RELATIONAL AND CONTEXTUAL ANTECEDENTS OF EMOTION REGULATION: INTEGRATING THE EASI MODEL INTO EXISTING EMOTIONAL LABOR MODELS By Sergio Miguel Marquez A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology—Master of Arts 2020 ABSTRACT RELATIONAL AND CONTEXTUAL ANTECEDENTS OF EMOTION REGULATION: INTEGRATING THE EASI MODEL INTO EXISTING EMOTIONAL LABOR MODELS By Sergio Miguel Marquez Emotional labor research within organizational psychology has largely benefitted by adopting process models of emotion regulation into emotional labor models. However, relational, contextual moderators and antecedents of the emotional labor process are sorely missing from extant research. This study sought to further explicate the role of these factors on emotion regulation using the Emotions as Social Information (EASI) process model. Using a simulated negotiation scenario, this study investigated the role of information processing (mediation by relative power, manipulated through number of opponents engaged) and opponent’s affect (angry vs. happy) on subsequent emotion regulation strategy adopted. Utilizing a sample of 326 undergraduate psychology students, results suggest that participants experienced emotion contagion and drew meaningful inferences about own performance from their opponent’s affect. However, only affective reactions -- not performance inference -- were related to participants’ emotion regulation. Further, relative power did not mediate the relationship between number of opponents and information processing. Exploratory analyses instead suggest information processing, epistemic motivation, experienced power, and relative power moderate the relationship between affective reactions and performance inference and subsequent emotion regulation. Theoretical and practical implications of results are discussed, along with future suggestions for following research. Copyright by SERGIO MIGUEL MARQUEZ 2020 ACKNOWLEDGEMENTS This thesis was completed thanks to the support from several invaluable family, friends, and mentors. First, I would like to thank my advisor, Dr. Daisy Chang, for her outstanding mentorship and guidance during this process. Daisy has dedicated a substantial number of hours reviewing my work and expecting the highest standards and quality, all while remaining positive and pleasant to work with. This thesis could not have been completed without her patience and trust in me and my work throughout. I would also like to thank Dr. Chris Nye and Dr. Brent Scott, who have been an excellent committee, further solidifying my developing research skills. I am also indebted to my mother and family, who always supply the love and support necessary to carry me through the difficulties of graduate school. They are the force that continue driving me forward, as everything I do is for them. Lastly, thank you to my several closest friends. Together, they have helped me either stay connected to my home in Miami or helped me find a home in Michigan. Ultimately, and most importantly, they have kept me sane. From spontaneous texts and phone calls to late night magic the gathering sessions, these shared experiences have been central to my amazing experience thus far in graduate school. iv TABLE OF CONTENTS LIST OF TABLES ........................................................................................................................ vii LIST OF FIGURES ..................................................................................................................... viii INTRODUCTION .......................................................................................................................... 2 Emotion Regulation ..................................................................................................................... 7 Emotional Labor .......................................................................................................................... 9 Emotional Labor as Emotion Regulation .................................................................................. 11 Antecedents of Emotion Regulation ......................................................................................... 13 The Emotions as Social Informational Framework (EASI Model) ........................................... 15 Affective Pathway ..................................................................................................................... 17 Inferential Pathway ................................................................................................................... 21 Moderators of the EASI model ................................................................................................. 24 Power ......................................................................................................................................... 26 METHOD ..................................................................................................................................... 36 Participants ................................................................................................................................ 36 Negotiation Task ....................................................................................................................... 37 Pilot Study ................................................................................................................................. 41 Measures .................................................................................................................................... 42 Perceptions of power. ............................................................................................................ 42 Perception of target emotion. ................................................................................................ 42 Actor affective reactions. ....................................................................................................... 43 Information processing. ......................................................................................................... 44 Need for cognitive closure (NFCC). ..................................................................................... 45 Performance inference. .......................................................................................................... 45 Antecedent-focused emotion regulation strategies. ............................................................... 46 Response-focused emotion regulation strategies. .................................................................. 46 Expression of naturally felt emotions (authentic display). .................................................... 46 Manipulation checks. ............................................................................................................. 46 Data Analysis ............................................................................................................................ 47 RESULTS ..................................................................................................................................... 48 Manipulation Checks ................................................................................................................. 48 Test of Hypotheses .................................................................................................................... 51 Exploratory Analyses. ............................................................................................................... 55 Positive Affect. ...................................................................................................................... 55 Information processing and epistemic motivation. ........................................................... 56 Experienced and relative power. ....................................................................................... 56 Negative Affect. .................................................................................................................... 57 Information processing and epistemic motivation. ........................................................... 57 Experienced and relative power. ....................................................................................... 58 Performance Inference. .......................................................................................................... 58 v Information processing and epistemic motivation. ........................................................... 58 Experienced and relative power. ....................................................................................... 59 Response-focused Emotion Regulation. ................................................................................ 60 Information processing and epistemic motivation. ........................................................... 60 Experienced and relative power. ....................................................................................... 61 Antecedent-focused Emotion Regulation. ............................................................................. 62 Information processing and epistemic motivation. ........................................................... 62 Experienced and relative power. ....................................................................................... 64 Authentic Display. ................................................................................................................. 65 Information processing and epistemic motivation. ........................................................... 65 Experienced and relative power. ....................................................................................... 66 DISCUSSION ............................................................................................................................... 68 Emotion Regulation and the EASI Framework ......................................................................... 69 Information Processing and Emotion Regulation ..................................................................... 71 Relative and Experienced Power and Emotion Regulation....................................................... 75 Theoretical and Practical Implications ...................................................................................... 76 Limitations and Future Directions ............................................................................................. 78 APPENDICES .............................................................................................................................. 80 APPENDIX A: Tables and Figures ........................................................................................... 81 APPENDIX B: Outline of Negotiation Scenario .................................................................... 104 APPENDIX C: Round 1, 2, and 3 Survey ............................................................................... 107 APPENDIX D: Final Round Survey ....................................................................................... 108 REFERENCES ........................................................................................................................... 111 vi LIST OF TABLES Table 1. Means, standard deviations, and bivariate correlations for study predictors. ................. 81 Table 2. Cells means for opponent affect and number of opponents manipulation. .................... 84 Table 3. Results for path model analysis investigating opponent’s affect on cognitive and affective factors and subsequent emotion-regulation. .................................................................. 85 Table 4. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s positive affect. ........................................... 86 Table 5. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s negative affect. .......................................... 87 Table 6. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s performance inference. ............................. 88 Table 7. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s response-focused emotion regulation ....... 89 Table 8. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s antecedent-focused emotion regulation .... 90 Table 9. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s authentic display. ...................................... 91 vii LIST OF FIGURES Figure 1. Proposed model ............................................................................................................. 92 Figure 2. The moderating role of information processing on the relationship between opponent affect participant’s positive affective state. .................................................................................. 93 Figure 3. The moderating role of information processing on the relationship between opponent affect participant’s negative affective state. .................................................................................. 94 Figure 4. The moderating role of information processing on the relationship between opponent affect participant’s performance inference ................................................................................... 95 Figure 5. The moderating effect of information processing on the relationship between opponent affect and participant’s positive affect. ......................................................................................... 96 Figure 6. The moderating effect of information processing on the relationship between opponent affect and participant’s negative affect. ........................................................................................ 97 Figure 7. The moderating effect of information processing on the relationship between opponent’s affect and their performance inference. ..................................................................... 98 Figure 8. The moderating effect of experienced power on the relationship between opponent’s affect and participant’s performance inference. ............................................................................ 99 Figure 9. The moderating effect of information processing on the relationship between negative affect and antecedent-focused emotion regulation. .................................................................... 100 Figure 10. The moderating effect of epistemic motivation on the relationship between participant’s performance inference and participant’s antecedent-focused emotion regulation. 101 Figure 11. The moderating effect of experienced power on the relationship between participant’s performance inference and participant’s antecedent-focused emotion regulation. .................... 102 Figure 12.The moderating effect of relative power on the relationship between participant’s performance inference and participant’s antecedent-focused emotion regulation. .................... 103 viii INTRODUCTION Recently, both academic and public domains are becoming increasingly aware of research concerning employee’s emotion regulation. Emotion regulation studies have received media coverage in outlets such as The Atlantic (“How ‘Service With a Smile’ Takes a Toll on Women”; Wingfield, 2016) and The Huffington Post (e.g. “Navigating Emotional Labor at Work”; McWilliams, 2017), along with a recent boon in both theoretical and practical advances by organizational researchers. The ethical and practical implications of jobs that mandate display rules fuel these investigations. Display rules, which constrain the range of emotions able to be expressed by employees, are prescribed by organizations in the hope that these rules will increase employee performance. The most recognized and common iteration of a display rule is “service with a smile,” which requires the display of positive emotions and the suppression of negative emotions. Conversely, display rules could also be neutral or negative (i.e. tax collectors or leaders may display negative affect to show frustration or lack of satisfaction toward targets, or judges may suppress all forms of emotional display to seem impartial). However, regardless the nature of the display rule, emotion regulation is used by employees to address the discrepancy between felt emotions and emotional display rules. This process requires either deliberate alignment of one’s emotions with display rules or the suppression of one’s emotions (Hochschild, 1983). Thus, emotion regulation is prevalent in any context that requires individuals to regulate emotional expressions in accordance with display rules and organizational goals. One reason for emotion regulation’s increased attention is due to the numerous outcomes across attitudinal, behavioral, and well-being domains associated with emotion regulation. Outcomes such as turnover intentions (Chau, Dahling, Levy, & Diefendorff, 2009), performance 2 (Richards & Gross, 2000), and organizational citizenship behaviors (Grant, 2013) have all been linked to emotion regulation. Additionally, emotion regulation has been tied to health outcomes such as increased cortisol (Kotsou, Nelis, Grégoire, & Milkolajczak, 2011), exhaustion (Hülsheger & Schewe, 2011; Kammeyer-Mueller, Rubenstein, Long, Buckman, Zhang, & Halvorsen‐Ganepola, 2013), and insomnia (Wagner, Barnes, & Scott, 2014). While these outcomes suggest urgency for organizational researchers to better understand the emotion regulation process, emotion regulation research has paid skewed attention toward outcomes. Subsequently, the relative nascence of this literature still leaves much to be desired. As noted by Grandey and Melloy’s (2017) review, factors such as emotional events (e.g. goal achievement, interpersonal behavior), role expectations (e.g., relational moderators such as power or intimacy), and target reactions (e.g. anger, happiness) remain largely uninvestigated by researchers. In other words, there is less literature available concerning the antecedents of emotion regulation. For instance, Grandey and Melloy (2017) just recently introduced relational characteristics to their emotional labor/emotion regulation model, citing empirical studies (e.g. Diefendorff & Greguras, 2009) which suggest that relational characteristics are likely to affect how emotion regulation functions differently across targets and contexts. The following paragraphs will highlight notable emotion regulation antecedents that have received either little or no empirical attention. One relational antecedent that has received some research attention is relative power. Diefendorff, Morehart, and Gabriel (2010) found that relational power differences affect the perceived degree of emotion regulation (e.g., suppression, de-amplification, amplification) required by the employee. They also found an interaction between the employee’s displayed emotion (happiness vs. anger) and relative power, such that focal employee displays of anger 3 were perceived as less appropriate for high power targets than low power targets. However, the question as to how perceived power differences may affect different emotion regulation strategy usage by employees remains unanswered. This question is relevant for jobs where the relational characteristics between the employee and the target may be significant, such as service providers interacting with their customers or doctors interacting with their patients. Accordingly, there are different bases for power (French & Raven, 1959). For instance, power may be derived by someone’s position within an organization or their access to valuable resources (e.g. expertise). Cues that signal power may also shift depending on context. Many jobs require simultaneous engagement with multiple targets (e.g. teachers, chief executives, clergy) rather than engagement on a one-to-one basis. Perceived relative power may distinguish emotion regulation between interacting with a single target compared to with multiple targets. Thompson and Peterson (1996) argue that there is a team-efficacy effect, whereby both teams and individuals perceive teams as having the advantage regarding resources and information processing over individual members. This preconception may theoretically lead to perceived power differences between parties within negotiations (Cohen & Thompson, 2011), such that individuals may differentially regulate their emotions depending on whether they are interacting with a team opposed to an individual. This contextual factor (i.e., number of targets) is relevant to many jobs and tasks, yet to be investigated within the emotion regulation literature. On a similar note, there are many other diverse job contexts that require regulating one’s emotions that are not addressed by the emotion regulation literature. Grandey and Melloy (2017) and another previous emotional labor review by Grandey and Gabriel (2015) both note how emotion regulation studies are investigated in non-diverse (most often service) contexts. Many studies use call centers, which feature short, scripted, low intensity interactions, and often control 4 for several different aspects of role and context to better understand emotion regulation directly (i.e., Tschan, Rochat, & Zapf, 2005). Further investigation is required to understanding meaningful differences in emotion regulation across occupations and roles. Jobs such as sales require negotiation, a task not often investigated within the emotion regulation literature, despite requiring significant emotion regulation (Van Kleef, 2009). Lastly, target’s emotional display has also been identified as a key antecedent of subsequent emotion regulation. Experimental studies using call center simulations (Gabriel & Diefendorff, 2015; Goldberg & Grandey, 2007) showed that negative emotional events (i.e. customer mistreatment) cause increased emotion regulation, and that this effect is stronger for response-focused strategies such as emotional suppression compared to antecedent-focused strategies (Sliter, Jex, Woldford, & McInnerney, 2010). On the other hand, positive emotional events tend to decrease the need for response-focused strategies but are positively related to subsequent antecedent-focused strategies through better felt mood (Totterdell & Holman, 2003). Grandey and Melloy (2017) note that while the relationship between emotional events and response-focused strategies (e.g., suppression) is more investigated, thus better understood, why employees engage in antecedent-focused regulation strategies with others during emotional events is theoretically unclear. Dissecting this relationship will help emotion regulation researchers understand why individuals differentially regulate their responses to certain events, thereby theoretically expanding how organizational researchers view emotion regulation as a process. However, there is no theoretical framework or mechanism commonly applied by emotional labor researchers to explain why employees differ in emotion regulation in response to target affect, leaving this question unanswered. 