empathic feeling for the target was taken in this study, so we cannot be sure that the empathy induction was successful or that empathy existed at the point of the opportunity to retaliate. There is no reason to expect empathy to inhibit aggression unless it is present when participants have the chance to aggress.

In the one study reported by Miller and Eisenberg in which self-reported empathic concern for the target of aggression was assessed and found to be present after provocation (Gaines, Kirwin, & Gentry, 1977), the empathy-inhibition association was highly significant. But the causal relationship is unclear in this study because assessment of empathy was based on retrospective self-reports after the opportunity to harm the victim. Those who harmed the victim less may have inferred that they felt more empathic concern, rather than empathic concern inhibiting their impulse to aggress. Overall, then, results of research prior to 1995 on the empathy-aggression relationship are far from clear.

More Recent Research

Three more recent lines of research provide clearer evidence that empathy-induced altruism can inhibit aggression. First, research on forgiveness has found that an important step in the forgiveness process is to replace feelings of anger toward a harm-doer with (p.167) empathic feelings (Fincham, Paleari, & Regalia, 2002; McCullough, Rachal, Sandage, Worthington, Brown, & Hight, 1998; McCullough, Worthington, & Rachal, 1997; Witvliet, Ludwig, & Vander Laan, 2001; Worthington, 1998). Of course, replacing feelings of anger with empathic feelings is often easier said than done.

Second, research suggests that empathic concern may be an important antidote to child abuse and neglect, as well as to sexual assault. Milner et al. (1995) examined the empathic responsiveness of mothers while they watched video clips of an infant who was (a) smiling and laughing, (b) looking around, or (c) crying. The mothers were from two matched groups, a group identified as being at high risk of physically abusing a child and a group identified as being at low risk. Low-risk mothers showed a significant increase in empathic concern while watching the crying infant, whereas high-risk mothers showed no reliable change regardless of whether the infant was laughing, looking around, or crying. Instead of empathy, high-risk mothers reported feeling more personal distress and hostility while watching the crying infant (see Frodi & Lamb, 1980, for parallel results using physiological measures). This pattern of response by the high-risk mothers is congruent with clinical reports that physical child abusers experience less empathic concern and more hostility in response to a crying child (de Paúl, Pérez-Albéniz, Guibert, Asla, & Ormaechea, 2008).

Turning from child abuse to neglect, de Paúl and Guibert (2008) provided a thoughtful analysis of the ways in which child neglect can result from breakdown of the process leading from (a) perception of need and valuing the other’s welfare to (b) empathic concern and, thereby, to (c) altruistic motivation. However, de Paúl and Guibert provided no direct evidence to support their analysis. Regarding sexual abuse, clinical interventions aimed at increasing empathy have been found to reduce the reported likelihood of abuse, rape, and sexual harassment by men identified as being at high risk for committing sexual assault (e.g., Schewe, 2002; Schewe & O’Donohue, 1993).

Third, in an intriguing and ambitious experiment, Harmon-Jones, Vaughn-Scott, Mohr, Sigelman, and Harmon-Jones (2004) sought to assess the effect of empathic concern on anger-related left-frontal cortical electroencephalographic (EEG) activity. In the initial phase of the experiment, Harmon-Jones et al. used a perspective-taking manipulation (remain objective; imagine the other’s feelings) to induce undergraduate men and women to experience either low or high empathic concern for another student who was suffering from multiple sclerosis. (Checks indicated that the manipulation was effective.) Later, this other student provided either (a) a harsh and insulting (aggression provoking) evaluation of an essay that the participant had written or (b) a neutral evaluation. EEG activity was recorded immediately after participants received the evaluation. Attitudes toward the other student were also measured. As predicted based on the empathy-altruism hypothesis, relative left-frontal cortical EEG activity, which typically increases after insult and is associated with aggressive behavior (and which increased in the low-empathy condition), was inhibited in the high-empathy condition. Hostile attitudes toward the other student were too. This experiment provides the clearest evidence to date that empathic concern can directly inhibit the desire to aggress, at least when the empathy is in place before provocation.

(p.168) Reduced Derogation and Blaming of Victims of Injustice

More broadly, empathic concern may be effective in counteracting a particularly subtle and insidious form of hostility: blaming the victims of injustice. In his classic work on the just-world hypothesis, Melvin Lerner (1970) found that research participants were likely to derogate an innocent victim of suffering. Lerner argued that this derogation was motivated by desire to maintain a belief in a just world—a belief that people get what they deserve and deserve what they get. Those who get undesirable outcomes must be undesirable people. Protecting one’s belief in a just world in this way can lead to what William Ryan (1971) called “blaming the victim.” Ryan suggested that we are likely to react to the victims of unjust discrimination and oppression by unconsciously blaming them: If they have less, they must be less deserving. Ryan further suggested that the prototypical victim blamer is someone who is fairly well-off financially but not entirely secure. By blaming the victims of poverty and social injustice, such people can reassure themselves about their own financial situation; they really do deserve their relative advantage.

Derogation and blaming the victim are all too common alternatives to caring about poverty and social injustice. These alternatives can lead to smug acceptance of the suffering of others as just and right. But empathy-induced altruism may counteract this tendency. In an important follow-up to Lerner’s classic experiments, Aderman, Brehm, and Katz (1974) found that perspective-taking instructions designed to induce empathic concern eliminated the tendency for participants to derogate an innocent victim.

Increased Cooperation and Care in Conflict Situations

There is also evidence that empathy-induced altruistic motivation can increase cooperation and care in conflict situations. Paradigmatic of such situations is a one-trial Prisoner’s Dilemma like the following (adapted from Rapoport and Chammah, 1965): Two people must each choose between two options—cooperate or defect—without knowing the other’s choice. If both choose to cooperate, each receives a payoff of +15; if both defect, each receives a payoff of +5. If one cooperates and the other defects, the former receives nothing and the latter receives a payoff of +25. Given these payoffs, if both people defect, they are each individually worse off (+5) than if they both cooperate (+15). On the other hand, it is best for each person (P) to defect regardless what the other (O) does. To illustrate, if O cooperates, P receives +25 by defecting but only +15 by cooperating; if O defects, P receives +5 by defecting but nothing by cooperating. There is irony—and fascination—in this simple dilemma.

If one faces a Prisoner’s Dilemma repeatedly over a number of trials, it is in one’s interest to cooperate on at least some trials. Strategies like tit-for-tat, where P cooperates on the first trial and then responds on every subsequent trial as O responded on the (p.169) previous trial, are likely to produce more overall personal gain than a strategy of relentless defection—even though defecting is optimal on each individual trial (Axelrod & Hamilton, 1981; Nowak, May, & Sigmund, 1995). However, in a one-trial situation, the situation in which the Prisoner’s Dilemma was originally conceived, tit-for-tat and other strategies for inducing reciprocity are irrelevant (Dawes, 1991). Why would anyone cooperate in a one-trial Prisoner’s Dilemma?

