If You Can't Join Them, Beat Them: Effects of Social Exclusion on Aggressive Behavior
[Interpersonal Relations and Group Processes]
Twenge, Jean M.1,4; Baumeister, Roy F.2; Tice, Dianne M.2; Stucke, Tanja S.3
1Department of Psychology, San Diego State University
2Department of Psychology, Case Western Reserve University
3Department of Psychology, University of Giessen, Giessen, Germany.
4Correspondence concerning this article should be addressed to Jean M. Twenge, Department of Psychology, San Diego State University, 5500 Campanile Drive, San Diego, California 92182-4611. Electronic mail may be sent to email@example.com or to firstname.lastname@example.org.
During part of the completion of this research, Jean M. Twenge was supported by National Institutes of Health National Research Service Award Postdoctoral Grant MH12329-01A2. Portions of this research were presented at the conference of the Society for Experimental Social Psychology, Atlanta, Georgia, October 2000. We thank Brad Bushman, Natalie Ciarocco, and Kip Williams for their help and ideas.
Received Date: March 1, 2001; Revised Date: June 19, 2001; Accepted Date: July 5, 2001
Social exclusion was manipulated by telling people that they would end up alone later in life or that other participants had rejected them. These manipulations caused participants to behave more aggressively. Excluded people issued a more negative job evaluation against someone who insulted them (Experiments 1 and 2). Excluded people also blasted a target with higher levels of aversive noise both when the target had insulted them (Experiment 4) and when the target was a neutral person and no interaction had occurred (Experiment 5). However, excluded people were not more aggressive toward someone who issued praise (Experiment 3). These responses were specific to social exclusion (as opposed to other misfortunes) and were not mediated by emotion.
Over the past several years, a series of shootings at American schools have demonstrated that young people who feel socially excluded sometimes turn violent. In fact, a careful study of the school shooting incidents found that almost all of the perpetrators experienced rejection and/or bullying by peers (Leary, Kowalski, Smith, & Phillips, 2001). These vivid incidents accurately reflect a more general pattern linking social exclusion with aggressive behavior. Garbarino (1999) found that many perpetrators of violence are young men who feel rejected by family members, peers, and society in general (see also Walsh, Beyer, & Petee, 1987). Developmental psychologists have shown that aggressive children have fewer friends and receive less acceptance from the peer group (Coie, 1990; Newcomb, Bukowski, & Pattee, 1993). Adults demonstrate this pattern as well: Single men commit more crimes than married men do, even when age is controlled (Sampson & Laub, 1990, 1993). Having at least one stable relationship (a marriage) seems to inoculate against criminal behavior.
A possible link between social exclusion and aggression may be important for understanding recent changes in American society. The past 35 years have seen parallel increases in rates of criminal, violent, and antisocial behavior, on the one hand, and rates of divorce, living alone, and other signs of social fragmentation, on the other (U. S. Bureau of the Census, 1998; see Twenge, 2000, in press, for discussion). Several authors have argued that these changes have led to a society in which people lack stable relationships and feel disconnected from each other. Putnam (1995, 2000) found that Americans are now less likely to join community organizations and visit friends than they once were. The proportion of the population living alone has nearly doubled in recent decades, from 13% in 1960 to 25% in 1997 (U. S. Bureau of the Census, 1998). In step with this rising tide of aloneness, violent crime has skyrocketed, property crime has increased, and people trust and help each other less than they once did (Fukuyama, 1999). In fact, Lester (1994) performed a time-series analysis and found that statistics measuring social integration (e.g., divorce, marriage, and birth rates) showed a nearly perfect correlation with homicide rates. Mental health professionals, educators, policy makers, law enforcement officers, and others who must deal with social problems may find it useful to know whether a loss of social bonds leads directly to aggressive and antisocial behavior.
Which way does the causal arrow point? Undoubtedly, violent and aggressive tendencies could cause rejection by others. Some eminent developmental psychologists, for example, have concluded that aggression elicits rejection by other children (e.g., Coie, 1990; Dodge, 1983; Newcomb et. al., 1993; Schuster, 2001). This principle might apply to adults as well. Aggressive, violent men might be less attractive marriage partners, and, in general, people do not want to associate with dangerous, law-breaking, and trouble-prone individuals. These patterns all suggest that aggression is a cause and aloneness a consequence.