5 The current study aims to investigate interaction- and context-related antecedents of emotion regulation through the lens of the Emotions as Social Information (EASI) model. Specifically, this proposal seeks to investigate whether emotion regulation strategies within the context of a negotiation scenario differ based on interaction with multiple targets compared to a single target, and whether target’s emotional display is positive or negative. The EASI model (Van Kleef, 2009) serves as the theoretical foundation for the current research because it proposes that emotions serve to disambiguate social interactions by conveying important social information. A target’s emotional display triggers two pathways: an affective pathway and an inferential pathway. The current study proposes that the displayed target emotion triggers the activation of the affective and inferential pathways in individuals, such that individuals will regulate their emotional display during an interaction according to their felt emotions (affective pathway) and their perceptions regarding their performance (inferential pathway). Tying these pathways as the mediating process between an emotional event (i.e., customer interaction) and subsequent emotion regulation will help explicate why individuals differ their emotion regulation strategies depending on whether an encounter was positive or negative. This study will also investigate how perceived power difference associated with an actor (i.e. an individual who emotionally regulates when interacting with the targets) interacting with multiple targets (i.e. individuals who actors regulate their emotions in response to the actor’s emotion display). simultaneously (vs. a single target) may have implications for their emotion regulation. The EASI framework offers insights for how relative power difference between the actor and the target may make the inferential versus the affective pathway differentially salient (Van Kleef, 2009; 2010). The proposed model is represented by Figure 1. Through the investigation of these questions, this study seeks to contribute to the literature in several ways. 6 First, this study aims to better understand how target’s emotional display affects subsequent employee’s emotion regulation, a novel application of the EASI model within the emotion regulation literature. Second, this study introduces two new mediating mechanisms (i.e., affective and inferential pathways) by which employees react to a target’s affective display. This responds to a call by Grandey and Melloy (2017) for research explaining why positive or negative customer events result in differential regulation of emotions. Third, this study offers a potential explanation as to why emotion regulation responses to a target’s affective display may differ according to perceived relative power differences. Although an included variable within emotion regulation models (Grandey & Gabriel, 2015; Grandey & Melloy, 2017), perceived power difference has received limited empirical attention. Lastly, this study introduces a new contextual domain for emotional labor: negotiation. Since the basis of power can vary depending on one’s context, it is problematic that both power (a relational antecedent of emotional regulation) and contextual moderators (such as a negotiation scenario) have been largely uninvestigated within the emotional labor literature. Many emotional labor studies control for factors independent of emotional labor, effectively removing the effects of context from the investigation of emotional labor. This study seeks to partly address this gap by further investigating emotion regulation within the context of a negotiation, a novel context that often features emotional labor. Emotion Regulation Emotion regulation refers to “the process by which individuals influence which emotions they have, when they have them, and how they experience and express them (Gross, 1998).” The emotion regulation process is inherently dynamic since individuals are posited to modulate the expression and experience of emotions from start to finish. There are many ways that individuals 7 can effortfully manage their emotions across physiological, behavioral, and cognitive domains, meaning that there are then different strategies that are more or less useful depending on when and how individuals wish to control their emotions (Gross, 1998). Emotion regulation strategies are broadly divided into two camps. Antecedent-focused strategies seek to alter one’s emotions before the emotion is generated (Gross, 1998). Within this camp there are a number of different strategies. Situation selection and modification strategies aim to alter the emotional environment an individual navigates. Situation selection refers to whether an individual will approach or avoid a situation depending on anticipated emotional responses (Gross, 1998). However, this requires significant foresight and potentially previous experience on an individual’s part, so this strategy may not always be an option. In this case, situation modification, which refers to altering a situation as a means to change the emotional impact (Gross, 1998), may be more appropriate. Attentional deployment is another strategy, which refers to the changing of one’s attentional focus away or toward an emotional stimulus (Gross, 1998). These strategies all collectively serve to help regulate emotions through the specific curation of the emotion information gathered. The last of the antecedent-focused strategies is cognitive change. Cognitive change differs from the previous strategies mentioned because it involves the manipulation of emotional perceptions and the meaning attached to the emotional stimulus. Reappraisal is one notable form of this process. Reappraisal involves the evaluation of an emotionally stimulating event in a less or non-emotionally impactful manner (Gross, 1998). Therefore, changing the meaning of an emotional event has been found to be a useful antecedent-focused strategy for managing one’s emotion regulation. The second camp of strategies is known as response-focused strategies. Per the name, these strategies involve the manipulation of one’s experiential, physiological or behavioral 8 reaction/responses to an emotional stimulus (Gross, 1998). While activities such as exercise, food and alcohol intake, or smoking cigarettes may help alter physiological and experiential emotional reactions, the most common and studied form of emotion regulation is behavioral. The emotion regulation strategy that involves total inhibition of emotional behavior is known as emotional suppression (Gross, 1998). There is also a less extreme form of behavioral, response- focused emotion regulation strategy that involves the amplification or de-amplification of emotional expressions (i.e., Cote & Morgan, 2002). These strategies involve exaggerating or down-playing emotional behaviors so as to hide to highlight certain emotional displays. Emotional Labor Much of the emotion regulation literature stems from cognitive psychology. However, organizational psychologists have studied emotion regulation in the context of a work environment under the title emotional labor. The concept of emotional labor was first posited by Hoschshild (1983) and originated within Sociology. The first definition of emotional labor was: “the management of feelings to create a publicly observable facial and bodily display (Hoschshild, 1983).” Emotional labor was theorized to manifest within the workplace through the combination of employees’ frequent interactions with the public, the management of their own and other’s emotions, and the fact that these emotions are often tracked and monitored by management. Given these factors, Hoschshild (1983) posited that emotional labor will always be psychologically harmful to employees since emotional labor “commoditizes” emotions and estranges individuals from their true feelings. Increased interest of emotions and affect within the workplace drew attention to the topic of emotional labor, procuring its eventual boom in research attention. Similar to emotion regulation, the strategies that employees use to regulate and manage their public identity is bifurcated into two strategies: deep acting or surface acting (Hoschshild, 9 1983). Surface acting refers to masking one’s felt emotions to perform expected emotions whereas deep acting refers to the modification of one’s mood to better align with expected emotional displays (Hoschshild, 1983). Display rules create mandatory emotional boundaries that cause emotional labor. Display rules refer to perceived or explicit emotional requirements of a job (Brotheride & Grandey, 2002). These rules serve as the standard to which employees are expected to emotionally display while interacting with individuals. Display rules are often enforced in either two ways: through explicit training materials and/or evaluated as part of one’s formal performance appraisal, or they could be socialized as informal norms by employees (Grandey & Melloy, 2017). Studies have consistently found a positive relationship between display rules and emotion regulation, supporting the assertion that display rules are the adopted standards which dictate employees’ degree of emotion regulation (Brotheridge & Grandey, 2002; Diefendorff & Richard, 2003). In terms of expected emotional valence, display rules can be integrative, differentiating, or masking in nature (Jones & Best, 1993). Those who are asked to display differentiating display rules must display anger or frustration as part of their job. Jobs which may require this display are cops or tax collectors who must appear angry to individuals in order to help them complete tasks such as collecting bills or telling someone to get down (Jones & Best, 1993). Masking is required for occupations where the appearance of emotional neutrality is necessary. Judges must seem impartial during deliberation so as to not be viewed as unfair (Jones & Best, 1993). Lastly, integrative display rules require the display rule of emotions such as happiness or empathy. This display rule is the most common and most investigated (Grandey & Melloy, 2017) amongst all the other display rules. This is because a range of occupations such as servers (e.g. making sure the customer is always pleasantly attended to during the service encounter) to nurses (e.g. 10 displaying empathy for individuals who are watching their loved ones go through chemotherapy) require employees to provide the best service to customers as possible, and often exacerbated by the fact that customer service is largely centered on the customer’s experience (Lopez, 2010). Emotional Labor as Emotion Regulation Emotional labor and emotion regulation are becoming further integrated. The first model to integrate emotional labor and emotion regulation was Grandey’s (2000) model. She likened response-focused and antecedent-focused strategies to surface acting and deep acting strategies, respectively. Antecedent-focused strategies were mapped onto deep acting through reappraisal, a strategy which specifically sought to alter one’s felt mood and subsequent expression toward customers. Response-focused strategies were mapped on surface acting through the strategy of emotional suppression, which refers to the total inhibition of one’s emotional display. However, this overlap is imperfect due to increased recognition of other emotion regulation strategies (apart from surface/deep acting or reappraisal/suppression) individuals use to perform emotional labor. This more nuanced view of emotion regulation strategies led researchers to challenge the initial presumption that emotional labor is always harmful to the employee. Research showed how the relationship between surface acting and deep acting and performance/well-being outcomes were found to vary depending on contextual and relational factors (Grandey, 2000; Gabriel & Grandey, 2015). The most recent model posited by Grandey and Melloy (2017) sought to further integrate emotional labor and emotion regulation, specifically seeking to expand and address the gaps in previous emotional labor/emotion regulation research. There are two main points to take away from this updated model. First, Grandey and Melloy (2017) introduced a dynamic component to emotional labor by incorporating a feedback loop into the revised model. The model helps 11 address the “chicken or egg?” problem associated with felt mood and whether it is a predictor or an outcome of emotional labor since research had previously found evidence for both claims. This feedback loop is also consistent with emotion regulation research which has viewed emotion regulation as a dynamic process where previous goal attainment or emotional information guides subsequent emotion regulation. The current study will use Grandey and Melloy’s (2017) emotion regulation as emotional labor process model as the base framework to understand how individuals regulate their emotions in response to emotional events (i.e., target affect) within the context of a negotiation scenario. The second distinction made by the Grandey and Melloy (2017) model is the level at which emotional labor is posited to occur. In this review, emotional labor was reconceptualized to occur at the event level. This distinction is important given that emotion regulation is more often tested and investigated at the event or episodic level, whereas research on emotional labor often focuses on aggregated individual experiences more generally. However, there is growing evidence suggesting that there are within-person fluctuations in emotional labor. For example, Gabriel, Daniels, Diefendorff, and Greguras (2015) found that individuals can be divided into 5 profiles of “emotion regulators.” Of these profiles, four showed fluctuation in the strategies used to navigate emotional events over time. Similarly, Diefendorff and Greguras (2009) showed that individuals may adopt different display rules in accordance with who they are regulating to. They found that individuals rated how appropriate different emotion regulation strategies (i.e., masking, expressing, neutralizing, qualifying, de-amplifying) are depending on the emotion regulation target. For instance, emotion regulation to customers involved greater neutralization and de-amplification compared to subordinates, which featured primarily de-amplification (Diefendorff & Greguras, 2009). This study shows that there are within person shifts on emotion 12 regulation depending on the interaction target. The event conceptualization then corresponds with research that has found that emotion regulation strategies differ depending on factors such as the nature of customer interaction and goal achievement (as suggested by control theory models of emotional labor). Thus, the investigation and conceptualization of emotional labor at the event/episodic level may serve as a powerful tool that may help researchers better understand the role of customer interactions and subsequent strategy usage. Antecedents of Emotion Regulation As briefly mentioned earlier, research on the antecedents of emotion regulation is sorely missing (Grandey & Gabriel, 2015; Grandey & Melloy, 2017). As of lately, a few important antecedents have been identified within the literature. The reconceptualization of emotional labor at the event level introduced customer interactions (i.e., pleasant or negative customer experiences) and goal achievement as antecedents of subsequent emotion regulation. For instance, pride-evoking events could result in increased response-focused regulation strategies to avoid seeming arrogant (Kalokerinos, Greenaway, Pedder, & Margetts, 2014) or decreased response-focused strategies in an effort to seem more competent (Grandey, Tam, & Brauburger, 2002). However, a number of other interpersonal and contextual antecedents were also introduced by the Grandey and Melloy (2017) model. The introduction of relational variables clarifies how emotion regulation may function differently across contexts and targets. One notable relational variable is relative power. Although theoretically relevant to emotional labor/emotion regulation, relative power has received scant investigation. One relevant study is by Diefendorff, Morehart, and Gabriel (2010). They found that relative power was related to the perceived degree of emotion regulation required. When participants interacted with a target with greater relative power, the use of response-focused strategies were rated by participants as being 13 more appropriate than antecedent-focused strategies (Diefendorff et al., 2010). Further, a moderation effect between relative power and discrete emotion (happiness vs. anger) was also found. Specifically, displays of anger were perceived as more inappropriate toward higher power targets compared to happiness by participants, resulting in a greater perceived need for emotional suppression by participants (Diefendorff et al., 2010). However, the question as to whether these perceptions actually lead to differential strategy usage is open. This distinction is important because an individual’s judgement of a strategy’s appropriateness may not exactly reflect what one would do in scenarios with great relative power difference. There were also contextual antecedents which have been identified and integrated into the emotion regulation process by Grandey and Melloy (2017). The emotional valence of an interpersonal event has been identified as an important antecedent of subsequent emotion regulation when individuals are bounded by integrative display rules. Diefendorff et al. (2008) found that negative interpersonal events were associated with greater deep acting and surface acting strategies. A causal link, through the use of a call center simulation experiment, between a negative interpersonal event and subsequent emotion regulation was found when individuals were expected to perform under integrative display rules (Gabriel & Diefendorff, 2015). The relationship between these negative interpersonal events and differential strategy usage was also found to vary. Following a negative interpersonal event, participants were more likely to use response-focused strategies compared to antecedent-focused strategies (again, under the context of positive display rules; Grandey, Dickter, & Sin, 2004; Sliter et al., 2010). Similarly, the previously discussed studies (i.e., Diefendorff, 2008; Gabriel & Diefendorff, 2015) found that participants were more likely to use response-focused strategies followed a negative event than antecedent-focused strategies. These findings are corroborated by other studies, which found that 14 increased response-focused regulation often followed work stressors (Grandey et al., 2002) and abusive supervision (Carlson, Ferguson, Hunter, & Whitten, 2012). These findings may be due to the fact that surface acting is perceived as more effective following a stressful event than deep acting (Grandey et al., 2004). Concerning positive interpersonal events, research has found that these events are also related to subsequent emotion regulation. In contrast to negative emotional events, positive emotional events lead to more usage of antecedent-focused strategies, as mediated by actor’s increased positive mood (Totterdell & Holman, 2003). However, there is still much room to further explicate and investigate the mechanisms responsible for these findings. The investigation of positive or negative affect as a mediator between an emotional event and subsequent emotion regulation is not consistent across studies. In fact, there is no consistently applied mechanism to explain the relationship between these emotional events and subsequent emotion regulation (Grandey & Melloy, 2017). The current research will help illuminate why positive/negative emotional events may lead to subsequent emotion regulation –– either in regard to increased or decreased degree of emotion regulation or through differential strategy usage. This question will help researchers understand emotion regulation as a process. The Emotions as Social Informational Framework (EASI Model) The EASI framework proposed by Van Kleef (2009; 2010) may be a helpful tool for addressing the gaps present in current emotional labor/emotion regulation research, especially concerning findings regarding relational factors. The groundwork of the EASI model is part socio-cognitive and part evolutionary. While emotions undeniably serve an important function in helping individuals (e.g. running away from danger due to fear or avoiding poison due to disgust), the fact that emotions are often expressed facially suggests that emotions also serve 15 important interpersonal and social functions. As noted by Van Kleef (2010), this idea is not new (e.g., Côté, 2005; ElfenBein, 2007; Fischer & Manstead, 2008). However, the primary goal of the EASI model is to help explicate the processes responsible for how an individual’s observation of another individual’s emotional display guides their subsequent response and how context may influence this process. Van Kleef (2010) asserts that social situations are inherently ambiguous because people often have limited insight into other’s feelings, goals, and intentions. People are left with having to use subtle emotional cues such as voice, bodily posture, and choice of words (Byron, 2008; Ekman & O’sullivan, 1991) to glean important information regarding a targets appraisal of a situation. Indeed, it has been found that people often use other’s emotions to make sense of ambiguous situations (Manstead & Fischer, 2001). Specifically, emotions have been found to provide valuable information in regard to the expresser’s inner feelings (Ekman, 1993), social motives (Fridlund, 1994), orientation toward others (Hess, Blairy, & Kleck, 2000), and appraisal of a situation (Manstead & Fischer, 2001). Since emotions are closely tied to an actor or target’s appraisal of goal satisfaction (i.e., either positive or negative goal attainment; Frijda, 1986; Lazarus, 1991), emotions provide invaluable information when seeking to understand how to appraise and evaluate a complex social situation. The utility of this theory for emotional labor/emotion regulation lies in the notion that a target’s reaction and emotional display may alter an individual’s subsequent emotion regulation through dual pathways. These assertions from the EASI model align with findings within the emotional labor/emotion regulation literature. Specifically, prior research and theory (Diefendorff & Gosserand, 2003; Diefendorff et al., 2008; Gabriel and Diefendorff, 2015) suggests that previous customer interactions serve as a useful feedback mechanism for future emotion regulation. However, the EASI model has the 16 advantage of explaining why a target’s emotional display (positive or negative) may result in actor’s differential use of antecedent- or response-focused strategies. The EASI model may help clarify how felt mood or previous goal attainment leads to specific emotion regulation strategy usage and how various interpersonal factors may moderate this relationship. Before specific interpersonal moderators are introduced, the two mechanisms thought to link a target’s emotional display to a person’s behavior will be explained below. Affective Pathway The affective pathway is responsible for shaping behaviors through felt mood and affective reactions (Van Kleef, 2009; 2010), functioning through mirror neurons and emotional contagion (Hatfield, Cacioppo, & Rapson, 1994). Activation of the affective pathway also influences behavior through one’s interpersonal impressions, such as increased favorability when a target displays happiness (Clark & Taraban, 1991) or decreased liking when a target displays anger (Van Kleef, De Dreu, Manstead, 2004). There is evidence that an individual’s affective reaction influences their subsequent behavior. For instance, Barsade (2002) found that the presence of a happy confederate within a group promoted increased cooperation and lower conflict by developing more pleasant emotions. This was compared to the presence a negative confederate, which created increased negative affect across group members and led to poorer cooperation. Similarly, Sy, Côté, and Saavedra (2005) found that leaders in a happy mood were more likely to inspire positive mood in a group, which led to greater coordination. This effect was reversed for leaders in a negative mood. Lastly, the role of affective reactions has been given greater attention in the context of a negotiation scenario. Examining e-Bay dispute incidents, Friedman et al. (2004) found that expressions of anger elicited reciprocal anger which led to lower chances of dispute resolution. Van Kleef and Côté (2007) found that a target’s expression 17 of anger led to an observer developing a strong negative emotional reaction, which motivated retaliatory and competitive behavior from the observer. Lastly, bargainers who faced an angry opponent were far more likely to deceive their opponent and make less generous offers, mediated by reciprocal anger on the participants part (Van Dijk, Van Kleef, Steinel, & Van Beest, 2008). This collection of evidence leads to the first set of hypotheses. Hypothesis 1: There will be a main effect of target emotion on participants’ positive affective reaction. Specifically, participants’ positive affective reaction will be (a) positively related to target’s level of happiness display and (b) negatively related to target’s level of anger display. Hypothesis 2: There will be a main effect of target emotion on participants’ negative affective reaction. Specifically, participants’ negative affective reaction will be (a) positively related to target’s level of anger display and (b) negatively related to target’s level of happiness display. Both the emotion regulation/emotional labor literature and EASI model suggest that an actor’s affective reaction plays an important role in the type of emotion regulation strategy adopted. This alignment helps build support for the application of the EASI model to the emotion regulation process. Previously, control theory models of self-regulation have been applied by emotional labor researchers to explain the relationship between emotional events and subsequent emotion regulation strategies (i.e., Diefendorff & Gosserand, 2003). Control theory posits a negative feedback loop involving an input function, a standard, a comparator, and an output function (Carver & Scheier, 1999; Klein, 1989). The role of the input function is to incorporate outside information into the feedback loop (through the process of perceiving one’s own behavior or emotional state). The standard refers to the emotion regulation goal (i.e., the display rule), while the comparator refers to the comparison made between the input and the standard. The output is a function of the difference (or lack of) signaled by the comparator. When there is a large discrepancy in the comparator, individuals