The way I make you feel: Social exclusion enhances the ability to manage others' emotions

The way I make you feel: Social exclusion enhances the ability to manage others' emotions

Journal of Experimental Social Psychology 60 (2015) 59–75 Contents lists available at ScienceDirect Journal of Experimental Social Psychology journa...

744KB Sizes 0 Downloads 23 Views

Journal of Experimental Social Psychology 60 (2015) 59–75

Contents lists available at ScienceDirect

Journal of Experimental Social Psychology journal homepage: www.elsevier.com/locate/jesp

The way I make you feel: Social exclusion enhances the ability to manage others' emotions Elaine O. Cheung ⁎, Wendi L. Gardner Northwestern University, United States

H I G H L I G H T S • • • • •

We examined potentially enhanced emotional intelligence after social exclusion. Excluded participants showed a heightened tendency to manage others' emotions. Exclusion enhanced emotion-management strategies, but not non-social strategies. This enhanced tendency to manage others' emotions may promote reconnection. Excluded participants were rated as more likable by coders and interaction partners.

a r t i c l e

i n f o

Article history: Received 21 October 2014 Revised 4 May 2015 Accepted 5 May 2015 Available online 8 May 2015 Keywords: Belonging Social exclusion Emotional intelligence Emotion regulation

a b s t r a c t Original conceptions of social exclusion focused upon the negative impact of exclusion on intelligent thought (Baumeister, Twenge, & Nuss, 2002). We propose that although exclusion may impair cognitive forms of intelligence, exclusion should enhance more socially relevant forms of intelligence, such as emotional intelligence. Specifically, we examined whether exclusion would enhance performance in one branch of emotional intelligence: the ability to manage others' emotions. Social exclusion heightened the number and breadth of strategies that participants used for managing others' emotions when responding to hypothetical scenarios (Study 1) and when responding to online pen pals (Studies 3 and 4). Furthermore, excluded participants were more effective at energizing an interaction partner in a face-to-face coaching interaction (Study 2) and were rated as more effective at managing their pen pal's emotions in an online pen pal exchange (Studies 3 and 4). Although exclusion heightened the number and breadth of emotion management strategies generated in a social task, exclusion did not heighten the number or breadth of nonsocial strategies (creative uses for common household items) generated in a comparison task (Study 4). Lastly, we found preliminary evidence suggesting that this enhanced emotion management after exclusion may serve to facilitate reconnection; excluded participants were liked more by their interaction partners (Study 2) and were rated to be more likable by objective coders (Studies 3 and 4). Altogether, these findings suggest that individuals may be more effective at managing others' emotions following social exclusion, and this greater effectiveness may promote reconnection. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Given the evolutionary importance of social connection for survival (Caporael, 2001) and contemporary importance for physical and psychological wellbeing (Baumeister & Leary, 1995; Maslow, 1968), it is not surprising that people experience social exclusion as a painful experience they are motivated to avoid (Eisenberger, Lieberman, & Williams, 2003; MacDonald & Leary, 2005). Accordingly, when people are confronted with social exclusion (e.g., ignored by a friend, not invited to a social event), they are often motivated to engage in behaviors ⁎ Corresponding author at: Northwestern University, Department of Psychology, Swift Hall 412, Evanston, IL 60208, United States. E-mail address: [email protected] (E.O. Cheung).

http://dx.doi.org/10.1016/j.jesp.2015.05.003 0022-1031/© 2015 Elsevier Inc. All rights reserved.

that restore social bonds and a feeling of social connection (Gardner, Pickett, Jefferis, & Knowles, 2005; Molden & Maner, 2013; Pickett & Gardner, 2005). However, one ironic consequence of social exclusion is that it has been found to deplete both cognitive and self-regulatory resources — the very resources that would be essential for navigating a complex world alone. For instance, Baumeister, Twenge, and Nuss (2002) found that participants who received feedback that they would likely end up alone in life showed impaired performance on IQ and Graduate Record Examination tests. Furthermore, Baumeister, DeWall, Ciarocco, and Twenge (2005) found that participants who were made to feel socially excluded (either through receiving feedback that they would likely end up alone in life or feedback that no one in their group wanted to work with them) subsequently showed decrements in performance on

60

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

self-regulation tasks. Why might social exclusion impair these abilities at a time when they might be especially important for ensuring survival? We propose that this impairment in cognitive and self-regulatory functioning may be best understood through a resource-conservation explanation (Muraven, Shmueli, & Burkley, 2006). Specifically, given the depleting consequences of exclusion for cognitive and selfregulatory resources, excluded individuals may be careful about conserving the diminished resources that remain. Excluded individuals may be particularly motivated to limit the allocation of their resources to acts that could serve the broader goal of reconnection. Consistent with this explanation, prior research has found that performance on self-regulatory tasks was spared for excluded individuals when those tasks were framed as diagnostic of their social ability (DeWall, Baumeister, & Vohs, 2008); the social framing of the task motivated excluded individuals to use regulatory resources they were unwilling to use with a nonsocial frame. Rather than framing a task as socially diagnostic, the present investigation focuses on tasks that are directly related to social ability. Specifically, we examine whether social exclusion enhances performance on emotional intelligence tasks (Mayer & Salovey, 1997; Mayer, Salovey, & Caruso, 2004). We posit that whereas social exclusion may impair cognitive forms of intelligence, social exclusion should enhance socially relevant forms of intelligence, such as emotional intelligence. In their model of emotional intelligence (Mayer & Salovey, 1997; Mayer et al., 2004), Mayer and colleagues have divided the construct of emotional intelligence into four branches of interrelated abilities: (1) the ability to accurately perceive emotion, (2) the ability to utilize emotion to facilitate thought, (3) the ability to understand emotion and emotional knowledge, and (4) the ability to manage the emotions of oneself and others. We believe that the four-branch model of emotional intelligence may provide a useful framework for informing the strategies that people may marshal after social exclusion, as some of these facets of emotional intelligence have been linked with positive social outcomes, such as having more satisfying social interactions (Brackett & Mayer, 2003; Brackett, Rivers, Shiffman, Lerner, & Salovey, 2006; Lopes, Salovey, & Straus, 2003; Lopes et al., 2004, 2011) and superior performance in leadership roles and organizational settings (for reviews, see Côté, 2014; Mayer et al., 2004). Although emotional intelligence has typically been studied as a stable individual difference, it is likely that performance in each of these four branches also varies as a function of their social motivation. Indeed, there has been a recent call by emotional intelligence researchers to understand the factors that determine both “when” and “why” people apply their emotional intelligence skills (see Ybarra, Rees, Kross, & Sanchez-Burks, 2012; Ybarra et al., 2013). In the current work, we sought to answer this call and examine whether social exclusion may actually heighten the more socially relevant facets of emotional intelligence. Although all four branches of emotional intelligence are interrelated and have been shown to be beneficial in different circumstances (see Côté, 2014; Mayer et al., 2004; Ybarra et al., 2012, 2013), the two that appear most directly relevant for facilitating social reconnection are the first branch (accurately perceiving emotion in others) and the fourth branch (managing emotions in the self and others), as these two branches are interpersonal in nature, directly involving the emotions of others. The first branch, the ability to accurately understand how others are feeling, is clearly socially adaptive, as accurate perception can help people navigate their social world in a way that enhances social outcomes (Elfenbein, Marsh, & Ambady, 2002; McArthur & Baron, 1983). Prior research examining responses to social exclusion strongly suggests that social exclusion enhances performance on this first branch of emotional intelligence. Specifically, when people feel insufficient levels of social inclusion (e.g., after a laboratory-induced exclusion, chronicallylonely individuals, individuals with dispositionally-high belonging needs), they show heightened attention to and accuracy in understanding both facial and vocal expressions of emotion (e.g., Bernstein, Sacco,

Brown, Young, & Claypool, 2010; Bernstein, Young, Brown, Sacco, & Claypool, 2008; Gardner et al., 2005; Pickett, Gardner, & Knowles, 2004; Sacco, Wirth, Hugenberg, Chen, & Williams, 2011). Accurately perceiving others' emotions may not be enough to ensure successful reconnection, however. People must also respond appropriately to others' emotions in a manner that promotes liking and rapport – or in other words – demonstrate skills from the fourth branch of emotional intelligence: managing others' emotions. The fourth branch of emotional intelligence encompasses this notion of sensitive and appropriate responsiveness. However, unlike the relatively automatic perceptual processes underlying the first branch of emotional intelligence, emotional perception, the deployment of the skills that underlie the fourth branch may require the very selfregulatory resources that were likely to have been depleted by the prior exclusion (Baumeister et al., 2002). Indeed, regulatory deficits resulting from social exclusion may underlie maladaptive, anti-social, and self-defeating responding (Twenge, Baumeister, Tice, & Stucke, 2001; Twenge, Catanese, & Baumeister, 2002). Thus, the current research seeks to test the extent to which social exclusion may heighten versus hinder the skills associated with the fourth branch of emotional intelligence. To the extent that excluded individuals marshal their remaining regulatory resources toward social tasks (DeWall et al., 2008), this motivation may be sufficient to enhance the ability to manage others' emotions, because skill in managing others' emotions can facilitate social connection. 2. Managing others' emotions facilitates social connection Managing others' emotions involves our ability to influence and change the emotional experience of others (Niven, Totterdell, & Holman, 2009). We use this ability in our day-to-day lives to help us form new relationships and strengthen current social bonds. For instance, we help our significant others savor the good news when they get a promotion at work, and likewise we cheer them up when they find out this promotion does not include a pay raise. We expected that the managing others' emotions branch of emotional intelligence to be an especially promising branch for facilitating reconnection, as the ability to manage others' emotions seems to play an important role in impacting the quality of our social relationships. For instance, people who score higher on the managing emotions (in the self and others) subscale of the Mayer Salovey Caruso Emotion Intelligence Test (MSCIET; Mayer, Salovey, & Caruso, 2002) tend to have higher quality interactions with their friends (Lopes et al., 2004; Lopes, Salovey, Côté, Beers, & Petty, 2005), and less conflict in their social relationships (Lopes et al., 2003). In addition to influencing social outcomes in the context of peer relationships, the ability to effectively manage others' emotions also influences social outcomes in workplace contexts. For instance, medical students who scored higher on the managing emotions branch of emotional intelligence showed superior performance in courses on communication and interpersonal sensitivity (Libbrecht, Lievens, Carette, & Côté, 2014). Moreover, the ability to effectively manage others' emotions is considered to be an important component of effective leadership (for reviews, see Ashkanasy & Humphrey, 2014; Humphrey, 2002; Pescosolido, 2002). Indeed, leaders who tend to be more skilled at managing others' emotions tend to have more positive interactions with their subordinates and tend to create workplaces that have more positive emotional climates (see Ashkanasy & Humphrey, 2014). Furthermore, the act of managing others' emotions seems to play an important role in both the formation and deepening of social bonds. For instance, Beckes, Simpson, and Erickson (2010) have theorized that the process of managing others' emotions may underlie the formation of social bonds. Specifically, they demonstrated that people may, in part, develop social attachments through a conditioning process whereby novel others become associated with the downregulation of distress and the up-regulation of feelings of comfort.

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

Indeed, in two studies in which they assessed whether participants would develop positive attachment associations with pictures of strangers, they found that attachment was facilitated by strangers who down-regulated a participant's distress. These findings suggest that the process of managing others' emotions may underlie even the initial development of social ties. Of course, in addition to playing a role in initial social attraction, acts of managing others' emotions seem to strengthen the quality of established relationships. For instance, Niven, Holman, and Totterdell (2012) found that acts of managing others' emotions in the workplace predicted later increases in perceptions of friendship and trust for both the regulator and the target of regulation. Interestingly, the relationship-strengthening effects of managing others' emotions seemed to be largely due to changes in the target's emotions, suggesting that effectiveness in managing others' emotions may play an important role for strengthening social bonds. In the present research, we sought to examine whether social exclusion would enhance the tendency to manage others' emotions. Given that social exclusion seems to motivate behaviors that can help facilitate subsequent re-inclusion and connection, and given that the process of managing others' emotions seems to play an important role in the formation and deepening of social bonds, we posited that social exclusion should increase attempts to manage others' emotions. We were further interested in exploring the potential social benefits of managing others' emotions for the previously-excluded individual. Although prior research in the emotional intelligence literature has established the adaptive advantage of managing others' emotions for relationship quality (Libbrecht et al., 2014; Lopes et al., 2003, 2004, 2005; Niven et al., 2012), this research has primarily considered emotional intelligence as a stable individual-difference construct. However, by highlighting that people's skill at socially relevant forms of emotional intelligence can vary as a function of their social motivation, we are able to examine the benefits of managing others' emotions in terms of temporarily enhancing the individual's likability. 3. Overview of the present research The purpose of the present research was to explore whether excluded individuals would show greater management of others' emotions. In four studies, we examined how a laboratory-induced exclusion influenced the number of attempts that participants' made to manage others' emotions as well as the breadth of strategies they used in their responding (Studies 1, 3 and 4 and 4). In addition, we also examined the influence of exclusion on potential effectiveness at managing others' emotions in a face-to-face coaching interaction (Study 2) and in an online pen pal exchange (Studies 3 and 4). To examine the specificity of these effects, we varied the social vs. nonsocial context of the experimental task to examine whether exclusion would also motivate participants to generate a greater number and breadth of nonsocial strategies (creative uses for common household items) in a comparison task (Study 4). Given that the proposed function of managing others' emotions is to facilitate reconnection, we also explored the potential social benefits of managing others' emotions following exclusion. Specifically, we examined whether excluded participants were liked more by their interaction partners after a face-to-face interaction (Study 2) and whether excluded participants' responses were rated as more likable by objective coders (Studies 3 and 4). 4. Study 1 The purpose of Study 1 was to explore whether people would show greater management of others' emotions after experiencing social exclusion. To induce feelings of social exclusion, participants completed a reliving task in which they either relived a past experience of social exclusion (exclusion condition) or they described the layout of their local grocery store (neutral control condition). We assessed the management

