The bystander effect and social control behavior: the effect of the presence of others on people's reactions to norm violations

Observers of deviant social behavior sometimes communicate disapproval directly or indirectly to the perpetrator of a deviant act. This reaction has been termed ‘social control’. Three field studies were conducted to explore the influence of the number of bystander-observers on the likelihood of social control. We predicted that the presence of others would inhibit people’s tendency to communicate their disapproval to the deviant but only if personal implication was low. In the first study, we measured participants’ perceptions of two fictive situations, one in which a deviant draws graffiti in an elevator of a shopping center and one in which a deviant litters in a small neighborhood park by throwing a plastic bottle in the bushes. As expected, participants considered both behaviors to be equally counternormative but felt personally more implicated by the littering behavior in the park. In Studies 2 and 3, the two situations were enacted with confederates of the experimenter. Naı̈ve bystanders served as participants, and social control was the primary dependent variable. Consistent with our hypothesis, we found evidence for a bystander effect in the low personal implication situation (‘graffiti in the elevator’) but not in the high personal implication situation (‘littering in park’). These results make clear that perceived personal implication moderates the extent to which people are inhibited by the presence of others when they decide whether they should exert social control or not. Copyright # 2002 John Wiley & Sons, Ltd. It is generally agreed that social norms are powerful determinants of human behavior (Asch, 1951; Cialdini, Kallgren, & Reno, 1991; Newcomb, 1961; Schachter & Singer, 1962; Schwartz, 1977; Sherif, 1936; Sherif, Harvey, White, Hood, & Sherif, 1961). Social norms are specific to particular groups, as each group creates its own standards for what attitudes or behaviors are acceptable and desirable. What is considered normative in one group may be seen as being a counternormative (deviant) behavior in another group. For example, in France it is considered polite to fill up a guest’s wine glass when it is still half full. A host who lets his or her guests empty their wine glasses is considered inattentive. In Germany, on the other hand, it is expected that hosts pour more wine only when the guest’s wine glass is empty. Filling up a wine glass that is still half full would be considered Received 24 February 2000 Copyright # 2002 John Wiley & Sons, Ltd. Accepted 23 April 2001 *Correspondence to: Markus Brauer, LAPSCO/CNRS, 34, Av Carnot, 63037 Clermont-Ferrand, France. E-mail: [email protected] to be an (impolite) attempt to encourage the guests to drink too much. Although social norms sometimes work against our need for individuation (Snyder & Fromkin, 1980) and our desire to maintain control over events in our lives (Burger, 1992), they also fulfill important functions in a group’s life: social norms contribute to the group’s identity (Brown, 1995), they create the feeling of belongingness (Smith & Mackie, 1995), they provide the individual with guidelines for his or her behavior in ambiguous situations (Sherif, 1936), they render more predictable others’ reactions (Cialdini, Reno, & Kallgren, 1990), and they structure situations that would otherwise be chaotic (e.g. the norm to drive on the right side of the road, the norm to form a line). Given how important social norms are for a group’s existence, it is not surprising that individuals generally conform to group norms (Asch, 1951; Sherif, 1936). Those who engage in deviant behaviors are often victims of negative sanctions by other group members, who exert pressure in order to obtain conformity (Schachter, 1951). This process is called ‘social control’ (Collins & Frey, 1992; Gibbs, 1981a,b; Liska, 1997). For the present purposes, and consistent with past literature, we define social control as any verbal or nonverbal communication by which individuals show to another person that they disapprove of his or her deviant (counternormative) behavior. It should be noted, however, that the term ‘social control’ has been used to designate a variety of behaviors related to social influence. A situation that illustrates our use of the term social control is one in which a couple observes another person throwing a plastic bottle in the bushes of a public park. In most Western societies, littering is considered a deviant behavior (Krauss, Freedman, & Whitcup, 1978; Reiter & Samuel, 1980). The couple may decide to ignore the littering. Or they may decide to communicate to the ‘deviant’ in some way that they disapprove of his or her behavior. They may give the deviant an angry look, they may make a comment to each other (but loud enough so that the deviant can hear), or they may confront the deviant directly and express their disapproval angrily. All these behaviors are called social control, because they attempt to encourage the deviant to conform to social norms. Social norms appear to be perpetuated by rewards for normative behaviors, and punishments for counternormative behaviors. Group members who behave normatively are seen as more intelligent, competent, confident, and sincere (Eisenger & Mills, 1968; Levinger & Schneider, 1969). Deviant group members, on the other hand, are in a difficult position: they are not liked by their fellow group members, their views are ignored, they are assigned undesirable tasks, they receive less money, and attempts are made to exclude them from the group (Dedrick, 1978; Festinger, 1950; Festinger & Thibaut, 1951; Schachter, 1951). Janis (1982), in his classic work on ‘groupthink’, reports the brutality with which dissenting group members are brought in line or ‘cut out of the loop’ in highly cohesive groups. With few exceptions, social control has been largely submitted to sociological analysis. Sociology textbooks generally devote entire chapters to deviance, the societal reactions to deviance (social control), and the necessity of social control for the perpetuation of social norms. Some sociologists even maintain that the concept ‘social norm’ can not be defined without reference to social control. According to the labeling perspective (Kitsuse, 1962; Lemert, 1972), behaviors can be considered deviant only if they generate disapproval by other group members. All behaviors that do not provoke social control are by definition normative. We would argue that this position is somewhat extreme because one could imagine behaviors which are clearly deviant but which, at the same time, are not sanctioned by social control because people are inhibited by the presence of others (e.g. the bystander effect demonstrated by Latané and Darley, 1968). However, it is certainly true that norms that prescribe some individual effort in order to protect a public good (e.g. the norm to carry the plastic bottle to the next trash can instead of throwing it in the bushes) are likely to disappear if deviant behaviors are not sanctioned by social control. Despite the important role of negative sanctions for the perpetuation of social norms, social psychologists have shown relatively little interest in the phenomenon of social control, at least in the 854 Peggy Chekroun and