will be motivated to alter their output such that 18 the subsequent input will be more closely aligned with the standard. If there is no discrepancy in the comparator, then the subsequent output will not be altered. Following the output, affective events such as customer mistreatment or customer satisfaction can be modeled as environmental “disturbances” which prompt an individual’s response (e.g., display positive emotion to address a complaint). The feedback loop begins anew once the individual’s response is re-appraised by themselves (i.e., input). This process continues until there is no more discrepancy or there is no standard to compare to (i.e., the encounter is over). Control theory has been applied by researchers to understand how the valence of a customer interaction may lead to differential strategy usage across emotional events. When asked to conform to integrative display rules, customer mistreatment leads to both negative affect and increased emotion regulation (Gabriel & Diefendorff, 2015; Goldberg & Grandey, 2007). However, the use of response-focused strategies (namely suppression) is far more common in these customer mistreatment scenarios (Rupp & Spencer, 2006; Sliter et al., 2010). An actor’s self-appraisal of their negative affect would signal a large discrepancy between their felt mood and the expected demands of the display rule (i.e., display happiness and suppress negative emotions; Grandey et al., 2013), resulting in this greater reliance on response-focused (e.g., suppression) strategies. Conversely, in the context of positive customer events and integrative display rules, Totterdell and Holman (2003) found that the relationship between a positive customer event and subsequent increased use of antecedent-focused strategies was mediated by felt mood. This relationship was posited to exist through the congruence of one’s mood with positive display rules, which calls for maintenance of one’s mood if congruent. These findings seemingly contradict predictions of control theory since Totterdell and Holman (2003) found that the use of antecedent-focused strategies increased despite the lack of discrepancy between felt 19 emotion and the display standard. However, upon closer inspection, the findings of this study may be interpreted differently for two reasons. First, participants aiming to maintain a good mood by leveraging previous customer interactions may have a different goal than simply meeting the display rules standard. That is, their antecedent-focused strategy usage likely reflects an effort to pursue a goal of positive mood maintenance rather a goal to meet the display rule standard. Second, Totterdell and Holman (2003) used a time-sampling method for data collection, recording participants’ emotion regulation and mood every two hours during a work day for two weeks. An individual’s positive mood may have originated from a prior encounter rather than reflecting their assessment of a current encounter. If the former is the case, then positive mood may be viewed as a spillover, rather than being considered as feedback concerning one’s current performance. Therefore, this study’s unique research design may inhibit complete integration of its finding into control theory tenants. Following the tenants of control theory, when participants are bounded by positive display rules, experiences of positive mood are largely congruent with expected displays and would signal little to no need for regulating one’s emotional display given the small amount of discrepancy (Diefendorff & Greguras, 2003). In other words, a lack of significant discrepancy would likely signal to an individual that their performance is adequate and would likely dedicate little to no effort in altering subsequent emotion regulation (for either antecedent- or response- focused strategies). This assertion is congruent with the EASI framework which also predicts that individuals are likely to continue the same effort and strategy when there is no perceived performance discrepancy (Van Kleef, 2010; Van Kleef et al., 2009). Further, when there is a large discrepancy between the individual’s goal and their performance (i.e., a positive display rule standard and high experienced negative affect in an individual), individuals are also more 20 likely to increase effort and emotion regulation to reduce the discrepancy (Diefendorff & Gosserand, 2003). The assertions of the EASI framework are parallel to the assertions of control theory, but control theory is a helpful framework for understanding how one’s behaviors function under display rules and how one’s affect guides subsequent emotion regulation within the same interaction episode. Therefore, when an individual is bounded by integrative display rules, the individual’s emotion regulation behavior will be affected in accordance to the congruence between felt mood and display rules. Thus, the following hypotheses are proposed: Hypothesis 3: Participant positive affective reaction is negatively related to participant response-focused emotion regulation strategies. Hypothesis 4: Participant positive affective reaction is negatively related to participant antecedent-focused emotion regulation strategies. Hypothesis 5: Participant negative affective reaction is positively related to participant response-focused emotion regulation strategies. Hypothesis 6: Participant negative affect reaction is positively related to participant antecedent-focused emotion regulation strategies. Inferential Pathway The inferential pathway is the other pathway the EASI framework posited to influence behavior. This pathway is posited to function through appraisal theories and social referencing (Van Kleef, 2009; 2010). The base assumption of this pathway is that emotional expressions convey important social information, such as the expresser’s feeling, attitudes, behavioral intentions, and relational orientation. The information conveyed by an emotional expression is largely context dependent, but the basic informational value of these emotions is generalizable (Van Kleef, 2009). As mentioned earlier, a person observing anger may act less cooperatively with whomever is displaying anger because of negative mood. However, as Van Kleef (2009) notes, appraisal theories suggest that the display of anger is most likely to occur when one’s own 21 goals are obstructed, whereas the display of happiness is likely to occur when one’s goals are met. Thus, displays of anger or happiness by a target (e.g., customer, manager) may help individuals infer information regarding their own performance. Van Kleef (2010) showed that individuals who observe anger from targets tend to infer that they did something wrong. Similarly, Van Kleef et al. (2009) found that work teams use the emotions displayed by their leader to infer their performance. Work teams inferred that their performance was not satisfactory when their leader showed anger, which led to team members’ greater effort and performance, Conversely, work teams inferred that their performance was adequate when their leader displayed happiness, which led members to reduce effort and performance. These findings collectively suggest that actors use the emotional information presented to them by a target to make performance-related inferences according to whether the information is positive or negative. Hypothesis 7: There will be a main effect of target emotion on participant’s performance inference. Specifically, participant’s performance inference will be (a) negatively related to target’s level of anger display and (b) positively related to target’s level of happiness display. Control theory also helps explain how one’s performance inference generated from a target’s affective display may be related to an actor’s emotion regulation. Within the control theory framework, individuals monitor the discrepancy between their display goal (e.g., regulate own display of emotions to match display rules) and actual emotion display. Actors who draw a low performance inference from target’s anger –– a large discrepancy in the comparator –– are more motivated to increase their emotion regulation in order to reduce the performance discrepancy. Conversely, if the emotional information provided by the target signals that the actor’s performance is satisfactory, then the actor is less motivated to engage in emotion regulation, leading to less emotion regulation. In support of this proposed pattern, Sliter et al. 22 (2010) sought to investigate the role of customer mistreatment on emotion regulation and found that negative events are marked by a significant increase in both response-focused and antecedent-focused emotion regulation. This finding suggests that actors may be attempting to regulate their emotions following a negative event signaling that performance had not been satisfactory. Further, the increase in emotion regulation following a negative event features significantly more use of response-focused strategies compared to antecedent-focused strategies (Grandey, Dickter, & Sin, 2004; Sliter et al., 2010). At the within-episode level, Diefendorff and Gabriel (2015) investigated the role of customer incivility on emotion regulation (i.e. deep and surface acting) on continuous measures (captured every 200 milliseconds). Actors were asked to partake in a call center simulation where they were then exposed to an instance of customer incivility. Diefendorff and Gabriel’s (2015) study found that immediately following customer incivility, actors were significantly more likely to engage in increased emotion regulation (from both camps of strategies). However, the recovery period –– marked by an increase in customer civility –– resulted in a significant reduction in both emotion regulation strategies. Other studies more directly support the link between the discrepancy size (i.e. negative or positive feedback) and subsequent output (i.e. effort). Interventionist approaches to goal setting (Locke & Latham, 1990) find that a negative feedback discrepancy lead to greater effort rather than lowering the standard for performance (Bandura & Cervone, 1983). In the presence of a clear negative feedback intervention, the presence of feedback helps drive motivation to reach the standard (Erez, 1977). Matsui, Okada, and Inoshita (1983) also investigated the role of feedback on subsequent effort. When participants were asked to complete a certain number of math problems in a fixed amount of time, only participants who were below their goal and showed a substantial discrepancy between the goal and progress benefitted from negative 23 feedback. These lower performing individuals were found to be more motivated, effortful, and completed problems quicker following negative feedback, even when they had otherwise believed the goal was not attainable (Matsui, Okada, and Inoshita, 1983). Other studies suggest that the degree of effort may differ following performance inference drawn from leaders. Van Kleef et al. (2009) found that positive displays of affect by a leader counterintuitively led to lower performance in followers because positive affect signaled that performance adequate, thus less effort should be dedicated to performance. This mechanism proposed by the EASI model will help clarify if individuals choose emotion regulation strategies depending on their perceptions of whether their performance is satisfactory. Thus, the following hypotheses are posited: Hypothesis 8: Participant’s performance inference are negatively related to participant antecedent-focused emotional regulation strategies. Specifically, participants’ high perceived performance inference (i.e. believe that they are performing well) will be negatively related to antecedent-focused emotion regulation strategies. Hypothesis 9: Participant performance inference are negatively related to participant response-focused emotional regulation strategies. Specifically, participants’ high perceived performance inference (i.e. believe that they are performing well) will be negatively related to response-focused emotion regulation strategies. Moderators of the EASI model According to the EASI model (Van Kleef, 2009; 2010), both the affective and inferential pathways simultaneously explain how individuals are influenced by a target’s affect. However, which pathway is more influential depends on two moderators. The first class of moderators concern the social-contextual factors which could affect the predictive power of performance inference and affective reactions. The first factor is the interdependence of the situation. Both cooperative and competitive situations are most likely characterized by low trust (Van Kleef et al., 2010). In these scenarios, the use of emotional information is especially important since 24 individuals largely rely on the actions and information provided by their opponent to infer opponent’s intentions, which can help them select the most advantageous strategy. Van Kleef et al. (2010) specifically found that the inferential pathway is more predictive of behavior in these interdependent contexts due to the lower trust across parties. The prevailing (cultural) norms of the situation are also an important moderator (Van Kleef, 2009; 2010). For example, display rules function as an overarching norm that guides the expected displays within the context. Van Kleef and Côté (2007) found that when the negotiation opponent displayed anger without the presence of a positive display rule, participants considered the anger display appropriate. This resulted in participants decreasing bargain demands over the course of 6 negotiation rounds, as participants inferred that their offer was not satisfactory to their opponent, which then resulted in the anger. On the other hand, when the opponent displayed anger under positive display rules, such display was viewed as inappropriate and fostered participants’ competitive, retaliatory behaviors. In this case, participants were more likely to refuse to waive their demands to opponents and reported greater desire to retaliate against their opponent. These aggressive reactions can be seen as driven by participants’ own negative affect elicited by the opponent’s inappropriate display of anger. The other moderator within the EASI framework refers to the individual’s degree of information processing. Van Kleef (2009; 2010) explains that the degree to which a target’s emotional display can influence the behavior of an actor is dependent on the actor’s motivation and ability to process the emotional information. The deeper the information processing, the stronger the predicted power of the inferential pathway. Meanwhile, the shallower the information processing, the greater relative predictive power of the affective pathway (Van Kleef, 2009). A number of factors have been identified as affecting an individual’s degree of 25 information processing. Van Kleef et al. (2004) found that high time pressure, low need for cognitive closure, and being in a high-power position all decrease the ability and motivation to process emotional information. Conversely, low time pressure, high need for cognitive closure, and low power all predict greater motivation and ability to process emotional information. Actors may only be as affected by a target’s emotions to the extent that they can perceive the target’s emotional display and then effectively draw inferences from this information. This is why factors which lower one’s ability to effectively process emotional information are tied to greater activation of the affective pathway, because individuals are more likely to be affected by the emotions of a target through automatic processes such as emotional contagion (Van Kleef et al., 2009). When individuals are able to process emotional information and draw inferences, the inferential pathway is most likely to predict subsequent behavior due to the individual’s ability to thoroughly process and react to information (Van Kleef 2009;2010; Van Kleef et al., 2009). Power The role of power as a moderator within the EASI framework will be further examined in the current research. Power is broadly defined as the capacity to exert influence on other people (Bacharach & Lawler, 1981) and is recognized as a moderator by both the EASI model and extant emotional labor/emotion regulation research. Within the EASI framework, power is posited to affect an individual’s degree of effortful information processing (i.e., epistemic motivation). Within the emotional labor/emotion regulation literature, Diefendorff and Greguras (2009) found that the degree to which an actor controls their emotional display is much higher when the actor is interacting with a supervisor compared to a coworker or subordinate, and that the expected degree of emotional control is highest for when interacting with customers. As discussed earlier, Diefendorff, Morehart, and Gabriel (2010) also showed that a target’s relative 26 power affects an actor’s hypothetical emotion regulation control. However, the investigation of power in emotion regulation is largely based on prior work showing that relative power shapes the freedom actors have in expressing their emotions, which is distinct from the EASI models reasoning for why power affects subsequent emotion regulation/emotional labor. French and Raven (1959) argue that there are numerous bases for power. Power can be signaled by someone’s position within an organization or the access to valuable information such as expertise. Therefore, there can be multiple cues to signal power. Accordingly, an individual’s power, and the cues that signal power, is largely contingent on the context. For instance, French and Raven (1959) posits that power can be conceptualized as coercive if the source of power comes from the dependence of one party on another. More specifically, power is coercive if one party is able to leverage punishment on the other party when the latter fails to appropriately respond to an influence attempt (e.g., a customer complaining to management if their request is not adequately addressed by the employee). Coercive power is most powerfully enacted in situations where the threat of conformity is strongest and is dependence on the influencer’s ability to threaten punishment due to nonconformity (French & Raven, 1959). Power can also be conceptualized as a party’s ability to reward or give positively-valanced stimuli, or remove negatively-valanced stimuli, from the other party. How these different forms of power manifest (and what constitutes a positive or negative stimulus) is largely contingent on the context in which the power difference takes play. The EASI framework specifically identifies some contexts where power may play an important role in the emotion regulation process (Van Kleef, 2009; 2010). In line with French and Raven’s assertions, Van Kleef (2009; 2010) stresses that the interdependence of the context makes power a salient factor. Negotiation scenarios represent a context where power is especially important throughout the bargain (Van Kleef, 2009; 2010). 27 Power within the context of a negotiation scenario is often conceptualized and operationalized as the availability of resources (Giebels, De Dreu & Van de Vliert, 1998; Pinkley, 1995) or greater leverage over resources (Lee, & Tiedens, 2001; Yukl, & Falbe, 1991). Parties must be able to leverage resources and rationalize with each other in order to reach a favorable outcome for themselves or, sometimes, both parties. Therefore, an imbalance of power within a negotiation scenario may be problematic for a party’s ability to reach their goals. One important cue that may signal differences in relative power in negotiation may be the contexts that require the actors to interact with multiple targets simultaneously. Several jobs require individuals to regulate their emotions while interacting with multiple targets simultaneously. For example, servers may have to wait multiple customers at the same table; employees have to give a presentation in front of a group; salespeople may have to bargain with multiple buyers at the same time. The distinction between regulating one’s emotions while interacting with a single target compared to interacting with multiple targets may be important because multiple targets may signal an imbalance of power. As previously discussed, French and Raven (1959) outlines how context can alter the cues which signal power. For instance, customer service scenarios may require servers to wait multiple targets simultaneously. In this service context, multiple targets (rather than a single target) can signal greater power through the group’s ability to reward or punish the server. Larger parties may hold greater power over a single person party through their ability to either offer greater tips (reward) or collectively withdraw tips (punish) than a single person. Larger parties may also hold coercive power through their power to punish servers who do not conform to display rules. Servers may feel as if they are in a worse position if they do not conform to expected service displays when interacting with a group rather than a single person due the group’s ability to punish the server more severely (i.e. multiple 28 complaints to management) than a single person. French and Raven (1959) similarly hypothesizes that an imbalance of power may arise between individuals due to differences in expertise and knowledge. Multiple targets may signal greater information processing and expertise over a single individual, which in turn could lead to a power advantage favoring the group. In contexts like a negotiation, actors who have to regulate to multiple targets may feel disadvantaged due to greater information sharing and pooled expertise on the party’s behalf. There is some evidence that interacting with multiple targets may signal a power imbalance in the context of a negotiation. Sally and O’Conner (2004) outline why team negotiators may be seen as having an advantage over solo negotiators. Negotiation is a complex task that requires sufficient cognitive effort. An individual may not have the same memory ability, creativity, simultaneous data processing that a team is able to less effortfully achieve. Therefore, teams can be seen as being advantageous in terms of information processing and resources compared to individuals (Sally & O’Conner, 2004). Cohen and Thompson (2011) similarly argue that there is a team efficacy effect whereby both individuals and teams perceive teams as having an advantage in a negotiation scenario. In their study, Thompson and Peterson (1996) found that this was the case when solo negotiators were asked to rate their relative power when bargaining against a negotiation team. Compared to solo negotiators bargaining against solo negotiators or team negotiators bargaining against solo negotiators –– who all perceived an advantage for their party –– solo negotiators perceived themselves at a significant disadvantage when bargaining against team negotiators (Thompson & Peterson, 1996). Patton and Balakrishnan (2011) corroborated these finding in their study. When solo negotiators bargained against a buying team, solo negotiators significantly attributed greater relative power to team negotiators. Meanwhile, solo negotiators did not perceive a difference in power when bargaining against a 29 solo negotiator. Outside the context of negotiation, Drach-Zahavy, Yagil, and Cohen (2017) found that health care providers relied more on surface acting when they were alone compared to when they were in teams. The theoretical justification for this finding was not explicitly investigated within the paper, but one possible suggestion offered by the authors is that the perception of being outnumbered by the patients and their family members resulted in greater stress and surface acting. Thus, this leads to the hypothesis: Hypothesis 10: There will be a main effect of number of targets participants interact with on perceived relative power. Specifically, participants who interact with multiple targets will report lower relative power (i.e. participants will perceive a power disadvantage relative to their targets) compared to participants who interact with a single target. The EASI framework details how power may first affect an actor’s reaction to a target’s emotional information. Being in a position of lower power is posited to result in greater epistemic motivation –– which is defined as one’s motivation to accurately understand the world (De Dreu, 2003). Those with high epistemic motivation are posited to process emotional information more thoroughly (De Dreu, 2003; Van Kleef, 2009; 2010). There is evidence to support this claim. Van Kleef et al. (2006) found that individuals in a low power position during a negotiation (high epistemic motivation) were more influenced by an opponent’s emotions (i.e., they conceded more often, offered higher bids) compared to individuals in higher power positions. Higher power individuals