61

of others' emotions by having participants list the various strategies they would use to cheer up a sad friend and calm down an angry friend. Multiple strategies may be used to manage another's emotion, ranging from providing emotional support, to encouraging cognitive reappraisal, to positive distraction, or decreasing arousal. Participants' responses were thus coded not only for the sheer number of attempts that they made to manage their friend's emotions, but also for the breadth of strategies they used. In the literature on individual emotion regulation, both the number of attempts for emotion management and the breadth of strategies used have been linked with positive outcomes for psychological adjustment (see Bonanno & Burton, 2013). First, the number of emotion regulation strategies that people use in response to stressful life events appears to be important for promoting emotional recovery (e.g., Lam & McBrideChang, 2007). For instance, after a campus mass shooting, students who made a greater number of emotion-regulation attempts after the shooting showed a more resilient trajectory of recovery relative to students who made fewer emotion-regulation attempts (Orcutt, Bonanno, Hannan, & Miron, 2014). Second, the diversity of emotion-regulation strategies used also appears to be important for promoting emotional recovery (e.g., Lougheed & Hollenstein, 2012). Oftentimes, when people are confronted with intense emotional events (e.g., divorce, death of a loved one, job loss, and financial stress), these events encompass a variety of challenges that may be best addressed using a diverse repertoire of strategies (e.g., Stroebe & Schut, 1999, 2010). Indeed, Bonanno and Burton have argued that accessing a diverse repertoire of emotion regulation strategies should increase the potential for effective regulation, by helping the individual flexibly tailor their responding to what may be most effective for the context at hand (see Bonanno & Burton, 2013). For instance, Bonanno and colleagues have found that students who were high in the ability to both enhance and suppress their emotional expressions showed greater long-term psychological adjustment 2 years after the September 11 terrorist attack compared with students who were high in either ability alone (enhancement or suppression) (Bonanno, Papa, Lalande, Westphal, & Coifman, 2004). Whereas this prior research had primarily focused on the effectiveness of using a greater number and breadth of emotionregulation strategies for individual emotion regulation, we reasoned that using a greater number and breadth of strategies should similarly be beneficial for the regulation of others' emotions. Accessing a greater number and broader repertoire of strategies when managing others' emotions should help the individual tailor their responding to what may be most effective for both the individual they are regulating, as well as the situational context. As such, we considered both the number of attempts made and the breadth of strategies used as rough indicators of participants' potential effectiveness at managing others' emotions. We expected excluded participants to show enhanced management of their friends' emotions than control participants, making a greater number of attempts to manage their friends' emotions for both scenarios and using a greater breadth of strategies in their responses. 4.1. Method 4.1.1. Participants One hundred and three participants (63 women, 40 men; Mage = 36.53, age range: 18–82 years) were recruited from Amazon's Mechanical Turk (MTurk) to participate in an online study for payment. Our sample size for this study was based on previous studies that had used similar reliving manipulations of exclusion on MTurk (e.g., Garczynski & Brown, 2014; Poon, Chen, & DeWall, 2013). We targeted approximately 50 participants per cell and we stopped data collection when we reached our pre-determined targets. Seven participants were excluded from the analyses: Six participants in the exclusion condition were excluded because they did not write about past experiences of

62

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

social exclusion in the reliving task; one participant in the control condition was excluded because they did not complete either the reliving task or the emotion management task. Altogether, our final sample resulted in a total of 96 participants.

4.1.2. Procedure Participants first completed a reliving task in which they were randomly assigned to write about one of the following experiences: a time they felt intensely socially excluded in some way — a time when they felt as if they did not belong (social exclusion condition) or the layout of their local grocery store (control condition). Similar reliving manipulations have been successfully used in past research to activate belonging needs (e.g., Maner, DeWall, Baumeister, & Schaller, 2007; Pickett et al., 2004). Following the reliving task, participants were asked to list the strategies they would use in two scenarios: a scenario in which they wanted to cheer up a sad friend and a scenario in which they wanted to calm down an angry friend. Participants responded in a free response format, listing as many or as few strategies as they wanted. Lastly, we assessed the following self-reported measures of emotional intelligence to examine the potential influence of dispositional emotional intelligence on emotion management: the Schutte Emotional Intelligence Scale (SEI; Schutte et al., 1998), the Emotion Regulation of Others and Self Scale (EROS; Niven, Totterdell, Stride, & Holman, 2011), and the Tromso Social Intelligence Scale (TSIS; Silvera, Martinussen, & Dahl, 2001). See Supplemental analyses.

4.1.3. Coding procedures Two coders who were blind to condition coded the responses for the number of attempts that participants made to manage their friend's emotions in each scenario (α's N .95), and the different types of strategies listed. The coders categorized participants' emotion-regulation strategies as follows: emotion-focused support (offering validation, understanding, sympathy, reassurance, consolation, encouragement etc.), problem-focused support (offering advice, feedback, suggestions, solutions, guidance about the problem etc.), cognitive reappraisal (offering a different perspective or a different interpretation of the situation to modify it's emotional impact, helping the individual reframe the emotional event, helping the individual derive meaning from the situation etc.), distraction (offering strategies for shifting attention away from the emotional event, offering suggestions for activities to help the individual get their mind off the problem), and strategies aimed at reducing physiological arousal (suggesting the individual take a few deep breaths, talking to the individual using a calm, soothing tone of voice, suggesting the individual drink a cup of tea). We calculated a breadth index of the number of different types of strategies that participants used in each response (α's N .91). We considered the breadth of strategies used to be distinct from the number of emotion-regulation attempts made, as participants could make a high number of attempts, but low breadth of strategies in their response. For example, consider the following response: “I would mirror what they are saying to let them know I understand,” “I would validate what they're saying and commiserate,” and “I would try to listen as long as it took for them to start calming down”. Although this individual made three total attempts at emotion management, they would receive a breadth score of one as they only used one type of strategy (i.e., emotional support) in their response. We considered both the number of attempts made and the breadth of strategies used as rough indicators of participants' potential effectiveness at managing others' emotions. Both of these indicators signify that the individual is drawing from a broad emotional repertoire when responding, which should help the individual respond flexibly and effectively across a wide range of social contexts (see Bonanno & Burton, 2013). Descriptive statistics and intercorrelations between the variables are presented in Table 1.

Table 1 Intercorrelations among and descriptive statistics for Study 1 variables.

1. Number of attempts 2. Breadth of strategies

M

SD

1

2

1.76 1.53

0.68 0.53

– .87⁎⁎⁎

– –

⁎⁎⁎ p b .001.

4.2. Results and discussion1 4.2.1. Number of attempts To determine whether excluded participants made a greater number of attempts to manage their friend's emotions than controls, we conducted a 2(Condition: Exclusion vs. Control) × 2(Emotional Scenario: Cheer-Up Sadness vs. Calm-Down Anger) mixed analysis of variance (ANOVA) on the number of attempts that participants made for each scenario. We found a main effect of condition, such that those in the exclusion condition made a greater number of emotion-management attempts (M = 2.09 95% CI [1.89, 2.29]2) than those in the control condition (M = 1.54 [1.37, 1.71]), F(1, 94) = 17.74, p b .001, η2p = .16. Additionally, we found a main effect of scenario, such that participants made a greater number of attempts for cheering up a sad friend (M = 1.93 [1.76, 2.10]) than for calming down an angry friend (M = 1.70 [1.55, 1.85]), F(1, 94) = 5.25, p = .02, η2p = .05. We did not find a Condition × Scenario interaction, F(1, 94) = 0.35, p = .56, η2p = .004. 4.2.2. Breadth of strategies To determine whether excluded participants utilized a broader repertoire of strategies than controls, we conducted a 2(Condition: Exclusion vs. Control) × 2(Emotional Scenario: Cheer-Up Sadness vs. CalmDown Anger) mixed ANOVA on the breadth of strategies that participants used for each scenario. We found a main effect of condition, such that those in the exclusion condition utilized a greater breadth of strategies (M = 1.79 [1.63, 1.95]) than those in the control condition (M = 1.36 [1.23, 1.49]) to manage a friend's emotions, F(1, 94) = 17.44, p b .001, η2p = .16. There was no main effect of scenario, F(1, 94) = 0.33, p = .57, η2p = .003, and no Condition × Scenario interaction, F(1, 94) = 0.33, p = .57, η2p = .003. 4.3. Discussion Findings from Study 1 suggest that heightened belonging needs may motivate individuals to manage others' emotions. Compared to their neutral counterparts, participants reminded of a prior exclusion made a greater number of attempts to cheer up a sad friend and to calm down an angry friend. Moreover, excluded individuals tended to use a greater breadth of strategies to manage a friend's emotions, which may increase their potential for effective action. Accessing a greater number and broader repertoire of strategies can provide individuals with more alternatives to choose from when responding to others, which may help them select strategies that may be most effective for the individual they are regulating and the context at hand (see Bonanno & Burton, 2013). One limitation of Study 1 was that our experimental paradigm was relatively contextually impoverished and offered little to no potential for social bonding or reacceptance. When people manage others' emotions in day-to-day life, they are oftentimes managing others' emotions in the context of face-to-face interactions whereby they are provided with rich details about the target individual and the situational context at hand. The rich details available in day-to-day life help individuals tailor their strategies to the target other and the situation. We sought to address this limitation in Study 2 by having participants take part in a face-to-face interaction with a stranger and by providing participants 1 2

We did not find a consistent pattern of gender effects across our studies. From this point forward, square brackets indicate 95% confidence intervals.

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

with rich details about the situational context (a negotiation for a signing bonus). In addition, we included a positive social condition (social acceptance) as a control condition in Study 2, to rule out the alternative hypothesis that our findings from Study 1 could be explained by priming social constructs more broadly as opposed to activating the need to belong specifically. 5. Study 2 In Study 2, we were interested in examining whether excluded individuals would be more effective at managing others' emotions in a faceto-face interaction and the potential social benefits of managing others' emotions for the excluded individual. In Study 2, we were interested in exploring the management of others' emotions in the context of a workplace coaching session. A great deal of the research that has explored the social benefits of emotional intelligence has explored these benefits in workplace and leadership contexts (e.g., Niven et al., 2012; for reviews, see Côté, 2014; Ybarra et al., 2012, 2013). Indeed, the ability to effectively manage others' emotions is considered to be an important component of effective leadership (for reviews, see Ashkanasy & Humphrey, 2014; Humphrey, 2002; Pescosolido, 2002), and skill at managing others' emotions tends to be associated with superior performance in leadership roles and organizational settings (for reviews, see Côté, 2014; Mayer et al., 2004). In the present research, we sought to explore the influence of social exclusion on emotion management in a workplace context. Specifically, we were interested in exploring the effectiveness of excluded individuals as coaches serving to increase feelings of energization and/or anger in an interaction partner. Indeed, prior research has found that a leader's skill at energizing subordinates was associated with superior employee performance and positive coworker exchanges (e.g., McColl-Kennedy & Anderson, 2002; Tsai, Chen, & Cheng, 2009). The strongest test of the present hypothesis is to examine how manipulated belonging need in one participant can influence outcomes in their interaction partner (in terms of altered emotional state and partner liking) after a face-to-face interaction. Thus, in the present study, we manipulated our independent variable in one participant and assessed our dependent variables in their interaction partner after a face-to-face interaction. Participants took part in a study on negotiation behavior in same-sex pairs. Upon arrival to the laboratory, participants were randomly assigned to serve one of two roles: the job candidate (who would be negotiating a signing bonus with a recruiter later on in the study) and the negotiation coach for the job candidate (who would serve to help the job candidate develop strategies to use in the negotiation in order to obtain the highest possible signing bonus). At the onset of the study, we manipulated belonging need in the negotiation coaches by having them complete a reliving task in which they either relived a past experience of social exclusion (social exclusion condition) or a past experience of social acceptance (positive, social control condition). After the reliving task, participants reviewed information about the negotiation scenario and their roles. At this point, negotiation coaches were given additional instructions to increase feelings of confidence, assertiveness, and even anger in their partner (the job candidate) during the coaching session. After the coaching session, we assessed how effective the coach was at managing the job candidates' emotions by assessing two indicators of effectiveness in the job candidate: (1) the job candidate's affect (both implicit and explicit) after the coaching session and (2) the job candidate's negotiation confidence after the coaching session. Additionally, we also assessed the job candidate's liking of their coach after the coaching session as a potential indicator of the coach's success at reconnection. We hypothesized that excluded coaches would be more effective at managing their partner's emotions than accepted coaches. We examined two potential ways in which excluded coaches could effectively manage their partner's emotions in the coaching session to encourage greater confidence and assertiveness: (1) excluded coaches could be

63

more effective at increasing their partner's levels of high arousal negative emotions (i.e., anger) and/or (2) excluded coaches could be more effective at increasing their partner's levels of high arousal positive emotions (i.e., energization). Both high arousal negative emotions (i.e., anger) and high arousal positive emotions (i.e., energization) have been found to predict superior negotiation performance (e.g., Carnevale & Isen, 1986; Tamir & Ford, 2012). We reasoned that coaches could be able to effectively help their partner's performance in the negotiation by increasing high arousal positive emotions, high arousal negative emotions, or both. In this particular setting, increasing feelings of anger might be seen as the most direct strategy for effective emotion management, as people hold the expectation that anger will be useful in a confrontational negotiation context, and are more likely to make attempts to increase their own anger before a negotiation (Tamir & Ford, 2012). However, a second way in which coaches could effectively manage their partner's emotions in the coaching session is to increase their partner's levels of high-arousal positive affect (i.e., energization). This second possibility may be considered to be a less direct coaching strategy, but it has the potential to provide greater benefits for both the coach and their interaction partner. Specifically, coaches may want to avoid increasing feelings of anger in their interaction partner, as increasing feelings of anger may ultimately undermine the broader goal of reconnection (e.g., if their partner associates feelings of anger with the coach, rather than the negotiation context). In this way, coaches may prefer to increase high arousal positive emotions like energization in their partner, as energization may hold similar benefits as anger in terms of promoting negotiation performance (see Carnevale & Isen, 1986), without the risk of undermining reconnection. We examined both potential possibilities for effective emotion management in this study. We hypothesized that the job candidates who interacted with excluded coaches would exhibit higher levels of energization and/or anger (both implicit and explicit) after the coaching session. We also hypothesized that the job candidates who interacted with excluded coaches would exhibit superior negotiation confidence after the coaching session. Lastly, given that the proposed function of managing others' emotions is to promote reconnection, we hypothesized that job candidates who interacted with excluded coaches should also report greater liking for their partner.