Cooperation in a One-Trial Prisoner’s Dilemma

Narrow versions of game theory and the theory of rational choice both predict no cooperation in a one-trial Prisoner’s Dilemma because each theory assumes that there is only one motive in play: material self-interest. Regardless of what the other person does, material self-interest is best served by defecting. However, broader versions of rational choice allow for forms of self-interest that can be served by cooperating, such as feeling good about oneself or avoiding pangs of guilt. These broader versions can account for the finding that as many as one-third to one-half of people placed in a one-trial Prisoner’s Dilemma cooperate.

What about empathy-induced altruistic motivation? The empathy-altruism hypothesis predicts that if one person in a Prisoner’s Dilemma is induced to feel empathic concern for the other, then this person should be even more likely to cooperate. In addition to the various forms of self-interest, this person should also be motivated by interest in the other’s welfare. And the other is always better off if one cooperates than if one defects.

An Initial Test

To provide an initial test of this prediction, Batson and Moran (1999) conducted an experiment in which undergraduate women faced a one-trial Prisoner’s Dilemma. These women learned at the outset that they would never meet the other woman participating in the dilemma (who was actually fictitious). Payoffs were the same as those just described, but were made concrete and real in the form of the number of raffle tickets (from 0 to 25) received. The prize was a $30 gift certificate at any store the winner chose.

All participants in the experiment were told that one factor being studied was type of interaction between participants prior to choosing, and that they were in a condition with indirect rather than face-to-face interaction. What participants were told indirect interaction meant differed across experimental conditions. One-third of the participants learned that it meant no communication would occur between themselves and the other woman. The other two-thirds learned that it meant one-way written communication, and that they had been randomly assigned to be the Receiver of the communication. As Receiver, they would read a note the other woman—the Sender—had written before knowing anything about the study. The note was to be about something interesting that happened to the Sender recently.

(p.170) The Sender’s note was always the same. It told of being down after suffering a breakup with her boyfriend. The note ended: “I’ve been kind of upset. It’s all I think about. My friends all tell me that I’ll meet other guys and all I need is for something good to happen to cheer me up. I guess they’re right, but so far that hasn’t happened.” It was assumed participants would think that giving the Sender more tickets and a better chance at the raffle by cooperating might cheer her up, whereas reducing her chances by defecting would not.

Perspective-taking instructions given prior to reading the note manipulated empathic concern for the Sender. Participants in a low-empathy condition were instructed to take an objective perspective toward what was described in the note. Those in a high-empathy condition were instructed to imagine how the Sender felt about what was described.

After reading the note from the assigned perspective (or reading no note), participants made their decision to cooperate or defect. Results revealed that cooperation was much higher among participants induced to feel empathic concern for the other woman (75 percent) than among those not induced to feel empathy—whether those in the no-communication condition (30 percent) or the communication/low-empathy condition (35 percent). (For other evidence of empathy-induced cooperation in dilemmas, see Cohen & Insko, 2008; Van Lange, 2008; Wade-Benzoni & Tost, 2009.)

A More Stringent Test

In a subsequent experiment, Batson and Ahmad (2001) used a similar procedure to conduct an even more stringent test of the ability of empathic concern to increase cooperation. Rather than the standard one-trial Prisoner’s Dilemma, in which participants make their decisions simultaneously without knowing what the other has done, Batson and Ahmad altered the procedure so that decisions were made sequentially. All participants learned that the other woman had been randomly chosen to go first, and that she defected. Thus, when each of the undergraduate women in this experiment made her decision, she knew that the other woman (again, actually fictitious) had already defected. This meant that possible payoffs for the participant were either to receive 5 tickets if she also defected (in which case, the other woman would receive 5 tickets as well) or to receive 0 tickets if she cooperated (in which case, the other woman would receive 25 tickets).

Predictions from game theory, from the theory of rational choice, and even from theories of justice and social norms are clear. In this sequential situation, there is no longer a dilemma at all; the only rational thing to do is to defect. Not only will defecting maximize your own outcome but it will also satisfy the norms of fairness and distributive justice. Moreover, there is no need to worry about feeling guilty should you defect and the other person cooperate, as can happen in a simultaneous-decision dilemma. The other woman has already defected. Not surprisingly, in the very few previous studies that even bothered to look at such a situation, the rate of cooperation was extremely low (around 5 percent—see Shafir & Tversky, 1992; Van Lange, 1999).

(p.171) The empathy-altruism hypothesis predicts that even in this sequential situation a dilemma remains for participants led to feel empathic concern for the defecting woman. For them, self-interest and fairness counsel defection, but empathy-induced altruism counsels cooperation. Results again patterned as predicted by the empathy-altruism hypothesis. In the absence of empathy—i.e., in the no-communication condition and the communication/low-empathy condition—cooperation was extremely low (0 percent and 10 percent, respectively). When empathy was induced, cooperation rose to 45 percent. Empathy-induced altruism was not strong enough to override other motives (self-interest, retribution, fairness) for all participants led to feel empathic concern, but it was strong enough to do so for almost half. Building on the same logic, Rumble, Van Lange, and Parks (2010) showed that empathy-induced altruism can also counteract the effects of unintended incidents of defection (“negative noise”) in an iterated social dilemma.

Results of these experiments suggest that empathy-induced altruism can add complexity to economic bargaining situations. When one feels empathic concern for the other, one’s interest lies not only in maximizing one’s own gains—or even in maximizing joint gains—but also in maximizing the other’s gains. Insofar as I know, the idea of using empathy to increase cooperation in a one-trial Prisoner’s Dilemma had not even been considered in any of the over 2,000 Prisoner’s Dilemma studies previously conducted. I suspect this was because no one thought empathy-induced altruistic motivation could increase cooperation. Yet clearly it can. Indeed, inducing empathy seems far more effective than most other techniques that have been proposed to increase cooperation in one-trial dilemmas.

More Positive Negotiations

Might induction of empathy-induced altruism be worth pursuing in real-world conflict situations, such as business or political negotiations? Or is allowing oneself to feel concern for the other’s welfare in these situations too big a risk to take? Think, for example, of negotiations between management and labor, between Catholics and Protestants in Northern Ireland, between Palestinians and Israelis, between Pakistanis and Indians. Empathy-induced altruism might prompt one to give ground. But it might also produce a better outcome for all. It might even save lives.

Research by Galinsky, Maddux, Gilin, and White (2008) suggests that empathic concern may have both of these effects on negotiations—(a) prompting one to give ground and (b) creating a more positive environment that may, in the long run, produce a better outcome for all. In one experiment, Galinsky et al. had M.B.A. students in a negotiations course pair up and engage in a 30-minute two-party negotiation exercise. One student played the role of a Job Candidate, and the other, a Recruiter. Eight issues were negotiated, including salary, work location, bonus, vacation time, and so on. Both students knew that some of these issues mattered more to the Candidate and some mattered more to the Recruiter. Joint gain could be maximized by being sensitive to which issue mattered most to whom and using this information to negotiate trade-offs. As a manipulation of (p.172) perspective taking, the student in the Recruiter role was randomly assigned to one of three sets of instructions: (a) consider your own role carefully, (b) try to understand what the Candidate is thinking, and (c) try to imagine what the Candidate is feeling. The imagine-feelings instructions were similar to instructions used in previous studies to induce empathic concern.