alternatively were not influenced by the counterpart’s emotions (i.e., made smaller bids and did not concede as often). Thus, power may be an important predictor of whether individuals will or will not be motivated to use the emotional information presented to them. Van Kleef, Oevis, Van Der Löwe , LuoKogan, and Keltner (2008) also showed that participants with lower power were more motivated to understand a conversation partner’s emotional cues compared to those with higher power. In their study, Van Kleef et al. (2008) paired two individuals who had never met prior and had them write down and 30 share a personal, emotionally distressing story to each other. Participants took a turn either playing the role of “talker” (convey the feelings evoked by their story) or a “listener” (attempt to understand the experience of the other participant and ask questions). Van Kleef et al. (2008) found that participants who self-reported themselves as having a greater sense of general power experienced less distress and compassion when listening to other’s stories. These effects remained significant after controlling for participants’ baseline happiness and decoding accuracy, suggesting that they are motivational in nature. This finding was further corroborated by physiological data, which found that high power participants relied more on autonomic processes to down-regulate their reactions to other’s distress. While one may reason that a significant increase in emotional information processing (i.e., higher epistemic motivation) means greater likelihood of being emotionally influenced by a target’s emotion cues, Van Kleef et al. (2009) showed that this is not the case. This study found that teams differentially leveraged their affective reactions and performance inference depending on their epistemic motivation. Teams with higher epistemic motivation (i.e. greater information processing) were more likely to rely on the emotional information provided by their leader for future behavior. This is compared to teams lower in epistemic motivation (i.e. lower information processing) who relied more on how they felt to shape future performance. As noted by the authors, these findings suggest that more automatic processes of emotional contagion may be bounded by epistemic motivation (Van Kleef et al., 2009). Thus, in situations when individuals perceive lower or no power difference, they may have lower on epistemic motivation, which in turn makes the target’s affect less influential to actors due to decreased motivation to process the interaction target’s emotional display. On the other hand, when individuals perceive higher power difference (specifically, there is a high-power difference disadvantaging the individual), they may have higher epistemic 31 motivation, which may make the target’s affective display more salient. Therefore, the following set of hypotheses are asserted: Hypothesis 11: There will be a main effect of participant’s perceived relative power on participant’s information processing. Specifically, participants who report lower relative power (i.e. perceive a power disadvantage relative to their interaction target) will report greater information processing compared to participants who report greater relative power. Hypothesis 12: The effect of target emotion on participants’ positive affective reaction will be moderated by participants’ information processing. Specifically, higher participant information processing will attenuate the relationship between target emotion and participant positive affective reaction. Hypothesis 13: The main effect of target emotion on participant negative affective reaction will be moderated by participant information processing. Specifically, higher participant information processing will attenuate the relationship between target emotion and participant negative affective reaction. Hypothesis 14: The main effect of target emotion on participant’s performance inference will be moderated by participant information processing. Specifically, higher participant information processing will strengthen the relationship between target emotion and participant’s performance inference. The EASI model next posits that the relative influence of either the affective pathway or inferential pathway on subsequent behavior will also be moderated by power. Actors in a low power position are more likely to rely on the inferential pathway than the affective pathway because inferences are best drawn under conditions of greater information processing depth (i.e., higher epistemic motivation/low power; Van Kleef, 2009; 2010). Thus, low power actors are more likely to utilize the emotional information provided by a target strategically. This argument has been supported by empirical studies. In their three-study paper, Van Kleef et al. (2004b) investigated whether actors would concede more to an angry opponent in comparison to a happy opponent under conditions hypothesized to increase epistemic motivation, such as lower time pressure, high dispositional epistemic motivation, and low power. Greater concession to angry opponents signaled that actors strategically “tracked” their opponents’ limits (Van Kleef et al., 32 2004) and inferred that targets would be less likely to concede, leading to increased concession and smaller demands by actors (i.e., greater reliance on inferential pathway). Van Kleef et al. (2004b) found support for these hypotheses. When actors processed information more deeply (i.e. lesser time pressure and low power), they conceded more to and demanded less from angry targets compared to actors who did not engage in deep process information (i.e., those under greater time pressure and high power). Sinaceur and Tiedens (2006) similarly found that participants who faced poor negotiation alternatives relative to the counterpart were more motivated to engage in thorough information processing. This led to increased likelihood for participants to infer that the target either had ambitious or lenient limits when targets displayed anger. For participants with lower information processing, a target’s emotional display did not shape subsequent behavior. This supports the idea that individuals draw strategic inferences regarding the target’s limits and ambitions when motivated to use emotional information, but not otherwise. While the Sinaceur and Tiedens (2006) study did not explicitly manipulate power in their study design, it is worth reiterating that the EASI model posits the direct, inverse relationship between information processing and power (Van Kleef, 2009; 2010). There are two studies which investigate the influential strength of the affective and inferential pathway under different epistemic motivation conditions. For example, Van Kleef et al. (2009) found that the affective display of a leader (anger vs. happiness) differentially influenced team performance depending of the team’s dispositional epistemic motivation. An angry display by a leader to a low epistemic motivation team resulted in lower performance because anger undermined team cooperation (affective pathway). However, the same display resulting in higher performance in teams with higher epistemic motivation team because leader’s anger signaled poor performance (Van Kleef et al., 2009). Similarly, Van Kleef and Cote (2007) showed that high power 33 negotiators were more influenced by target’s displays of inappropriate anger compared to appropriate displays of anger and non-emotional displays. On the other hand, low power negotiators develop a more strategic route by conceding more often regardless of the appropriateness of anger displayed by the target (inferential pathway; Van Kleef & Cote, 2007). Taken all together, the following hypotheses test whether differential pathway reliance occurs as proposed by Van Kleef’s (2009; 2010) model and whether low relative power perceptions will lead to greater reliance on the inferential pathway through greater information processing. This leads to the following research question and hypotheses: Hypothesis 15: The main effect of participants’ positive affect on their antecedent- and response-focused emotion regulation strategies will be moderated by participant’s information processing. Specifically, higher participant information processing will attenuate the negative relationship between participant’s positive affect and participant antecedent- and response-focused emotion regulation. Hypothesis 16: The main effect of negative affect on participant antecedent- and response-focused emotion regulation strategies will be moderated by participant information processing. Specifically, higher participant information processing attenuates the positive relationship between participant’s negative affect and participant antecedent- and response-focused emotion regulation. Hypothesis 17: The main effect of performance inference on participant antecedent- and response-focused emotion regulation strategies will be moderated by participant information processing. Specifically, higher participant information processing will strengthen the negative relationship between performance inference and participant antecedent- and response-focused emotion regulation. Authentic Display In addition to investigation emotion regulation (i.e., response-focused and antecedent- focused emotion regulation), this study investigated the moderation effects of information processing on the relationship between participant’s positive affect, negative affect, and performance and participants’ display of naturally felt emotions. Prior research has found that display of naturally felt emotions, antecedent-focused, and response-focused emotion regulation 34 are distinct strategies used by individuals to display organizationally desired emotions (Diefendorff et al., 2005). Diefenforff et al. (2005) further showed that authentic display of emotion was used more often than surface acting and deep acting, suggesting that emotional labor research should include measures of authentic display because it is a prominent emotion regulation strategy. Authors further show that authentic display was predicted by both dispositional (e.g., agreeableness and extroversion) and situational factors (e.g., negative display rules, frequency of interactions), noting that dispositional factors were stronger predictors of participant’s authentic display. Therefore, this purpose of including authentic display along with antecedent- and response-focused emotion regulation strategies was to explore whether the hypothesized effects of affective and inferential predictors similarly function for displays of naturally felt emotions. Therefore, the two following research questions are proposed: Research question 1: What is the role of participants’ positive affect, negative affect, and performance inference on the expression of naturally felt emotions (authentic display)? Research question 2: Does information processing moderate the relationship between participants’ positive, negative, and performance inference and their expression of naturally felt emotions (authentic display)? 35 Participants METHOD Participants were recruited through a participant pool of undergraduate students at Michigan State University, and received course credit for participation. A power analysis was conducted with .05 level of significance, and a moderate effect size (d=.3) with four groups (2 X 2 ANOVA design). The a priori power analysis showed that the minimum number of participants required would be N = 280. Data were collected from 640 participants who completed the experiment. In order to ensure maximum validity, participants were asked whether they believed they were negotiating with real participants (yes/no/maybe). Only data from participants who reported yes or maybe were included in analyses. Of these 640 participants, 270 (42.2%) reported that they did not believe they were negotiating with real opponents, 215 (33.6%) reported that they believed their opponents were real, and 155 (24.2%) reported that they maybe believed opponents were real, resulting in 370 participants who reported yes or maybe to believing their opponents were real. In order to match participant responses on the simulated negotiation task software and participants responses on the online Qualtrics survey, participants were assigned a unique 8-digit ID following the negotiation scenario to be subsequently reported on the Qualtrics survey. Of the 370 participants who responded yes/maybe to the manipulation, 326 (88%) correctly entered the unique ID to the Qualtrics survey, resulting in a final sample of 326 participants whose data can be linked to their responses on the negotiation simulation app. Participants were majority self- reported female (72%) and between the ages of 18-20 (89%). Further, 62% of participants were White, 10% were Black, 15% Asian, 4% Hispanic or Latino, 5% Multi-racial, and 4% self- identified as a different race or ethnicity. 36 Negotiation Task The task for this study was an application designed to simulate an online negotiation scenario with fake participants. The scenario used was based on a roleplaying simulation from Harvard’s Program on Negotiation (PON) created for the purpose of training and teaching negotiation. The scenario chosen is titled “West Wind in Pine Hills.” The negotiation scenario involves a bargain between a small town (Pine Hills) with a private company (West Wind) seeking to expand onto the territory of Pine Hills. This scenario was selected for several reasons. First, this scenario allowed for natural adaptation of number of targets by either having participants engage with a single representative or multiple representatives of West Wind. Second, the scenario does not take longer than an hour, and thus the scenario is meant to be resolved quickly. This scenario is also meant to be a good introduction for participants who are not familiar with negotiation according to Harvard’s PON website. Therefore, there is little expertise required to engage with the task and the task is easy to understand for those new or unfamiliar with negotiation. Lastly, this scenario sets both parties in the negotiation on equal footing initially. West Wind wishes to provide a stipend to Pine Hills in exchange for the ability to build on Pine Hills property and the mayor of Pine Hills can use this negotiation as leverage for re-election as mayor of Pine Hills. Therefore, both parties (West Wind and Pine Hills) are on equal footing at the start of this negotiation because both parties have access to important resources that the other party needs for their goals. This equal footing addresses the potential confound of unequal bargaining resources that may signal unequal power between parties. Therefore, the nature of the task allows for clear interpretation regarding whether the presence of one or many opponents corresponds with hypothesized power differences. Participants prior to the task were told they will be negotiating with other undergraduate participants during the 37 scenario, such that participants do not have reason to believe that either party has an advantage in terms of expertise over the other. Appendix A summarizes a full outline of the simulated negotiation task. Experimental sessions allowed for up to eight participants to sign-up. Participants were instructed to meet at the experiment site where two research assistants were present to run the experiment. In order to maximize the face validity of the negotiation task, research assistants were instructed to split participants across two rooms (up to four participants per room). Once participants were split, they were asked to enter an IP domain on an internet browser to access the negotiation simulation application on their laptops. Once they entered this address, participants were presented with a consent form. After they provided consent, participants were sent to the following page explaining the general context of the negotiation scenario, their role in the negotiation, and their display rules for the negotiation. Broadly, the negotiation scenario involves two parties: Pine Hills, a rural town, and West Wind, a U.S.A wind energy company. West Wind is seeking to propose a development project within the town of Pine Hills. The parties are then meeting to discuss two issues: the number of wind turbines to be built within the town and the annual community benefit stipend that West Wind will contribute to Pine Hills. This is where participants were introduced to their role as the mayor of Pine Hills (Riley Marks). This role involves negotiating with either a single or a panel of representative(s) of West Wind. Participant’s goal for the negotiation is to meet two criteria on the previously mentioned issues: less than 10 turbines on Pine Hills property and more than $100,000 in the community benefit stipend. Participants were told they cannot accept offers above 10 turbines or less than $100,000 in stipend. Moreover, participants were advised to obtain an outcome as favorable (i.e., lower number of turbines and higher number of stipend) as 38 possible. Following the introduction of the goal, participants were given guidelines and rules to abide throughout the negotiation to best achieve successful a negotiation outcome. These guidelines served to introduce the positive display rule to the participants (i.e. “aim to display positive emotions and inhibit the showing of negative emotions throughout the negotiation to best ensure success”) and help data collection (i.e. “aim to use proper grammar and syntax to clearly present your offer to the other party”). Participants were told they will be negotiating with other participants in the other room and asked to pay close attention to the scenario, their role in the negotiation, and their goals. Participants were allowed 5 to 7 minutes to review this information and encouraged to take notes of the information presented. A printout was also available to participants during the negotiation upon request. Once participants were given sufficient time to review the information, research assistants told participants to not continue onto the following segment until the other room was ready. Research assistants then exited the room and waited 20 to 30 seconds before re-entering and telling participants that the other room is ready to begin and that they too can begin. Participants then began the negotiation scenario in the following window. The chat interface began by notifying participants that their negotiation partner(s) joined the chat room, serving to make the of number of targets salient. A sidebar also displayed all participants in the room during the negotiation. The first round of the negotiation is the introduction of the West Wind negotiation party. A pre-determined response was sent to each participant regardless of condition. Following this response, participants introduced themselves to the other party. After introductions, the West Wind party presented the first offer to participants. The first offer was largely misaligned with both criteria of the participant’s negotiation goal such that participants cannot reasonably accept the offer. The West Wind 39 party’s response in their first offer introduced the affect manipulation (positive vs. negative affect). Participants were then asked to either accept or decline West Wind’s offer and if declining, to present a counteroffer. Following the participant’s counteroffer, the round concluded. Participants then completed a survey assessing their felt emotion, performance inference, and their perceptions of the target’s affect. This same iteration occurred for two additional rounds. On the last round, the offer presented by the opponents met one criterion (i.e., 100,000 dollar stipend offered by West Wind) and participants were asked to accept or decline the final offer. Once participants committed to an offer decision, the experiment ended. A final survey assessing perceived power (i.e., relative to the opponent(s) and throughout the negotiation), emotion regulation (antecedent-, response-focused emotion regulation, and authentic display), need for cognitive closure, and information processing was then administered. Once participants completed the final survey and were debriefed, they were provided a unique 8-digit ID assigned from the app to enter on a Qualtrics survey. This Qualtrics survey asked participants for their 8-digit ID (used to match responses from the survey to the app), participant’s prior work history, whether they believed the negotiation opponents were real, and whether the instructions of the negotiation scenario and experiment were clear. Participant demographics were collected using SONA’s prescreen responses. These prescreen responses were captured when undergraduate participants sign up for the SONA human resource pool software. In these prescreen responses, participants were asked to self-report several questions that capture demographic variables such as their age, Race/ethnicity, gender. These prescreen responses were used to for demographic data. 40 Pilot Study A pilot study involving fifteen undergraduate students was used to assess whether the simulated negotiation application and experimental procedures led participants to believe they were negotiating with other undergraduate students. First, participants completed the procedure described above. Participants were then asked whether they believed they were negotiating with real opponents. Of these 15 participants, 4 responded maybe, 1 responded yes, and 10 responded no. Those who reported “maybe” or “no” were then asked to “…please explain below what made you believe the participants may not be real” and “what would have made the participants seem more convincing as real people?” After the pilot data collection session, four research assistants were provided participants’ qualitative responses to the questions above and qualitative responses during the simulated negotiation. Research assistants used this data to re-write the simulated opponent’s script to better align with participant’s suggestions for making opponents seem real. Research assistants and I met after one week with revised scripts for each participant condition (happy/angry and single/multiple). Each script was projected onto a screen and then evaluated regarding how well the opponent’s affect was expressed and how believable the script was. Following this discussion, scripts were re-written. The updated script was then implemented and used for data collection, resulting in a final breakdown of 42% participants reporting they did not believe the opponents were real, 24% reporting that they maybe believed, and 34% reporting they believed opponents were real. 41 Measures Perceptions of power. Perceptions of power was assessed using Van Kleef, De Dreu, Pietroni, and Manstead’s (2006) measures of experienced power and relative power (see Appendix B1). The experienced power measure assessed an individual’s total sense of a power during the negotiation using 9 items. An example item is “During the negotiation, I felt that I was in control of the situation”, (1 = totally disagree and 7 = totally agree). These 9 items were averaged into an index of total experienced power (α =.75). Refer to Appendix B1 for items. The relative power scale assesses an individual’s sense of power compared to their opponent during the negotiation using 9 items. An example item is: “Who do you think had the strongest position in the negotiation?”, (1 = totally the other to 7 = totally myself). Items were averaged into an index of relative power (α =.90). Refer to Appendix B2 for items. Perception of target emotion. To assess whether participants perceived a difference in target affect across conditions, individual’s perception of the target’s emotion was assessed. These items are based on Van Kleef et al.’s (2006) perception of opponent emotion measure. To assess the perceived anger of the target, participants rated how angry and irritated their opponent seemed. An example item is: “the other party appeared angry during the negotiation,” (1 = totally disagree to 7 = totally agree). Items were then averaged into a perceived opponent anger index (α = .95). Refer to Appendix D for items. To assess the perceived happiness of the target, participants were asked to rate how happy, joyful, and enthusiastic their opponent seemed. An example item is: “the other party appeared joyful during the negotiation,” (1 = totally disagree to 7 = totally agree). Items were averaged into a perceived happiness index (α = .95). Perceptions of target emotion were used as 42 manipulation checks to assess whether participants responded according to their assigned condition. Refer to Appendix D for items. Actor affective reactions. Participants’ own experience of anger and happiness were measured using Weiss, Suckow, and Cropanzano’s (1999) assessment of emotions questionnaire. Weiss et al. (1999) developed their measure based off Shaver, Schwartz, Kirson, and