5.1. Method 5.1.1. Participants Eighty-two dyads (164 total participants: 103 women, 61 men; Mage = 18.92, age range: 18–25 years) were recruited from an introductory psychology course to participate in this study for course credit. Our sample size for this study was based on previous studies using similar reliving manipulations of exclusion in the laboratory (e.g., Bernstein et al., 2008). We targeted approximately 40 dyads per cell. We first stopped data collection at 55 dyads due to limited subject availability. Following analyses, we collected an additional 29 dyads over the following academic quarter before analyses were repeated. Seven dyads were excluded from the analyses: Five dyads were excluded from the exclusion condition (three dyads were excluded because the negotiation coaches did not write about past experiences of social exclusion in the reliving task, one dyad was excluded for not staying on task for the majority of the coaching session, and one dyad was excluded because of the accidental scheduling of a mixed-sex dyad) and two dyads were excluded from the acceptance condition (one dyad was excluded for not staying on task for the majority of the coaching session and the other dyad was excluded because the participants had known each other previously and also did not stay on task for the majority of the coaching session). Altogether, our final sample resulted in a total of 75 dyads.

64

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

5.2. Materials 5.2.1. Explicit affect In order to assess participants' explicit affect, we had participants respond to eight items assessing the following four emotional quadrants: explicit energization (energized, lively; α = .81), explicit calm (calm, content; α = .61), explicit anger (angry, fed up; α = .74), and explicit sadness (sad, gloomy; α = .81). Participants were asked to rate the extent to which they currently felt each item on a 7-point Likert-type scale, from 1 (not at all) to 7 (extremely). 5.2.2. Implicit affect We assessed participants' implicit affect after the interaction with a modified version of the Implicit Positive and Negative Affect Test (IPANAT; Quirin, Kazén, & Kuhl, 2009). The IPANAT is an implicit measure of affect that is sensitive to state variations in positive and negative affect. In this task, participants were presented a series of nonsense words (e.g., “SAFME”) that were supposedly from an artificial language. Participants were asked to rate the emotional tone of each nonsense word by rating the extent to which each of eight emotional items “fit” with the word on a scale from 1 (doesn't fit at all) to 7 (fits very well). The original IPANAT used six items to capture implicit levels of positive and negative mood (happy, energetic, cheerful, tense, helpless, and inhibited). In order to better capture the emotions that were most relevant to our research question at hand, we modified the IPANAT to assess both valence (positive vs. negative) and arousal (high vs. low) dimensions, resulting in the following four emotional quadrants: implicit energization (energized, excited; α = .62), implicit anger (angry, determined; α = .63), implicit calm (calm, content; α = .55), and implicit sadness (sad, inhibited; α = .49). 5.2.3. Partner liking index We assessed participants' liking of their partners by having participants rate their partners on the following dimensions: liking, friendliness, smoothness of interaction, and rapport, using a 7-point Likerttype scale, from 1 (not at all) to 7 (extremely). We also included the one-item Inclusion of Other in Self Scale (Aron, Aron, & Smollan, 1992) as an additional indicator of partner liking. Since the reliability of all the partner-liking items was high (α = .84), we standardized and aggregated all items into one composite measure of partner liking. 5.3. Procedure Participants were run in same-sex pairs for each experimental session. This study employed a 2(Role in Dyad: Negotiation Coach vs. Job Candidate) × 2(Dyad Condition: Excluded Coach vs. Accepted Coach) experimental design. Upon arrival to the laboratory, the experimenter informed participants that the purpose of the study was to examine the strategies that people use in job negotiations, and that they would be taking part in a negotiation later on in the study. Each participant in the pair was randomly assigned to one of two roles: the job candidate (who would be negotiating their signing bonus with a recruiter later on in the study) and the negotiation coach for the job candidate (who would serve to help the job candidate develop strategies to use in the negotiation in order to obtain the highest signing bonus possible). The experimenter then informed the participants that the recruiter would be completing the first part of the study in “company headquarters”, which was located in another lab downstairs. After assigning participants their role for the negotiation, the experimenter informed participants there was some extra time while they waited for the recruiter to review the information about the negotiation and make their initial offer. In the meantime, participants were asked to take part in a separate pilot study on autobiographical memory. In actuality, this autobiographical memory task served as our key belonging manipulation of interest. We manipulated belonging need in the negotiation coaches by having them complete a reliving task in which

they were randomly assigned to either write about a past experience of social exclusion (social exclusion condition) or a past experience of social acceptance (positive, social control condition). In order to increase the salience of our belonging manipulation (to ensure the belonging prime would be powerful enough to affect behavior in a face-to-face interaction), we had participants first describe four instances in which they felt socially excluded (vs. socially accepted). Next, participants were asked to select the one experience that made them feel the most intensely socially excluded (vs. socially accepted), the experience that affected them most to this day, and spend 3 min writing an essay describing that experience in detail (following Lerner & Keltner, 2001). All job candidates (regardless of dyad condition) completed a reliving task in which they wrote about the layout of their local grocery store. We reasoned that writing about the layout of their local grocery store should keep job candidates in a relatively neutral mood prior to the coaching session. Following the reliving task, participants reviewed an information packet outlining the negotiation scenario and his or her role in the negotiation (adapted from Galinsky, Mussweiler, & Medvec, 2002). Participants were told that the job candidate would be negotiating a signing bonus at a prestigious consulting firm in New York City. The job candidate was a competitive applicant, having received a high GPA from Northwestern University, completed several internships at prestigious consulting firms, and won an elite fellowship that was only granted to 10 students nationwide for outstanding scholarship and leadership experience. Furthermore, the job candidate already had a back-up job offer and a $10,000 signing bonus with another consulting firm in Chicago. Lastly, the job candidate had $20,000 in student loans to take into consideration when negotiating their signing bonus, with the aim of paying off as much of the student loans (if not all) if possible. In addition to the information about the negotiation scenario, the negotiation coaches were given privileged information about the consulting industry (e.g., average signing bonuses and maximum signing bonuses offered) that they could use to in helping the job candidate devise strategies for effective negotiation. Crucially, embedded in this privileged information were regulation instructions for the coaches to increase feelings of confidence, assertiveness, or even anger in their partner during the coaching session. After reviewing the materials, the negotiation coaches were given a few minutes to prepare strategies for helping the job candidate justify their position in the negotiation. After both participants had finished reviewing the information about the negotiation and their roles, the experimenter led them to a separate room where the coaching session was to take place. Participants were seated facing each other at a table and given 3 min to discuss whether the job candidate should accept or reject the recruiter's initial offer and to come up with strategies for how the job candidate should justify their position in the negotiation. We videotaped these interactions using a hidden camera to record participants' interaction behavior during the coaching session (see Supplemental analyses). Following the coaching session, the experimenter led participants back to their separate cubicles for the remaining portion of the study. First, the job candidate recorded a counteroffer video in response to the recruiter's initial offer, in which they stated their counteroffer for the signing bonus and provided justification for why they thought they deserved that amount for the signing bonus. Second, all participants (both job candidates and negotiation coaches) completed measures of their post-interaction affect (both implicit and explicit) and measures of partner liking (described above). Third, all participants completed a measure of global vs. local processing (Navon, 1977) that was neither central to the present research question at hand, nor related to the experimental manipulation in the coach or negotiation outcomes of the candidates. As such, we do not discuss this measure further. 5.4. Results and discussion In the present study, we manipulated our independent variable in the negotiation coaches (excluded coach vs. accepted coach) and

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

assessed our dependent variables in the job candidates after the coaching session. Specifically, we assessed: (1) the job candidate's affect, both explicit and implicit, after the coaching session (as an indicator of the coach's effectiveness at managing others' emotions), (2) the job candidate's negotiation confidence after the coaching session (as an additional indicator of the coach's effectiveness at managing others' emotions), and (3) the job candidate's liking of their coach after the coaching session (as an indicator of the coach's potential success at reconnection). Descriptive statistics and intercorrelations between the variables are presented in Table 2. 5.4.1. Affect (both implicit and explicit) We hypothesized that excluded coaches would be more effective than accepted coaches at managing their partner's emotions. Specifically, we hypothesized that excluded coaches would be more effective than accepted coaches at increasing their partners' levels of energization and/or anger during the coaching session. To test whether job candidates' affect after the interaction varied as a function of dyad condition, we conducted a 2(Dyad Condition: Excluded Coach vs. Accepted Coach) × 2(Affect Type: Implicit vs. Explicit) × 2(Valence: Positive vs. Negative) × 2(Arousal: High Arousal vs. Low Arousal) mixed ANOVA on job candidates' affect scores. We standardized participants' responses for the two different affect measures (implicit and explicit) for the analyses. We did not predict or find significant main effects of dyad condition, F(1, 73) = 1.63, p = .21, η2p = .02, affect type, F(1, 73) = 0.05, p = .83, η2p = .001, valence, F(1, 73) = 0.05, p = .82, η2p = .001, or arousal, F(1, 73) = 0.01, p = .99, η2p = .001. Furthermore, we did not predict or find a significant Affect-Type × Valence Interaction, F(1, 73) = 0.01, p = .98, η2p = .001, Affect-Type × Arousal interaction, F(1, 73) = 0.04, p = .83, η2p = .001, Valence × Arousal interaction, F(1, 73) = 0.03, p = .86, η2p = .001, Affect-Type × Valence × Arousal interaction, F(1, 73) = 0.01, p = .98, η2p = .001, or an Affect-Type × Valence × Dyad Condition interaction, F(1, 73) = 0.04, p = .84, η2p = .001. We found a marginally significant Dyad Condition × Valence interaction, F(1, 73) = 3.45, p = .07, η2p = .04. Tests of simple effects revealed that job candidates who interacted with excluded coaches reported higher levels of positive affect after the coaching session (M = 0.15 [− 0.04, 0.33]) than job candidates who interacted with accepted coaches (M = −0.19 [−0.39, 0.02]), p = .02, η2p = .25. There were no differences in negative affect between job candidates who interacted with excluded coaches or accepted coaches, p = .67. Furthermore, we found a marginal Dyad Condition × Affect-Type interaction, F(1, 73) = 3.38, p = .07, η2p = .04. Tests of simple effects revealed that job candidates who interacted with excluded coaches reported higher levels of overall implicit affect (M = 0.12 [−0.09, 0.33]) than job candidates who interacted with accepted coaches (M = − 0.15 [− 0.39, 0.09]), p = .09. There were no differences in overall explicit affect

65

between job candidates who interacted with excluded coaches or accepted coaches, p = .90. Our central hypothesis of interest was that excluded coaches would be more effective at increasing high-arousal affect (i.e., anger and/or energization) in their partner. Although we did not find our predicted Dyad Condition × Arousal interaction, F(1, 73) = 0.01, p = .92, η2p = .001, we did find a marginally significant Dyad Condition × Arousal × AffectType three way interaction, F(1, 73) = 3.06, p = .08, η2p = .04, implying the results differed across implicit and explicit measures of affect. Tests of simple effects revealed a significant Dyad Condition × Affect-Type two-way interaction for high arousal affect, F(1, 73) = 0.02, p = .01, η2p = .08, but not for low arousal affect, F(1, 73) = 0.03, p = .86, η2p = .001. Specifically, for high arousal affect, we found the predicted pattern of effects within our implicit affect measure, such that job candidates who interacted with excluded coaches reported higher levels of high arousal implicit affect (M = 0.18 [−0.08, 0.44]) than job candidates who interacted with accepted coaches (M = −0.23 [−0.52, 0.07]), p = .04. We did not find any differences in high arousal affect for our explicit affect measure, p = .37. Lastly, we did not find a significant Dyad Condition × Affect-Type × Valence × Arousal four-way interaction, F(1, 73) = 2.07, p = .15, η2p = .03. Taken together, these findings suggest that excluded coaches seem to be effective at increasing high-arousal affect in their partner during the coaching session, though we only found the predicted differences with our implicit affect measure and not our explicit affect measure. One potential explanation for why we did not find any differences on our explicit affect measure may be that our explicit affect measure may not have been sensitive enough to detect such changes. Alternatively, our explicit affect measure could have yielded a Type II error, which tends to be more common than expected when testing a true hypothesis (see Stukas & Cumming, 2014). Although we found that excluded coaches were more effective at increasing high-arousal affect in their partner on the implicit affect measure more generally, we were also interested in exploring whether excluded coaches were more likely to tune this high-arousal affect toward greater positivity (i.e., energization) versus negativity (i.e., anger) more specifically. As such, we conducted focused contrasts on participants' implicit affect scores comparing the effects of dyad condition within each emotional quadrant separately (i.e., implicit energization, implicit anger, implicit calm, and implicit sadness). As seen in Fig. 1, we found that job candidates' implicit energization scores differed by dyad condition, such that job candidates who interacted with excluded coaches showed higher levels of implicit energization after the coaching session (M = 3.99 [3.78, 4.20]) than job candidates who interacted with accepted coaches (M = 3.55 [3.32, 3.79]), t(73) = − 2.73, p = .008, d = 0.63. We did not find any differences by dyad condition for the remaining three emotional quadrants (all ps N .20). Altogether, these findings suggest that excluded coaches were more effective at managing their partner's emotions by increasing their partner's levels of implicit energization. This

Table 2 Intercorrelations among and descriptive statistics for Study 2 variables (job candidate responses).