Dyads with Recruiters assigned to one of the latter two sets of instructions each produced greater joint gain than dyads with Recruiters assigned to consider their own role. For Recruiters who focused on the Candidate’s thoughts, the difference was highly significant; for those who focused on the Candidate’s feelings, the difference was marginal. More interesting was how the greater joint gains were achieved. Recruiters who focused on the Candidate’s thoughts got more of what they wanted than did Recruiters who focused on the Candidate’s feelings. On the other hand, Candidates negotiating with a Recruiter focused on their feelings got more of what they wanted than did Candidates negotiating with a Recruiter focused either on the Recruiter’s role or (non-significantly) on their thoughts.

These results led Galinsky et al. to conclude that, when negotiating, it is more effective “to ‘think for’ than to ‘feel for’ one’s adversaries” (Galinsky et al., 2008, p. 383). Recruiters who imagined the Candidate’s feelings—which presumably induced empathic concern (we cannot be sure because no measures of emotion were taken)—gave ground, benefiting the Candidate at cost to self. Recruiters who were able to get inside the adversary’s head and strategy, as a skilled chess-player might, got more of what they wanted.

But these results were in the short term, in a single negotiation. What about long-term effects in situations where negotiators interact over time? In such situations, it seems possible that the goodwill produced by giving ground might tip the scales in favor of negotiators who imagine feelings rather than those who imagine thoughts. In the long run, those who imagine feelings, not thoughts, might end up with the better overall outcome.

Consistent with this possibility, in a different negotiation exercise Galinsky et al. (2008) found that Sellers who negotiated with an empathic Buyer (one focused on the Seller’s feelings) were significantly more satisfied with the way they were treated during the negotiation than were Sellers who negotiated with a Buyer focused on their thoughts. This was true even though agreement on a sale was (non-significantly) more likely to be achieved in the latter case than the former. Whether these feelings of satisfaction would produce more productive subsequent negotiations seems worth exploring in future research.

Reduced Intergroup Conflict

When one thinks of trying to use empathy-induced altruism to reduce intergroup conflict (e.g., conflict between religious, racial, or ethnic groups), two problems immediately loom. First, intergroup relations often have a history of disdain and mistrust, if not outright hostility. To feel empathic concern requires that one be other-oriented, valuing the other’s welfare and attending with sensitivity to the other’s plight. In the face of such (p.173) history, is other-oriented sensitivity not too much to ask? Second, both empathic concern and altruistic motivation are interpersonal processes. One feels and cares for another individual or individuals. Is it possible to feel for a group? Underscoring this second concern, research on the identified-victim effect has found that when an individual in need is one of a group of individuals with similar needs, both empathy and willingness to help are diminished (Kogut & Ritov, 2005b, Experiment 3; Small, Lowenstein, & Slovic, 2007).

To address these two problems, a key element in strategies using empathy-induced altruism to reduce real-world intergroup conflict has been to work from the interpersonal to the intergroup level by providing personalizing contact with one or more outgroup members. Through such contact, members of one group are led to deal with members of the other group on a personal basis, not simply as one of them. (As I am using the term, “personalizing” refers to the nature of one’s interaction with members of the outgroup, not to perception of them along a dimension of similarity to self instead of group membership; see Brewer, 1988, and Miller, 2002, for discussions of “personalization” in this second, perceptual sense.)

Personalizing contact addresses the second problem noted above by inducing empathic concern at the individual not the group level. Such contact should encourage empathic concern for members of the outgroup in two ways. First, it should increase the likelihood of accurately perceiving outgroup members’ needs—their hopes and fears. Second, if positive, personalizing contact should increase the likelihood of valuing outgroup members’ welfare. As discussed in Chapter 2, these two conditions are the antecedents of empathic concern. Personalizing contact can also address the first problem if care is taken that the contact occur in a non-adversarial and low-threat situation, one in which mistrust and conflict are either not evoked or, better yet, are counterproductive.

How can personalizing contact among individuals on opposite sides of an intergroup conflict be achieved? Obviously, it is not easy. More is required than simply bringing the antagonists together. Mere contact is likely to invite further hostility and aggression (Pettigrew, 1998). One structural technique that has proved especially effective in creating non-adversarial personalizing contact, and thereby reducing intergroup conflict and hostility, is to introduce a superordinate goal (Sherif, Harvey, White, Hood, & Sherif, 1961). A superordinate goal is something that both sides in the conflict want but can attain only if the two sides join forces and work together. Potential antagonists find themselves united in the effort to reach a common goal. Strange bedfellows, perhaps, but bedfellows nonetheless.

Think of the psychological consequences. When working together toward a common goal, hostility and aggression are counterproductive. Instead, members of one group must attend to and understand what members of the other group value—what they want and need. And to coordinate efforts in pursuit of the goal, members of each group must attend to the perspective of those in the other group. In combination, these two consequences should increase feelings of empathic concern for members of the outgroup. (Note that these effects on empathy do not require that group members give up their own group identity in order to pursue the superordinate goal—Dovidio, Gaertner, & Saguy, 2009.)

(p.174) Sherif et al. (1961) provided the classic demonstration of effectiveness of superordinate goals in reducing intergroup conflict. In their Robber’s Cave experiment, superordinate goals were used to eliminate the open hostility that had erupted between competing groups of 12- to 14-year-old boys at a summer camp. Although this experiment dramatically demonstrated the effectiveness of superordinate goals, it revealed little about the psychological process through which they work. The following examples of programs designed to reduce intergroup conflict suggest that empathy may play an important role.

Conflict Resolution Workshops, Peace Workshops, and Peace Camps

Stephan and Finlay (1999) pointed out that the induction of empathy is often an explicit component of techniques used in conflict resolution workshops, peace workshops, and peace camps. In conflict-resolution workshops, 3-6 leading figures on opposing sides of an international conflict are brought together in a non-threatening, neutral situation for a brief workshop (rarely lasting more than a week). The confidential, off-the-record interaction is designed to encourage (a) better understanding each other’s position and (b) finding a path toward a mutually beneficial negotiated settlement. The exchange is guided by trained facilitators who establish ground rules and agenda. Perhaps the best-known examples of such workshops are those organized by Herbert Kelman and his colleagues that have brought together Israeli and Palestinian representatives (Kelman, 1990, 1997; Kelman & Cohen, 1986; Rouhana & Kelman, 1994; also see Burton, 1986, 1987; Fisher, 1994).