O’Connor’s (1987) 200 emotion word dictionary. Four primary emotion words were used to assess emotion: anger, aggravated, happiness, and joyful. Per Weiss et al. (1999), each emotion word was preceded by the question stem “Please indicate how you feel about what just happened within the negotiation,” (1 = not at all to 7 = very much). Participants rated affective reactions each round and reliabilities were calculated by averaging alphas across each round. For the happiness index, α = .94. For the anger index, α = .87. See Appendix E for items. Actor’s affective reaction was also assessed using Watson et al.’s (1998) positive affect and negative affect scale (PANAS). Participants rated the extent to which they felt an emotion during the negotiation scenario from 1 = very slightly or not at all to 5 = extremely. For a list of words, refer to Appendix K. Ultimately, an EFA revealed that actor’s positive affective reaction should be calculated using the Weiss et al. (1999) measure solely. EFAs of the four positive affect items across the Weiss and Watson scales all suggested a two-factor structure. The first factor was composed of the two Weiss items, that accounted for 50% of the variance (factor loadings for these items ranged from .87 and 1.03). The second factor was composed of the Watson (i.e., PANAS) items, which accounted for an average of 21% of the variance (factor loadings ranged from .34 to .76). Because the Weiss items were most aligned with the intended construct measured (i.e., positive affective state) and the PANAS items indicated a second, unintended structure with lower loadings, the Weiss items were utilized as positive affective state 43 indicators for each round. On the other hand, an EFA revealed the four negative affect items composed of the Weiss et al. (1999) and Watson et al’s (1998) measure resulted in a single factor structure. This resulted in four items (2 from each scale) used to calculate participant’s negative affective reactions for each round. Across these items, factor loadings ranged from .80 to .92 and items accounted for 75% of variance in the factor. Actor’s positive and negative affective reactions were measured at the end of each round, for all three rounds. Each round then has a corresponding aggregate score for actor’s positive and negative affective reactions. Round one, two, and three positive and negative affective reactions were then averaged to represent overall positive and negative affective reactions throughout the negotiation scenario, respectively. Correlations show that round scores were significantly correlated with subsequent and prior rounds. The correlation between round one positive affect and round two positive affect was r = .52, p < .05, the correlation between round two and round three positive affect was r = .52, p < .05, and the correlation between round one and round three positive affect was r = .37, p < .05. The correlation between round one negative affect and round negative affect was r = .67, p < .05. The correlation between round two and round three negative affect was r = .62, p < .05, and the correlation between round one and round three negative affect was r = .37, p < .05. Together, these correlations support the aggregation of round affective responses into a single composite value for both positive and negative affect. These aggregated positive and negative affective reactions were then used for bivariate, moderation, and SEM analyses. Information processing. Information processing was assessed using four items from Van Kleef et al. (2004). These items assess the degree to which participants seek and process the information given to them throughout the negotiation scenario. Items were averaged into an 44 index of information processing. Example items from Van Kleef et al. (2004) include: “During the negotiation, I tried to consider all the available information before placing an offer” and “During the negotiation, I made my offers without thinking too much (reverse scored),” (1 = not at all to 7 = very much; α = .59). See Appendix H for items. Need for cognitive closure (NFCC). Need for cognitive disclosure is often used as an assessment of one’s epistemic motivation (e.g. Van Kleef et al., 2004). Need for cognitive closure was assessed using seven items from Webster and Kurglanski’s (1994) NFCC validated scale. Items were adapted for the use of a negotiation scenario and averaged as an index of NFCC. Example items include: “I felt uncomfortable because the other party’s meaning, or intentions, were unclear to me”; “I enjoyed the uncertainty of going into a new situation without knowing what might happen, such as this negotiation”, (1 = strongly disagree to 5 = strongly agree; α = .79). Refer to Appendix I for items. Performance inference. Four items from Van Kleef et al. (2009) were used to assess participant’s performance inference each round. These items assess the inference that participants draw from a target’s emotional display regarding their performance. An example item is: "I feel that the other party was satisfied with my performance", (1 = strongly disagree to 5 = strongly agree). Items were adapted for a negotiation scenario and averaged into an index of total performance inference. One item was unintentionally excluded during the creation of the negotiation application, resulting in three final performance inference items. Participants rated their performance inference each round, but because opponents only respond to participant’s counter offers in the second and third rounds of the negotiation, participant’s performance inference in the first round were not included in the round averages. Correlations show that round one performance inference was related to round two performance inference (r = .48) and 45 round three performance inference (r = .31). Round two and three performance inference were also strongly related (r = .62). Correlations support that round one performance inference is least related to the other rounds. Rather, round one performance inference was included as a control in analysis. Therefore, scale scores and reliabilities were calculated by averaging alphas for round two and three specifically (α = .77). Antecedent-focused emotion regulation strategies. Antecedent-focused regulation strategies were assessed using five items from Gross and John (2003). An example item is “I changed what I was thinking about when I wanted to feel more positive emotion (such as joy or amusement)”, (1 = strongly disagree to 7 = strongly agree; α =. 87). See Appendix F for items. Response-focused emotion regulation strategies. Response-focused regulation strategies were assessed using four Gross and John response-focused, suppression emotion regulation items. Example items from the Gross and John (2003) measure are: “During the negotiation, I controlled my emotions by not expressing them”; “When I felt negative emotions, I made sure not to express them to the other party”, (1 = strongly disagree to 7 = strongly agree; α = .81). See Appendix H for items. Expression of naturally felt emotions (authentic display). To assess whether participants regulated their emotions or leveraged their naturally felt emotions during the negotiation, Diefendorff et al.’s (2004) expression of naturally felt emotions scale will be administered. Example items from the Diefendorff et al. (2004) scale include: “the emotions I expressed to the other party came naturally”, (1 = strongly disagree to 5 = strongly agree; α = .87). See Appendix J for sample items. Manipulation checks. Two manipulation checks were included. As mentioned earlier, perception of target’s emotion served as a manipulation check for target affect. Participants were 46 also asked to answer whether they believed if they were negotiating with real opponents. Data from opponent’s who responded “yes” or “maybe” were used for data analyses to maximize validity. Data Analysis In order to assess whether participants reacted to experimental conditions in the expected direction, A 2 (happy vs. angry opponents) by 2 (single vs. three opponents) ANOVA was conducted to investigate effects on perceptions of target positive, negative affect, own positive and negative affective reactions, and relative and experienced power. Next, structural equation modeling (SEM) was used to test proposed hypotheses. Analyses were run in R-studio using the lavaan package (version 0.6-5), which features a default maximum likelihood estimate. The justification for SEM is the ability to test the proposed model and hypotheses, along with helping evaluate the relative weight of the affective and cognitive pathways proposed by the EASI theory. To test hypothesized interaction effects of relative and experienced power, need for cognitive closure, and information processing, separate regressions were run for positive affect, negative affect, performance inference, response-focused emotion regulation, antecedent-focused emotion regulation, and authentic display. 47 RESULTS Descriptive statistics (mean and standard deviation) and intercorrelations between variables are provided in Table 1. There are some notable correlations. First, deep acting was positively correlated with surface acting (r = .17, p < .05) and authentic display (r = .14, p <.05), and surface acting was negatively correlated with authentic display (r = -.15, p < .05). Deep acting was further related to experienced power (r = .13, p < .05) and relative power (r = .13, p < .05), and negatively related to need for cognitive closure (r = -.15, p < .05). Surface acting was negatively related to experienced power (r = -.12, p < .05). Authentic display was positively related to experienced power (r = .19, p < .05), but not relative power (r = .07, p >.05). Further, authentic display was negatively related to need for cognitive closure (r = -.20, p < .05), and positively related to performance inference (r = .16, p < .05). Manipulation Checks The first set of analyses involved testing whether participants reacted accordingly per condition. Specifically, participants in the happy opponent condition are expected to report significantly greater levels of target positive affect and significantly lower levels of negative affect compared to the angry condition. Participants in the multiple opponent condition were hypothesized to report significantly lower levels of experienced and relative power compared to participants in the single opponent condition. No interactive effects between conditions were expected on any of these outcomes. Refer to Table 2 for a display of marginal means for each cell. A chi-square test was further conducted to investigate whether the proportion of individuals who reported “yes,” “no,” or “maybe” differed by condition. That is, was each condition equally believable by participants? Results suggest significant differences in “yes”, “no”, “maybe” responses by conditions, X2(6, N = 569) = 14.24, p = .03. The happy single opponent condition 48 had the highest “yes” response rate (36%), followed by the single angry opponent condition (26%), multiple happy opponents (21%), and finally multiple angry opponents (18%). The multiple angry opponent condition had the highest rate of “no” responses (28%), followed by multiple happy opponents (26%), single angry opponent (23%), and single happy opponent (23%). Participants in the happy condition reported significantly greater levels of perceived target positive affect (M = 4.18, SE =.08) compared to those in the angry condition (M =2.43, SE = .08), F(1, 322) = 245.15, p < .05. Participants in the happy condition also reported significantly lower levels of perceived negative affect (M =3.58, SE = .08) compared to those in the angry condition (M = 5.53, SE = .09), F(1, 322) = 259.34, p < .05. There were no significant differences in target’s positive affect between participants in the single opponent condition (M = 3.37, SD = .08) compared to the multiple opponent condition (M = 3.24, SE =.08), F(1, 322) = 2.54, p > .05. Participants in the single opponent condition, however, reported significantly lower levels of perceived target negative affect (M = 4.44, SE = .08) compared to the multiple opponent condition (M = 4.67, SE = .09), F(1, 322) = 6.00, p < .05. For perceived target positive affect, a significant interaction between affect and number of opponents emerged, F(1,322) = 6.22, p < .05. Specifically, participants in the happy single opponent condition (M = 4.39, SE = .10) reported higher opponent’s positive affect compared to participants in the multiple opponent condition (M = 3.98, SE = .11); the same pattern did not emerge for the angry condition (Angry, single opponent, M = 2.35, SE = .12; angry multiple opponent M = 2.51, SE = .12). An interaction between target affect and number of opponents also emerged for perceived target negative affect, F(1,322) = 10.74 , p < .05. In this interaction, similar to the previous interaction, participants in the single, happy opponent condition reported lower ratings of target negative 49 perceptions (M = 3.26, SE = .11) compared to participants in the multiple happy opponent condition (M = 3.90, SE = .12). Again, the same pattern did not emerge for the angry condition Angry, single opponent, M = 5.62, SE = .13; angry multiple opponent M = 5.45, SE = .14).Together, these results suggest that participant’s perceptions of opponents affect were in the expected direction. However, there were differences in target affect perceptions between the solo and multiple opponent conditions, suggesting some contamination of the number of opponent factor on participants’ perceived target affect. Therefore, these results suggest controlling for the number of opponents in subsequent analyses. Next, the effects of opponent affect and number of opponents on relative and experienced power were investigated. There were significant differences in relative power between participants in the happy and angry condition, with participations in the happy condition (M = 4.28, SE = .10) reporting significantly lower levels of relative power compared to those in the angry condition (M = 4.60, SE =.11), F(1,322) = 4.59 , p < .05. There was also a significant difference in relative power between participants in the single (M = 4.21, SE = .10) and multiple opponent (M = 4.66, SE =.11) conditions, F(1,322) = 9.40 , p < .05. However, these differences were in the reverse hypothesized direction, such that participants in the solo opponent condition reported lower levels of relative power compared to participants in the multiple opponent condition. There was no significant interaction in conditions on relative power, F(1,322) = .98 , p > .05. A similar trend of findings emerged for experienced power. There was no significant difference in experienced power perceptions between the happy (M = 3.32, SE = .05) and angry (M = 3.29, SE = .05) opponent condition, F(1,322) = .20 , p > .05. A significant difference in experienced power was found between the single (M = 3.22, SE = .05) and multiple opponent (M = 3.38, SE = .05) conditions, F(1,322) = 4.87 , p < .05. Similar to relative power, this difference 50 was in the reverse hypothesized direction, such that participants in the solo opponent condition reported lower levels of power compared to the multiple opponent condition. There was no significant interaction in conditions on experienced power F(1,322) = .31, p > .05. Test of Hypotheses Hypotheses 1 through 17 were investigated utilizing a path model. Refer to Figure 1 for visualization of the model tested. Fit indices for this model were CFI = .94, RMSEA = .07, SRMR = .08, X2 = 115.53, df = 44. A path model was used to investigated hypotheses because the alternative SEM including latent factors for emotion regulation and interaction effects featured much worse model fit (vs. CFI = .72, RMSEA = .09, SRMR = .09, X2 = 2213.34, df = 753) and featured a non-positive covariance matrix. In both models, several covariances were estimated for theoretical reasons. Positive affective reactions, negative affective reactions, and performance inference were allowed to covary. Antecedent-focused emotion regulation, response-focused emotion regulation, and authentic display were also allowed to covary. Further, the interaction terms between information processing and affective reactions and performance inference were allowed to covary with the other interaction terms, and the mediators themselves. Lastly, information processing was allowed to covary with the interaction terms to account for multi-collinearity effects. Table 3 shows the full model, and coefficients corresponding to the estimated paths Hypotheses 1 and 2 proposed that target’s affective display will be positively related to participant’s affective reaction and negatively related to participant’s affective reaction. For analyses, the affect manipulation was coded 0 = angry opponent, 1 = happy opponent. Results indicate that opponent’s affective display was positively related to positive affect (β = .69, p < 51 .05) and negatively related to negative affect (β = -.87, p < .05). Thus, Hypotheses 1 and 2 were supported. Hypotheses 3 and 4 proposed that participants’ positive affective reactions will be negatively related to both response- and antecedent-focused emotion regulation. Results show that positive affective reactions were not related to response-focused emotion regulation (β = - .08, p = .43) and positively related to antecedent-focused emotion regulation (β = .15, p < .05). Further, the relationship between positive affect and authentic display was also investigated and found to be non-significant (β = -.01, p = .93). Therefore, Hypothesis 3 was supported, and Hypothesis 4 was not supported. Hypotheses 5 and 6 proposed that participant’s negative affective reaction will be positively related to both response- and antecedent-focused emotion regulation. Results showed that the relationships between negative affect and response-focused (β = .10, p = .30) and antecedent-focused (β = .09, p = .22) emotion regulation were non-significant. A significant, negative relationship between negative affect and authentic display was found (β = -.15, p < .05). Thus, both Hypothesis 5 and 6 were not supported. Hypotheses 7, 8, and 9 pertained to the relationship between opponent affect and participant’s performance inference, and the relationship between this inference with antecedent- and response-focused emotion regulation. As mentioned, the round 1 measure of performance inference was excluded in the calculation of the performance inference average across rounds. Round 1 performance inference was included as a control. Results show that opponent affect was positively related to participant’s performance inference (β = .79, p < .05), such that participants exposed to positive opponent reported greater performance inference (i.e., believed they were performing well) compared to participants who faced angry opponents. However, performance 52 inference had no significant relationship with response-focused emotion regulation (β = .04, p = .67), antecedent-focused emotion regulation (β = .05, p = .51), nor authentic display (β = .05, p = .43). Taken together, Hypothesis 7 was supported, and Hypotheses 8 and 9 were not supported. Hypothesis 10 proposed that the number of negotiation targets would be negatively related to perceptions of relative power (number of opponents manipulated was coded 0 = single opponent, 1 = multiple opponents). However, this was not the case. Instead, participants in the condition with greater number of opponents reported significantly higher perceptions of relative power (β = .32, p < .05). Therefore, Hypothesis 10 was not supported. Hypothesis 11 proposed that relative power will be positively related to participant’s information processing. This hypothesis was also not supported (β = .00, p = .92). Thus, Hypothesis 11 was not supported. Hypotheses 12, 13, and 14 investigate the moderating effect of information processing on the relationship between opponent’s affect and participant’s negative affect, positive affect, and performance inference. The main of effect of information processing was significant for positive affect (β = -.28, p < .05) and performance inference (β = -.14, p < .05), but not for negative affect (β = .13, p = .08). This suggests that as participants information processing increased, their positive affect and performance inference decreased. The interaction between opponent’s affect and information processing was significant for positive affect (β = .47, p < .05), negative affect (β = -.32, p < .05), and performance inference (β = .21, p < .05). The trend of these interactions was the same. There was no relationship between opponent affect and positive affective state, negative affective state, and performance inference when information processing was low. However, the relationship between opponent affect and positive affective state and performance inference was positively related – and negatively related for opponent affect and negative affective state – when information processing was high. Figure 2, Figure 3, and Figure 4 53 illustrates these effects for positive affective state, negative affective state, and performance inference, respectively. Because the proposed role of information processing was to attenuate the relationship between opponent affect and both positive affective and negative affective states, H12 and H13 were not supported. H14 was supported. Hypotheses 15, 16, and 17 similarly investigated the moderating effect of information processing on the relationship between participant’s negative affect, positive affect, and performance inference on subsequent emotion regulation (i.e., antecedent- and response-focused emotion regulation. Originally, these hypotheses proposed a mediated moderation effect of number of opponents -- through relative power and subsequent information processing -- on the relationship between target’s affective display and participant’s positive affect, negative affect, and performance inference. This moderation effect was also proposed for the relationships between the mediators and antecedent- and response-focused emotion regulation. However, bootstrapped mediation analyses suggest both the indirect (β = .03, p = .16) and direct effects (β = .02, p = .81) of number of opponents on information processing through relative power were non-significant. However, the interaction between information processing and mediators on subsequent emotion regulation was nevertheless investigated. The main effect of information processing was not significant for response-focused emotion regulation (β = .05, p = .46). However, the main effect of information processing was significant for antecedent-focused emotion regulation (β = .27, p < .05) and authentic display (β = .12, p < .05). These results suggest as information processing increased, participants engaged in greater antecedent-focused emotion regulation and authentic display. Next, the moderation effects of information processing on the relationship between positive affect and emotion regulation were investigated. Results suggest information processing did not moderate these 54 relationships for antecedent-focused (β = -.06, p = .41), response-focused emotion regulation (β = .06, p = .49), or authentic display (β = .03, p = .69). Results also suggest information processing did not moderate the relationship between negative affect and emotion regulation for antecedent-focused (β = -.12, p = .08), response-focused emotion regulation strategies (β = .14, p = .11), or authentic display (β = -.01, p = .86). Lastly, information processing did not moderate the relationship between performance inference and emotion regulation for antecedent-focused (β = .12, p = .06), response-focused emotion regulation (β = .11, p = .17), and authentic display (β = .02, p = 72). Altogether, results suggest H15, H16, and H17 were not supported. Exploratory Analyses. Because the number of opponent manipulation did not work as intended, the relationship between information processing and relative power was not significant, and the mediation analyses do not show mediation, the moderation effects of information processing, epistemic motivation, experienced power, and relative power were further explored using regression analyses. Separate models investigating the moderation role of these predictors for positive affect, negative affect, performance inference, emotion-regulation, and authentic display are investigated next. Positive Affect. First, the moderating effects of information processing, epistemic motivation, experienced power, and relative power were investigated for positive affect. Information processing and epistemic motivation were posited to attenuate the positive relationship between opponent’s affect and participant’s felt positive affect, because they represent a boundary for more automatic contagion processes. The expected relationship