1. Implicit energization 2. Implicit calm 3. Implicit anger 4. Implicit sadness 5. Explicit energization 6. Explicit calm 7. Explicit anger 8. Explicit sadness 9. Partner liking (z-score) 10. Counteroffer amount 11. Reasons for justification † p ≤ .10. ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001.

M

SD

1

2

3

4

5

6

7

8

9

10

11

3.80 3.43 3.65 3.15 4.04 4.47 2.47 2.35 0.04 19,376.71 5.53

0.72 0.68 0.80 0.69 1.29 1.06 1.28 1.30 0.76 9685.25 2.47

– .16 .50⁎⁎⁎ .24⁎ .19† .04 −.10 −.07 .14 −.02 −.05

– – .39⁎⁎ .15 .15 .19 −.02 −.12 .10 −.20 −.01

– – – .42⁎⁎⁎ .12 −.01 .25⁎

– – – – .32⁎⁎ .01 .17 .11 −.08 .03 −.03

– – – – – .25⁎ −.26 −.58⁎⁎⁎ .22† .27⁎ .18

– – – – – – −.53⁎⁎⁎ −.39⁎⁎

– – – – – – – .62⁎⁎⁎ .08 −.13 −.06

– – – – – – – – −.21† −.09 −.12

– – – – – – – – – .22† .08

– – – – – – – – – – .16

– – – – – – – – – – –

.11 .05 −.16 −.15

.15 .12 .24⁎

66

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

Accepted Coach

Implicit Affect (1 to 7)

4.35 Excluded Coach

4.15 3.95 3.75 3.55 3.35 3.15 2.95 2.75 Implicit Energization

Implicit Anger

Fig. 1. Study 2: Job candidates' implicit energization and implicit anger scores as a function of dyad condition (excluded coach vs. accepted coach). Higher numbers indicate higher levels of implicit affect. Error bars indicate 95% confidence intervals.

strategy of increasing a partner's level of high-arousal positive affect (i.e., energization) is a sensible strategy, as coaches may not have been comfortable increasing feelings of high arousal negative affect (i.e., anger) in their partners — especially given that these partners were strangers. Furthermore, increasing feelings of anger in one's interaction partner may be a particularly risky strategy for excluded coaches to adopt, as increasing feelings of anger may ultimately undermine the broader goal of successful reconnection (e.g., if their partner associates feelings of anger with the coach, rather than the negotiation context). 5.4.2. Negotiation confidence To examine whether job candidates who interacted with excluded coaches were more energized and confident negotiators than job candidates who interacted with accepted coaches, we conducted independent-sample t-tests on two measures of negotiation confidence: (1) the amount of money that job candidates requested for their counteroffer and (2) the number of reasons that job candidates provided as justification for why they thought they deserved more money. We did not find any differences by dyad condition for the amount of money that job candidates' requested as their counteroffer, t(71) = 0.83, p = .41, d = 0.19. However, as predicted, we found that the number of reasons that job candidates provided as justification did significantly differ as a function of dyad condition, such that job candidates who interacted with excluded coaches provided a greater number of reasons for justification (M = 6.02 [5.18, 6.87]) than job candidates who interacted with accepted coaches (M = 4.91 [4.25, 5.57]), t(72) = −1.97, p = .053, d = 0.47, as might be expected if they were feeling more energized. 5.4.3. Partner liking Given that the proposed function of managing others' emotions is to facilitate reconnection, we examined whether excluded coaches were indeed liked better by their partners than accepted coaches. To explore this, we conducted an independent-sample t-test examining whether job candidates' partner liking ratings differed as a function of dyad condition (excluded coach vs. accepted coach). We found that partner liking ratings did differ as a function of dyad condition, such that job candidates who interacted with excluded coaches tended to like their partners more (M = 0.20 [− 0.01, 0.42]) than job candidates who interacted with accepted coaches (M = − 0.18 [− 0.45, 0.09]), t(71) = −2.20, p = .03, d = 0.51. In Study 2, we conducted a strong test for investigating whether excluded individuals would be more effective at managing others' emotions and whether this greater effectiveness at managing others' emotions would serve to promote reconnection. Specifically, we examined how manipulated belonging need in one participant influenced outcomes (altered emotional state and partner liking) in another

participant (their interaction partner) after a face-to-face interaction. As predicted, we found that excluded participants were more effective at managing their partner's emotions (as assessed by both implicit affect and behavior in the negotiation) during the coaching session and this greater effectiveness at managing others' emotions seemed to promote reconnection. Specifically, we found that job candidates who interacted with excluded coaches exhibited greater levels of high-arousal affect on the implicit affect measure, and that this high-arousal affect seemed to be tuned toward greater positivity (i.e., implicit energization). Furthermore, job candidates who interacted with excluded coaches exhibited superior negotiation effectiveness, providing greater justification for their position in the negotiation after the coaching session. Lastly, we found that job candidates who interacted with excluded coaches also reported greater liking for their partner, supporting the notion that managing others' emotions may promote positive social outcomes for the excluded individual. Taken together, these findings suggest that excluded participants were more effective at managing their partner's emotions during the coaching session and this greater effectiveness at managing others' emotions may help the excluded individual garner liking from others. One limitation of Study 2 was that we provided all participants with the information they needed to energize their partner for the negotiation. Specifically, we gave all coaches the same information about the qualifications of the job candidate (e.g., impressive awards and GPA) as well as the negotiation context (e.g., the industry norms for signing bonuses and the job candidates' student loans). Although providing all participants with the same information about the job candidate and the negotiation context allowed us to have high levels of experimental control, one limitation was that we likely constrained the strategies that participants could have used for managing their partner's emotions. Furthermore, another limitation of Studies 1 and 2 was that in both studies, we gave participants explicit instructions to manage another person's emotions. As a result, we cannot draw the conclusion that excluded individuals spontaneously manage others' emotions as a strategy to facilitate social connection. Indeed, it could simply be the case that excluded individuals sought to perform well on the experimental tasks, perhaps as a strategy to repair their self-esteem after social exclusion. We sought to address both of these limitations in Studies 3 and 4 by examining whether excluded individuals would spontaneously manage others' emotions in the context of an online pen pal exchange. Given that emotional intelligence has been linked with superior skill at navigating a broad spectrum of social contexts (e.g., Côté et al., 2011; Lopes et al., 2004; Mayer et al., 2004), we sought to explore whether excluded individuals would similarly show greater management of others' emotions across a wide range of emotional contexts. As such, in Studies 3 and 4, we had participants respond to multiple different pen pals in an online pen pal exchange, each of whom was experiencing a different emotional event in their lives. Lastly, in Studies 3 and 4 we did not give participants explicit instructions to manage others' emotions, so that we could assess whether participants would spontaneously manage others' emotions as a strategy to promote reconnection. 6. Study 3 In Study 3, we sought to explore whether excluded individuals would spontaneously manage others' emotions, and whether this enhanced tendency to manage others' emotions would translate across a broad spectrum of emotional contexts. Once again, we induced feelings of social exclusion by having participants complete a reliving task. Next, participants took part in an online pen pal exchange in which they exchanged letters with three ostensible pen pals, each of whom wrote letters describing a different emotional event in their lives (pen pal 1 suffered the death of a beloved pet, pen pal 2 was anxious about her recent unemployment, and pen pal 3 was celebrating the birth of his first grandchild).

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

As in Study 1, we assessed emotion management by coding the number of attempts that participants made in their letters to manage their pen pals' emotions as well as the breadth of strategies they used in their responding. Additionally, we also assessed effectiveness at managing other's emotions by having coders blind to condition rate participants' letters for their likely effectiveness at the specific emotion management task most appropriate to their pen pal's emotional needs (e.g., comforting sadness, soothing anxiety, and encouraging the savoring of happiness). Lastly, we were interested in exploring potential social benefits of managing others' emotions; in particular, whether managing others' emotions would increase the likability of the previously excluded individual. As a result, we had coders rate participants' response letters for their general likability. Consistent with Study 1, we expected participants who had been reminded of a social exclusion to make a greater number of attempts to manage their pen pals' emotions and to use a broader repertoire of strategies in their response letters relative to control participants. We expected them to attempt to down-regulate negative emotions, but work to maintain or up-regulate positive emotions. Furthermore, we expected the letters written by excluded participants to be coded as more effective at managing their pen pals' emotions than the letters written by control participants. Given that the one important function of managing others' emotions is to facilitate social connection (Libbrecht et al., 2014; Lopes et al., 2003, 2004, 2005; Niven et al., 2012), we explored whether managing others' emotions could potentially help excluded individuals garner liking from others. Thus, we hypothesized that response letters written by excluded participants would be coded as more likable than response letters written by control participants. 6.1. Method 6.1.1. Participants One hundred and eighteen participants (72 women, 46 men; Mage = 33.93, age range: 18–64 years) were recruited from Amazon's Mechanical Turk (MTurk) to participate in an online study for payment. Following the sample size determination of Study 1, we targeted approximately 50 participants per cell (after excluding participants from analyses). We stopped data collection when we reached our predetermined targets. Seventeen participants were excluded from the analyses: Ten participants were excluded from the exclusion condition (four participants did not write about past experiences of social exclusion in the reliving task, two participants indicated they could not recall a past experience of exclusion, two participants indicated they were no longer affected by the experience in their reliving essays, and two participants did not write response letters in the online pen pal exchange) and seven participants were excluded from the control condition (all seven participants did not complete the reliving task). Altogether, our final sample resulted in a total of 101 participants. 6.1.2. Procedure Participants first completed the reliving task from Study 1 in which they were randomly assigned to write about one of the following experiences: a time they felt intensely socially excluded in some way — a time when they felt as if they did not belong (social exclusion condition) or the layout of their local grocery store (control condition). Following the reliving task, participants were informed that the experimenters were interested in developing a new “pen pal” type of social networking site and that they would be taking part in a pilot pen pal exchange. Participants were told that they were paired up with three pen pal buddies who already wrote introductory letters to the participant, and they were asked to write a response letter to each pen pal. All participants received letters from three ostensible pen pals who each described a recent emotional event in their lives: Kim was feeling sad because her dog passed away, Ruth who was feeling anxious because she was recently laid off and needed a job to support her four children,

67

and Roger who was feeling happy because of the birth of his first grandchild. Participants were asked to write letters in response to each pen pal. Lastly, participants completed the Schutte Emotional Intelligence Scale (SEI; Schutte et al., 1998) to examine the potential influence of dispositional emotional intelligence on emotion management. See Supplemental analyses.

6.2. Coding of pen pal letters Upon completion of the study, we had two pairs of coders who were blind to experimental condition code the pen pal letters for the number of attempts that participants made to manage their pen pal's emotions, for the breadth of strategies that participants used in their responses, for their effectiveness at emotion management, and for their likability. Following Study 1, the first pair of coders first coded the pen pal letters for the number of attempts made per letter (α's N .78) as well as for the different types of strategies that participants used to manage their pen pal's emotions. We calculated a breadth index of the number of different types of strategies that participants used in each letter (α's N .71). Following Study 1, we considered both the number of attempts and the breadth of strategies used as rough proxies for potential effectiveness. Next, as a richer assessment of participants' potential effectiveness at managing emotions, we had the second pair of coders rate how effective each participant's response was at managing their pen pal's emotions on a 5-point scale from 0 (not at all effective) to 4 (extremely effective) (α's N .85). Finally, we had the first pair of coders recode the letters and rate the likability of each participant's pen pal letter on a scale from 1 (not at all likable) to 7 (extremely likable) (α's N .71). Descriptive statistics and intercorrelations between the variables are presented in Table 3.

6.3. Results and discussion 6.3.1. Number of attempts To determine whether excluded participants made a greater number of attempts at emotion management in their response letters than controls, we conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on the number of attempts made in participants' response letters to manage their pen pals' emotions. We found a main effect of condition, such that those in the exclusion condition made a greater number of attempts at managing their pen pals' emotions in their response letters (M = 2.77 [2.41, 3.13]) relative to those in the control condition (M = 2.15 [1.82, 2.48]), F(1, 99) = 6.37, p = .01, η2p = .06, replicating Study 1's findings. Additionally, we found a main effect of pen pal emotion, F(2, 198) = 94.34, p b .001, η2p = .49, such that people made a greater number of attempts to manage the emotions of their pen pal when responding to the two negative emotion scenarios, the sad and the anxious pen pals (Msad = 3.03 [2.69, 3.38] and Manxious = 3.08 [2.75, 3.41] respectively) relative to the happy pen pal (M = 1.26 [1.05, 1.46]), both ps b .001. There were no significant differences in the number of attempts made in the response letters to manage the emotions of the sad and anxious pen pals (p = 1.00). Lastly, we did not find a Condition × Pen Pal Emotion interaction, F(2, 198) = .84, p = .43, η2p = .008.

Table 3 Intercorrelations among and descriptive statistics for Study 3 variables.