Although aggressive tendencies may cause rejection, the reverse is also plausible: Perhaps rejection and social exclusion are potential causes of aggressive behavior. The school shooters apparently regarded their actions as responses to rejection by others; thus they believed that these rejection experiences elicited their violent actions. Writers sympathetic to disaffected youth have repeatedly proposed that feeling excluded by mainstream society causes young people to adopt a sweeping disregard for its conventions and expectations. This, in turn, produces a willingness to break its laws (and other rules) and even to engage in violent behavior (e.g., Jankowski, 1991; McCall, 1995; Shakur & Scott, 1994).
Ultimately, the observational, correlational, and anecdotal evidence cannot effectively establish whether rejection causes aggressive behavior. The present investigation was therefore designed to provide experimental tests. The central hypothesis was that rejection and exclusion increase aggressive behavior.
Thwarting the Need to Belong
The present work is based on the assumption that human beings are powerfully and deeply motivated to form stable, lasting connections with other people. On the basis of a broad literature review, Baumeister and Leary (1995) concluded that much human behavior is motivated by a "need to belong." Some authors have proposed that the need to belong is rooted in the evolutionary history of humanity. Social exclusion may have hampered reproductive success and often led to death due to lack of food sharing, the difficulty of hunting alone, and inadequate protection from animal and human enemies (e.g., Ainsworth, 1989; Axelrod & Hamilton, 1981; Barash, 1977; Bowlby, 1969; Buss, 1990, 1991; Hogan, Jones, & Cheek, 1985; Moreland, 1987). Social exclusion usually leads to negative emotional experiences such as anxiety, depression, loneliness, and feelings of isolation (Baumeister & Leary, 1995; Baumeister & Tice, 1990; Gardner, Gabriel, & Diekman, 2000; Gardner, Pickett, & Brewer, 2000; Leary, 1990; Leary & Downs, 1995; Williams, Cheung, & Choi, 2000).
Prosocial, nonaggressive behavior might therefore be linked to the fundamentally social nature of the human organism, which thrives primarily in a network of stable, supportive relations. Although maintaining good social relationships may benefit the self-interest of the individual generally, there are many specific situations in which salient, immediate self-interest is opposed to collective well-being. Perhaps for this reason, concepts of virtue and social desirability center on the need to do what is best for everyone even if it runs contrary to the immediate, selfish inclinations of the individual (e.g., Baumeister & Exline, 1999; Hogan, 1973). Freud (1930) proposed that the superego is an essential prerequisite to civilized life; it emerges as a capacity to thwart instinctual or selfish interests in order to pursue actions that are valued by the group. For example, some authors have argued that children must be socialized into helping others (e.g., Cialdini, Kendrick, & Baumann, 1981). Without the socializing influence of a group, prosocial behavior might fade, and aggressive behavior, the more instinctual and impulsive tendency, might emerge instead. Thus, exclusion from a group might lessen or overwhelm restraints against aggressive behavior. This would fit the evidence we review linking social disintegration to aggression, from the rising divorce rate to the school shootings.
A contrary prediction could also be made, however. If people have a need to belong and act logically to satisfy that need, social exclusion should cause behaviors designed to make the self more attractive to others. By this argument, social exclusion should decrease aggression against others and increase prosocial and appeasing behaviors. Williams and colleagues (Williams, 1997; Williams et. al., 2000; Williams & Sommer, 1997) have demonstrated that being ostracized by others leads to increased conformity and, at least for women, decreased social loafing. This seems to be an eminently rational strategy as well, especially if aggressive behavior causes social rejection; in that case, rejected people would refrain from aggressive behavior to avoid further rejection. Hence, we had to acknowledge the competing prediction that social exclusion would elicit less aggressive behavior.
The possible effects of mood and emotion must also be addressed. If social exclusion is linked to aggressive behavior, does negative mood mediate this effect? Negative affect has been connected to increases in aggression (e.g., Berkowitz, 1989). Hence, one could well predict that rejection causes negative affect, which in turn causes an increase in aggressive behavior.
Again, though, the opposite prediction is plausible. Bad moods and aversive emotional states may motivate an increase in prosocial behavior because such positive actions often produce an improvement in mood. The negative state relief model proposes that negative affect should lead to an increase in prosocial behavior, because the act of helping draws people (at least adults) out of bad moods (e.g., Cialdini & Fultz, 1990; Cialdini & Kendrick, 1976; for a dissenting view, see Carlson & Miller, 1987). Hence, several of our studies investigate the possible role of emotion in mediating aggressive behavior.