Immediate goals of these workshops are for each side to understand the perspective of the other side and to begin to trust them. The long-range goal is superordinate—to find a mutually acceptable peaceful solution to the conflict (Kelman, 2005). To these ends, participants are encouraged to express their hopes and fears and to listen to one another’s concerns, actively adopting the perspective of those on the other side but not losing track of real differences. In Kelman’s (1997) words, “Out of these interactions, participants develop increasing degrees of empathy, of sensitivity and responsiveness to the other’s concerns, and of working trust, which are essential ingredients of the new relationship to which conflict resolution efforts aspire” (p. 219).

Peace workshops and camps are typically designed for the young people (teenagers) of warring factions. Workshops often last only 3-4 days; camps may last a month or more. In these workshops and camps, participants from the two sides of the conflict live together, spend free time together, exchange views in dialogue sessions under the direction of trained leaders, take part in structured exercises, and share cultural experiences. These activities provide personalizing contact, superordinate goals, and awareness of outgroup needs. They encourage cross-group friendships, perspective taking, and empathic concern for outgroup members.

One well-known example is the workshop program for Jewish and Arab youth at Neve Shalom/Wahat al Salam (the Hebrew and Arabic names for the same community) (p.175) (Bargal & Bar, 1992; also see Bar-On & Kassem, 2004). Less well-known, but quite interesting because of a one-year follow-up assessment of attitudes and behavior toward the outgroup, was a 4-day peace workshop in Sri Lanka that brought together Sinhalese (majority) and Tamil (minority) youth (Malhotra & Liyanage, 2005). After one year, participants in this workshop expressed more understanding of and concern for the well-being of members of the other group (on a version of Davis’s, 1994, Empathic Concern scale modified to be specific to the other group) than did either of two comparison groups—(a) youth who were nominated for the workshop but did not take part due to budget cuts and (b) youth from demographically similar schools not involved in nominating students. After completing the follow-up questionnaire packet, members of each group were given a chance to donate part of the payment money to a program designed to help poor children of the outgroup. On average, workshop participants donated more than did those in the non-participant groups.

Jigsaw Classroom

The Jigsaw Classroom is a learning technique originally developed in the 1970s by Elliot Aronson and his colleagues to try to overcome racial tension and animosity in desegregated schools in Austin, Texas (Aronson, 2004; Aronson, Blaney, Stephan, Sikes, & Snapp, 1978). This technique is especially revealing of the role of empathy processes in the use of superordinate goals to improve intergroup relations. In a Jigsaw Classroom, students spend part of their school day in racially/ethnically mixed groups (ideally, 5-6 students per group). Each group is given a learning task, and each member of the group has one, but only one, part of the information the group needs to complete the task. As a result, each person in the group must rely on the contribution of every other person to succeed. After about eight weeks the groups are dissolved, new groups are formed, and each student must learn to work effectively with 4-5 more students in a new racially/ethnically mixed group. After another eight weeks, new groups are formed again, and so on.

Aronson et al. (1978) reported that liking for fellow group members increased as a result of the jigsaw experience; so did helping. Unfortunately, Aronson et al. did not report the effect specifically on interracial liking or helping. However, in an earlier study, Weigel, Wiser, and Cook (1975) did report effects of interdependent, ethnically mixed (European-, African-, and Mexican-American) student workgroups on cross-ethnic liking, conflict, and helping. Results of that study indicated that working together in interdependent groups significantly increased both cross-ethnic liking and helping behavior; it also reduced cross-ethnic conflict (also see Johnson & Johnson, 1987).

Why does cooperative interaction in jigsaw groups increase liking and helping? Aronson et al. (1978) suggested that perspective taking (which they called empathy) was “one of the crucial mechanisms underlying the effects” (p. 118—also see Aronson & Bridgeman, 1979). Supporting this suggestion, Diane Bridgeman (1981), in dissertation research under Aronson’s direction, tested students’ ability to adopt the perspective of (p.176) characters in brief stories, seeing the story situation from the character’s point of view rather than their own. She found that students from a Jigsaw Classroom were better at this perspective-taking task than were students from a traditional classroom. Apparently, perspective-taking abilities learned in jigsaw groups generalize. This finding suggests that the ability of empathy-induced altruism to increase cooperation may extend to conflict situations beyond the one in which empathic concern is initially induced. Further, research suggests that programs like the Jigsaw Classroom, which involve learning cooperatively in racially or ethnically mixed groups, increase cross-group friendships, especially close friendships (see Paluck & Green, 2009, and Stephan & Stephan, 2001, for reviews; for other examples of programs designed to enhance empathy in educational settings, see Eisenberg & Morris, 2001).

Writing in the wake of the tragic shooting at Columbine High School in Littleton, Colorado, Aronson (2004) affirmed:

I believe that, if the jigsaw method had been widely used in Littleton, the Columbine massacre might never have occurred, and those 15 people would still be alive. Admittedly, that is a bold statement—one not usually made by academicians. And, of course, it can never be proved. But I have a high degree of confidence because 31 years of research on the jigsaw method have made it undeniably clear: The jigsaw process builds empathy [i.e., perspective taking], and students in jigsaw classrooms are more open to one another, more compassionate, and more tolerant of diversity than students in traditional classrooms. (p. 486)

Roots of Empathy Project

The Roots of Empathy project developed by Mary Gordon (2005) and implemented in primary-school classrooms (kindergarten through Grade 8) in Canada and Australia was not explicitly designed to reduce intergroup conflict. It was designed to develop empathy—including “emotional literacy” (the ability to “find the humanity in one another”) and perspective taking—as a means to (a) increase collaboration and civility and (b) reduce aggression and bullying (Gordon, 2005, p. 8). However, Gordon saw this empathy as the key to overcoming intergroup conflict as well:

Normally, our differences define and separate us, providing the fodder for marginalization, bullying, and exclusion. If we were to listen to the language of the groups who are in a “hate relationship” with another group, they somehow manage to speak of the other group as less human—or so different that there can be no basis for human exchange….

The Roots of Empathy classroom is creating citizens of the world—children who are developing empathic ethics and a sense of social responsibility that takes the position that we all share the same lifeboat. These are the children who will build a more caring, peaceful and civil society, child by child. (2005, pp. xvi–xvii)

The Roots of Empathy project is novel in its approach. The core of the program is a visit to the classroom monthly throughout the school year by a mother (or sometimes a father, or both) and infant from the community. Pupils ring a green blanket on which the parent places the infant. They observe the infant and the parent-infant interaction, interact (p.177) with the baby themselves, and ask the parent questions about what the infant has learned since the last visit. The idea is that “the relationship between the parent and child is a template for positive, empathic human relationships” (Gordon, 2005, p. 6) and that observing the baby’s development and the parent-infant interaction will encourage perspective taking and valuing of the infant’s welfare. Using the parent-infant interaction as a catalyst for empathy development is quite congruent with possibility presented in Chapter 2 that the biological substrate for empathic concern lies in parental nurturance and tenderness.

A trained Roots of Empathy instructor guides the family visits and meets with the class prior to and after each visit, providing basic information about infant development, encouraging pupils to imagine what the infant is thinking and feeling, and extending this perspective taking to the pupils themselves and to peers. When a Roots of Empathy classroom is racially or ethnically diverse, explicit attention is given to bringing in parents and infants from the different groups represented in the class in order to provide a basis for intergroup perspective taking and affection.

Evaluation research assessing effectiveness of the Roots of Empathy project suggests that the program increases children’s emotional development and perspective-taking skills and reduces aggression (Schonert-Reichl, 2005). Compared to children who have not experienced a Roots of Empathy classroom, children who have were rated by both teachers and peers as more advanced in emotional and social understanding. This understanding was, in turn, associated with reduced aggression and increased helping, sharing, and cooperation.

Improving Attitudes Toward—and Action on Behalf of—Stigmatized Groups

It may be possible to use empathy-induced altruism to improve attitudes toward, and action on behalf of, stigmatized groups. Moreover, this may be possible without organizing carefully orchestrated face-to-face contact and introducing superordinate goals. Consider books such as Uncle Tom’s Cabin (Stowe, 1852/2002), Manchild in the Promised Land (Brown, 1965), House Made of Dawn (Momaday, 1968), One Flew Over the Cuckoo’s Nest (Kesey, 1962), The Color Purple (Walker, 1982), and Borrowed Time (Monette, 1988). Think of movies such as A Raisin in the Sun, The Elephant Man, Rain Man, and Longtime Companion. Think of the TV documentaries such as Eyes on the Prize and Promises. Each of these works, and many similar ones, appear designed to improve attitudes toward a stigmatized group—a racial or ethnic minority, an outgroup, or people with some social stigma, disability, or disease. Creators of works like these seem to share two beliefs. First is the belief that by inducing us to imagine the thoughts and feelings of a member of a stigmatized group as he or she attempts to cope, we can be led to value this person’s welfare and to feel empathic concern. Second, that these empathic feelings will generalize, leading us to feel more positively toward the group as a whole. Are they right?

(p.178) Improving Attitudes

To make the attitude-change process implicit in these books, movies, and documentaries explicit, let me outline a three-step model of how empathic concern can serve to improve attitudes toward a stigmatized group (also see Batson, Polycarpou et al., 1997):

  • Step 1. Induce adoption of the perspective of a member of a stigmatized group as he or she describes stigma-related needs. Perspective taking should increase empathic concern for this person.

  • Step 2. This empathic concern should lead to increased valuing of the group-member’s welfare (through the backward inference described in Chapter 2).

  • Step 3. Valuing the group-member’s welfare should generalize to valuing the welfare of the stigmatized group as a whole, producing more positive beliefs about, feelings toward, and concern for the group.

This model poses two empirically testable questions: 1. Can perspective taking be used to arouse empathic concern for the needs of a member of a stigmatized group? 2. If so, will the increased valuing produced by this empathic concern generalize to the group as a whole? The answer to each of these questions seems to be yes—as long as membership in the stigmatized group is a salient aspect of the need for which empathy is induced.

In a series of three experiments, Batson, Polycarpou et al. (1997) successfully used perspective-taking instructions to induce empathic concern for a member of a stigmatized group and, thereby, to improve attitudes toward the group as a whole. Each experiment employed this strategy with a different stigmatized group—people with AIDS (Experiment 1), homeless people (Experiment 2), and to provide an extreme test, convicted murderers (Experiment 3). In each of the experiments, the effect of perspective taking on improved attitudes was clearly mediated by self-reported empathic concern.

Especially interesting was the effect on attitudes toward murderers of inducing empathic concern for a convicted murderer. When attitudes were assessed in the laboratory immediately after the empathy induction, there was only a non-significant trend for research participants in the high-empathy condition to report more positive attitudes toward murderers than participants in the low-empathy condition. But when attitudes were assessed in an unrelated telephone interview 1-2 weeks later, participants who had been induced to feel empathic concern for the convicted murderer in the lab reported significantly more positive attitudes toward murderers in general than participants who had not. Apparently, the high-empathy participants resisted letting their empathic feelings for one murderer influence their attitudes toward murderers in general when these attitudes were assessed immediately and participants were aware of the influence. Later, with their guard down, the effect on attitudes surfaced. Similar long-term effects of empathy on attitude change were reported by Clore and Jeffrey (1972) in a study of attitudes toward the physically disabled.

In related research, inducing empathic concern (including empathic anger) for a member of a racial or ethnic minority has improved attitudes toward the minority group (Dovidio, Johnson, Gaertner, Pearson, Saguy, & Ashburn-Nardo, 2010; Dovidio, ten (p.179) Vergert, Stewart, Gaertner, Johnson, Esses, Rick, & Pearson, 2004; Esses & Dovidio, 2002; Finlay & Stephan, 2000; Vescio, Sechrist, and Paolucci, 2003). Inducing empathic concern for a gay man has improved attitudes toward homosexuals (Vescio & Hewstone, 2001). More broadly, attitude effects of participation in the role-play simulations of discrimination that are often used in educational settings, such as the “Blue Eyes-Brown Eyes” simulation developed by Elliott (Peters, 1987), have been interpreted as being a result of empathy (Byrnes & Kiger, 1990; Weiner & Wright, 1973). And the more positive intergroup attitudes that result from friendship with an ethnic outgroup member have been interpreted as being a result of empathic concern (Brown & Hewstone, 2005; Pettigrew, 1997, 1998). Underscoring the wide applicability of empathy-induced attitude change, Schultz (2000) found that empathic concern felt for animals being harmed by pollution improved attitudes toward protecting the natural environment, and Berenguer (2007) found that pro-environmental attitudes produced by empathy carried over into action on behalf of the environment. Even video computer games have been designed to foster empathy and, thereby, increase concern for the welfare of others (Belman & Flanagan, 2010).

Paluck (2009) conducted an ambitious year-long field experiment in Rwanda to test the effect of a radio soap opera designed to promote reconciliation between Tutsi and Hutu. Along with didactic messages about the roots and prevention of prejudice, the program presented characters wrestling with problems known to all Rwandans, such as cross-group friendships, overbearing leaders, poverty, and memories of violence. The story line featured the struggles of a young cross-group couple who pursue their love in the face of community disapproval and who start a youth coalition for peace and cooperation. The story, especially the young couple’s struggles, seemed to produce intrinsic valuing, perspective taking, and empathic concern. Follow-up measures indicated that these effects generalized, producing increased perspective taking and feelings of concern for a range of people in Rwandan society. Compared to individuals who listened to a soap opera focused on health issues, those who listened to the reconciliation soap opera were more accepting of cross-group marriage and more willing to trust and to cooperate with others in their community, including members of the other group. Paluck (2009) concluded:

The dramatic narrative form of the radio program may have provoked emotional and imaginative processes critical to the changes observed…. Listeners’ emotional empathic reactions to the soap opera characters may have transferred onto the real-life counterparts of the groups the characters represented (measured by the increased empathy for real-life Rwandans—prisoners, genocide survivors, the poor, and leaders). (p. 584)


Do these more positive attitudes manifest themselves in action on behalf of the stigmatized group? Batson, Chang, Orr, and Rowland (2002) provided evidence that they do. Inducing empathic concern for a convicted heroin addict and dealer led to increased budget allocations to help drug addicts. Importantly, the increase in helping occurred even (p.180) when it was clear that the help would not benefit the particular heroin addict for whom empathy was induced. Again reflecting the breadth of applicability, Shelton and Rogers (1981) found that empathic concern induced while watching a video clip showing whales increased readiness to help save whales in general. Apparently, those induced to feel empathy for a member of an outgroup—stigmatized or not—are willing to put their money where their mouth is.

Perceptual/Cognitive Effects of Perspective Taking

Not all of the effects of perspective taking on improved attitudes toward stigmatized groups are a result of empathy-induced altruistic motivation. Research suggests that there are perceptual/cognitive effects as well. Moreover, the perceptual/cognitive effects seem to be different for the two forms of perspective taking identified in Chapter 1. Imagining how a member of a stigmatized group feels about his or her situation (an imagine-other perspective) has been found to lead to situational rather than dispositional attributions for this person’s difficulties (e.g., Regan & Totten, 1975). This attributional shift can, it seems, generalize to the group as a whole, leading to more positive attitudes toward the group (Vescio et al., 2003). Imagining oneself in the place of a member of a stigmatized group (an imagine-self perspective) has been found to reduce negative stereotyping of the group member and the group as a whole (Galinsky & Ku, 2004; Galinsky & Moskowitz, 2000).

These perceptual/cognitive effects appear to be distinct from the effects mediated by empathic concern (Vescio et al., 2003). Whether the perceptual/cognitive effects carry sufficient motivational force to affect behavior, as do the effects mediated by empathic concern, remains an open question. Some research suggests that they may not (Dovidio et al., 2010).

Pragmatic Considerations

Improving attitudes by inducing empathy through novels, movies, and documentaries is likely to be easier, at least initially, than trying to improve attitudes through other methods, such as face-to-face intergroup contact. Why? First, as the novels and movies listed earlier show, it is quite possible for a skilled writer to induce empathic concern for a member of a stigmatized group, either a real or a fictional member (Harrison, 2008; Oatley, 2002; Zillmann, 1991; also see Batson, Chang et al., 2002). Second, this concern can be induced in low-cost, low-risk situations. Rather than the elaborate arrangements required to create direct, cooperative, personal contact, books and TV can lead us to feel empathy for a member of a stigmatized group as we sit comfortably in our own home.

Third, media-generated experiences can be controlled to ensure that they are positive and empathy-inducing far more readily than can live, face-to-face contact. Finally, as long as membership in the stigmatized group is a salient feature of the need for which empathy is induced, the attitude change does not seem vulnerable to sub-typing, whereby (p.181) attitudes improve only toward one or a small subset of “exceptional” members of the group (the exceptions that prove the rule). Sub-typing has been found to plague cognitive approaches to attitude change, such as learning stereotype-inconsistent information about an individual group member (Brewer, 1988; Pettigrew, 1998).

For these four reasons, media-generated empathy-induced attitude change looks promising as a first step toward more positive attitudes and action on behalf of stigmatized groups, and research supports its promise (Graves, 1999; Hayes & Conklin, 1953; Paluck, 2009; Slater, 2002; Strange, 2002). This first step can—and should—be followed with direct, personalizing contact and introduction of superordinate goals, lest one simply understand and feel for imagined or abstract outgroup members, not real ones. (For further discussion of the use of empathy-induced altruism to improve intergroup attitudes and relations, see Batson & Ahmad, 2009b.)

Finally, it should be noted that religion also has long sought to expand the circle of care. In Western religions, the faithful are admonished to “love your neighbor as yourself” (Leviticus 19:18) and that “the stranger who sojourns with you shall be to you as the native among you, and you shall love him as yourself” (Leviticus 19:34), and even to “love your enemies, and do good to those who hate you” (Luke 6:27). In the contemplative traditions of the East, such as Tibetan Buddhism, compassion meditation involves, first, reminding oneself that all beings wish to avoid suffering and to know happiness and, second, cultivating the desire for that wish to be granted not only for those near and dear but also for strangers and, again, even enemies (Ricard, 2006). At this point, the success of these different religious efforts is either questionable (Batson, Floyd, Meyer, & Winner, 1999) or yet to be systematically studied.

More Positive Close Relationships

Intrinsic valuing of the other in a friendship, romantic relationship, marriage, or family relationship should set the stage for feeling empathic concern when that other is in need. The resulting altruistic motivation directed toward having the other’s need relieved should, in turn, make for a more positive relationship. There is much evidence that greater intrinsic value (i.e., love) predicts relationship satisfaction and relationship longevity (Berscheid & Reis, 1998), but research that has explored the role of empathy-induced altruistic motivation in producing these effects is quite limited. Researchers have more often focused on personal needs met through the relationship (Berscheid, 1983; Kelley, 1979; Rusbult, 1980), on compliance with normative expectations for relationship-appropriate behavior (Clark & Mills, 1979), and on gaining a safe haven and secure base from which to operate (Bowlby, 1969; Mikulincer & Shaver 2003). The limited research on empathy-induced altruism in close relationships does, however, suggest positive effects.


Beginning with friendship, Schlenker and Britt (1997) extended research on impression management to show that people will selectively present information about a friend to a (p.182) third person in order to promote the friend’s interests. Research participants were asked to describe a same-sex friend to a person of the opposite sex whom they believed the friend regarded either as extremely attractive or as unattractive. Even though the friend would not learn of their description, participants tended to present the friend as having attributes the attractive person liked, but as not having attributes the unattractive person liked. Thus, participants promoted their friend’s chances with the attractive person, while communicating to the unattractive person that the friend was “not your type.” Schlenker and Britt (1997) suggested that the motivation to look out for a friend in this way could be altruistic, although they recognized that egoistic motives could also account for their results.

Providing some evidence for the role of empathy-induced altruism in friendship development, Crocker and Canevello (2008) found that first-semester university students who self-reported other-oriented, compassionate goals for their developing relationships with friends and roommates (rather than self-image-enhancing goals) also reported more closeness, support, and trust in their relationships. These findings are suggestive, but clearly, much more research is needed on the role of empathy and altruism in friendships. To date, there has been remarkably little.

Romantic Relationships

Turning to romantic relationships, most of the research that has considered the role of empathy-induced altruism has looked at caregiving in the broad framework of attachment theory (Bowlby, 1969; Mikulincer & Shaver 2003). The research has focused on antecedents and consequences of sensitive and responsive care. Relying on self-report measures, Feeney and Collins (2001, 2003) found, first, people who report that the help and support they provide to their romantic partner is altruistically rather than egoistically motivated also report that their care is more sensitive and responsive. Their partners report this as well, but to a lesser degree. Feeney and Collins found, second, partner assessment of the sensitivity of the care received was associated with partner satisfaction with the relationship both in the present and (again, to a lesser degree) two to three months later. Similarly, using daily reports Maisel and Gable (2009) found need-responsive social support by one partner was associated with less sadness and anxiety in the second partner, and with more positive ratings of relationship quality.