1. Number of Attempts 2. Breadth of Strategies 3. Rated Effectiveness 4. Rated Likability ⁎⁎⁎ p b .001.

M

SD

1

2

3

4

2.43 1.41 1.82 4.28

1.26 0.61 0.90 1.28

– .87⁎⁎⁎ .92⁎⁎⁎ .86⁎⁎⁎

– – .88⁎⁎⁎ .84⁎⁎⁎

– – – .86⁎⁎⁎

– – – –

68

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

6.3.2. Breadth of strategies To determine whether excluded participants utilized a broader repertoire of emotion-regulation strategies in their response letters than controls, we conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on the breadth of strategies used in participants' response letters. We found a main effect of condition, such that those in the exclusion condition utilized a greater breadth of strategies in their response letters to manage their pen pal's emotions (M = 1.56 [1.39, 1.74]) relative to those in the control condition (M = 1.29 [1.13, 1.45]), F(1, 99) = 5.15, p = .025, η2p = .05, again replicating Study 1's findings. Additionally, we found a main effect of pen pal emotion, F(2, 198) = 91.63, p b .001, η2p = .48, such that people used a greater breadth of strategies when responding to the sad and the anxious pen pals (Msad = 1.77 [1.60, 1.95] and Manxious = 1.69 [1.55, 1.84] respectively) relative to the happy pen pal (M = .81, [.69, .93]), both ps b .001. There were no significant differences between the sad and anxious pen pals in the breadth of strategies used (p = .95). Lastly, we did not find a Condition × Pen Pal Emotion interaction, F(2, 198) = .98, p = .38, η2p = .01. 6.3.3. Subjective effectiveness To determine whether excluded participants wrote letters that were rated as more effective at managing their pen pal's emotions than controls, we conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on effectiveness ratings. We found a main effect of condition, such that those in the exclusion condition wrote letters that were rated as more effective at managing their pen pal's emotions (M = 2.09 [1.84, 2.34]) than those in the control condition (M = 1.59 [1.36, 1.82), F(1, 99) = 8.25, p = .005, η2p = .08. There was a main effect of pen pal emotion, F(2, 198) = 35.56, p b .001, η2p = .26, such that people wrote letters that were rated as more effective when down-regulating the sad and the anxious pen pals (Msad = 2.02 [1.80, 2.24] and Manxious = 2.20 [1.99, 2.40] respectively) relative to up-regulating the happy pen pal (M = 1.30 [1.08, 1.52]), both ps b .001. There were no significant differences between the sad and anxious letters in effectiveness (p = .18). Lastly, we neither predicted nor found a Condition × Pen Pal Emotion interaction, F(2, 198) = 0.17, p = .85, η2p = .002. 6.3.4. Likability Given that the proposed function of managing others' emotions is to facilitate reconnection, we examined whether excluded participants' response letters were rated as higher in likability than the letters written by control participants. We conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on likability ratings. As expected, we found a main effect of condition, such that those in the exclusion condition wrote letters that were rated as more likable (M = 4.73 [4.38, 5.09]) than those in the control condition (M = 3.90 [3.58, 4.23), F(1, 99) = 11.64, p = .001, η2p = .10. There was a main effect of pen pal emotion, F(2, 198) = 14.16, p b .001, η2p = .13, such that participants' letters to sad pen pals were rated as more likable (M = 4.69 [4.41, 4.97]) than participants' letters to anxious and happy pen pals (Manxious = 4.10 [3.84, 4.37] and Mhappy = 4.16 [3.87, 4.45] respectively), both ps b .001. There were no significant differences between the anxious and happy letters in likability (p = 1.00). Lastly, we neither predicted nor found a Condition × Pen Pal Emotion interaction, F(2, 198) = 0.13, p = .88, η2p = .001. 6.3.5. Mediation analysis for effectiveness We examined whether our main effect of condition on subjective effectiveness ratings was mediated by the greater number of attempts and the breadth of strategies that excluded individuals used in their pen pal letters to manage their pen pal's emotions. Note that we had two separate pairs of coders coding participants' response letters (one pair of coders counted the number and breadth of strategies that participants used in their response letters and a separate pair of coders rated

participants' response letters for their effectiveness at managing the pen pal's emotions) to avoid biases in our coders' ratings. As shown in Fig. 2, condition significantly predicted ratings of effectiveness, b = 0.50, t(101) = 2.87, p = .005. We next established that condition significantly predicted our hypothesized mediators: the number of attempts made (b = 0.62, t(101) = 2.52, p = .01) and the breadth of strategies used (b = 0.27, t(101) = 2.27, p = .025) in participants' letters to manage their pen pal's emotions. Then, we established that both our hypothesized mediators significantly predicted ratings of effectiveness: the number of attempts made (b = 0.43, t(101) = 8.27, p b .001) and the breadth of strategies used (b = 0.52, t(101) = 4.76, p b .001). Furthermore, when both the number of attempts made and the breadth of strategies used were entered into the regression simultaneously with condition, the total direct effect of condition on effectiveness was reduced to nonsignificance, b = 0.09, t(101) = 1.40, p = .16. Following Preacher and Hayes (2008) recommendations, we employed bootstrapping procedures with 1000 resamples, using the biased corrected and accelerated approach, to assess whether the number and breadth of strategies significantly mediated the effect of condition on effectiveness. As predicted, when both mediators (the number of attempts made and the breadth of strategies used) were considered together, the 95% confidence interval for the combined indirect effect (b = 0.40 [0.08, 0.71]) did not contain 0, indicating significant mediation at p b .05. In addition, when examining each of the indirect paths separately, the 95% confidence intervals for the number of attempts made (b = 0.27 [0.06, 0.52]) and the breadth of strategies used (b = 0.14 [0.03, 0.31]) both did not contain 0, indicating significant mediation at p b .05 for both variables. Altogether, these findings suggest that excluded individuals may be more effective at managing others' emotions, because they tend to make a greater number of attempts at managing others' emotions and they also tend to use a broader repertoire of emotion management strategies in their responses. In Study 3, excluded participants spontaneously exhibited greater management of others' emotions across a broad spectrum of emotional contexts. Even though these participants were not explicitly instructed to attempt to manage other's emotions (as in Studies 1 & 2), nor given specific information that could be used to manage the other's emotions (as in Study 2), we found that excluded participants wrote letters that contained a greater number of attempts at managing their pen pal's emotions (consistent with Study 1) and that they were seemingly more effective and likable (consistent with Study 2). Excluded participants wrote letters that were rated as more effective at managing their pen pal's emotions, and this greater effectiveness was mediated by both the greater number of attempts and the greater breadth of strategies that excluded participants used in their letters to manage their pen pal's emotions. Lastly, consistent with the notion that one adaptive function of managing others' emotions is to facilitate social connection (Libbrecht et al., 2014; Lopes et al., 2003, 2004, 2005; Niven et al., 2012), we found that excluded participants wrote letters that were rated as more likable than letters written by control participants. Altogether, these findings provide evidence suggesting that excluded individuals may exhibit greater management of others' emotions, and that the ability to effectively manage others' emotions may potentially garner greater liking from others. One limitation of our studies thus far is that we exclusively relied on experimental tasks that were social in nature. As such, we cannot rule out the alternative explanation that excluded individuals may have been seeking to distract themselves from their own negative affect by immersing themselves in subsequent experimental tasks (a strategy for managing negative emotions in the self). We sought to rule out this alternative explanation in Study 4 by varying the social vs. nonsocial context of the experimental tasks. If excluded individuals were simply motivated to distract themselves from their own negative affect, then we would expect them to show enhanced performance relative to control participants across both social and nonsocial tasks. By contrast, if excluded individuals were motivated to manage others' emotions as a strategy to facilitate reconnection, then we would expect them to

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

69

Number of Attempts b = 0.43***

Breadth of Strategies

b = 0.62*

b = 0.27*

b = 0.52***

Condition

Effectiveness

(0 = Control;

b = 0.50** (0.09)

Ratings

1 = Exclusion) Fig. 2. Mediational analysis for Study 3. Note. *p b .05, **p b .01, ***p b .001.

show selectivity in their responding, showing enhanced performance on social tasks, but not on nonsocial tasks. Lastly, we included a negative nonsocial control condition (reliving a past experience of intellectual failure) in Study 4 in order to rule out the alternative hypothesis that our findings were a result of eliciting negative affect broadly as opposed to activating belonging needs specifically. 7. Study 4 The purpose of Study 4 was to replicate the findings from Study 3, and to examine whether exclusion would also motivate participants to generate a greater number and breadth of strategies in a nonsocial task (generating creative uses for common household items). In Study 4, we also included a negative nonsocial control condition (reliving a past experience of intellectual failure) to rule out the alternative hypothesis that negative affect more broadly was driving the effects. Once again, we induced feelings of social exclusion by having participants complete a reliving task in which they either relived a past experience of social exclusion (social exclusion condition) or a past experience of intellectual failure (negative, nonsocial control condition). Next, participants were randomly assigned to complete either a social task or a nonsocial task. Participants in the social task condition completed the online pen pal exchange from Study 3, in which they exchanged letters with three ostensible pen pals, each of whom wrote letters describing a different emotional event in their lives (pen pal 1 suffered the death of a beloved pet, pen pal 2 was anxious about her recent unemployment, and pen pal 3 was celebrating the birth of his first grandchild). By contrast, participants in the nonsocial task condition completed a nonsocial measure of cognitive breadth (Guilford's (1967) alternative uses task). In this task, participants were asked to list as many possible uses as they could think of for three common household items (a brick, a paperclip, and a newspaper). For both of these tasks, we coded the number of attempts that participants made and the breadth of strategies used in their responding. Following Study 3, we explored the down-stream social consequences of emotion management for participants' effectiveness at managing others' emotions and their likability. As such, we had coders, blind to experimental condition, rate participants' pen pal letters for their likely effectiveness at managing their pen pal's emotions and for their general likability. Consistent with Studies 1 and 3, we expected participants who had been reminded of a social exclusion to make a greater number of attempts to manage their pen pals' emotions and to use a broader repertoire of strategies in their response letters relative to control participants. In addition, we expected that the greater emotion management exhibited by excluded individuals in the pen pal task would be associated with positive down-stream social consequences for their rated effectiveness and likability (consistent with Study 3). Specifically we expected the letters written by excluded participants to be coded as more effective at managing their pen pals' emotions than control

participants. Furthermore, we expected that response letters written by excluded participants to be coded as more likable than response letters written by control participants (consistent with Study 3). Lastly, we expected the influence of exclusion on the number and breadth of strategies generated to be limited to social tasks. Specifically, we expected excluded participants to make a greater number of attempts and to use a broader range of emotion management strategies in the social task (emotion management in an online pen pal exchange). By contrast, we did not expect excluded participants to differ from control participants in terms of the number of attempts made or the breadth of categories listed in the nonsocial task (generating creative uses for common household items). 7.1. Method 7.1.1. Participants Two hundred and fifty nine participants (152 women, 107 men; Mage = 33.93, age range: 18–67 years) were recruited from Amazon's Mechanical Turk (MTurk) to participate in an online study for payment. Following the sample size determination of Studies 1 and 3, we targeted approximately 50 participants per cell (after excluding participants from analyses). We stopped data collection when we reached our predetermined targets. Forty-eight participants were excluded from the analyses (19 from the social exclusion condition and 29 from the intellectual failure condition; 23 from the social task condition and 25 from the nonsocial task condition). Nineteen participants in the social exclusion condition were excluded from the analyses for the following reasons: seven participants did not write about past experiences of social exclusion, seven participants indicated they were no longer affected by the experience in their reliving essays, and five participants indicated in an attention check at the end of the study that they did not complete the study carefully. Twenty-nine participants in the intellectual failure condition were excluded from the analyses for the following reasons: twenty participants did not write about past experiences of intellectual failure in the reliving task, four participants indicated they were no longer affected by the experience in their reliving essays, and five participants indicated in an attention check at the end of the study that they did not complete the study carefully. Altogether, our final sample resulted in a total of 211 participants. 7.1.2. Procedure Participants first completed a reliving task in which they were randomly assigned to write about one of the following experiences: a time they felt intensely socially excluded in some way — a time when they felt as if they did not belong (social exclusion condition) or a time they experienced intense failure in an intellectual domain — a time when they felt as if they were not very smart (negative, nonsocial control condition). Following the reliving task, participants were randomly assigned to complete one of two types of tasks: a social, emotion management

70

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

task (social task condition) or a nonsocial, cognitive breadth task (nonsocial task condition). Participants in the social task condition completed the online pen pal exchange from Study 3, in which they received pen pal letters from three ostensible pen pals, each of whom wrote letters describing a different emotional event in their lives: Kim was feeling sad because her dog passed away, Ruth who was feeling anxious because she was recently laid off and needed a job to support her four children, and Roger who was feeling happy because of the birth of his first grandchild. Participants were asked to write letters in response to each pen pal. By contrast, participants in the nonsocial task condition completed a nonsocial measure of cognitive breadth (Guilford's (1967) alternative uses task). In this task, participants were asked to list as many possible uses as they could think of for three common household items (a brick, a paperclip, and a newspaper).

7.1.3. Coding Upon completion of the study, we had three pairs of coders who were blind to experimental condition code participants' responses. For both the social and the nonsocial tasks, we coded for the number of attempts that participants made in their responses and the breadth of categories that participants used in their responses.

7.1.3.1. Coding for the social task (emotion management in an online pen pal exchange). Following Study 3, we had one pair of coders who were blind to experimental condition code the pen pal letters for the number of attempts that participants made to manage their pen pal's emotions (α's N .84) and the breadth of strategies that they used in their responses (α's N .71). Next, we had a separate pair of coders code participants' letters for their effectiveness at emotion management (α's N .86), and their general likability (α's N .65). Descriptive statistics and intercorrelations between the variables are presented in Table 4.

7.1.3.2. Coding for the nonsocial task (generating creative uses for common household items). We had a third pair of coders code participants' responses in the alternative uses task for the number of attempts that participants made (α's N .97) and for the breadth of categories that participants used in their responses (α's N .94).

7.2. Results and discussion The current study offered (1) an opportunity to replicate our Study 3 findings with a negative (instead of neutral) control condition, (2) an extension to examine whether exclusion would also motivate participants to generate a greater number and breadth of nonsocial strategies in a comparison task and (3) an opportunity to compare the two types of tasks (social vs. nonsocial), all in a between subjects design. Thus, we will first present analyses focused upon the potential replication of Study 3 — including number and breadth of strategies, effectiveness, likability, and the mediation analyses presented in Study 3, next the analyses of the nonsocial alternative uses task, and finally a comparison of the two.

Table 4 Intercorrelations among and descriptive statistics for Study 4 variables.

1. Number of attempts 2. Breadth of strategies 3. Rated effectiveness 4. Rated likability ⁎⁎⁎ p b .001.

M

SD

1

2

3

4

2.74 1.45 2.09 4.28

1.15 0.43 0.73 0.88

– .85⁎⁎⁎ .79⁎⁎⁎ .69⁎⁎⁎

– – .74⁎⁎⁎ .63⁎⁎⁎

– – – .86⁎⁎⁎

– – – –

7.3. Social task (emotion management in an online pen pal exchange) 7.3.1. Number of attempts To determine whether participants who relived a past experience of exclusion made a greater number of attempts at emotion management in their response letters than participants who relived a past experience of intellectual failure, we conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on the number of attempts made in participants' response letters to manage their pen pals' emotions. As predicted, we found a main effect of condition, such that those in the exclusion condition made a greater number of attempts at managing their pen pals' emotions in their response letters (M = 2.97 [2.69, 3.25]) relative to those in the failure condition (M = 2.45 [2.13, 2.77]), F(1, 111) = 5.87, p = .02, η2p = .05, replicating the findings from Studies 1 and 3, and implying the findings were not limited to a neutral control condition. Additionally, we found a main effect of pen pal emotion, F(2, 222) = 83.41, p b .001, η2p = .43, such that people made a greater number of attempts to manage the emotions of their pen pal when responding to the two negative emotion scenarios, the sad and the anxious pen pals (Msad = 3.05 [2.75, 3.34] and Manxious = 3.49 [3.16, 3.82] respectively) relative to the happy pen pal (M = 1.58 [1.40, 1.76]), both ps b .001. Furthermore, people made a greater number of attempts to manage the emotions of the anxious pen pal relative to sad pen pal (p = .02). Lastly, we did not find a Condition × Pen Pal Emotion interaction, F(2, 222) = 0.32, p = .73, η2p = .003. 7.3.2. Breadth of strategies To determine whether participants who relived a past experience of exclusion utilized a broader repertoire of emotion-regulation strategies in their response letters than participants who relived a past experience of intellectual failure, we conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on the breadth of strategies used in participants' response letters. We found the predicted main effect of condition, such that those in the exclusion condition utilized a greater breadth of strategies in their response letters to manage their pen pal's emotions (M = 1.53 [1.42, 1.63]) relative to those in the control condition (M = 1.35 [1.23, 1.47]), F(1, 111) = 4.87, p = .03, η2p = .04, again replicating the findings from Studies 1 and 3. Additionally, we found a main effect of pen pal emotion, F(2, 222) = 105.15, p b .001, η2p = .49, such that people used a greater breadth of strategies when responding to the sad and the anxious pen pals (Msad = 1.69 [1.55, 1.83] and Manxious = 1.73 [1.61, 1.84] respectively) relative to the happy pen pal (M = .89 [.83, .95]), both ps b .001. There were no significant differences between the sad and anxious pen pals in the breadth of strategies used (p = 1.00). Lastly, we did not find a Condition × Pen Pal Emotion interaction, F(2, 222) = 1.43, p = .24, η2p = .01. 7.3.3. Subjective effectiveness To determine whether participants who relived a past experience of exclusion wrote letters that were rated as more effective at managing their pen pal's emotions than participants who relived a past experience of intellectual failure, we conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on effectiveness ratings. Consistent with the findings from Study 3, we found a main effect of condition, such that those in the exclusion condition wrote letters that were rated as more effective at managing their pen pal's emotions (M = 2.25 [2.07, 2.42]) than those in the control condition (M = 1.89 [1.69, 2.10]), F(1, 111) = 6.81, p = .01, η2p = .06. There was a main effect of pen pal emotion, F(2, 222) = 8.93, p b .001, η2p = .07, such that people wrote letters that were rated as more effective when down-regulating the sad and the anxious pen pals (Msad = 2.21 [2.06, 2.36] and Manxious = 2.15 [1.99, 2.32] respectively) relative to up-regulating the happy pen pal (M = 1.85 [1.65, 2.04]), both ps b .001. There were no significant differences

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

between the sad and anxious letters in effectiveness (p = 1.00). Lastly, we neither predicted nor found a Condition × Pen Pal Emotion interaction, F(2, 222) = 0.35, p = .70, η2p = .003.

71

number and broader repertoire of emotion management strategies in their responses. 7.4. Nonsocial task (alternative uses for common household items)

7.3.4. Likability We examined whether excluded participants' response letters were rated as higher in likability than the letters written by control participants. We conducted a 2(Condition: Exclusion vs. Control) × 3(Pen Pal Emotion: Sad vs. Anxious vs. Happy) mixed ANOVA on likability ratings. Consistent with the findings from Study 3, we found a main effect of condition, such that those in the exclusion condition wrote letters that were rated as more likable (M = 4.45 [4.24, 4.67]) than those in the control condition (M = 4.05 [3.81, 4.30]), F(1, 111) = 6.05, p = .01, η2p = .05. There was no main effect of pen pal emotion, F(2, 222) = 0.16, p = .85, η2p = .001, and no Condition × Pen Pal Emotion interaction, F(2, 222) = 0.33, p = .72, η2p = .003. 7.3.5. Mediation analysis for effectiveness Following Study 3, we examined whether our main effect of condition on subjective effectiveness ratings was mediated by the greater number of attempts and the breadth of strategies that excluded individuals used in their pen pal letters to manage their pen pal's emotions. As shown in Fig. 3, condition significantly predicted ratings of effectiveness, b = 0.35, t(113) = 2.61, p = .01. We next established that condition significantly predicted our hypothesized mediators: the number of attempts made (b = .52, t(113) = 2.43, p = .02) and the breadth of strategies used (b = 0.18, t(113) = 2.21, p = .03) in participants' letters to manage their pen pal's emotions. Then, we established that both our hypothesized mediators significantly predicted ratings of effectiveness: the number of attempts made (b = 0.37, t(113) = 5.34, p b .001) and the breadth of strategies used (b = 0.37, t(101) = 2.05, p = .04). Furthermore, when both the number of attempts made and the breadth of strategies used were entered into the regression simultaneously with condition, the total direct effect of condition on effectiveness was reduced to nonsignificance, b = 0.09, t(113) = 1.07, p = .28. Following Preacher and Hayes (2008) recommendations, we employed bootstrapping procedures with 1000 resamples, using the biased corrected and accelerated approach, to assess whether the number and breadth of strategies significantly mediated the effect of condition on effectiveness. As predicted, when both mediators (the number of attempts made and the breadth of strategies used) were considered together, the 95% confidence interval for the combined indirect effect (b = 0.26 [0.06, 0.46]) did not contain 0, indicating significant mediation at p b .05. In addition, when examining each of the indirect paths separately, the 95% confidence intervals for the number of attempts made (b = 0.19 [0.06, 0.34]) and the breadth of strategies used (b = 0.07 [0.01, 0.21]) both did not contain 0, indicating significant mediation at p b .05 for both variables. These findings provide further evidence suggesting that excluded individuals may be more effective at managing others' emotions, because they tend to use a greater

7.4.1. Number of attempts To determine whether the number of attempts that participants made in the alternative uses task differed as a function of condition (exclusion vs. failure), we conducted an independent-sample t-test on the average number of attempts that participants made per item. Participants who relived a past experience of social exclusion (M = 4.83, SD = 1.99) did not differ from participants who relived a past experience of intellectual failure (M = 4.76, SD = 1.71) on the average number of attempts that participants made per item, t(96) = −0.19, p = .85, d = 0.04. 7.4.2. Breadth of categories To determine whether the breadth of categories that participants generated in the alternative uses task differed as a function of condition (exclusion vs. failure), we conducted an independent-sample t-test on the average breadth of categories listed in their responses. Participants who relived a past experience of social exclusion (M = 4.25, SD = 1.64) did not differ from participants who relived a past experience of intellectual failure (M = 4.07, SD = 1.50) on the average breadth of categories listed in their responses, t(96) = −0.56, p = .58, d = 0.11. 7.5. Comparing the effects of exclusion across social vs. nonsocial tasks 7.5.1. Number of attempts One of our key predictions in Study 4 was that participants who relived a past experience of social exclusion should make a greater number of attempts in their responses relative to participants who relived a past experience of intellectual failure, but that this enhanced performance should be limited to tasks that are social in nature. To test this hypothesis, we first standardized participants' responses for the two different task types (social task vs. nonsocial task). Then, given the focused nature of our predictions, we conducted a One-Way ANOVA with planned contrasts testing the hypothesized pattern: exclusion social task coded 1, exclusion nonsocial task coded 0, failure social task coded − 1, and failure nonsocial task coded 0. The omnibus test, F(3, 209) = 2.22, p = .08 was marginally significant, and the planned contrast testing our predicted pattern was significant, t(206) = 2.49, p = .01. The pattern of standardized means can be seen in Fig. 4. 7.5.2. Breadth of strategies A second key prediction that we had in Study 4 was that participants who relived a past experience of social exclusion should use a greater breadth of categories in their responses relative to participants who relived a past experience of intellectual failure, but that this enhanced performance should be limited to tasks that are social in nature. We

Number of Attempts b = 0.37***

Breadth of Strategies

b = 0.52*

b = 0.18*

b = 0.37*

Condition (0 = Control;

Effectiveness b = 0.35* (0.09)

1 = Exclusion) Fig. 3. Mediational analysis for Study 4. Note. *p b .05, ***p b .001.

Ratings

72

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

0.6 Failure Condition

Number of Attempts (z-score)

0.4

Exclusion Condition

0.2 0 -0.2 -0.4 -0.6 Social Task

Non-social Task

Fig. 4. Study 4: Number of attempts as a function of reliving condition (exclusion vs. failure) and task type (social vs. nonsocial). Error bars indicate 95% confidence intervals.

In addition, we found evidence suggesting that this exclusionenhanced tendency to manage others' emotions appears to be limited social contexts. Specifically, participants who relived a past experience of exclusion generated a greater number and breadth of emotion management strategies in a social task relative to participants who relived a past experience of intellectual failure. By contrast, participants who relived a past experience of exclusion did not differ from participants who relived a past experience of failure in terms of the number or breadth of nonsocial strategies generated in a comparison task (generating creative uses for common household items). These findings are consistent with prior research suggesting that excluded individuals selectively show enhanced performance on tasks when they are social in nature (e.g., DeWall et al., 2008; Gardner, Pickett, & Brewer, 2000; Maner et al., 2007). Altogether, these findings provide evidence suggesting that excluded individuals may exhibit greater management of others' emotions, that these effects appear to be limited to social domains, and that the ability to effectively manage others' emotions can potentially garner greater liking from others. 8. General discussion

thus repeated the analysis above with the breadth of categories that participants made in their responses. The omnibus test was not significant, F(3, 209) = 1.89, p = .10, but the contrast was significant, t(206) = 2.28, p = .02. The pattern of standardized means can be seen in Fig. 5. Taken together, the results of Study 4 presented both a replication of the effects in Study 3 and two valuable extensions. We replicated the findings from Study 3 using a negative, nonsocial control condition, providing further evidence that social exclusion enhances the tendency to spontaneously manage others' emotions. Specifically, participants who relived a past experience of social exclusion wrote response letters that contained a greater number of attempts and used a broader range of strategies to manage their pen pals' emotions relative to participants who relived a past experience of intellectual failure (consistent with Studies 1 and 3). Furthermore, the enhanced emotion management exhibited by excluded participants appeared to promote positive downstream outcomes for the excluded individual, such that their letters were rated as more effective and likable (consistent with Studies 2 and 3). The greater rated effectiveness of excluded participants' letters was mediated both by the greater number of attempts made and the greater breadth of strategies used to manage their pen pal's emotions (consistent with Study 3). By including a negative, nonsocial control condition (reliving a past experience of intellectual failure), we were able to rule out the alternative hypothesis that these effects were driven by negative affect alone.

Breadth of Categories (z-score)

0.6 0.4

Failure Condition

0.2

Exclusion Condition

0 -0.2 -0.4 -0.6 Social Task

Non-social Task

Fig. 5. Study 4: Breadth of categories as a function of reliving condition (exclusion vs. failure) and task type (social vs. nonsocial). Error bars indicate 95% confidence intervals.