between opponent’s affect and positive affect will be strengthened by experienced power and relative 55 power because greater power is posited to decrease motivation to process information, leading to greater reliance on affective cues by the participant. Refer to Table 4 for summary of results. Information processing and epistemic motivation. The moderation effect of information processing was first investigated. In this model opponent’s affect was positively related to positive affect (β = .82, p < .05). The main of effect of information processing was also significant (β = -.28, p < .05). This suggests that as participants information processing increased, their positive affect decreased. Lastly, the interaction between opponent’s affect and information processing was significant (β = .47, p < .05). Simple slope analyses showed that opponents affect had no relationship with positive affect when information processing was low (β = .15, p = .23). However, participants with greater information processing reported lower positive affect facing angry opponents and greater positive affect facing happy opponents (β = 1.49, p = .00). Figure 5 illustrates this effect. Next, the moderation effects of epistemic motivation (need for cognitive closure) was investigated. In this model, opponents’ affect was positively related to positive affect (β = .77, p < .05). The main effect of epistemic motivation was also significant (β = -.25, p < .05). This suggests that as participants epistemic motivation increased, their positive affect decreased. Lastly, the interaction between opponent’s affect and epistemic motivation was not significant (β = .01, p = .89). Experienced and relative power. The moderation effect of experienced power was then investigated. Increased experienced power was posited to strengthen the relationship between opponent’s affect and participants’ positive affective reaction. Opponents’ affect was positively related to positive affect (β = .81, p < .05). The main effect of experienced power on positive affect was not significant (β = .04, p = .62). Lastly, the interaction effect of experienced power and opponent affect was not significant (β = .05, p = .61). For relative power, opponents’ affect 56 was positively related to positive affect (β = .83, p <.05). Like experienced power, the main effect of relative power was not significant (β = .09, p = .24). Lastly, the interaction between relative power and opponents’ affect was not significant (β = -.09, p = .39). Negative Affect. Next, the moderating effects of information processing, epistemic motivation, experienced power, and relative power were investigated for negative affect. Information processing and epistemic motivation were posited to attenuate the negative relationship between opponent’s affect and participant’s felt negative affect because they represent a boundary for more automatic contagion processes. The relationship between opponent affect and negative affect will be strengthened by perceived experienced power and relative power because greater power is posited to decrease motivation to process information, leading to greater reliance on affective cues by the participant. Refer to Table 5 for summary of results. Information processing and epistemic motivation. Opponent affect was negatively related to negative affect (β = -.85 p < .05). Information processing was not significantly related to negative affect (β = .10, p = .10). Lastly, the interaction between opponent’s affect and information processing was significant (β = -.27, p < .05). Refer to Figure 6 for visualization. Simple slopes revealed that when information processing was low, opponent’s affect was negatively related to negative affect (β = -.45, p < .00), with this effect being much larger for those with high information processing (β = -1.25, p < .00). The next moderation effect investigated involved epistemic motivation. In this model opponent’s affect was negatively related to negative affect (β = -.81, p <.05). The main of effect of epistemic motivation was not significant (β = .06, p = .32). Lastly, the interaction between opponent’s affect and epistemic motivation was not significant (β = .11, p = .21). 57 Experienced and relative power. The moderation effect of experienced power was investigated next. Experienced power is posited to strengthen the relationship between opponent’s affect and participants’ negative affective reaction. Opponents’ affect was negatively related to negative affect (β = -.85, p <.05). The main effect of experienced power on negative affect was not significant (β = .04, p = .50). Lastly, the interaction effect of experienced power and opponent affect was not significant (β = -.14, p = .10). For relative power, opponents’ affect was negatively related to negative affect (β = -.85, p < .05). Like experienced power, the main effect of relative power was not significant (β = .00, p = .96). Lastly, the interaction between relative power and opponents’ affect was not significant (β = .01, p = .92). Performance Inference. The following models investigated the moderating effects of information processing, epistemic motivation, experienced power, and relative power on the relationship between opponent’s affect and participant’s performance inference. Information processing and epistemic motivation are hypothesized to strengthen the relationship between opponents affect and performance inference because they represent increased information seeking by the participant, presumably about their performance. On the other hand, experienced and relative power are hypothesized to attenuate the relationship between opponent’s affect and performance inference because participants are less motivated to draw inferences about their performance when their power perception is high. As mentioned, performance inference in round 1 were included as controls. Refer to Table 6 for summary of results. Information processing and epistemic motivation. The first model investigated the effects of information processing. Opponents’ affect was positively related to performance inference (β = .65, p < .05), suggesting that participants with happy opponents believed they performed better. Information processing was negatively related to performance inference (β = - 58 .11, p < .05), suggesting that as participants’ information processing increased, their performance inference decreased. The interaction of opponents’ affect and information processing also emerged significant (β = .16, p < .05). Simple slopes show that opponent’s affect was positively related to performance inference when information processing was low (β = .41, p < .05), and this positive relationship was stronger for those with high information processing, but to a larger extent (β = .89, p < .05). Refer to Figure 7 for visualization of this effect. For epistemic motivation, opponents’ affect was positively related to performance inference (β = .63, p < .05). The main effect of epistemic motivation on performance inference was also significant (β = -.11, p < .05). That is, as participant’s epistemic motivation increased, their performance inference decreased. However, the interaction between epistemic motivation and opponent affect was not significant (β = .01, p = .91). Experienced and relative power. The moderation effects of experienced and relative power were next investigated. For the experienced power model, opponent affect was positively related to performance inference (β = .65, p < .05). The effect of experienced power was not significant (β = .00, p = .97). Lastly, the interaction effect between experienced power and opponent’s affect was significant (β = .16, p < .05). This interaction suggests that opponents’ affect was positively related to performance inference when experienced power was low (β = .55, p < .05), and this positive relationship was stronger when experienced power was high (β = .75, p < .05), albeit a larger effect. This interaction is not consistent with the expectation that low power strengthens the relationship between opponent’s affect and participant’s performance inference. Refer to Figure 8 for visualization. For relative power, the main effect of opponent’s affect and performance inference was significant (β = .65, p < .05), and the main effect of relative power on performance inference was 59 not significant (β = -.03, p = .54). Further, the interaction effect of opponent’s affect and relative power was not significant (β = .10, p = .16). Response-focused Emotion Regulation. The following models investigated the moderation effects of information processing, epistemic motivation, experienced power, and relative power on the relationship between opponent affect, positive affect, negative affect, and performance inference on response-focused emotion regulation strategies. Information processing (and epistemic motivation, by extension) is posited to attenuate the relationship between positive affect and negative affect and response-focused emotion regulation. On the other hand, greater information processing is posited to strengthen the relationship between performance inference and response-focused emotion regulation. Like the prior models, the reverse relationship is expected for both experienced and relative power because these factors are posted to decrease motivated information processing, thereby increasing the influence of affect on behavior. Refer to Table 7 for summary of results. Information processing and epistemic motivation. The first model investigated is information processing. In this model, opponent’s affect was negatively related to response- focused strategies (β = -.36, p = .05), suggesting that happy opponents were associated with less response-focused emotion regulation. The main effect of information processing on response- focused emotional regulation was not significant (β = .19, p = .14). Further, the interaction effect between information processing and opponent affect was also not significant (β = -.24, p = .19). The main effects of positive affect (β = -.08, p = .43), negative affect (β = .08, p = .42), and performance inference (β = .04, p = .62) were not significantly related to response-focused emotion regulation. Lastly, the interaction effect between information processing and positive affect (β = .05, p = .60) and negative affect (β = .08, p = .40) were not significant. Further, the 60 interaction between information processing and performance inference was not significant (β = .15, p = .09). Similar trends emerged for epistemic motivation. Opponent affect was negatively related to response-focused strategies at the .1 level (β = -.32, p = .07). The main effect of epistemic motivation was not significant (β = -.05, p = .70). The interaction effect of epistemic motivation and opponent’s affect was also not significant (β = .15, p = .43). The main effect of positive affect (β = -.11, p = .24), negative affect (β = .04, p = .72), and performance inference was not significant (β = .03, p = .71). The interaction effects of epistemic motivation and positive affect (β = .03, p = .75), negative affect (β = -.09, p = .35), and performance inference were also not significant (β = -.03, p = .75). Experienced and relative power. For experienced power, opponent’s affect was negatively related to response-focused emotion regulation at the .1 level (β = -.33, p = .06). Experienced power was also significant related at the .1 level (β = -.22, p = .06). The moderating effect of experienced power and opponent’s affect was not significant (β = .16, p = .39). The main effect of positive affect (β = -.09, p = .34), negative affect (β = .08, p = .42), and performance inference (β = .05, p = .58) and response-focused emotion regulation was not significant. Further, the moderation effect of experienced power and positive affect (β = .06, p = .48), negative affect (β = .05, p = .58), and performance inference (β = .08, p = .30) were also not significant. The final model investigated the moderating effects of relative power. Again, opponent’s affect was negatively related to response-focused strategies at the .1 level (β = -.34, p = .06). The main effect of relative power was not significant (β = -.18, p = .16). The interaction between relative power and opponent affect was also non-significant (β = .22, p = .23). The main effects of positive affect (β = -.09, p = .36), negative affect (β = .08, p = .40), and performance inference 61 were not significant (β = .02, p = .83). Lastly, the interaction effects of relative power and positive affect (β = .06, p = .45), negative affect (β = .05, p = .56) were non-significant. Lastly, the interaction between relative power and performance inference was not significant (β = .13, p = .10). Antecedent-focused Emotion Regulation. The following models investigate the moderation effects of information processing, epistemic motivation, experienced power, and relative power on the relationship between opponent affect, positive affect, negative affect, and performance inference on antecedent-focused emotion regulation strategies. Information processing (and epistemic motivation, by extension) is posited to attenuate the relationship between positive affect and negative affect and antecedent-focused emotion regulation. On the other hand, greater information processing is posited to strengthen the relationship between performance inference and antecedent-focused emotion regulation. Like the prior models, the reverse relationship is expected for both experienced and relative power because these factors are posted to decrease motivated information processing, thereby increasing the influence of affect on behavior. Refer to Table 8 for summary of results. Information processing and epistemic motivation. The first model investigated was information processing. In this model, opponent’s affect was not related to antecedent-focused strategies (β = -.20, p = .15). The main effect of information processing on antecedent-focused emotion regulation was significant (β = .33, p < .05), suggesting that greater information processing was associated with greater antecedent-focused emotion regulation. The interaction effect between information processing and opponent affect was not significant (β = -.11, p = .42). The main effects of positive affect (β = .15, p < .05) was significantly related to antecedent- focused emotion regulation. This suggests that greater positive affect was associated with greater 62 antecedent-focused strategies. Negative affect (β = .09, p = .26) and performance inference (β = .05, p = .51) were not significantly related to antecedent-focused emotion regulation. Lastly, the interaction effect between information processing and positive affect (β = -.06, p = .39) and performance inference (β = -.10, p = .14) were not significant. However, the interaction between information processing and negative affect was significant (β = -.15, p = .05). This interaction suggests that the relationship between negative affect and antecedent-focused emotion regulation was non-significant when information processing is high (β = -.13, p = .36). However, the relationship between negative affect and emotion regulation is positive when information processing is high (β = .31, p < .05). This relationship is consistent with expectations in that greater information processing attenuated, or in this case nullified, the relationship between affect and emotion regulation when information processing is high. Refer to Figure 9 for visualization. Similar trends emerged for epistemic motivation. Opponent affect was not related to antecedent-focused strategies (β = -.18 p = .21). The main effect of epistemic motivation was not significant (β = -.14, p = .16). The interaction effect of epistemic motivation and opponent’s affect was also not significant (β = .03, p = .83). The main effect of positive affect (β = .17, p < .05) was significant. However, the main effect for negative affect (β = .08, p = .29), and performance inference was not significant (β = -.01, p = .84). The interaction effects of epistemic motivation and positive affect (β = -.02, p = .83) and negative affect (β = -.12, p = .12) were not significant. However, the interaction between epistemic motivation and performance inference was significant (β = -.21, p < .05). Simple slopes show that the relationship between performance inference and antecedent-focused emotion regulation was positive when epistemic motivation was low (β = .16, p = .07), but negative when epistemic motivation was high (β = -.19, p = .07). 63 This interaction is consistent with expectations in that greater epistemic motivation reversed the relationship between performance inference and emotion regulation. Refer to Figure 10 for visualization. Experienced and relative power. For experienced power, opponent’s affect was not related to antecedent-focused emotion regulation (β = -.20, p = .16). Experienced power was also not related (β = -.15, p = .12). The moderating effect of experienced power and opponent’s affect was not significant (β = .00 p = .99). The main effect of positive affect was significant at the .1 level (β = .14, p = .06). Negative affect (β = .08, p = .42), and performance inference (β = -.01, p = .92) emotion regulation was not related to antecedent-focused emotion regulation. Further, the moderation effect of experienced power and positive affect (β = -.11, p = .14), negative affect (β = -.08, p = .29) were not significant. However, the interaction between experienced power and performance inference was significant (β = .12, p < .05). This interaction suggests the relationship between performance inference and antecedent-focused emotion regulation negative when experienced power is low (β = -.09, p = .31) and positive when experienced power is high (β = .07, p = .39). This is consistent with expectations in that the relationship between performance inference and emotion regulation is reversed when experienced power is high, and consistent with prior performance inference findings for epistemic motivation. Refer to Figure 11 for visualization. The final model investigated the moderating effects of relative power. Opponent’s affect was not related to emotion regulation (β = -.15, p = .29). The main effect of relative power was significant (β = .25, p < .05), suggesting that greater relative power was linked to greater antecedent-focused emotion regulation. The interaction between relative power and opponent affect was non-significant (β = -.13, p = .38). Next, the main effects of positive affect (β = .15, p 64 = .05), negative affect (β = .07, p = .34), and performance inference were not significant (β = - .01, p = .91). Lastly, the interaction effects of relative power and positive affect (β = -.10, p = .11), negative affect (β = -.01, p = .90) were non-significant. However, the interaction between relative power and performance inference was significant (β = .21, p < .05). Simple slopes suggests the trend for this interaction is that the relationship between performance inference and antecedent-focused emotion regulation was negative when relative power was low (β = -.30, p < .05) and positive when relative power was high (β = .28, p < .05). Again, this relationship is in the expected direction and consistent with the prior experienced power interaction. Refer to Figure 12 for visualization. Authentic Display. The final analyses investigated the moderation roles of information processing, epistemic motivation, experienced power, and relative power on participants’ authentic display. These relationships are exploratory and not hypothesized, and instead serve as supplementary research questions. Therefore, there are no expected directions or results when investigating factors and authentic display. Refer to Table 9 for summary of results. Information processing and epistemic motivation. The first model investigated is information processing. In this model, opponent’s affect was not related to authentic display (β = .14, p = .29). The main effect of information processing on antecedent -focused emotion regulation was not significant (β = .01, p = .95). The interaction effect between information processing and opponent affect was not significant (β = .20, p = .14). The main effects of positive affect (β = -.01, p = .86) was not significantly related to authentic display. Negative affect was significantly related to authentic display at the .1 level (β = -.14, p = .06). Lastly, performance inference was not significantly related to authentic display (β = .05, p = .45). Lastly, the interaction effect between information processing and positive affect (β = .03, p = .63), 65 negative affect (β = .03, p = .67), and performance inference (β = -.01, p = .85) were not significant. For epistemic motivation, opponent’s affect was not related to authentic display (β = .12, p = .37). The main effect of epistemic motivation on antecedent-focused emotion regulation was significant at the .1 level (β = -.18, p = .06). The interaction effect between epistemic motivation and opponent affect was not significant (β = .05, p = .71). The main effects of positive affect (β = .01, p = .84) was not significantly related to authentic display. Negative affect was significantly related to authentic display (β = -.15, p = .05), suggesting that greater levels of negative affect was negatively associated with participant’s authentic display. Lastly, performance inference was not significantly related to authentic display (β = .01, p = .83). Lastly, the interaction effect between epistemic motivation and positive affect (β = -.03, p = .65), negative affect (β = -.03, p = .64), and performance inference was not significant (β = -.13, p = .07). Experienced and relative power. The following model investigated experienced power. Opponent’s affect was not related to authentic display (β = .14, p = .30). The main effect of experienced power on authentic display was significant at the .1 level (β = .15, p = .08). The interaction effect between experienced power and opponent affect was not significant (β = .08, p = .57). The main effects of positive affect (β = .00, p = .99) was not significantly related to authentic display. Negative affect was significantly related to authentic display (β = -.16, p < .05), suggesting that greater levels of negative affect was negatively related to participant’s authentic display. Performance inference was not significantly related to authentic display (β = .01, p = .88). Lastly, the interaction effects between experienced power and positive affect (β = .02, p = .78), negative affect (β = .03, p = .66), and performance inference (β = .06, p = .29) were not significant. 66 The final model investigated the moderating effects of relative power on authentic display. Opponent’s affect was not related to emotion regulation (β = .15, p = .25). The main effect of relative power was significant (β = .23, p < .05), suggesting that greater relative power was linked to greater authentic display. The interaction between relative power and opponent affect was not significant (β = -.23, p = .1). The main effect of positive affect (β = -.01, p = .85) and performance inference was not significant (β = .05, p = .47). The main effect of negative affect on authentic display was significant (β = -.16, p < .05). Lastly, the interaction effects of relative power and positive affect (β = -.03, p = .66), negative affect (β = .00, p = .95), and performance inference were non-significant (β = .06, p = .27). 