Social Exclusion and Prosocial and Antisocial Behavior



Social Exclusion and Prosocial and Antisocial Behavior

Dianne M. Tice, Jean M. Twenge, and Brandon Schmeichel

During the 1990s, a series of violent incidents at American schools took the lives of a series of schoolchildren. The murderers were themselves pupils in the schools, and the common thread seemed to be that they felt themselves to be rejected by their peers. In some cases, a romantic rejection appeared to be the precipitating incident, whereas others felt chronically left out by the dominant cliques. These incidents created the impression that social rejection can be a stimulus that leads to violent, antisocial behavior.

The link between social exclusion and antisocial behavior is suggested by data more systematic than vivid news stories. Many violent young people are men who feel that their families and peer groups have rejected them (Garbarino, 1999). The majority of violent crimes are committed by young men who are single and who often lack other social ties. Indeed, the likelihood of committing a crime drops significantly when a man marries, and it rises again if he gets a divorce (Sampson & Laub, 1990, 1993; cf. Wright & Wright, 1992). Sometimes these unattached young men band together to form a club or gang, and these are often some of the most violent organizations in civil society (e.g., Jankowski, 1991). At the broadest level, homicide rates across society as a whole correlate well with statistics that measure social integration, such as marriage and divorce rates (Lester, 1994).

The link between antisocial behavior and social exclusion is familiar to developmental psychologists. Children who are rejected by their peers are more likely than other children to threaten and physically attack others, as well as being more disruptive in class and other settings (Newcomb, Bukowski, & Pattee, 1993). Bullies tend to have fewer friends than other children and to be less liked and accepted by them (Coie, 1990).

Yet the developmental research raises an important question about the direction of causality. Does the aggression cause the social rejection, or vice versa? Many experts believe that the aggression is the cause: Children do not like aggressive children, and so they reject them. Although this is quite plausible and may be an important part of the story, our work has been concerned with the opposite possibility, namely that social exclusion and rejection contribute to causing violent, aggressive, and antisocial behavior.

The effects of social exclusion can be seen in the broader context of the human motivation to form social bonds with other people. Baumeister and Leary (1995) have argued that the human “need to belong” is one of the most powerful and pervasive motivations. According to their analysis, it consists of two components. The first is the motivation to have frequent interactions with other people, especially interactions that are either pleasant or neutral (as opposed to aversive, conflictual interactions). The second component involves an overarching structure of relatedness, in which the persons experience mutual concern and caring for each other’s welfare. People find both the interactions and the relationship structure appealing and to some extent satisfying, but the need to belong is not fully satisfied unless both components are present. In other words, people want to have frequent interactions with someone with whom they have an ongoing relationship.

If the need to belong is indeed a basic and powerful motivation, then social exclusion is a direct blockage that thwarts this motivation. To feel rejected or excluded is thus to experience a frustration of something one earnestly desires. On that basis, one could well predict that socially excluded people will exhibit a variety of signs of disturbance. It may be quite difficult to be indifferent to having one’s most powerful needs and desires thwarted.

Then again, one could in principle make the opposite prediction — namely, that social exclusion would produce a shift toward more prosocial behavior. This line of reasoning would emphasize the notion of adaptation. If you have been rejected by a friend or a group, and your need to belong has therefore been frustrated, the adaptive response would be to redouble your efforts to secure social acceptance. Prosocial behavior, including cooperation, friendliness, generosity, and helpfulness, would seemingly be the best strategy for obtaining such acceptance. This line of reasoning follows a standard pattern in motivation, in which people respond to a goal blockage by finding an alternate path to reach that goal.

Thus, it is reasonable to think that social exclusion should ideally elicit ever more adaptive efforts to behave in a prosocial manner so as to be accepted by others. Why, then, might we predict the opposite? We had two lines of reasoning. First, we thought that social exclusion and rejection might produce strong negative emotions, and these in turn could lead to antisocial behavior. It is well known that anger can promote aggression, and indeed aggression has been linked to a variety of unpleasant emotional states (Berkowitz, 19xx). Social exclusion has been shown to cause anxiety along with a variety of other negative emotions (Baumeister & Tice, 1990), and anxiety too tends to produce an assortment of undesirable rather than adaptive behaviors. Our first line of reasoning therefore emphasized emotional states as mediating between social exclusion and antisocial behavior.