The present investigation examined whether social exclusion would enhance performance on the fourth branch of emotional intelligence, the ability to manage the emotions of others. Given that one important function of managing others' emotions is to facilitate social connection (Libbrecht et al., 2014; Lopes et al., 2003, 2004, 2005; Niven et al., 2012), we hypothesized that excluded individuals may marshal this ability, making both more numerous and more varied attempts to regulate the emotions of another person, as a potential strategy for promoting reconnection. In Study 1, we found preliminary evidence suggesting that excluded participants may show greater management of others' emotions. Specifically, we found that excluded participants made a greater number of attempts to cheer up a sad friend and to calm down an angry friend. Furthermore, excluded participants used a greater breadth of strategies in their responding. In Study 2, we conducted a strong face-to-face test of our hypothesis investigating whether excluded individuals would be more effective at managing others' emotions. Specifically, we examined how manipulated belonging need in one participant influenced the emotional state of their interaction partner after a brief interaction. As predicted, we found evidence suggesting that excluded participants were more effective at energizing an interaction partner during a faceto-face coaching session (Study 2). In Studies 3 and 4, we found that excluded participants spontaneously managed their pen pals' emotions in an online pen pal exchange, that they showed this greater emotion management across a broad spectrum of emotional contexts, and that this contributed to likability. Specifically, when responding to multiple different pen pals who were each experiencing different emotional events in their lives, excluded participants wrote response letters that contained a greater number and breadth of strategies for managing these emotions and their letters were rated by objective coders to be more effective at managing their pen pals' emotions (Studies 3 and 4). Furthermore, these effects appear to be limited to social contexts, such that exclusion heightened the number and breadth of emotion management strategies generated in a social task (an online pen pal exchange), but did not heighten the number and breadth of nonsocial strategies generated in a comparison task (listing creative uses of common household items) (Study 4). One adaptive advantage of the fourth branch of emotional intelligence (the ability to manage others' emotions) is that it can serve to facilitate social connection (Libbrecht et al., 2014; Lopes et al., 2003, 2004, 2005; Niven et al., 2012). Given the social benefits of managing others' emotions effectively, we propose that excluded individuals may attempt to manage others' emotions as a way to foster social connection with others. Indeed, we found some preliminary support for this claim, such that excluded participants were actually liked more by

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

their partners after a face-to-face interaction (Study 2) and were rated to be more likable by objective coders (Studies 3 and 4). Taken together, these findings suggest that social exclusion may motivate attempts to manage others' emotions, and that these attempts to manage others' emotions may help facilitate reconnection by garnering greater liking from others. Our findings are especially striking in light of prior research demonstrating the depleting consequences of social exclusion for cognitive and self-regulatory resources (Baumeister et al., 2002, 2005). Given that social exclusion is linked with self-regulatory depletion, one might expect excluded individuals to be especially poor at regulating others' emotions. That excluded individuals in our studies showed superior management of others' emotions highlights the power of our belonging need. That is, even in spite of self-regulatory depletion, excluded individuals were willing to marshal their remaining self-regulatory resources toward fostering social connection (see also DeWall et al., 2008). The present investigation not only contributes to our understanding of when and how people can adaptively respond to social exclusion (e.g., DeWall, 2010; Gardner et al., 2000, 2005; Lakin, Chartrand, & Arkin, 2008; Mead, Baumeister, Stillman, Rawn, & Vohs, 2011; Wang, Zhu, & Shiv, 2012; Williams & Sommer, 1997), it also answers the recent call for researchers to examine the contextual and motivational factors that influence people's skill at emotional intelligence (see Ybarra et al., 2012, 2013). Whereas most prior emotional intelligence research has considered emotional intelligence as a stable, individual-difference construct, in the present research, we demonstrate that people's skill at socially relevant forms of emotional intelligence can vary as a function of their social motivation. Lastly, the present research contributes to the growing literature on interpersonal emotion regulation. Whereas most research on social emotion regulation has explored interpersonal emotion regulation processes in the context of established relationships — particularly attachment relationships (e.g., Bowlby, 1969; Coan, 2011), there has been limited research exploring how these processes can facilitate the formation of novel social bonds (e.g., Beckes et al., 2010). The present research provides further support for the role of interpersonal emotion regulation in the formation of novel social bonds. 9. Limitations and future directions One limitation of the present research is that we relied on a reliving task across all studies as our sole manipulation of social exclusion. We posited that reliving actual past experiences of exclusion may be more personally significant for individuals compared with other exclusion manipulations where individuals are typically excluded by unknown strangers in the experiment (e.g., Gardner et al., 2000; Maner et al., 2007; Williams & Jarvis, 2006). Nevertheless, future research would benefit from employing different exclusion manipulations – particularly manipulations that involve direct (as opposed to relived) experiences of exclusion – to provide convergent evidence for the current findings. Another limitation of the present research is that the use of a positive, social control condition was limited to Study 2. Given that our outcome variables in Study 2 were different than those assessed in Studies 1, 3, and 4 (assessing feelings and outcomes in participants' interaction partners rather than directly assessing participants' spontaneously generated strategies), the most proximal evidence of exclusion-enhanced management of others' emotions remains untested by a manipulation free of the social–nonsocial confound. Future research would benefit from including a positive social control condition when directly assessing the influence of social exclusion on the strategies that participants use to manage others' emotions. A potentially more important limitation of the present research is that in all our studies, we explicitly provided participants with information about others' current (Studies 1, 3 and 4) or desired (Study 2) emotional states. Although providing participants with explicit information about others' emotions allowed for a more controlled test of our

73

hypotheses, future research would benefit from examining whether excluded participants would spontaneously show greater effectiveness at managing others' emotions in more naturalistic contexts where others' emotions may be communicated in more subtle and indirect ways (e.g., vocal tone, facial expressions). Not only would this future direction provide a more naturalistic account of whether excluded individuals will spontaneously manage others' emotions as a strategy to facilitate social connection, but it also has the potential to highlight how social exclusion can simultaneously influence both branch one (accurately perceiving others' emotions) and branch four (managing others' emotions) of emotional intelligence, and the potential interplay between the two branches. In the present research, we found preliminary evidence suggesting that excluded individuals were more effective at managing others' emotions. Future research should explore why excluded individuals may be more effective at managing others' emotions. That is, future research should explore whether excluded individuals are more likely to engage in specific strategies or behaviors that facilitate effective responding. We attempted to explore this question using data from Study 2. Specifically, in Study 2, we videotaped the coaching sessions using a hidden camera and coded participants' interaction behavior during the coaching session to explore which coaching and/or interaction behaviors may be associated with greater regulatory effectiveness. However, none of our coded coaching and interaction behaviors seemed to predict greater coaching effectiveness (see Supplemental analyses). Given that each job candidate may have idiosyncratic responses to different coaching and/or interaction behaviors, we may not have found consistent patterns of coaching or interaction behavior across participants as coaching effectiveness may have resulted from coaches tailoring their responding to what may be most effective for their specific interaction partner at hand. Consistent with this explanation, Bonanno and colleagues have suggested that adaptive emotion regulation does not involve the use of one “superior” emotion-regulation strategy over another, but rather involves the ability to flexibly tailor emotion-regulation strategies to the context at hand (see Bonanno & Burton, 2013). Future research should extend this reasoning to strategies for managing others' emotions, and explore whether excluded individuals would be more likely to tailor their responding to others in their social context. The current work found that exclusion selectively enhanced performance on social tasks, having little impact on nonsocial cognitive tasks. Future research should explore other potential boundary conditions to the present findings. For instance, Maner and colleagues have proposed that social exclusion should only motivate affiliative responding to the extent that others are perceived as realistic sources of reconnection (Maner et al., 2007). As such, we would not expect excluded individuals to exhibit greater management of others' emotions when they do not perceive others as realistic sources of reconnection. For instance, we would not expect our findings to generalize to individuals who tend to chronically anticipate negative social evaluation from others (e.g., socially anxious individuals, chronically lonely individuals, and people who are dispositionally high in the fear of negative evaluation). Such individuals are likely to perceive novel social partners as sources of potential rejection rather than potential sources of reconnection, and as such, might withdraw from the interaction rather than attempt to manage others' emotions. Furthermore, we might not expect excluded individuals to attempt to manage the emotions of the perpetrators who excluded them, as excluded individuals would not perceive their perpetrators as realistic sources of potential reconnection. However, if the exclusion was warranted (e.g., if the excluded individual had committed a transgression) and they perceived that the perpetrator was receptive to repairing the bond, perhaps the excluded individual would be especially likely to redouble their efforts at managing the perpetrator's emotions to repair the relationship. In such circumstances, excluded individuals' enhanced tendency to manage others' emotions may prove especially helpful for repairing their relationships.

74

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75

10. Conclusion The present investigation represents one of the first forays examining the motivational determinants of the emotional intelligence. Specifically, the current work adopted an emotional intelligence framework to explore whether social exclusion would enhance performance in one branch of emotional intelligence, the ability to manage the emotions of others. In four studies, we found evidence suggesting that excluded individuals showed greater management of others' emotions. Furthermore we found preliminary evidence suggesting that this enhanced tendency to manage others' emotions may be associated with positive social outcomes for the excluded individual. Excluded participants were liked more by their interaction partners after a face-to-face coaching session and were rated as more likable in the context of an online pen pal exchange. Taken together, these findings suggest that social exclusion may enhance the ability to manage others' emotions, and that this enhanced ability may evoke greater liking from others. Acknowledgments The authors wish to thank Leah Grodinsky, Samuel Jahangir, John Michael Kelly, Emily Moses, Jonathon Roullard, Kate Shepherd, Hayley Buch, Trevor Harty, Lianna McDaniel, Michaela Choy, Elizabeth Quinn, Danbee Chon, Carrie Phillips, Emily Blim, and Mehreen Itret for their assistance in data collection and coding. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jesp.2015.05.003. References Aron, A., Aron, E.N., & Smollan, D. (1992). Inclusion of other in the self scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63, 596–612. Ashkanasy, N.M., & Humphrey, R.H. (2014). Leadership and emotion: A multilevel perspective. In D.V. Day (Ed.), The Oxford handbook of leadership and organizations (pp. 783–804). New York, NY: Oxford University Press. http://dx.doi.org/10.1093/ oxfordhb/9780199755615.013.038. Baumeister, R.F., DeWall, C., Ciarocco, N.J., & Twenge, J.M. (2005). Social exclusion impairs self-regulation. Journal of Personality and Social Psychology, 88(4), 589–604. http://dx. doi.org/10.1037/0022-3514.88.4.589. Baumeister, R.F., & Leary, M.R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529. http://dx.doi.org/10.1037/0033-2909.117.3.497. Baumeister, R.F., Twenge, J.M., & Nuss, C.K. (2002). Effects of social exclusion on cognitive processes: Anticipated aloneness reduces intelligent thought. Journal of Personality and Social Psychology, 83(4), 817–827. http://dx.doi.org/10.1037/0022-3514.83.4.817. Beckes, L., Simpson, J.A., & Erickson, A. (2010). Of snakes and succor: Learning secure attachment associations with novel faces via negative stimulus pairings. Psychological Science, 21(5), 721–728. http://dx.doi.org/10.1177/0956797610368061. Bernstein, M.J., Sacco, D.F., Brown, C.M., Young, S.G., & Claypool, H.M. (2010). A preference for genuine smiles following social exclusion. Journal of Experimental Social Psychology, 46(1), 196–199. http://dx.doi.org/10.1016/j.jesp.2009.08.010. Bernstein, M.J., Young, S.G., Brown, C.M., Sacco, D.F., & Claypool, H.M. (2008). Adaptive responses to social exclusion: Social rejection improves detection of real and fake smiles. Psychological Science, 19(10), 981–983. http://dx.doi.org/10.1111/j.14679280.2008.02187.x. Bonanno, G.A., & Burton, C.L. (2013). Regulatory flexibility: An individual differences perspective on coping and emotion regulation. Perspectives on Psychological Science, 8(6), 591–612. http://dx.doi.org/10.1177/1745691613504116. Bonanno, G.A., Papa, A., Lalande, K., Westphal, M., & Coifman, K. (2004). The importance of being flexible: The ability to both enhance and suppress emotional expression predicts long-term adjustment. Psychological Science, 15(7), 482–487. http://dx.doi.org/ 10.1111/j.0956-7976.2004.00705.x. Bowlby, J. (1969). Attachment and loss, Vol. 1: Attachment (2nd ed.). New York, NY: Basic Books. Brackett, M.A., & Mayer, J.D. (2003). Convergent, discriminant, and incremental validity of competing measures of emotional intelligence. Personality and Social Psychology Bulletin, 29(9), 1147–1158. http://dx.doi.org/10.1177/0146167203254596. Brackett, M.A., Rivers, S.E., Shiffman, S., Lerner, N., & Salovey, P. (2006). Relating emotional abilities to social functioning: A comparison of self-report and performance measures of emotional intelligence. Journal of Personality and Social Psychology, 91(4), 780–795. http://dx.doi.org/10.1037/0022-3514.91.4.780.