67 DISCUSSION Research on antecedents of emotion regulation and emotional labor is scarce. As process models of emotional labor continue to develop, increased focus on these antecedents and the mechanisms tying them to subsequent emotional labor is critical for continued investigation of emotion regulation in the workplace. This study sought to contribute to this effort by applying the EASI dual-process framework to investigate the relationship between a negotiation opponent’s affect on participant’s emotion regulation (i.e., antecedent- and response-focused emotion regulation, and authentic display). The number of opponents was also manipulated to investigate the role of unequal power perceptions, information processing, and the relationship between affective and cognitive factors on emotion regulation. Results suggest that participants experienced emotion contagion from opponents (i.e., participants with happy opponents reported greater positive affective states compared to those with angry opponents, and vice-versa.) and also drew inferences regarding their performance from opponent’s affect (i.e., participants with happy opponents reported a greater performance inference than those with an angry opponent). However, only significant relationships between participants’ affective state and emotion regulation were found, and participants’ performance inference was not related to emotion regulation strategies. The relationship between the number of opponents manipulation did not result in the expected effect; rather, participants facing multiple opponents felt more empowered. However, results suggest that the degree of participants’ information processing altered the relationship between opponent’s affect and participant’s affective state and performance inference. Moreover, participant’s epistemic motivation, experienced power, and relative power moderated the relationship between these factors and emotion regulation, suggesting that participants engaged 68 in different levels of emotion regulation in response to their own affective state depending on perceptions of their power, generally and relative to their opponent. The following segments discuss the implications for future emotion regulation research, practical implications, and limitations and future research directions. Emotion Regulation and the EASI Framework Results from this study corroborate the assertions of the EASI framework (Van Kleef, 2009; 2010). Participants’ affective state were largely influenced by their opponent’s affective display, such that happy opponents increased participant’s positive affective state and angry opponents triggered increased participant’s negative affective state. These findings are consistent with prior research that a target’s affective display affects observer’s reactions through mechanisms such as emotion contagion and mirror neurons (Hatfield, Cacioppo, & Rapson, 1994). These findings are also consistent prior research concerning contagion effects of customer mistreatment on employees’ affective state. Specifically, customer mistreatment leads to lower positive affect states and greater negative affect states on employees (Rupp & Spender, 2006; Totterdell & Holman, 2003). Results also supported the assertion that observers used target’s affective display to draw inferences about their performance, aligning with prior work linking the relationship between a target’s affect and observer’s performance inference (e.g., Keltner & Haidt, 1999; Van Kleef et al., 2009; Weiss & Cropanzano, 1996). Altogether, this study provides further evidence supporting the EASI model’s assertions that target’s affective displays influence observer’s own affective states and also convey information used to form expected performance inference. In investigating the relationship between affective and cognitive factors and subsequent emotion regulation, results suggest that affective factors were predictive of emotion regulation 69 while cognitive factors were not. Participants’ positive affective state was found to be positively related to antecedent-focused emotion regulation and negatively related to response-focused emotion regulation. A negative affective state was, on the other hand, negatively related to authentic display. Regardless, performance inference was found to be unrelated to emotion regulation strategies. These results provide interesting insight on emotion regulation through a dual-process lens. Participants’ antecedent- and response-focused emotion regulation seemed to be driven by a positive affective state. This is partly consistent with Totterdell and Holman’s (2003) study that found participant’s happiness and enthusiasm positively related to antecedent- focused emotion regulation. However, they did not distinguish between participant’s positive and negative affective states. Other prior research has found that response-focused strategies are far more likely to occur following negative events (Diefendorff et al., 2008; Grandey et al., 2002). Results also show a negative affective state consistently, negatively predicted authentic display. Results suggest a dual-goal structure to positive display rules: displaying positive emotions and affect, and suppressing/masking negative emotions and affect. Participants’ antecedent-focused emotion regulation was predicted by a positive affective state, suggesting that participants leveraged positive affective reactions to help them adhere to display rules. On the other hand, a low positive affective state was related to greater response-focused strategies and a negative affective state was negatively related to participants' authentic display. Together, these findings suggest a greater negative affect state was associated with suppression, which is consistent with prior research that has found that asking participants to display positive affect and emotions lead to increased deep acting and asking participants to suppress negative affect and emotions lead to greater surface acting (Kammeyer-Mueller et al., 2013; Mesmer-Magnus et al., 2012). 70 Information Processing and Emotion Regulation Hypotheses posited that participants who faced several opponents would feel less powerful than participants facing a single opponent. This is because the former group was hypothesized to perceive that several opponents deciding on a bargain together have an information and decision-making advantage compared to those negotiating alone. However, in the current study, participants who faced several opponents instead perceived themselves in a more powerful negotiation position. There are two potential methodological reasons for this effect. First, the sample was drawn on a university participant pool featuring primarily psychology undergraduate students. Thompson’s (2011) study -- showing that bargaining teams are perceived as having an advantage over solo bargainers -- was found in a sample of experienced negotiators. Participants in Thompson’s study were presumably more familiar with negotiating than this study’s undergraduate sample. Thus, participants in Thompson’s study were likely to be more aware of the negotiation process and why multiple opponents represented a power imbalance. Second, the experimental task automatically presented bargains that approached participant’s bargaining goal each round. Participants in the multiple opponent condition likely perceived that as the negotiation continued, their opponents’ offer was getting closer to meeting their negotiation threshold (in the final round, the opponent’s offer met one of the participants two demands). These participants likely noticed and believed that they persuaded several opponents to come to an agreement and reasoned they were more powerful and influential. Nevertheless, the moderating role of information processing on opponent’s affect on participants’ affective states and performance inference was investigated. Results suggest higher levels of information processing strengthened the effect of opponent affect and participants’ 71 positive affective state, negative affective state, and performance inference. These results support the assertions of the EASI model and prior studies. Van Kleef (2009; 2010) proposed that individuals react to emotion displays based on the degree to which they are motivated and able to process information. So, those with low information processing are less likely to be influenced by displays of affect because they are less motivated to understand and use the associated information displayed to them. Van Kleef et al. (2008) showed that participants with lower information processing reported less distress and compassion in reaction to a personal, emotionally distressing story by someone else. Van Kleef et al. (2006) also showed that individuals in higher power positions did not concede bids as often and made smaller bids during a negotiation compared to those with higher information processing. Our study then supports propositions of the EASI model (Van Kleef, 2009;2010), in that the influence of a target’s affect on inferences and affective state varied on the individual’s motivation to use emotional displays (and their information) to their advantage. Information processing moderated the relationship between performance inference and negative affective state and antecedent focused emotion regulation in exploratory analyses. The relationship between a negative affective state and antecedent-focused emotion regulation was not significant when information processing was high, but a negative affective state was positively related to emotion regulation when information processing was low (refer to Figure 6 for this effect). The positive relationship between a negative affective state and emotion regulation aligns with Diefendorff and Gosserand’s (2003) application of control theory to emotional labor. For these participants, a greater negative affective state likely signaled a discrepancy between one’s expected emotion display and internal emotions, triggering a self- regulation process aimed at aligning internal emotions with external expectation demands. 72 However, the effect of a negative affective state on antecedent-focused emotion regulation was negligible when information processing was high, offering support that greater information processing attenuates the effect of affective reactions on behavior. This finding aligns with Van Kleef’s (2009) assertions that automatic contagion processes may be undermined when greater effort and focus is dedicating to evaluating information, rendering the affective pathway as non- predictive for behavior, a finding that has been corroborated by prior research (Van Kleef et al., 2009). Need for cognitive closure was also measured to capture participant’s epistemic motivation. Theoretically, epistemic motivation and information processing are thought to co- occur: as epistemic motivation increases (i.e., their desire to understand more about their situation and environment), so does information processing (i.e., sense-making) (De Dreu et al., 2003; Van Kleef et al., 2004; Van Kleef, 2009;2010). However, bivariate correlation results and path analyses did not support this relationship. There are some potential explanations for these findings. The information processing items may not be capturing motivated sense-making and information seeking. For example, these two items: “During the negotiation, I hardly thought about my goals” and “During the negotiation, I made my offers without thinking too much” potentially capture both if participants engaged in information processing during the negotiation and participant’s overall engagement during the experiment. The other two information processing items are more focused on participant’s motivated information seeking during the negotiation, but may still be contaminated by with low engagement to the task (i.e.,“During the negotiation, I paid a lot of attention to the information about the negotiator’s intentions”; “During the negotiation, I tried to consider all the available information before placing an offer”). On the other hand, the epistemic motivation scale seemed to measure dispositional, 73 rather than situational, epistemic motivation (i.e, “I dislike unpredictable situations such as this negotiation”; “I don't like situations that are uncertain such as this negotiation.”), which may not accurately capture epistemic motivation during the experiment but still offer important insight on participant’s propensity to engage in deeper information processing and sense making. The interaction effect of epistemic motivation and performance inference on antecedent-focused emotion regulation suggests that performance inference was negatively related antecedent- focused emotion regulation when epistemic motivation was high, and positively related when epistemic motivation was low. One plausible explanation is those with low information processing may have dedicated greater focus to antecedent-focused strategies when they believed they were performing well. Grandey and Melloy (2017) explain that although the expected relationship between positive events at work and emotion regulation is negative, prior emotional labor work has found that positive work events can result in greater antecedent-focused emotion regulation (Totterdell & Holsman, 2003). In this study, low information processing participants may have engaged in emotion regulation to remain congruent with the display rule, and opting for antecedent-focused strategies when they perceived their opponents as happy with their offers. Notably, prior research suggests that response-focused strategies are most prevalent immediately following customer mistreatment (Gabriel & Diefendorff, 2015) and that these strategies are perceived more effective than antecedent-strategies for handling mistreatment (Grandey et al., 2004). Therefore, low information processing individuals seemed to have relied on greater antecedent-strategies when they are performing well, and moved away from these strategies when they perceived their performance as poor. On the other hand, those with greater epistemic motivation may have inferred that antecedent-focused emotion regulation was not necessary when performance is satisfactory. In this case, this relationship more closely aligns with 74 assertions of the EASI model. Specifically, the EASI model (Van Kleef, 2009; 2010) suggests that greater epistemic motivation (i.e., information processing) would lead to greater reliance on the inferences drawn about one’s performance and subsequent behavior. In this case, inferring that one’s performance is satisfactory was associated with less emotion-regulation -- likely because participant’s perceived that their opponents were satisfied with the offers presented. This result would then align with prior research showing that teams with a high average epistemic motivation decreased effort when leaders displayed positive affect, because this display suggested that performance was satisfactory (Van Kleef et al., 2009). Relative and Experienced Power and Emotion Regulation The effects of experienced power and relative power were also investigated. Power is posited to affect an individual’s desire and motivation to process an opponent’s affective display, and their subsequent performance inference and opponent’s intentions. First, results suggest that relative power was associated with greater response-focused emotion regulation and authentic display. Therefore, results suggest that antecedent-focused emotion regulation and displaying one’s authentic emotions were emotion regulation strategies adopted when individuals felt empowered relative to their opponent. These results suggest that participants were more willing to engage in strategies that are more time consuming and anticipatory (in the case of antecedent- focused regulation), or believe displaying their authentic emotions were appropriate to display. The interaction between experienced power and performance inference on antecedent-focused emotion regulation offers support for this assertion, such that those with lower experienced power perceptions engaged in less antecedent-focused strategies compared to those with high power perceptions. Further, relative power seemed to be an important moderator when considering the relationship between performance inference and antecedent-focused emotion 75 regulation. For antecedent-focused emotion regulation, participants who perceived themselves as performing well were more likely to adopt antecedent-focused emotion regulation strategies when relative power perceptions were high. On the other hand, low power participants performed lesser antecedent-focused emotion regulation when performance inference was high. Overall, greater relative power was associated with greater antecedent-focused emotion regulation and lower power with lesser antecedent-focused emotion regulation when performance inference is high. These results also align with the previous interactions suggesting participants' emotion regulation was decreased when performance inferences was high for participants with greater information processing, likely because they inferred that their performance is satisfactory and that emotion regulation is not required (Van Kleef, 2009a; 2009b; 2010). On the other hand, those with lower information processing seemed to have continued engaging in emotion regulation even when their performance inference was high, potentially because antecedent-focused strategies tend to be adopted during positive interpersonal events. Theoretical and Practical Implications These findings hold theoretical and practical implications for emotion regulation research. First, results seem to support a dual-goal structure of positive display rules, with an individual’s positive affective state being a significant driver for antecedent-focused emotion regulation strategy engagement and negative affective states associated with lower authentic display. Thus, future studies should consider the unique effects of both positive and negative affective states and reactions on emotion regulation (or lack thereof, in the case of authentic display). Positive and negative affective states may lead to different, respective outcomes and relationships with emotion regulation depending on the intensity of reactions. Results also 76 suggest that affective states drive subsequent emotion regulation and that emotion regulation is not significantly driven by one’s performance inference. These results may have some relevance for the application of control theory to emotion regulation (Diefendorff & Gosserand, 2003). Control theory has been applied to emotional labor, suggesting that poorer a performance inference led to greater emotion regulation, driven by the discrepancy between one’s affective state and the display rule standard. However, these results suggest that one’s performance inference did not lead to lesser regulation. Rather, emotion regulation is engaged if the display rule is followed, regardless of one’s performance inference. These results may hold some practical significance for those interested in emotion regulation training. First, results can be used to show the importance of considering one’s affective reactions and inferences when encountering a pleasant or rude customer. Specifically, greater attunement to affective states under conditions that limit information processing can help individuals mitigate stronger reactions during episodes of customer mistreatment. Future research could potentially investigate whether individuals can be trained to rely on performance inference and inferential processing to mitigate affective reactions and transform them into helpful performance implications. Results also help show whether response-focused or antecedent-focused strategies are engaged depending on whether an individual’s interaction partner is in a higher or lower relative standing. Keeping participants in a low power position seems to lead to greater reliance on response-focused strategies, and less authentic display or antecedent-focused strategies. As shown by previous research, response-focused strategies tend to be often related to negative outcomes, while antecedent-focused emotion regulation and authentic displays are considered less stressful and effortful to engage. 77 Limitations and Future Directions There are several limitations to this study. As noted earlier, this study’s sample was drawn from undergraduate psychology students who participated for class credit. Although participants were not asked whether they have prior negotiation experience, it is likely that these students do not have negotiation experience. Therefore, generalizing these results to samples with greater negotiation experience is likely premature. These results also feature participants who reported either believing or maybe believing their opponents were real. To best investigate emotion regulation strategy adoption with simulated opponents, responses from participants reporting “maybe” could represent confounding effects. Future studies should seek to replicate results in a sample that are experienced with negotiation. Further, rather than relying on simulated opponents, future studies ought to investigate with participants negotiating with another real participant and capture emotion regulation strategies from both parties. There are also some limitations with the negotiation task used during the study. First, as noted early, the simulated opponents always came closer to reaching the participant’s bargaining goals each round. Therefore, participants did not have a real influence over their opponents outcome. In the case of this study, offers that increasingly approach the participant’s goals likely gave the impression that whatever emotion regulation strategy adopted was effective and inflated perceptions of power. Future studies should investigate whether the effects remain whether offers are random or increasingly further away from participant’s goal throughout negotiation rounds. Another limitation is that the task measured power perceptions, epistemic motivation, and emotion regulation at one time point after the negotiation. In this case, only differences in affect and performance inference can be observed between rounds. Trends associated with decreasing bargain demands from opponents and the effects of these trends on emotion 78 regulation and power perceptions are unable to be considered, limiting causal interpretation of these factors throughout negotiation rounds. Future studies should investigate the implications of increasing or decreasing experienced power and bargaining demands during a negotiation on emotion regulation. Finally, the proportion of participants who believed each condition varied, suggesting that the most believable condition was those with happy opponents, and those with angry opponents less so. These findings are likely because participants were asked to display positive affective states and suppress negative affective states, suggesting that participants were likely dubious of those who were displaying emotions and affect counter to what they asked to display. Lastly, participants were only instructed to follow positive/integrative positive display rules. Emotion regulation and emotional labor research already feature several studies investigating the role of positive display rules. Studies investigating the effects of negative or neutral display rules are not frequent. This is a significant limitation given that negotiations often feature several different affective displays from both parties. Retaliation through anger or enthusiasm to elicit likeness from an opponent are two strategies that can be perceived as being effective during a negotiation, likely because negotiations are often interdependent and feature quick reactions from opponents to other’s affective display. Therefore, the application of a positive display rule to a negotiation scenario is not common in live settings. Future studies should investigate emotion regulation strategies during negotiations without imposing display rules to best evaluate the effects of affective reactions and perceived performance inference in response to their opponents display. This design would lend itself to greater face validity from participants and greater generalizability to those familiar with negotiating. 79 APPENDICES 80 APPENDIX A: Tables and Figures Table 1. Means, standard deviations, and bivariate correlations for study predictors. Information Processing 1. Antecedent-Focused Emotion Regulation 2. Response-Focused Emotion Regulation 3. Authentic Display 4. 5. Experienced Power 6. Relative Power 7. Need for Cognitive Closure 8. Performance Inference (average) 9. Negative Affect (average) 10. Positive Affect (average) 11. Target Negative Affect (average) 12. Target Positive Affect (average) 13. Performance Inference - round 1 14. Performance Inference - round 2 15. Performance Inference - round 3 16. Pos. affect - round 1 17. Pos. affect - round 2 18. Pos. affect - round 3 19. Neg affect - round 1 20. Neg affect - round 2 21. Neg affect - round 3 22. Target negative affect - round 1 23. Target negative affect - round 2 24. Target negative affect - round 3 25. Target pos. affect - round 1 26. Target pos. affect - round 2 27. Target pos. affect - round 3 28. Power Manipulation 29. Affect Manipulation Note. Bolded values are significant at the .05 level. Table 1 (cont’d) Mean (SD) 5.14(1.09) 5.11(1.48) 3.79(1) 5.42(1.43) 3.3(0.64) 4.4(1.4) 2.87(0.82) 2.9(0.81) 3.26(0.87) 3.98(1.01) 4.43(1.5) 3.42(1.35) 3.11(0.76) 2.89(0.94) 2.91(0.85) 3.68(1.18) 3.92(1.38) 4.33(1.23) 3.26(1) 3.27(1.14) 3.25(0.96) 3.42(1.78) 4.36(1.84) 4.5(1.47) 4.23(1.56) 3.51(1.62) 3.32(1.38) 0.46(0.5) 0.44(0.5) 1 2 3 4 5 6 7 8 9 10 11 1 0.11 0.19 0.07 -0.2 0.16 -0.22 0.14 -0.21 0.15 0.12 0.16 0.12 0.18 0.11 0.05 -0.23 -0.21 -0.1 -0.13 -0.24 -0.13 0.16 0.16 0.12 -0.04 -0.16 1 0.15 0.09 -0.09 0.03 -0.06 -0.05 0 -0.01 0.16 0.04 0.02 -0.07 -0.02 -0.03 -0.11 -0.08 0.04 -0.02 0 -0.01 -0.06 -0.04 0.04 0 0 1 0.68 -0.36 0.15 -0.03 0.07 0.03 -0.05 0.17 0.16 0.11 0 0.06 0.11 -0.02 -0.03 -0.04 -0.01 0.02 0.03 0.03 -0.09 0.01 0.12 -0.02 1 -0.29 0.01 0.08 0 0.15 -0.16 0.09 0.05 -0.03 -0.05 0.04 0.01 0.11 0.06 0.02 0.17 0.14 0.12 -0.11 -0.2 -0.09 0.17 0.12 1 -0.28 0.21 -0.2 0.15 -0.08 -0.27 -0.27 -0.22 -0.1 -0.15 -0.23 0.14 0.19 0.19 0.14 0.15 0.1 -0.08 -0.08 -0.08 -0.03 0.16 1 -0.46 0.44 -0.55 0.54 0.44 0.91 0.89 0.18 0.42 0.44 -0.24 -0.45 -0.47 -0.42 -0.49 -0.52 0.39 0.48 0.5 -0.02 -0.48 1 -0.6 0.66 -0.52 -0.24 -0.45 -0.38 -0.41 -0.57 -0.44 0.81 0.92 0.78 0.54 0.65 0.52 -0.45 -0.52 -0.41 0.1 0.49 1 -0.53 0.62 0.19 0.39 0.4 0.75 0.87 0.76 -0.4 -0.55 -0.56 -0.45 -0.52 -0.44 0.55 0.59 0.53 0 -0.4 1 -0.83 -0.18 -0.48 -0.52 -0.37 -0.46 -0.44 0.47 0.62 0.55 0.67 0.93 0.88 -0.58 -0.77 -0.73 0.1 0.66 1 0.17 0.14 0.09 0.13 0.13 -0.15 0.04 0.02 0.05 0.09 -0.09 0.1 0.06 0 0.04 0.07 0.01 0.02 -0.01 0.05 0.03 0.06 0.11 -0.01 -0.09 -0.08 0.04 0.05 1 -0.15 0.06 -0.12 -0.1 0.06 -0.1 0.1 -0.14 0.07 -0.09 -0.01 -0.11 -0.06 -0.07 -0.17 -0.08 0.09 0.12 0.04 0.11 0.08 0.03 -0.13 -0.11 -0.04 0.05 0.15 81 Information Processing 1. Antecedent-Focused Emotion Regulation 2. Response-Focused Emotion Regulation 3. Authentic Display 4. 