The second line of reasoning starts with a somewhat darker view of human nature. By this view, people generally have a broad assortment of antisocial and selfish motivations, which they normally keep in check. Freud (1930) proposed that civilized society is only made possible by virtue of a system of inner psychological restraints that prevent people from acting on their aggressive and other antisocial impulses. Recent work has confirmed that positive, prosocial outcomes are strongly linked to effective self-control (e.g., Gottfredson & Hirschi, 1990; Tangney & Baumeister, 2000).

Aggressive and antisocial behavior would thus constitute a regular, recurrent possibility, and it is only prevented by means of the system of inner controls. In essence, the person must learn to sacrifice the quest for immediate satisfaction of all impulses in order to live together with others in an equitable social group. We reasoned that these controls may depend on feeling that one belongs to the group, because that feeling would make the sacrifices seem worth while. Social exclusion would however undermine the feeling of belonging (almost by definition) — and, with it, the motivation to restrain oneself from acting on selfish, antisocial impulses. In essence, social exclusion sets free the antisocial impulses that were always there.

The two lines of reasoning are not mutually exclusive. Negative affect is well established as having the power to weaken self-regulation and inner restraints (Baumeister, Heatherton, & Tice, 1994). There are probably multiple ways in which emotional distress impairs self-control (e.g., Tice, Bratslavsky, & Baumeister, in press). Negative affect could thus contribute to undermining the inner controls, thereby allowing the antisocial behavior to emerge.

Research Manipulations of Social Exclusion

The research on which this presentation is based has relied chiefly on laboratory experiments. Social exclusion has been manipulated in two ways. In one procedure, we invite a group of participants (typically college students) to take part in a group discussion in order to get acquainted. They are then separated into individual rooms. Each is told that the next part of the study will involve working in pairs, and each is asked to select two people from the group with whom he or she would most like to work.

The experimenter then returns to visit each participant. By random assignment, each is told one of two stories, both of which lead to the conclusion that the participant will have to work alone during the next phase. Half the participants are told that every other person in the group chose to work with them. The others are told that no one chose to work with them. If elaboration is needed, the experimenter explains that the design for the study involved having each person work with one person who chose him (or her) and one person who did not, and so the unanimity of selection renders this impossible. Hence each person is told that he or she will have to work alone – but the reason is crucially different. Half believe it is because they were rejected by everyone, whereas the others were chosen by everyone. In this way, we seek to assess the consequences of social rejection.

The other procedure (called the future status procedure) involves running participants individually. The participant is first given a personality inventory. The experimenter takes the completed response sheet away and returns later, ostensibly with the results from that test. To establish credibility, the experimenter gives each participant accurate feedback about his or her score on introversion and extraversion. Following this, however, the experimenter gives bogus feedback about future social relations, by random assignment. In the crucial ("future alone") condition, participants are told that people with their personality profile have been found to end up alone in life. The experimenter says that people with that profile often have a good network of social relations while they are in their twenties, because this is the age at which people are constantly forming new ties, but as they get older these relationships drift apart and are not replaced with new ones. As a result, the person is likely to spend more and more time alone as he or she gets older.

Several control groups have been used in different studies. In one ("future belonging"), the experimenter says simply that the participant is likely to be surrounded by a good social network "of people who care about you" throughout life. In another, no forecast about the future is made at all.

In yet another ("misfortune control") group, people are told that they are likely to be accident prone later in life, which will entail having many mishaps that will lead to broken bones and other injuries. This last control group is important because it involves an aversive prediction – but not one that affects belongingness. We thought it possible that the "future alone" manipulation might have negative effects, not because of social exclusion per se, but simply because participants are led to expect bad things will happen to them. If so, then the misfortune control should produce similar effects.

Social Exclusion and Aggression

Using the manipulations described in the previous section, we sought to examine effects of social exclusion on a variety of prosocial and antisocial behaviors. Our first series of studies examined aggression. The central hypothesis was that social exclusion would increase aggression.

One form of aggression consists of giving a bad evaluation to someone, especially in a context in which that person suffer adverse consequences of the bad evaluation. Specifically, we told participants that the other person was applying for a position as a research assistant in the psychology department and so we wanted evaluations from research participants that would aid in the department’s decision as to who would be the best person to hire. In that context, a bad evaluation would presumably damage the person’s chances of obtaining the job he or she desired. The procedure of using bad, damaging evaluations as a form of aggression has been used in prior studies of aggression (e.g., Kulik & Brown, 1979; Ohbuchi, Kameda, & Agarie, 1989; O’Neal & Taylor, 1989; for a review, see Baron & Richardson, 1994, pp. 64-66).

Our first experiment scheduled participants in same-sex pairs of strangers. They were put into separate rooms and asked to fill out a long questionnaire, which was presented as a measure of personality. They were also instructed to write a brief essay expressing their opinion on abortion. The experimenter then took the questionnaire away (ostensibly for scoring) and took the essay away (ostensibly to let the other participant evaluate it). The participant was given an essay to read and was led to believe it had been written by the other participant just as the participant wrote one. In fact the essay given to each participant was a standard one that the experimenter had prepared and used for all participants. The only adjustment was that the experimenter had two essays, one on each side of the issue, and each participant was given the essay advocating the position opposite to his or her own. In that way, each participant was led to believe that the other participant held the view opposed to his or her own (and thus was a dissimilar person). The participant was asked to evaluate the essay.

The future status procedure was then administered. Most participants received bogus feedback about their responses to the questionnaire, and the feedback included saying that in the future they would be either alone much of the time, or generally embedded in a rich social network, or prone to accidents and injuries. There was also a no-feedback control group.

After this manipulation, the experimenter pretended to exchange the essay evaluations. In reality these were standard evaluations. Almost everyone received a bad evaluation. It consisted of low numerical ratings on all the dimensions (organization, writing style, etc.), and in the space for comments added the handwritten summary “one of the worst essays I’ve read!” There was however one “positive control” condition in which people received no feedback about their future and then received a favorable evaluation of their essay.

Following this, the participant was told that the other person had applied for a position as research assistant and was asked to evaluate that person. This constituted the measure of aggression. Thus, all participants were provoked by the other person by means of an insulting evaluation, and this evaluation came on the heels of the feedback about the participant’s future status. The target of aggression was not involved in the social exclusion manipulation in any way, however.

The results of this study suggested that social exclusion can increase aggressive tendencies. People who received the future alone manipulation were significantly more critical of the other person than participants in any other condition. In absolute terms, the future alone participants gave very negative evaluations (with a mean of 26 on a 100-point scale), whereas the other negative feedback conditions had means around the midpoint of 50. The positive control condition was the only one in which participants gave favorable evaluations, with a mean rating of 78 out of a maximum 100 possible.

Thus, excluded people gave more negative ratings than people in any other condition, even though most participants received an identical provocation from the person who became the target of their aggression. Accepted and accident prone people did not attack the other person with anywhere near the intensity that the excluded people used. Social exclusion also produced significantly more aggression than the provocation (i.e., receiving a bad evaluation of one’s essay) alone. Because of the importance of that conclusion, we conducted a replication of those two cells alone, and it confirmed that social exclusion plus provocation elicits higher aggression than the same provocation without any exclusion.

These first two studies thus showed that an experience of social exclusion — even one that simply forecasts being alone many years in the future — is sufficient to increase aggression under some circumstances. Specifically, the circumstances included a significant provocation. One might therefore interpret these findings by saying that social exclusion potentiates an aggressive response to provocation. These findings are already important, but they raised the question of whether social exclusion would also increase aggression toward someone who had not provoked and insulted the excluded person.

Our third experiment therefore examined the effects of social exclusion on aggression toward someone who praises rather than insults the participant. The procedure was almost identical to the first study, with one important change: Nearly everyone received a positive, praising evaluation of his or her essay, consisting of favorable numerical ratings and a handwritten comment “A very good essay!” The only exception was that we included a replication of the future alone/negative evaluation condition from the first experiment.

This study did not find that social exclusion led to any increase in aggression toward the praising person. All participants who received positive evaluations on their essays gave the other person (who had praised them) very positive ratings in connection with the job application. These did not show any variation as a function of the future status manipulation. In other words, the people who were told they would end up alone in life were just as favorable and friendly toward the other person as the people who were told they would have many friends or would be accident prone. In fact, the mean ratings in these conditions were nearly identical.

The only exception was the condition in which people received the insulting, negative evaluation of their essay. This condition replicated the finding of the first experiment and showed a very negative, aggressive evaluation.

The implication of these studies is that social exclusion can potentiate aggressive responses to a provocation, but it does not appear to produce an aggressive tendency toward someone who is kind and friendly. Excluded people seem to respond well to those who treat them favorably, even if they become extra nasty toward others who offend them.

To increase generality, we conducted a fourth experiment with quite different procedures. We used the group rejection manipulation (in which each person learns that everyone, or no one, in the group wanted to work with him or her). We also used a more standard measure of aggression than the negative job evaluation. Specifically, following Bushman and Baumeister (1998), we gave people the opportunity to deliver blasts of aversive, stressful noise to the other participant. This was done ostensibly as part of a reaction time competition. People were told that they would have to respond as fast as possible, and whichever of the two participants was slower would receive a blast of noise. Each person was permitted to set the intensity and duration of noise that would be delivered to the other person if the other person was slower.

The opponent in the reaction time task (who was thus also the target of aggression) was explicitly presented as someone who was not part of the group who had accepted or rejected the participant. Rather, the experimenter said that since the participant could not continue in the group experiment, the participant would do an entirely different experiment with different people. The participant did write an essay on abortion and receive a negative evaluation from the other person, thus constituting a provocation. Participants did not however see or evaluate any essay by that person, which would help eliminate any possibility that perceived dissimilarity would contribute to the aggressive behavior.

The fourth experiment provided valuable converging evidence. Rejected participants were significantly more aggressive than accepted ones. Again, the target of the aggression was someone who had provoked the participant by means of the negative essay evaluation (but was not involved in the social exclusion). Thus, using these different procedures, it was again found that social exclusion could potentiate aggressive responses to provocation.

At this point we felt rather secure in concluding that social exclusion produces an increase in aggression toward someone who insults and provokes the person but does not lead to aggression toward a friendly, praising person. One large question remained: What about neutral persons? That is, excluded people are aggressive toward their enemies and friendly toward their friends, but which is the exception? This question seemed important because it would indicate the scope of generality of the findings. One possibility was that excluded people would simply have a chip on their shoulder and would respond with hostility toward someone who provoked them but would be perfectly friendly toward everyone else. The other, darker possibility was that excluded people would be aggressive toward almost everyone, only making exceptions for someone who was explicitly nice to them.

Our fifth and final aggression experiment examined this by dispensing with the aggressive provocation. In other respects, the procedure was the same as in the fourth experiment. It used the group rejection manipulation and the noise-blast procedure for measuring aggression. The only change was to eliminate the essay writing and evaluation procedure. That way, the opponent in the noise-blast game was a neutral, seemingly innocent person who had not presented as either friend or enemy.

The results of this study pointed toward the darker, more disturbing conclusion. Rejected participants were significantly more aggressive than accepted participants toward the neutral partner.