Moving beyond general ratings of the quality of caregiving, several studies have employed a procedure developed by Simpson, Rholes, and Nelligan (1992), in which the caregiving of one partner in a romantic relationship is observed when the other partner is about to undergo a stressful experience. As a stressful experience, Feeney and Collins (2001) had the other partner prepare and give a videotaped speech for evaluation by peers. Level of the speech-giver’s need was manipulated by providing the caregiving partner with information concerning how nervous the speech-giver felt about the speech, either very nervous (high need) or not (low need). The caregiver then had an opportunity to write a private note to the speech-giver. Content of the note served as a behavioral measure of caregiving. Self-reports by caregiving partners of their general empathic (p.183) tendencies and altruistic motivation had been collected approximately one week earlier. These self-reports were associated with more sensitive caregiving—i.e., with providing a level of emotional support in the note (as rated both by the speech-giver and by the researchers) that was sensitive to the speech-giver’s apparent level of need.

Obviously, caution is needed in interpreting the results of these studies that rely on self-reported motives. First, there is the general issue of whether people know and will honestly report their motives (Chapter 4; Nisbett & Wilson, 1977), especially motives for providing care in a romantic relationship. As Kelley (1983) aptly put it:

The rules for showing altruism [in a romantic relationship] … are well known to ordinary people and therefore afford the basis for favorable self-presentations that may misrepresent a person’s true motives. There is much to be gained from convincing our partners that we are attuned to their interests and willing to put them before our own. And there is even more to be gained, for many persons, from convincing themselves of their beneficent motives. (p. 285)

Second, there is the more specific issue of the motives included under the umbrella label of “altruistic.” In the study reported by Feeney and Collins in 2001, participants’ motives were assessed by presenting them with the phrase: “On occasions when I help my partner, I generally do so because…” Participants then rated a number of possible motives (1 = strongly disagree, 6 = strongly agree). An example of the sixteen items used to assess altruistic motivation is: “I love my partner and am concerned about my partner’s well-being.” Such a motive seems clearly altruistic, but whether individuals’ ratings of agreement with this statement are a valid indicator of the degree to which they are altruistically motivated is less clear. In the study reported by Feeney and Collins in 2003, altruistic motivation was assessed by this item and six others. Unfortunately, the other six were not as clearly altruistic (e.g., “I can’t stand to see my partner hurting”), casting further doubt on the validity of this measure.

In an as yet unpublished follow-up to the Feeney & Collins (2001) research, Collins, Ford, Guichard, Kane, and Feeney (2008, Study 1) used the same stressor—giving a videotaped speech—and the same manipulation of level of need. However, they moved beyond self-report assessment of general tendencies toward empathic concern and altruistic motivation to assess empathic concern and altruistic motivation in the specific situation. To assess empathy, they measured both partner focus (i.e., perspective taking) and situation-specific empathic concern felt by the caregiving partner for the speech-giver. To assess altruistic motivation, they measured how often the caregiver checked for a message from the speech-giver requesting help with the speech. They also measured the caregiver’s willingness to forego working on enjoyable puzzles in order to give the speech in the partner’s place. Collins, Ford, Guichard, Kane, and Feeney (2010) summarized results from this study as follows: Secure individuals (those low in relationship anxiety and avoidance) in the caregiver role, compared to insecure individuals,

showed clear evidence of responsiveness. They experienced more emotional empathy and were more cognitively focused on their partner (thinking about their partner’s feelings, being distracted by thoughts of their partner while working on their own puzzle task) in the high-need condition (when they were led to believe that their partner was highly distressed about an (p.184) upcoming speech task) than in the low-need condition (when they were led to believe that their partner was not at all distressed about the task). They also provided more behavioral support in the high-need condition, as evidenced by an increase in the number of times they check a computer monitor for messages from their partner and by an increased willingness to switch tasks with their partner (volunteering to give the speech in place of their partner). (Collins et al., 2010, p. 382)

From this report we cannot be sure that empathic concern mediated the effect of security on vigilance and helping. However, research by Mikulincer et al. (2005) suggests that such mediation is likely. Using subliminal as well as supraliminal primes of relationship security, Mikulincer et al. found consistent evidence that priming relationship security led to increased empathic concern and, as a result, to increased willingness to help a person in need—even a stranger with whom the helper had no close relationship. Both Mikulincer et al. (2005) and Collins et al. (2010) concluded that relationship security enables individuals to shift from self-focus and self-concern to other-oriented empathic concern and altruistic motivation.

Finally, as noted earlier, empathic concern for one’s relationship partner after the partner transgresses has been found to be a strong predictor of forgiveness in a variety of close relationships—romantic relationships, family relationships, and friendships (McCullough et al., 1997, 1998). Fincham et al, (2002) found that empathic emotion predicted forgiveness of an imagined transgression in long-term marriages, especially among husbands. Willingness to forgive should make for more positive—and enduring—relationships.

Better Health for the Altruistic Helper

Evidence reviewed thus far suggests that when you experience empathy-induced altruistic motivation, those who are the target of your empathic concern can benefit in various ways. Can you benefit as well? Quite possibly. The evidence is largely circumstantial at this point, but empathy-induced altruism may contribute to your psychological and even physical health.

Circumstantial Evidence

Anecdotal testimony to the elixir of altruism runs the gamut from Dickens’s classic fictional depiction of the happiness and self-fulfillment experienced by Ebenezer Scrooge after his transformation in A Christmas Carol (1843/1913) to the following words of an elderly widow who adopted a dog from her local Humane Society:

Six months a widow, I had found my empty home unbearable. Adopting Mandy was my answer, and I became her savior as well…. She was malnourished and dehydrated. The steadfast love and loyalty we have given each other has been a quiet joy unlike any other! (Cohen & Taylor, 1989, p. 2)

There is also circumstantial evidence from more systematic research. Several studies suggest that adolescents who volunteer to tutor feel better about themselves (e.g., Yogev (p.185) & Ronen, 1982), although other studies have found that volunteering has no effect on adolescents’ depression, self-esteem, and well-being (e.g., Johnson, Beebe, Mortimer, & Snyder, 1998). Newman, Vasudev, and Onawola (1985) conducted a survey of older adults (55 to 85 years old) who volunteered to help in schools and found that 65 percent reported improved life satisfaction, 76 percent reported feeling better about themselves, and 32 percent reported improved mental health. The Changing Lives of Older Couples Study produced evidence that giving support to others reduces risk of mortality and also increases resiliency in the face of grief over spousal loss (Brown, Nesse, Vinokur, & Smith, 2003; Brown, Smith, Schultz, Kabeto, Ubel, Poulin, Yi, Kim, & Langa, 2009). Midlarsky and Kahana (1994, 2007) suggest that such effects may be due, in part, to feeling other-oriented empathic concern and to altruistic motivation.

Among adults in general, Thoits and Hewitt (2001) found that the number of hours a person engaged in volunteer activities correlated positively with reported self-esteem, life satisfaction, and physical health—and correlated negatively with depression. Luks (1991) collected self-report testimonials from over 3,000 volunteers in a range of settings across the U.S., all of whom were regularly involved in helping others. Analogous to feelings experienced during and after vigorous exercise, many of these volunteers reported feeling a “high” while helping—a sense of stimulation, warmth, and increased energy—and a “calm” afterward—a sense of relaxation, freedom from stress, and enhanced self-worth. These reports more often came when the help involved close personal contact with a person in need not previously known, suggesting that empathic concern and altruistic motivation may play a role. Such effects were less frequent when the help involved either obligatory helping of family and friends or anonymous donations of time and money (Luks, 1991).

Reviewing much of this circumstantial evidence of health benefits from helping, Dovidio, Piliavin, Schroeder, and Penner (2006) raised several methodological concerns. First, most of these studies, especially those examining psychological rather than physical benefits, rely on self-reports. It is not clear how valid these self-reports are. Second, the research is almost exclusively correlational. In some cases, attempts are made to control factors other than helping that might produce the benefits in question (socioeconomic status, prior health, staying busy, social interaction, and so on), and in some cases longitudinal data are collected. But one can never be sure that all of the relevant other factors have been adequately controlled. And even when longitudinal data are collected, rarely have cross-lagged correlations been adequately tested (Campbell & Stanley, 1966). Thus, what is actually being measured often remains unclear, and what is causing observed associations does too.

In an effort to respond to such concerns, Oman (2007) focused on six studies that made some attempt to control confounds when testing the hypothesis that volunteering increases longevity among older adults (55 years old or older). His conclusion was encouraging:

Although these studies do not agree in the precise details of their findings, the overall pattern seems clear: Volunteering is associated with substantial reductions in mortality rates, and these reductions are not easily explained by difference in demographics or socioeconomic (p.186) status, or by prior health status or other types of social connections and social support, or by prior level of physical activity and exercise. (Oman, 2007, pp. 25–26)

Is Empathy-Induced Altruism the Cause?

Still, whether the reported health benefits are due to helping and, more specifically, to altruistically motivated helping, is far from clear. Many if not all of the benefits may be a result of (a) displaying competence and control (Langer, 1989), (b) increased social connection (Thoits & Hewitt, 2001), or (c) focusing on something outside oneself (Midlarsky, 1991; Schwartz, 2007). If so, caring for another person’s welfare may be one way to produce these effects, but not the only way. Joining a tennis team, a bird-watching group, or a bridge club might be as effective. In spite of some optimistic claims—e.g., “It seems that human beings are wired to do well by doing good” (Post, 2007, vi)—it is not yet clear that empathy-induced altruism deserves credit for the health benefits associated with “doing good.”

That said, the possibility certainly deserves serious consideration. Beyond surveys, where might we find relevant evidence? If, as suggested in Chapter 2, empathy-induced altruism has its biological roots in parental nurturance and tenderness, then one might expect this altruism to be associated with oxytocin release, and oxytocin release has, in turn, been found to be associated with beneficial effects on the immune system and on response to stress (Carter 2007; Marques & Sternberg, 2007). So, to the extent that it is a generalized expression of parental nurturance, empathy-induced altruism may tap neurochemical resources that promote health.

One potentially fruitful context beyond parent-child relations for exploring possible neurochemical links is care for animals. There is preliminary evidence that positive, caring human-dog interactions produce oxytocin release in both the human and the dog (Odendaal & Meintjes, 2003). Care for animals, especially pets, has long been thought to promote psychological and physical health, producing greater meaning in life, less stress, lower blood pressure, and even longer life (Allen, 2003; Dizon, Butler, & Koopman, 2007). In particular, care for companion animals has been associated with health benefits for the sick and lonely in nursing homes and for prison inmates (Netting, Wilson, & New, 1987). As one young woman prisoner explained, “In prison time is endless, yet with a dog to love, time has meaning” (Cohen & Taylor, 1989, p. 62). However, the degree to which health benefits of human-animal interaction are a result of care for the animal as opposed to care by the animal is, as yet, unclear. So is the role of oxytocin and other neurochemicals in mediating these benefits.

Two Boundary Conditions

Even if empathy-induced altruism proves to be an antidote for depression, meaninglessness, and tension, two potentially important boundary conditions need to be noted. First, just as too much of almost any medicine can do you harm, too much selfless concern for others may lead to “caregiver burnout” (Maslach, 1982). It appears that both (p.187) volunteer and professional helpers—AIDS buddies, hospice workers, doctors, social workers, and therapists—who take on too heavy a load of other people’s burdens may find they run dry and have nothing more to give (Omoto & Snyder, 2002; Schultz, Williamson, Morycz, & Biegel, 1991). Our capacity to experience empathic concern is not a bottomless well. Perhaps we can feel only so much compassion before we go numb. If so, there may be biological and psychological limits to the health-giving properties of empathy-induced altruism.

A second boundary condition is even more fundamental. It is possible that those who turn to altruism as an antidote for depression, meaninglessness, and tension will find it does not work. To use altruism as yet another self-help cure—providing a means to the ultimately self-serving ends of gaining more meaning and better health—involves a logical and psychological contradiction. As soon as benefit to the other becomes an instrumental means to gain these self-benefits, the motivation shifts from altruistic to egoistic. So, if it is empathy-induced altruistic motivation—rather than simply helping behavior—that produces the health benefits noted above, intentional pursuit of these benefits may be doomed to failure. Altruism may enhance health as an unintended consequence but not be useful as a strategy to produce them. (See Batson, 1991, and Wallach and Wallach, 1983, for further discussion of this issue.)


More, more sensitive, and less fickle help for individuals in need. Less aggression. Less child abuse and neglect. Reduced sexual assault. More forgiveness. Less derogation and blaming of victims of injustice. Increased cooperation in conflict situations—including one-trial Prisoner’s dilemmas, negotiations and bargaining, political disputes, and tension between racial and ethnic groups in schools. More positive attitudes toward stigmatized groups. Increased willingness to help these groups. Increased concern for endangered species. More sensitive and responsive care in friendships, romantic relationships, and marriages. More happiness and increased self-esteem. A sense of fulfillment and meaning in life. Less stress. Increased longevity.

The list of potential benefits of empathy-induced altruism for which there is at least preliminary empirical evidence is impressive. Empathy-induced altruism is, it seems, a potentially powerful force for good. But it is no panacea. Empathy-induced altruism can create problems as well as cure them. To fully understand the role of altruism in human life, we need to recognize and appreciate its liabilities. Only then can we tap its power responsibly.