Caporael, L.R. (2001). Parts and wholes: The evolutionary importance of groups. In C. Sedikides, & M.B. Brewer (Eds.), Individual self, relational self, collective self (pp. 241–258). Philadelphia: Psychology Press. Carnevale, P.J., & Isen, A.M. (1986). The influence of positive affect and visual access on the discovery of integrative solutions in bilateral negotiation. Organizational Behavior and Human Decision Processes, 37(1), 1–13. http://dx.doi.org/10.1016/ 0749-5978(86)90041-5. Coan, J.A. (2011). The social regulation of emotion. In J. Decety, & J.T. Cacioppo (Eds.), Oxford handbook of social neuroscience (pp. 614–623). New York: Oxford University Press. Côté, S. (2014). Emotional intelligence in organizations. Annual Review of Organizational Psychology and Organizational Behavior, 1(1), 459–488. http://dx.doi.org/10.1146/ annurev-orgpsych-031413-091233. Côté, S., DeCelles, K.A., McCarthy, J.M., Van Kleef, G.A., & Hideg, I. (2011). The Jekyll and Hyde of emotional intelligence: Emotion-regulation knowledge facilitates both prosocial and interpersonally deviant behavior. Psychological Science, 22(8), 1073–1080. http://dx.doi.org/10.1177/0956797611416251. DeWall, C. (2010). Forming a basis for acceptance: Excluded people form attitudes to agree with potential affiliates. Social Influence, 5(4), 245–260. http://dx.doi.org/10. 1080/15534511003783536. DeWall, C., Baumeister, R.F., & Vohs, K.D. (2008). Satiated with belongingness? Effects of acceptance, rejection, and task framing on self-regulatory performance. Journal of Personality and Social Psychology, 95(6), 1367–1382. http://dx.doi.org/10.1037/ a0012632. Eisenberger, N.I., Lieberman, M.D., & Williams, K.D. (2003). Does exclusion hurt? An fMRI study of social exclusion. Science, 302(5643), 290–292. http://dx.doi.org/10.1126/science.1089134. Elfenbein, H., Marsh, A.A., & Ambady, N. (2002). Emotional intelligence and the recognition of emotion from facial expressions. In L. Barrett, & P. Salovey (Eds.), The wisdom in feeling: Psychological processes in emotional intelligence (pp. 37–59). New York, NY: Guilford Press. Galinsky, A.D., Mussweiler, T., & Medvec, V.H. (2002). Disconnecting outcomes and evaluations: The role of negotiator focus. Journal of Personality and Social Psychology, 83(5), 1131–1140. http://dx.doi.org/10.1037/0022-3514.83.5.1131. Garczynski, A.M., & Brown, C.M. (2014). Temporal perspective moderates self-reported responses to social exclusion. Journal of Experimental Social Psychology, 53, 40–50. http://dx.doi.org/10.1016/j.jesp.2014.02.003. Gardner, W., Pickett, C., & Brewer, M. (2000). Social exclusion and selective memory: How the need to belong influences memory for social events. Personality and Social Psychology Bulletin, 26, 486–496. http://dx.doi.org/10.1177/0146167200266007. Gardner, W., Pickett, C., Jefferis, V., & Knowles, M. (2005). On the outside looking in: Loneliness and social monitoring. Personality and Social Psychology Bulletin, 31, 1549–1560. http://dx.doi.org/10.1177/0146167205277208. Guilford, J. (1967). The nature of human intelligence. New York, NY, US: McGraw-Hill. Humphrey, R.H. (2002). The many faces of emotional leadership. The Leadership Quarterly, 13(5), 493–504. http://dx.doi.org/10.1016/S1048-9843(02)00140-6. Lakin, J.L., Chartrand, T.L., & Arkin, R.M. (2008). I am too just like you: Nonconscious mimicry as an automatic behavioral response to social exclusion. Psychological Science, 19(8), 816–822. http://dx.doi.org/10.1111/j.1467-9280.2008.02162.x. Lam, C.B., & McBride-Chang, C.A. (2007). Resilience in young adulthood: The moderating influences of gender-related personality traits and coping flexibility. Sex Roles, 56(3–4), 159–172. http://dx.doi.org/10.1007/s11199-006-9159-z. Lerner, J.S., & Keltner, D. (2001). Fear, anger, and risk. Journal of Personality and Social Psychology, 81(1), 146–159. http://dx.doi.org/10.1037/0022-3514.81.1.146. Libbrecht, N., Lievens, F., Carette, B., & Côté, S. (2014). Emotional intelligence predicts success in medical school. Emotion, 14(1), 64–73. http://dx.doi.org/10.1037/a0034392. Lopes, P.N., Brackett, M.A., Nezlek, J.B., Schütz, A., Sellin, I., & Salovey, P. (2004). Emotional intelligence and social interaction. Personality and Social Psychology Bulletin, 30(8), 1018–1034. http://dx.doi.org/10.1177/0146167204264762. Lopes, P.N., Nezlek, J.B., Extremera, N., Hertel, J., Fernández‐Berrocal, P., Schütz, A., et al. (2011). Emotion regulation and the quality of social interaction: Does the ability to evaluate emotional situations and identify effective responses matter? Journal of Personality, 79(2), 429–467. http://dx.doi.org/10.1111/j.1467-6494. 2010.00689.x. Lopes, P.N., Salovey, P., Côté, S., Beers, M., & Petty, R.E. (2005). Emotion regulation abilities and the quality of social interaction. Emotion, 5(1), 113–118. http://dx.doi.org/10. 1037/1528-3542.5.1.113. Lopes, P.N., Salovey, P., & Straus, R. (2003). Emotional intelligence, personality, and the perceived quality of social relationships. Personality and Individual Differences, 35(3), 641–658. http://dx.doi.org/10.1016/S0191-8869(02)00242-8. Lougheed, J.P., & Hollenstein, T. (2012). A limited repertoire of emotion regulation strategies is associated with internalizing problems in adolescence. Social Development, 21(4), 704–721. http://dx.doi.org/10.1111/j.1467-9507.2012.00663.x. MacDonald, G., & Leary, M.R. (2005). Why does social exclusion hurt? The relationship between social and physical pain. Psychological Bulletin, 131(2), 202–223. http://dx. doi.org/10.1037/0033-2909.131.2.202. Maner, J.K., DeWall, C., Baumeister, R.F., & Schaller, M. (2007). Does social exclusion motivate interpersonal reconnection? Resolving the ‘porcupine problem’. Journal of Personality and Social Psychology, 92(1), 42–55. http://dx.doi.org/10.1037/00223514.92.1.42. Maslow, A.H. (1968). Toward a psychology of being. New York: Van Nostrand. Mayer, J.D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey, D.J. Sluyter, P. Salovey, & D.J. Sluyter (Eds.), Emotional development and emotional intelligence: Educational implications (pp. 3–34). New York, NY: Basic Books. Mayer, J.D., Salovey, P., & Caruso, D.R. (2002). Mayer-Salovey-Caruso Emotional Intelligence Test: MSCEIT. Item booklet. MHS.

E.O. Cheung, W.L. Gardner / Journal of Experimental Social Psychology 60 (2015) 59–75 Mayer, J.D., Salovey, P., & Caruso, D.R. (2004). Emotional intelligence: Theory, findings, and implications. Psychological Inquiry, 15(3), 197–215. http://dx.doi.org/10.1207/ s15327965pli1503_02. McArthur, L.Z., & Baron, R.M. (1983). Toward an ecological theory of social perception. Psychological Review, 90(3), 215–238. http://dx.doi.org/10.1037/0033-295X.90.3.215. McColl-Kennedy, J.R., & Anderson, R.D. (2002). Impact of leadership style and emotions on subordinate performance. The Leadership Quarterly, 13(5), 545–559. http://dx. doi.org/10.1016/S1048-9843(02)00143-1. Mead, N.L., Baumeister, R.F., Stillman, T.F., Rawn, C.D., & Vohs, K.D. (2011). Social exclusion causes people to spend and consume strategically in the service of affiliation. Journal of Consumer Research, 37(5), 902–919. http://dx.doi.org/10.1086/656667. Molden, D.C., & Maner, J.K. (2013). How and when exclusion motivates social reconnection. In C. DeWall (Ed.), The Oxford handbook of social exclusion (pp. 121–131). New York, NY, US: Oxford University Press. Muraven, M., Shmueli, D., & Burkley, E. (2006). Conserving self-control strength. Journal of Personality and Social Psychology, 91(3), 524–537. http://dx.doi.org/10.1037/00223514.91.3.524. Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive Psychology, 9(3), 353–383. http://dx.doi.org/10.1016/0010-0285(77)90012-3. Niven, K., Holman, D., & Totterdell, P. (2012). How to win friendship and trust by influencing people's feelings: An investigation of interpersonal affect regulation and the quality of relationships. Human Relations, 65(6), 777–805. http://dx.doi.org/10.1177/ 0018726712439909. Niven, K., Totterdell, P., & Holman, D. (2009). A classification of controlled interpersonal affect regulation strategies. Emotion, 9(4), 498–509. http://dx.doi.org/10.1037/a0015962. Niven, K., Totterdell, P., Stride, C.B., & Holman, D. (2011). Emotion Regulation of Others and Self (EROS): The development and validation of a new individual difference measure. Current Psychology: A Journal for Diverse Perspectives on Diverse Psychological Issues, 30(1), 53–73. http://dx.doi.org/10.1007/s12144-011-9099-9. Orcutt, H.K., Bonanno, G.A., Hannan, S.M., & Miron, L.R. (2014). Prospective trajectories of posttraumatic stress in college women following a campus mass shooting. Journal of Traumatic Stress, 27(3), 249–256. http://dx.doi.org/10.1002/jts.21914. Pescosolido, A.T. (2002). Emergent leaders as managers of group emotion. The Leadership Quarterly, 13(5), 583–599. http://dx.doi.org/10.1016/S1048-9843(02)00145-5. Pickett, C.L., & Gardner, W.L. (2005). The social monitoring system: Enhanced sensitivity to social cues as an adaptive response to social exclusion. In K.D. Williams, J.P. Forgas, W. von Hippel, K.D. Williams, J.P. Forgas, & W. von Hippel (Eds.), The social outcast: Ostracism, social exclusion, rejection, and bullying (pp. 213–226). New York, NY, US: Psychology Press. Pickett, C.L., Gardner, W.L., & Knowles, M. (2004). Getting a cue: The need to belong and enhanced sensitivity to social cues. Personality and Social Psychology Bulletin, 30, 1095–1107. Poon, K.T., Chen, Z., & DeWall, C.N. (2013). Feeling entitled to more ostracism increases dishonest behavior. Personality and Social Psychology Bulletin, 39(9), 1227–1239. http://dx. doi.org/10.1177/0146167213493187. Preacher, K.J., & Hayes, A.F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. http://dx.doi.org/10.3758/BRM.40.3.879. Quirin, M., Kazén, M., & Kuhl, J. (2009). When nonsense sounds happy or helpless: The Implicit Positive and Negative Affect Test (IPANAT). Journal of Personality and Social Psychology, 97, 500–516.

75

Sacco, D.F., Wirth, J.H., Hugenberg, K., Chen, Z., & Williams, K.D. (2011). The world in black and white: Ostracism enhances the categorical perception of social information. Journal of Experimental Social Psychology, 47(4), 836–842. http://dx.doi.org/10.1016/ j.jesp.2011.03.001. Schutte, N.S., Malouff, J.M., Hall, L.E., Haggerty, D.J., Cooper, J.T., Golden, C.J., et al. (1998). Development and validation of a measure of emotional intelligence. Personality and Individual Differences, 25(2), 167–177. http://dx.doi.org/10.1016/S0191-8869(98) 00001-4. Silvera, D.H., Martinussen, M., & Dahl, T.I. (2001). The Tromsø Social Intelligence Scale, a self-report measure of social intelligence. Scandinavian Journal of Psychology, 42(4), 313–319. http://dx.doi.org/10.1111/1467-9450.00242. Stroebe, M., & Schut, H. (1999). The dual process model of coping with bereavement: Rationale and description. Death Studies, 23(3), 197–224. http://dx.doi.org/10.1080/ 074811899201046. Stroebe, M., & Schut, H. (2010). The dual process model of coping with bereavement: A decade on. Omega: Journal of Death and Dying, 61(4), 273–289. http://dx.doi.org/10. 2190/OM.61.4.b. Stukas, A.A., & Cumming, G. (2014). Interpreting effect sizes: Toward a quantitative cumulative social psychology7, 711–722. http://dx.doi.org/10.1002/ejsp.2019. Tamir, M., & Ford, B. Q. (2012). When feeling bad is expected to be good: Emotion regulation and outcome expectancies in social conflicts. Emotion, 12(4), 807–816. http:// dx.doi.org/10.1037/a0024443. Tsai, W., Chen, H., & Cheng, J. (2009). Employee positive moods as a mediator linking transformational leadership and employee work outcomes. International Journal of Human Resource Management, 20(1), 206–219. http://dx.doi.org/10.1080/ 09585190802528714. Twenge, J.M., Baumeister, R.F., Tice, D.M., & Stucke, T.S. (2001). If you can't join them, beat them: Effects of social exclusion on aggressive behavior. Journal of Personality and Social Psychology, 81(6), 1058–1069. http://dx.doi.org/10.1037/ 0022-3514.81.6.1058. Twenge, J.M., Catanese, K.R., & Baumeister, R.F. (2002). Social exclusion causes self-defeating behavior. Journal of Personality and Social Psychology, 83(3), 606–615. http://dx.doi.org/ 10.1037/0022-3514.83.3.606. Wang, J., Zhu, R., & Shiv, B. (2012). The lonely consumer: Loner or conformer? Journal of Consumer Research, 38(6), 1116–1128. http://dx.doi.org/10.1086/661552. Williams, K., & Jarvis, B. (2006). Cyberball: A program for use in research on interpersonal ostracism and inclusion. Behavior Research Methods, 38, 174–180. Williams, K.D., & Sommer, K.L. (1997). Social ostracism by coworkers: Does rejection lead to loafing or compensation? Personality and Social Psychology Bulletin, 23(7), 693–706. http://dx.doi.org/10.1177/0146167297237003. Ybarra, O., Kross, E., Lee, D., Zhao, Y., Dougherty, A., & Sanchez-Burks, J. (2013). Toward a more contextual, psychological, and dynamic model of emotional intelligence. In A.B. Bakker (Ed.), Advances in positive organizational psychology (pp. 167–187). Bingley, UK: Emerald Group Publishing. http://dx.doi.org/10. 1108/S2046-410X(2013)0000001010. Ybarra, O., Rees, L., Kross, E., & Sanchez-Burks, J. (2012). Emotional intelligence. In K.S. Cameron, & G.M. Spreitzer (Eds.), The Oxford handbook of positive organizational scholarship (pp. 201–214). New York, NY: Oxford University Press.