The conclusion from our aggression studies is that social exclusion does cause an increase in aggression. Excluded people were more aggressive than others toward someone who provoked them and toward neutral, innocent persons. The increase in aggression was large in every study, and in several of them it was double the traditional criterion (.80 standard deviations) for a large effect size. Their hostility was thus not limited to the people who had excluded them or even to new people who provoked and insulted them. The only exception was that excluded people were not aggressive toward a new person who treated them in a friendly, praising manner. Also, the results were not simply a matter of bad feedback producing aggression, because the people who received the unpleasant feedback that their futures would contain accidents and injuries were no more aggressive than the people who received positive feedback. Apparently there is somethings special and distinctly upsetting about social exclusion.

We did not find that emotion mediated these results. The five studies used a variety of measures of emotion. The effects on emotion were very small and in some cases not even significant. The future misfortune condition tended to produce mood and emotion ratings that were the same as the future alone manipulation, even though the behavioral effects were quite different. Careful mediation analyses (following procedures outlined by Baron & Kenny, 1986) consistently contradicted the hypothesis that emotional distress mediated the aggression. Thus, social exclusion did not make people clearly or visibly upset, but it did make them clearly aggressive.

Other Antisocial Behaviors

Next, we sought to investigate whether social exclusion would produce a broader range of antisocial behaviors, in addition to aggression. Our first study examined test-taking behavior. We reasoned that college students would be quite familiar with procedures for fair and honest testing, and so we gave them an opportunity to cheat. Specifically, each participant was told to take a test in private. The experimenter set a timer and told the participant to stop when the timer bell rang. Crucially, however, she said that she would be elsewhere in the laboratory and unable to hear the timer ring, and so it was up to the participant to stop working when the bell rang. She said that the participant should leave the room and come find her as soon as the bell rang. This afforded an opportunity for the participant to continue working past the bell, thereby gaining an unfair advantage. In reality, the experimenter sat outside the testing room and was able to hear the bell. She started a stopwatch when the bell went off and measured how long it was after that that the participant came out of the room to get her. This constituted the measure of cheating. The cheating measure followed the future status manipulation of social exclusion.

The results indicated that social exclusion made people more willing to break the rules and disobey instructions (to their own advantage). Participants who received the future alone manipulation worked significantly longer than all other participants on the test. The future accepted people did not differ significantly from the misfortune control condition, and so the results were not simply due to hearing any unpleasant forecast about one’s future. Rather, it was specifically learning that one would end up alone in life that appeared to make people more willing to take illicit extra time on the exam.

Next, we examined antisocial and prosocial behavior on the prisoner’s dilemma game. In a pair of studies, people played ten trials of this game. They were led to believe they were playing against another person, but actually they played with the computer following a preprogrammed plan. The game requires each person to choose between a cooperative and an antagonistic response. The cooperative response produces a good outcome if both players use it, but it leaves one open to be exploited and defeated by the other. The antagonistic response protects one against exploitation and can potentially bring the largest reward (if one uses it when the other person makes the cooperative move), but if both players use the antagonistic move, they both end up losing. Put another way, mutual cooperation produces a good result, unilateral cooperation produces a big loss (while the unilateral antagonistic move produces a big gain), and mutual antagonism produces a poor result for both. The game has been widely used to examine whether people choose cooperative, prosocial strategies or selfish, antagonistic strategies.

We found that social exclusion produced significant shifts toward the antagonistic responses. People who received the future alone manipulation made significantly (and substantially) fewer cooperative responses than people who received either the future accepted or the misfortune control manipulation.

The two studies we ran differed as to how the computerized opponent behaved on the first trial. In one study, the computerized opponent started off with an antagonistic move. After that, the computerized opponent was programmed to follow a tit-for-tat strategy, which meant that its response on each trial would be whatever the participant had done on the preceding trial. The only exceptions were that the computerized opponent was programmed to give antagonistic responses on trials 5 and 9, regardless of what the participant had done previously. This was done to prevent the game from being simply an endless cycle of mutual cooperation. In this study, the effects were huge (nearly two standard deviations). Socially accepted participants made antagonistic responses on about four of the ten trials, whereas socially excluded ones did so on eight out of ten.

The second study changed the opening move of the computerized opponent, because we thought that the high degree of antagonism in the previous study might have been due to the antagonistic response by the computerized opponent on the first move. Therefore we had the computerized opponent start off with a cooperative move. Although overall there was a shift toward greater cooperation than what we found in the previous study, the results still showed a very large increase in antisocial behavior caused by social exclusion. Socially excluded (future alone) participants gave on average more than twice as many antagonistic responses as the socially accepted ones.

Thus, social exclusion appears to cause an increase in a variety of antisocial behaviors, and not just aggression. Excluded people were more willing than others to cheat on a test and make antagonistic, uncooperative moves on a mixed-motive game. Moreover, as in the aggression studies, emotional distress failed to mediate the results.

Prosocial Behavior

We have also conducted some studies examining prosocial behavior. Doing good deeds has long been recognized as a way of making oneself appealing to others, and so it would seemingly be an adaptive and rational strategy for an excluded person to adopt so as to gain social acceptance. Good deeds also help overcome bad moods (Cialdini et al., 1973; Manucia, Baumann, & Cialdini, 1984), and so if people feel bad after being socially excluded they might seek to cheer themselves up by helping others or performing prosocial acts. However, our previous findings with aggression and antisocial behavior led us to doubt that social exclusion would in fact produce such a desirable outcome, and we predicted that excluded people would become no more (and possibly less) willing to perform prosocial acts.

In a first study, we administered the group rejection manipulation. When the experimenter told the participant that he or she could not continue as part of the group study (because everyone, or no one, had chosen to work with the participant), the experimenter said that the participant could therefore leave at once — but, alternatively, the experimenter needed some pilot data for future studies and would appreciate it if the participant would be willing to do one, two, or three brief experiments in the remaining time. The measure was how many of these the participant volunteered to do. We thought this request would be especially appealing to the socially excluded individuals, because it presented the opportunity to do a favor for a high-status person (the experimenter) and thus might enable the participant to make a very desirable friend — which would presumably help offset the impact of having been rejected by everyone in the group of peers.

Contrary to that hope, we did not find any rise in willingness to help among the excluded participants. Indeed, they became less willing to help. Nearly all the rejected participants refused to do any of the additional procedures, and the mean of 0.3 out of three favors reflects a very negative and minimal response to the request for favors. In contrast, the socially accepted participants were much more generous with their time, consenting to do an averge of nearly two (1.7) out of the maximum three additional studies.

Those findings suggested that social exclusion made people less willing to give help in response to a direct request for a favor. An alternative explanation was simply that the excluded people wanted to escape from the situation as fast as possible. Hence we conducted a second study that did not involve any difference in time. In this study, we gave each participant a series of coins as an ostensible reward for performance on an initial task. The purpose was simply to ensure that each person had some money. Then came the social exclusion manipulation. Last, when it was time for the participant to leave, the experimenter invited the participant to make a donation to the Student Emergency Fund, which helps students in financial need when they face personal crises. The donations were ostensibly anonymous.

Once again, social exclusion reduced prosocial behavior. Future alone participants donated significantly less money to the good cause than did participants in the other conditions.

Conclusion

The results of these studies suggest that social exclusion produces a significant shift toward antisocial behavior and away from prosocial behavior. Socially excluded participants became more aggressive toward other people generally, only making an exception for someone new who treated them nicely. They were more willing than others to cheat on an examination by disobeying the instructions about when to stop. They were less cooperative and more self-serving and antagonistic on a prisoner’s dilemma game. They were less willing to offer help in response to a request for a favor, and they were less generous in response to a request for cash donations.

The main surprise in these findings was the lack of mediation by emotional distress. Socially excluded people did not report feeling bad. Rather, their emotional self-reports consistently depicted neutral, emotionless states. Moreover, their self-reported emotions did not mediate any of the behavioral effects.

The lack of emotional mediation suggests that the antisocial effects of social exclusion should be understood as the release of pre-existing impulses. It may be that people frequently have selfish, antisocial impulses, but they restrain these because of their commitment to the social community. When they feel excluded from the social community, however, they cease to see any reason to restrain themselves, and so they become more willing to act in antisocial ways.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download