5. Experienced Power 6. Relative Power 7. Need for Cognitive Closure 8. Performance Inference (average) 9. Negative Affect (average) 10. Positive Affect (average) 11. Target Negative Affect (average) 12. Target Positive Affect (average) 13. Performance Inference - round 1 14. Performance Inference - round 2 15. Performance Inference - round 3 16. Pos. affect - round 1 17. Pos. affect - round 2 18. Pos. affect - round 3 19. Neg affect - round 1 20. Neg affect - round 2 21. Neg affect - round 3 22. Target negative affect - round 1 23. Target negative affect - round 2 24. Target negative affect - round 3 25. Target pos. affect - round 1 26. Target pos. affect - round 2 27. Target pos. affect - round 3 28. Power Manipulation 29. Affect Manipulation Note. Bolded values are significant at the .05 level. 15 1 0.16 0.33 0.47 -0.15 -0.31 -0.5 -0.37 16 1 0.52 0.3 -0.43 -0.36 -0.25 -0.38 17 1 0.52 -0.38 -0.6 -0.43 -0.41 18 1 -0.14 -0.34 -0.65 -0.28 19 20 21 22 1 1 0.34 1 0.61 0.5 1 0.67 0.37 0.51 12 13 1 0.16 0.45 0.53 0.45 0.53 0.51 -0.36 -0.49 -0.46 -0.67 1 0.48 0.31 0.19 0.18 0.1 -0.27 -0.22 -0.1 -0.26 14 1 0.62 0.17 0.43 0.33 -0.27 -0.49 -0.36 -0.39 82 Table 1 (cont’d) Information Processing 1. Antecedent-Focused Emotion Regulation 2. Response-Focused Emotion Regulation 3. Authentic Display 4. 5. Experienced Power 6. Relative Power 7. Need for Cognitive Closure 8. Performance Inference (average) 9. Negative Affect (average) 10. Positive Affect (average) 11. Target Negative Affect (average) 12. Target Positive Affect (average) 13. Performance Inference - round 1 14. Performance Inference - round 2 15. Performance Inference - round 3 16. Pos. affect - round 1 17. Pos. affect - round 2 18. Pos. affect - round 3 19. Neg affect - round 1 20. Neg affect - round 2 21. Neg affect - round 3 22. Target negative affect - round 1 23. Target negative affect - round 2 24. Target negative affect - round 3 25. Target pos. affect - round 1 26. Target pos. affect - round 2 27. Target pos. affect - round 3 28. Power Manipulation 29. Affect Manipulation Note. Bolded values are significant at the .05 level. 23 24 1 0.64 -0.56 -0.8 -0.57 0.11 0.67 28 27 26 25 1 -0.47 -0.56 -0.79 0.07 0.51 1 0.73 0.52 -0.04 -0.56 1 0.62 -0.06 -0.67 1 -0.05 -0.49 29 -- 0.01 -- 83 Table 2. Cells means for opponent affect and number of opponents manipulation. Target Pos. Affect Target Neg. Affect Pos. Affect Neg Affect Experienced Power Relative Power Note. N = 326 Number of Opponents Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Single Multiple Happy condition Mean (SE) 4.39 (.10) 3.98 (.11) 3.26 (.11) 3.90 (.12) 4.39 (.09) 4.28 (.10) 2.73 (.08) 3.06 (.08) 3.25 (.06) 3.38 (.07) 3.98 (.14) 4.58 (.16) Angry condition Mean (SE) 2.35 (.12) 2.51 (.12) 5.62 (.13) 5.45 (.14) 3.44 (.11) 3.61(.11) 3.76 (.09) 3.70 (.09) 3.19 (.07) 3.39 (.08) 4.45 (.15) 4.74 (.17) 84 Table 3. Results for path model analysis investigating opponent’s affect on cognitive and affective factors and subsequent emotion- regulation. Predictor Number of Opponents Predictor Relative Power Predictor Affect manipulation IP IP*Opponent Affect Performance inference (round 1) Coefficient 0.32 Coefficient 0.00 Relative Power SE 0.11 Information Processing (IP) SE 0.03 p-value 0.00* p-value 0.92 Positive affect Negative affect Performance inference Coefficient 0.69 -0.28 0.47 -- SE 0.10 0.08 0.10 -- p-value Coefficient 0.00* 0.00* 0.02* -- -0.87 0.13 -0.32 -- SE 0.09 0.07 0.10 -- p-value Coefficient 0.00* 0.08† 0.00* -- 0.79 -0.14 0.21 0.31 SE 0.09 0.07 0.09 0.04 p-value 0.00* 0.04* 0.02* 0.00* Response-focused Emotion Antecedent-focused Emotion Authentic Display Regulation Regulation Predictor Affect manipulation Positive Affect Negative Affect Performance Inference IP IP * Positive Affect IP * Negative Affect IP * Performance Inference Note. N = 326; coefficients represent values from path model with all paths simultaneously included; p < .05*, p < .10† p-value Coefficient 0.17 0.05 0.22 0.51 0.00 0.08† 0.41 Coefficient -0.32 -0.08 0.10 0.04 0.05 0.14 0.06 p-value Coefficient 0.06† 0.43 0.30 0.69 0.46 0.11 0.49 -0.18 0.15 0.09 0.05 0.27 -0.12 -0.06 SE 0.13 0.07 0.08 0.07 0.06 0.07 0.07 SE 0.17 0.10 0.10 0.09 0.07 0.09 0.09 0.11 -0.01 -0.15 0.05 0.12 -0.01 0.03 0.06 0.06† 0.02 0.13 0.07 0.07 0.07 0.05 0.07 0.07 0.06 0.11 0.08 0.17 -0.12 SE p-value 0.38 0.93 0.05 0.43 0.02 0.86 0.69 0.72 85 Table 4. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s positive affect. Predictor Intercept Opponent Affect Number of Opponents Main Effect of Moderator Opponent affect*Main Effect Information Processing β 3.49 .82 .06 -.28 .47 SE p-value .09 .10 .10 .07 .10 .00 .00 .56 .00 .00 Note. N = 326, p < .05, p < .10† F = 22.22 r2 = .22 Epistemic Motivation Experienced Power Relative Power SE p-value .09 .10 .10 .07 .12 .00 .00 .93 .04 .89 β 3.54 .77 .01 -.15 .01 17.88 .18 β 3.52 .81 .00 .04 .05 16.07 .17 SE p-value .10 .10 .10 .07 .10 .00 .00 .98 .62 .61 SE .09 .10 .10 .08 .10 β 3.50 .83 .01 .09 -.09 15.98 .17 .00 .00 .93 .24 .39 p-value 86 Table 5. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s negative affect. Predictor Intercept Opponent Affect Number of Opponents Main Effect of Moderator Opponent affect*Main Effect Information Processing p-value .00 .00 .08† .10† .00 β 3.67 -.85 .14 .10 -.27 SE .07 .08 .08 .06 .08 Epistemic Motivation p-value .00 .00 .04 .32 .21 β 3.64 -.81 .17 .06 .11 SE .07 .08 .08 .06 .08 Experienced Power β p-value .00 3.65 .00 -.85 .05 .17 .04 .50 .10† -.14 SE .07 .08 .08 .06 .08 Relative Power β 3.66 -.85 .16 .00 .01 SE .08 .09 .09 .06 .09 p-value .00 .00 .06† .96 .92 F = 29.93 r2 = .27 29.19 .27 27.17 .25 26.12 .25 Note. N = 326, p < .05, p < .10† 87 Table 6. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s performance inference. Predictor Intercept Opponent Affect Number of Opponents Main Effect Opponent affect*Main Effect Performance inference round 1 β 2.56 .65 -.06 -.11 .16 .28 F = 35.81 r2 = .36 Information Processing Epistemic Motivation p-value .00 .00 .28 .04 .91 .00 p-value .00 .00 .39 .05 .03 .00 SE .06 .07 .07 .05 .07 .04 β 2.59 .63 -.08 -.11 .01 .26 36.74 .36 SE .07 .07 .07 .05 .07 .04 Experienced Power β p-value .00 2.58 .00 .65 .21 -.09 .00 .97 .02 .16 .00 .27 37.17 .37 SE .06 .07 .07 .05 .07 .04 Relative Power β 2.59 .65 -.10 -.03 .10 .28 35.02 .35 SE .07 .08 .07 .05 .07 .04 p-value .00 .00 .22 .54 .16 .00 Note. N = 326, p < .05, p < .10† 88 Table 7. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s response-focused emotion regulation Predictor Intercept Opponent Affect Number of Opponents Main effect of moderator Positive Affect Negative Affect Performance Inference Opponent Affect * Moderator Positive Affect * Moderator Negative Affect * Moderator Performance Inference * Moderator β 4.89 -.36 .20 .19 -.08 .08 .04 -.24 .05 .08 .15 p-value Information Processing SE .14 .18 .14 .13 .10 .10 .09 .18 .09 .10 .09 .00 .05 .18 .14 .43 .42 .69 .19 .60 .40 .09† F = 1.98 r2 = .06 Note. N = 326, p < .05, p < .10† Epistemic Motivation Experienced Power Relative Power β 4.88 -.32 .24 -.05 -.11 .04 .03 .15 .03 -.09 -.03 1.87 .06 SE .14 .18 .15 .13 .10 .10 .09 .19 .09 .10 .10 p-value β .00 .07† .11 .70 .24 .72 .71 .43 .75 .35 .75 4.82 -.33 .27 -.22 -.09 .08 .05 .16 .06 .05 .08 2.38 .07 SE .14 .18 .15 .12 .09 .10 .09 .18 .09 .09 .07 p-value β .00 .06† .07† .06† .34 .42 .58 .39 .48 .58 .30 4.87 -.34 .23 -.18 -.09 .08 .02 .22 .06 .05 .13 2.52 .07 SE .14 .18 .15 .13 .09 .10 .09 .18 .08 .09 .08 p-value .00 .06† .12 .16 .36 .40 .83 .23 .45 .56 .10† 89 Table 8. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s antecedent-focused emotion regulation Epistemic Motivation Experienced Power Relative Power β 5.13 -.18 .05 -.14 .17 .08 -.01 .03 -.02 -.12 -.21 SE .11 .14 .12 .10 .08 .08 .07 .15 .07 .08 .08 p-value β .00 .21 .65 .16 .03 .29 .84 .83 .83 .12 .00 5.18 -.20 .03 .15 .14 .08 -.01 .00 -.11 -.08 .12 1.83 .05 SE p-value .11 .14 .12 .10 .08 .08 .07 .15 .07 .07 .06 .00 .16 .81 .12 .06† .34 .92 .99 .14 .29 .04 β 5.17 -.15 -.01 .25 .15 .07 -.01 -.13 -.10 -.01 .21 2.72 .08 SE .11 .14 .12 .10 .07 .08 .07 .15 .06 .07 .06 p-value .00 .29 .94 .01 .05 .34 .91 .38 .11 .90 .00 Predictor Intercept Opponent Affect Number of Opponents Main effect of moderator Positive Affect Negative Affect Performance Inference Opponent Affect * Moderator Positive Affect * Moderator Negative Affect * Moderator Performance Inference * Moderator β 5.18 -.20 .04 .33 .15 .09 .05 -.11 -.06 -.15 -.10 p-value Information Processing SE .14 .18 .14 .13 .10 .10 .09 .18 .09 .10 .09 .00 .15 .70 .00 .05 .26 .51 .42 .39 .05 .14 F = 3.61 r2 = .10 2.08 .06 Note. N = 326, p < .05, p < .10† 90 Predictor Intercept Opponent Affect Number of Opponents Main effect of moderator Positive Affect Negative Affect Performance Inference Opponent Affect * Moderator Positive Affect * Moderator Negative Affect * Moderator Performance Inference * Moderator β 3.74 .14 -.04 .01 -.01 -.14 .05 .20 .03 .03 -.01 p-value Information Processing SE .10 .13 .11 .10 .07 .07 .07 .14 .07 .07 .06 .00 .29 .68 .95 .86 .06† .45 .14 .63 .67 .85 β 3.72 .12 -.04 -.18 .01 -.15 .01 .05 -.03 -.03 -.13 SE .10 .13 .11 .09 .07 .07 .07 .14 .07 .07 .07 p-value β .00 .37 .68 .06† .84 .05 .83 .71 .65 .64 .07† 3.75 .14 -.09 .15 .00 -.16 .01 .08 .02 .03 .06 SE p-value .10 .13 .11 .09 .07 .07 .07 .14 .07 .07 .06 .00 .30 .40 .08† .99 .03 .88 .57 .78 .66 .29 β 3.72 .15 -.07 .23 -.01 -.16 .05 -.23 .03 .00 .06 2.47 .07 SE .10 .13 .11 .10 .07 .07 .07 .14 .06 .07 .06 p-value .00 .25 .52 .02 .85 .03 .47 .10† .66 .95 .27 Table 9. Effects of opponent’s affect, information processing, NFCC, experienced power, relative power, and their interaction on participant’s authentic display. Epistemic Motivation Experienced Power Relative Power Note. N = 326, p < .05, p < .10† F = 2.62 r2 = .08 3.07 .09 3.29 .09 91 Figure 1. Proposed model. Solid paths denote hypothesized relationships; dashed paths denote exploratory relationships. 92 Figure 2. The moderating role of information processing on the relationship between opponent affect participant’s positive affective state. 93 Figure 3. The moderating role of information processing on the relationship between opponent affect participant’s negative affective state. 94 Figure 4. The moderating role of information processing on the relationship between opponent affect participant’s performance inference 95 Figure 5. The moderating effect of information processing on the relationship between opponent affect and participant’s positive affect. 96 Figure 6. The moderating effect of information processing on the relationship between opponent affect and participant’s negative affect. 97 Figure 7. The moderating effect of information processing on the relationship between opponent’s affect and their performance inference. 98 Figure 8. The moderating effect of experienced power on the relationship between opponent’s affect and participant’s performance inference. 99 Figure 9. The moderating effect of information processing on the relationship between negative affect and antecedent-focused emotion regulation. 100 Figure 10. The moderating effect of epistemic motivation on the relationship between participant’s performance inference and participant’s antecedent-focused emotion regulation. 101 Figure 11. The moderating effect of experienced power on the relationship between participant’s performance inference and participant’s antecedent-focused emotion regulation. 102 Figure 12.The moderating effect of relative power on the relationship between participant’s performance inference and participant’s antecedent-focused emotion regulation 103 APPENDIX B: Outline of Negotiation Scenario 1) Consent form 2) General Context of Study a) “West Wind is a U.S wind energy company. They are proposing a new development project in the rural town of Pine Hills. However, Pine Hill's residents are wary of West Wind. This is due to a prior, negative experience with a different private sector led energy company. Residents felt that this previous company tried to take advantage of Pine Hills residents. Subsequently, the initial proposal for the project by West Wind was widely opposed by residents. Resident's main concern was due to the fact that the construction of this project will require roads to be constructed on adjacent municipal land. Other concerns included the visibility of the windmills along with potential disruption of wildlife and undue noise. West Wind and Pine Hills have two additional issues left to further negotiate. The first being the number of wind turbines to be constructed. The second is the amount of annual community benefit payments that West Wind will contribute to Pine Hills for its own use." 3) Introduction of participant’s role a) “In this exercise, you will be playing ' Riley Marks,' the representative of Pine Hills. As Riley Marks, you will be negotiating with representative(s) of West Wind to work out the last two issues. You plan to run as mayor of Pine Hills next year. Given the concerns of Pine Hill residents, this is an important opportunity for you to demonstrate the ability to represent the concerns of the community. Therefore, you do not want to be in the position of advocating for an agreement that is ultimately unpopular with the community. Citizens are exceptionally worried by the rumor that West Wind is planning on building 20 wind mills. Many residents have told you that anything above 10 turbines will be too much, while others have argued that the project should be scrapped all together. This leads you to believe that 10 turbines are the maximum politically acceptable. On the other hand, every resident acknowledges the benefits that will come with the annual benefit payments offered by West Wind since there are many local services, including local education, that will benefit from the additional funds. Accordingly, you agree that $100,000 in annual payments would be the minimum needed to impact the locals school or be felt in the community. b) “Your first priority is to secure as much in annual community benefits payments as possible. Remember that these represent substantial, direct financial benefits for the Pine Hills residents. The second priority is to lower the number of turbines that West Wind establishes. You need to remember that you want to make a politically smart agreement while understanding that each additional turbine potentially creates additional disturbance.” 4) Details of negotiation a) “For this exercise, you will be negotiating with Jordan Garcia/ Jordan Garcia., Skyler Jones, Payton Hernandez, who will be played by other participant(s) in another room. They will be representing the West Wind Company. The West Wind project is supported 104 by the state's governor given that renewable energy has been identified as an important goal for the state. West Wind has decided to revise the project with local's concerns in mind in hopes to improve relations with Pine Hills. Specifically, West Wind has agreed to reroute proposed access roads to reduce truck noise and traffic. This comes at a considerable cost to the company ($200k+ annually). Second, West Wind would like to establish a monitoring/advisory panel including Pine Hills residents to help minimize local impact on wildlife and ensure efficient and safe practices. As mentioned previously, West Wind Co. is looking to negotiate two issues. The first being the number of wind turbined to be constructed. The second is the amount of annual community benefit payments that West Wind will contribute to Pine Hills for its own use. i) Participants first learn they will be negotiating with one or multiple people here 5) Introduction of display rule a) “Before you begin, here are some helpful rules to follow to help you navigate this negotiation: i) Displaying positive emotions throughout the negotiation (e.g. use emojis, positive situation framing) is a helpful strategy for successful negotiation ii) Additionally, displaying negative emotions (e.g. sarcasm, frustration, anger) is an ineffective strategy while navigating the negotiation iii) Directly addressing the party by their first name is helpful strategy (e.g. the use of personal pronouns such as I, we) for a successful negotiation iv) Bargains are best expressed using correct grammar, syntax, and full sentences, to most clearly express ideas to the other party. v) Explanation of the bargain being proposed is also helpful for the other party by reducing the chance of miscommunication throughout the negotiation. 6) Negotiation starts in separate chat window a) Party introduction. i) Automated negotiator(s) enter chat room one at a time. (1) Automated introductions (2) Notification that each member joins the chat room b) Participant introduction – participants introduce themselves. c) First round of negotiation i) Party presents offer (1) Next condition (Anger vs. Happiness condition) presented/revealed here. ii) Participant asked to respond and either accept or decline the offer. (1) Accept / Decline (2) Participants will then be asked to fill out a paragraph explaining why with a counter bargain. iii) Appendices C,D,E, and K presented after response. d) Second round of negotiation i) Party presents offer. ii) Participant asked to respond and either accept or decline the offer. (1) Accept / Decline (2) Participants will then be asked to fill out a paragraph explaining why with a counter bargain. iii) Appendices C,D,E, and K presented after response. e) Third Round of Negotiation 105 i) Party presents offer. ii) Participant asked to respond and either accept or decline the offer. (1) Accept / Decline (2) Participants will then be asked to fill out a paragraph explaining why with a counter bargain. iii) Appendices C,D,E, and K presented after response. 7) Final page and measures (Appendices F, G, H, I, & J) 8) Debrief form 9) Personal ID for Qualtrics survey 106 APPENDIX C: Round 1, 2, and 3 Survey Performance Inference Measures From Van Kleef et al. (2009) For each of the following statements, please indicate to what degree you agree on the scale provided below (1= strongly agree, 5= strongly disagree). -"I feel that the other party was satisfied with my performance" -"I feel that the other party had expected more of me" (reverse-scored) -"I feel that the other party thought I had performed poorly" Perception of Target emotion From Van Kleef et al. (2006) Please indicate to what degree you believe their opponent had been XXX during the negotiation … (1=totally disagree to 7=totally agree) -Happy -Enthusiastic -Joyful -Angry -Irritated Own anger and own happiness From Weiss, Sucko, Cropanzo (1999) Please indicate how you feel about what just happened within the negotiation… (1=not at all to 7=very much_ - Happiness - Joyful - Anger - Aggravated Postive Affect and Negative Affect Scale (PANAS) From Watson, Clark, and Tellegen (1988) This scale consists of a number of words that describe different feelings and emotions. Read each item and then mark the appropriate answer in the space next to that word. Indicate to what extent you have felt this way during the negotiation. Use the following scale to record your answers: 1=very slightly or not at all, 5 = extremely. - Upset - Irritable - Enthusiastic - Interested - Determined 107 APPENDIX D: Final Round Survey Perceptions of Power From VanKleef, De Dreu, Pietroni, Manstead, 2006 For each of the following statements, please indicate to what degree you agree on the scale provided below (1= strongly agree, 5= strongly disagree). Experienced Power: 1. During the negotiation I felt that I depended on the buyer (reverse scored) 2. I felt that I had a strong negotiation position 3. During the negotiation I felt that I needed the buyer to make a good deal (reverse scored) 4. I felt powerful in the negotiation 5. I felt that I needed the buyer badly to earn an acceptable amount of points (reverse scored) 6. During the negotiation I did not feel dependent on the buyer 7. The fact that I had an alternative offer made me feel relaxed 8. The fact that I had an alternative offer gave me a sense of power in the negotiation 9. During the negotiation, I felt that I was in control of the situation For each of the following statements, please indicate the degree to which you think the question applies to you on the scale provided below (1= definitely the other, 7= definitely myself). Relative power: 1. Who do you think had the strongest position in the negotiation? 2. Who do you feel had most influence on the course of the negotiation? 3. Who do you feel had the most power in the negotiation? 4. Who do you think had the best basis to negotiate? 5. Who do you feel had the best negotiation position? 6. Who do you feel was most in control of the situation? 7. Who do you feel was the most powerful person in the negotiation? 8. Who do you think was most dependent on the other during the negotiation? 9. Who do you feel needed the other most during the negotiation? 108 Antecedent-focused Strategies From Gross and John (2003) For each of the following statements, please indicate to what degree you agree on the scale provided below (1=totally agree, 7= totally disagree). 1. I controlled my emotions by changing the way I thought about the negotiation. 2. When I wanted to feel less negative emotion, I changed the way I was thinking about the negotiation. 3. When I wanted to feel more positive emotion, I changed the way I was thinking about the negotiation. 4. I changed what I was thinking about when I wanted to feel more positive emotion (such as joy or amusement) 5. I changed what I was thinking about when I wanted to feel less negative emotion (such as sadness or anger) Response-focused strategies From Gross and John (2003) For each of the following statements, please indicate to what degree you agree on the scale provided below (1= totally agree, 7= totally disagree). 1. During the negotiation, I controlled my emotions by not expressing them. 2. When I felt negative emotions, I made sure not to express them to the other party. 3. I kept my emotions to myself during the negotiation. 4. When I felt positive emotions, I made sure to not to express them to the other party. Information Processing From De Dreu et al. (2004) For each of the following statements, please indicate to what degree you agree on the scale provided below (1= totally agree, 7= totally disagree). 1. “During the negotiation, I hardly thought about my goals,” reverse scored; 2. “During the negotiation, I paid a lot of attention to the information about the negotiator’s intentions” 3. “During the negotiation, I made my offers without thinking too much,” reverse scored 4. “During the negotiation, I tried to consider all the available information before placing an offer” 109 Need for Closure scale From Webster and Kurglanski (2004) For each of the following statements, please indicate to what degree you agree on the scale provided below (1= strongly disagree, 5 = strongly agree). 1. I felt uncomfortable because the other party’s meaning, or intentions, were unclear to me. 2. I dislike unpredictable situations such as this negotiation. 3. I don't like situations that are uncertain such as this negotiation. 4. I don't like to go into a situation without knowing what I can expect from it, like this negotiation. 5. I liked to know what the other party was thinking of me throughout the negotiation. 6. When I was confused about an important issue during the negotiation, I felt very upset 7. I enjoyed the uncertainty of going into a new situation without knowing what might happen, such as this negotiation. Expression of Naturally Felt Emotions Scale From Diefendorff, Croyle, Gosserand (2004) For each of the following statements, please indicate to what degree you agree on the scale provided below (1= strongly agree, 5= strongly disagree). 1. The emotions I expressed to the other party were genuine. 2. The emotions I expressed to the other party matched what I spontaneously felt. 3. The emotions I expressed to the other party came naturally. 110 REFERENCES 111 REFRENCES Bacharach, S. B., & Lawler, E. J. (1981). Bargaining: Power, Tactics and Outcomes. Jossey- Bass Inc., 433 California Street, San Francisco, CA 94104. