ON THE SOCIAL INFLUENCE OF EMOTIONS IN GROUPS 279 settings,deviance can frustrate effective coordination,whereas .Thus g.thereby inc sing chances a group may leac ng re n study.which s who were facing a happ Leary.004).We offer wo explanations for this difference in e examin d the processn for ng social exc mbers conformed more.This finding eptance as they are aimed at showing the effe s o onsistent with al perspectve.be exclusio groupand therefore have ssoithin the ng.where even if the ce Theoretical Implications By highlighting these effects of anger and happiness in a infating the likelihood of a sion.Prentice-Dun eroup context.we coniribute to 982 consiste rijda&Me 1994:Keltner&Haidt.1999:Park ial h vior,as there is nothing to gain by n 996 y to the have been des 20074 group mpact of this ang n the deviant individual ha Anothe the deviar indiv re n tend to similar otion thereby ini The fac four levels anger in re define group bo Second hich is the opposite o members (Study5).In order to te apart the antisocia Our findings spea ocial effe against cach c In any cas and fomp n of the ns thei darie ddition: ly.our e third avior than w tion of an emotional express on in shaping behavior points to the uch as vith Stud y 5 a0.2011 e tha Our fin inte emotions in in Noleworthy n this respect is that Studies3..andare.toour Wiest.&Swartz 1994)on the basis of their interpersonal effects.settings, deviance can frustrate effective coordination, whereas it does not in competitive settings. Thus, in cooperative settings (but not in competitive settings), one can show commitment to the shared goal by conforming, thereby increasing chances of being re-accepted. In Study 4, we replicated these results in a cooperative group interaction study, which showed that partic￾ipants who were faced with an angry majority felt more rejected and therefore had relatively less influence on their group out￾comes than participants who were facing a happy or neutral majority. Finally, in Study 5, we examined the process in yet more detail and found that only peripheral and not prototypical members conformed more after an angry response. This finding is consistent with our motivational perspective, because periph￾eral members could show good group membership by conform￾ing, whereas prototypical members are safely bound within the group and therefore have less to gain by conforming. Theoretical Implications By highlighting these effects of anger and happiness in a group context, we contribute to the growing literature on the social functions of emotions (e.g., Fischer & Manstead, 2008; Frijda & Mesquita, 1994; Keltner & Haidt, 1999; Parkinson, 1996; van Kleef, 2009) and specifically the functions that emotions have in groups. Although the angry reactions that may arise in groups in response to deviance have been described previously (Marques et al., 1998; Nemeth & Wachtler, 1974), the impact of this anger on the deviant individual has never been investigated. By analyzing the effects of emotions in this context, we have shown that a majority’s anger and happiness lead the deviant individual to feel rejected and accepted, re￾spectively, thereby influencing the behavior of the deviant. The motivational effects of receiving an angry reaction can be so profound that they persist over time, resulting in conformity to the group norm even several weeks later (Study 5). In their seminal article detailing the social functions of emo￾tions at four levels of analysis, Keltner and Haidt (1999) pro￾posed three social functions of emotions at the group level of analysis. First, collectively shared emotions help individuals define group boundaries and identify group members. Second, emotions may help individuals define and negotiate group￾related roles and statuses. Third, collective emotional behavior may help groups resolve potential problems. Our findings speak to the first two functions by showing that emotional expressions have consequences for a target group member’s perception of his or her inclusionary status in the group, which has implica￾tions for that person’s status within the group and for group boundaries. Additionally, our results support the third function by showing that emotions are functional in regulating individ￾ual members’ behavior in order to achieve collective goals. We have demonstrated that the effectiveness of anger in bringing about behavioral change in deviant group members is limited by a number of situational factors, such as the perceived compet￾itiveness of the situation (Study 3) and the extent to which a deviant group member is prototypical (Study 5). We believe that considering the functionality of group mem￾bers’ emotions in bringing about behavioral change in fellow group members may shed light on many intragroup processes. Noteworthy in this respect is that Studies 3, 4, and 5 are, to our knowledge, the first studies that show that feeling rejected by a group can lead to changes in behavior that are aimed at regain￾ing acceptance in the same group. Although previous research has shown that being socially excluded from a group may lead to conformity in a different group (e.g., Williams et al., 2000), it is usually found that exclusion provokes anger and leads people to derogate or aggress against the group that has ex￾cluded them (e.g., Bourgeois & Leary, 2001; Buckley, Winkel, & Leary, 2004). We offer two explanations for this difference. First, most paradigms that are used for studying social exclu￾sion offer few possibilities for the excluded person to seek re-acceptance, as they are aimed at showing the effects of exclusion on one’s functioning. For instance, in the cyberball paradigm (e.g., Williams et al., 2000), participants have no behavioral options while being ostracized. Thus, even if they would want to, participants have no means to seek re￾acceptance. Second, the studies in which participants could act toward the group they were excluded from were structured in a way that participants were not accountable for their actions, thus inflating the likelihood of aggression (e.g., Prentice-Dunn & Rogers, 1982). Also, consistent with our motivational per￾spective, this may have simultaneously reduced the likelihood of prosocial behavior, as there is nothing to gain by conforming if one’s actions cannot be identified. Thus, we think that a crucial factor in inducing conformity to the group that one is excluded from is whether there is a possibility to show that one is a valuable group member (Hollander, 1960; Klein et al., 2007). Another reason why it is noteworthy that we found conformity to the same group after an angry reaction from the majority is that emotions expressed by one person tend to trigger similar emotions in other people (e.g., Hatfield et al., 1994). Thus, the angry reaction from the majority might have produced anger in the deviant individual (see also Molden et al., 2009). The fact that we observed conformity in a situation that was likely to elicit anger and the accompanying aggressive tendencies contrasts with Smart Richman and Leary’s (2009) argument that feeling anger in re￾sponse to social exclusion is always accompanied by antisocial responses. If this were the case, our participants should have conformed less after an angry reaction, which is the opposite of what we found in cooperative settings (Studies 3 and 4) and for peripheral members (Study 5). In order to tease apart the antisocial effects of feeling anger and the prosocial effects of felt rejection, future research could fruitfully pit prosocial and antisocial re￾sponses directly against each other. In any case, these effects suggest that in the situations studied here, participants’ interpreta￾tion of the majority’s anger was more influential in shaping their behavior than was their reciprocal anger (van Kleef, 2009; see also Chow, Tiedens, & Govan, 2008). This potency of the interpreta￾tion of an emotional expression in shaping behavior points to the importance of considering inferential processes in describing the interpersonal effects of emotions, a conclusion that is in keeping with recent theoretical developments (e.g., emotions as social information [EASI] theory; see van Kleef, 2009; van Kleef et al., 2010, 2011). Our findings also contribute to the understanding of the inter￾personal effects of emotions. According to van Kleef et al. (2010), emotions may be categorized into families (see also Roseman, Wiest, & Swartz, 1994) on the basis of their interpersonal effects. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ON THE SOCIAL INFLUENCE OF EMOTIONS IN GROUPS 279