Markus Brauer Copyright # 2002 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 32, 853–867 (2002) sense used here (i.e. people’s reactions to counternormative behaviors in public). Schachter (1951) documented the manner by which group members respond to a deviant confederate: they first try to win him over, but when they see that they cannot succeed they ignore his views, they assign him to undesirable tasks, and they suggest that he be excluded from the group. Dedrick (1978) showed that bystanders exert more social control when the deviant is neutral rather than apologetic, and when a third person is hostile rather than friendly toward the deviant. Kiesler, Kiesler, and Pallak (1967) showed that when people are alone with the deviant they exert more social control when they believe that they have to work together with the deviant in the future than when they believe that they will never see the deviant again. However, when people are not alone with the deviant they are less likely to exert social control when they are committed to future interaction. The purpose of this article is to examine a factor that might be involved in the production of social control behavior: the number of bystander-observers. The groundbreaking work by Latané and Darley (1968, 1970) demonstrated that the presence of other people decreases our willingness to help a victim in an emergency situation. Later work showed that this ‘bystander effect’ is not restricted to emergency situations, but occurs in a variety of settings: in the presence of others, people are less likely to take a coupon for a free cheeseburger (Petty, Williams, Harkins, & Latané, 1977), pick up coins and pencils in an elevator (Latané & Dabbs, 1975), answer the door (Levy et al., 1972), report a broken tape recorder (Misavage & Richardson, 1974), or leave a large tip in a restaurant (Freeman, Walker, Bordon, & Latané, 1975). Applied to the enforcement of social norms, these past findings suggest that people are less likely to exert social control when in the presence of other bystander-observers. Several past studies examined the effect of the presence of others on helping behavior in the case of a theft, and also found a bystander effect (Howard & Crano, 1974; Latané & Darley, 1970; Latané & Elman, 1970). Although there is no doubt that a theft is a deviant behavior, it is unclear whether an intervention by a naive participant can be qualified as social control. We do not know if the person’s motivation is to help the victim or to enforce a social norm against crime. Nevertheless, these findings are convergent evidence for the hypothesis that the bystander effect extends to the domain of social control: the greater the number of people present in the situation the smaller should be the individual probability of intervening. Latané and Nida (1981) discuss three processes that account for the bystander effect. Audience inhibition refers to the risk of embarrassment if the situation turns out not to be an emergency. Social influence is the process by which each individual takes the inaction of others as an indicator that the expected pattern of behavior is not to do anything. Finally, diffusion of responsibility refers to the fact that costs are associated with intervention and costs are associated nonintervention. In the presence of other people the costs of nonintervention are shared, and nonintervention becomes more likely. All three processes are likely to operate independent of whether bystanders notice a victim in an emergency situation or they observe a deviant performing a counternormative behavior. Despite the overwhelming evidence for the bystander effect, some variables have been shown to moderate this effect. For example, Cramer, McMaster, Bartell, and Dragna (1988) showed that the presence of others does not inhibit the tendency to help an injured victim in an emergency when participants feel personally implicated in the situation. The women who participated in their study were either registered nurses or general education students. As each participant was escorted to the experimental room, she passed a confederate who was dressed up as a workman and was standing on a ladder. Once inside the experimental room, the participant worked either alone or in the presence of a second confederate who pretended to be another participant. Shortly after, there was a loud crash in the hallway, followed by a thud and moans of pain. The second confederate remained inactive and continued to work on the experimental task. When they were alone in the laboratory, both the nurses and the students intervened with an equally high probability. When they were in the presence of an inactive second participant, the intervention rate dropped drastically among general education students Social control and the bystander effect 855 Copyright # 2002 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 32, 853–867 (2002) but stayed about the same among registered nurses. Obviously, when individuals feel personally implicated and have the impression that it is their personal duty to intervene, the bystander effect disappears. Note, however, that the feeling of being personally implicated does not affect people’s reactions when they are alone. The purpose of the set of studies presented below is to examine whether personal implication plays a similar role in social control behavior. If personal implication is low, we expect to find the classic bystander effect: people’s tendency to negatively sanction deviant behaviors should be inhibited by the presence of others. However, if personal implication is high the bystander effect should disappear. We examined people’s reactions to two types of deviant behaviors. In one situation, a confederate drew graffiti on the walls of an elevator of a shopping mall. In the other situation, a confederate threw a plastic bottle in the bushes of a small neighborhood park. In each case, the confederate performed the deviant behavior in the presence of one, two, or three naive participants. We predicted that personal implication would be relatively low in the situation with the graffiti in the elevator. The elevator belongs to the corporate business who runs the shopping center. As such, it is clearly private property that is being used temporarily by the bystander. In addition, a professional cleaning personnel is responsible for the cleanliness of the shopping center. We also predicted that personal implication would be relatively high in the situation with the deviant littering in the small neighborhood park. Informal interviews had indicated that people cherished this park like their front yard. In France, there is a rising consciousness about environmental issues, and public institutions promote the protection of the environment as each individual’s personal concern. Finally, the maintenance of the park is paid by the inhabitants’ local taxes because the city pays the maintenance workers’ salaries. The first study reported below verified that the two situations actually differed on the dimension of personal implication. The second study tested our central hypothesis according to which we should find evidence for the bystander effect in the situation ‘graffiti in the elevator’ but not in the situation ‘littering in park’. Finally, the third study replicated the finding in the park with a different confederate.