We conducted a series of experimental studies to test the hypothesis that social exclusion and rejection cause behavior to become more aggressive. In these studies, we first manipulated people's perception of acceptance versus rejection. In three studies, this involved giving people feedback about their likely future relationships, ostensibly on the basis of a personality test. Some were told that they would likely end up alone for much of their adult lives (unlike others, who were told that their future would involve a rich network of personal relationships). In two other studies we manipulated social rejection more directly by telling participants that either no one in a group or everyone had chosen them as a desirable partner for a collaborative task. After these manipulations, we measured aggression in two different ways: giving someone a damaging negative evaluation (Experiments 1-3) or administering blasts of aversive, stressful noise (Experiments 4 and 5). We also sought to assess mediation by negative affect using two different affect measures.
Experiment 1 provides a direct test of the hypothesis that social exclusion leads to aggressive behavior. Participants were first given bogus feedback on a personality test. In the crucial condition (the future alone condition), this feedback provided the basis for predicting that the person would end up alone in life. The comparison group was told that their personality profile indicated a future with a rich and strong network of interpersonal relationships (the future belonging condition). We also included three control conditions. The first (the misfortune condition) gave people a forecast of an unpleasant but not lonely future: Participants were told that their personality profile predicted an adult life that would involve being accident prone. The second and third control conditions involved no forecast whatsoever regarding either future social relationships or future proneness to accidents (positive control and negative control conditions).
Next, most participants were provoked by receiving an ego threat. More precisely, they received feedback ostensibly written by another participant that described their writing and opinions as substandard, disorganized, and unpersuasive. Participants in one additional control group (the positive control group) received praise rather than criticism of their essay.
We measured aggression by giving participants the opportunity to damage another person's chances of getting a desirable job. This constitutes a measure of aggression because the evaluation has the potential to thwart the person's personal and career goals, a retaliation just as meaningful as physical harm. Similar techniques using job-relevant evaluations have been used to measure aggression in many previous studies (e.g., Kulik & Brown, 1979; Ohbuchi, Kameda, & Agarie, 1989; O'Neal & Taylor, 1989; for a review, see Baron & Richardson, 1994, pp. 64-66). If social exclusion causes aggressive behavior, then people who received the future alone manipulation should deliver the most negative ratings of the candidate.
Participants were 47 undergraduate students participating to fulfill a course requirement in introductory psychology. There were 22 men and 25 women; the sample was 75% White and 25% racial minority. Average age was 18.9 years. Participants were randomly assigned among conditions, except that the negative control condition was run later than were the others. We include it here for the sake of presenting a complete design and to permit tentative comparisons with other conditions, but because it was run separately all such comparisons must be regarded with caution. Experiment 2 was conducted to replicate Experiment 1 with complete random assignment.
Materials and procedure.
The participants signed up in single-sex groups of 2 and were placed in different rooms. They were first asked to fill out a personality questionnaire (the Eysenck Personality Questionnaire; Eysenck & Eysenck, 1975) and write an essay expressing their opinion on the abortion issue (they were required to choose one side on the issue). They then evaluated the "other person's" essay, always receiving an essay expressing views that were opposite to their own on abortion; these essays were actually written by the experimenter.
In three conditions, false feedback on the personality test then followed (participants in the positive control and negative control conditions were not given any feedback on their future outcomes). To gain credibility, the experimenter first gave an accurate assessment of the participant's extraversion score, providing correct feedback about whether the score was high, medium, or low on this scale. The experimenter used this as a segue into reading a randomly assigned "personality type" description. One of three descriptions was read. In the future alone condition, the participant was told,
You're the type who will end up alone later in life. You may have friends and relationships now, but by your mid 20s most of these will have drifted away. You may even marry or have several marriages, but these are likely to be short-lived and not continue into your 30s. Relationships don't last, and when you're past the age where people are constantly forming new relationships, the odds are you'll end up being alone more and more.
In contrast, people in the future belonging condition were told,
You're the type who has rewarding relationships throughout life. You're likely to have a long and stable marriage and have friendships that will last into your later years. The odds are that you'll always have friends and people who care about you.
Last, we included a misfortune control condition in which people were told,
You're likely to be accident prone later in life-you might break an arm or a leg a few times, or maybe be injured in car accidents. Even if you haven't been accident prone before, these things will show up later in life, and the odds are you will have a lot of accidents.
This condition was intended to describe a negative outcome that was not connected with relationships or social exclusion.
Participants then received (bogus) feedback that they believed came from the other participant. In the positive control condition they received positive feedback, with good ratings on the evaluation categories (e.g., organization, writing style) and the summary "A very good essay!!" In the four other conditions (future belonging, future alone, misfortune control, negative control), the "other person" gave the participants poor ratings and wrote "One of the worst essays I've read!!" This method has been used in several studies; the negative feedback conditions are designed to elicit aggression (e.g., Bushman & Baumeister, 1998).
All participants were told that the other participant (who had evaluated their essay) had applied to be a research assistant in the department. The experimenter explained that the position was very competitive, so the department was trying to get several evaluations of each candidate. Participants completed an evaluation form on which they rated the other participant on a scale ranging from 1 (strongly disagree) to 10 (strongly agree) on 10 statements. (e.g., "The applicant is friendly," "The applicant is open-minded," and "If I were in charge of hiring research assistants, I would hire the applicant"). Summed together, these responses formed a total evaluation with a possible range of 10-100. The internal reliability of this measure was high, with Cronbach's [alpha] = .90. Because the items were phrased in a positive way, a low score indicates a negative evaluation and a high expression of aggression, whereas a high score shows a positive evaluation and a low expression of aggression. This evaluation score served as the dependent variable.
After participants completed the evaluation form, they were fully debriefed. They were told that the feedback about their extraversion score was true but that the further feedback was a randomly assigned description. They were also informed that they had not really received an evaluation of their essay and that the negative (or positive) feedback was randomly assigned. Particular care was taken to ensure that participants in the future alone group understood that the prediction was random and not true for them. No participant was permitted to leave the laboratory until he or she expressed aloud that he or she understood that the manipulations were assigned at random and thus had no basis in fact. The experimenter also apologized for giving the feedback.
Results and Discussion
Are people more aggressive and critical toward others when they believe they will be alone later in life, or do they attempt to remedy the situation by being more kind? The results suggest that social exclusion led to a marked increase in aggression toward the issuer of an insult. The dependent measure was the negativity of ratings of the other person in connection with his or her application for the job. Table 1 depicts the means. One-way analysis of variance (ANOVA) reveals significant variation among the five conditions, F(4, 42) = 13.93, p < .001. The future alone and positive control groups were significantly different from all of the other conditions at p < .05 or lower. Another ANOVA reveals that there was also significant variation among the three conditions that received bogus feedback about the future, F(2, 25) = 8.08, p < .002.
[Help with image viewing]
Table 1 Aggressive Evaluations Given to Candidate, Experiment 1
The crucial prediction was that participants in the future alone condition would be the most aggressive (i.e., most negative in their evaluations of the job applicant who had offended them). As Table 1 shows, this was dramatically supported. Participants who believed they would be alone later in life evaluated the person who had provoked them very negatively, giving that person an average score of only 26 on the 10-100 scale. Those in the future belonging and misfortune conditions gave neutral evaluations close to the midpoint of 55 (these two conditions did not differ significantly in their evaluation scores, nor did they differ significantly from the negative control group, which gave a neutral to negative evaluation, with an average of 44). Thus, the future alone participants chose to retaliate against the person who insulted them, whereas those in the future belonging and misfortune conditions were neutral, not retaliating but not praising, either. Participants in the positive control condition received a favorable evaluation of their essay, and they reciprocated by rating the partner very positively, with a mean rating of 78.
The greater aggression of the future alone group was not due to simply hearing bad news: The future alone group was significantly more aggressive than the misfortune group, t(17) = 3.89, p < .001. They were also more aggressive than participants in the future belonging condition, t(16) = 3.34, p < .005, and those in the negative control group (who received a negative essay evaluation but no future prediction), t(16) = 2.66, p < .02.
When we compare the future alone and future belonging groups with each other, the effect size is d = 1.59. Cohen (1977) defined an effect size greater than 0.80 as large-and the present effect size is double that size. Thus, the difference in aggressive behavior between the low and high belongingness conditions was very large. Even though the manipulation consisted merely of giving people bogus and vague information about what would happen in the distant future, it appears to have had a very powerful impact on their responses to their immediate situation. Anticipating a lonely future made people sharply more harsh and aggressive toward someone who had recently criticized them.
The findings of Experiment 1 provide encouraging evidence that social exclusion causes an increase in aggressive behavior. However, confidence and generalizability are reduced because one of the control groups (which received no forecast about the future but which was presented with an insulting provocation) was run separately from the others, violating the principle of random assignment. Given the importance of random assignment, we felt it necessary to conduct a partial replication of Experiment 1.
In this experiment, we included a future alone group and a negative control group, both of which received negative essay evaluations. This should demonstrate that the aggressive response was elicited by something more than simply being insulted by the target. The prediction was that aggression would be higher among people who were told they would end up alone in life, as compared with the other (control) participants.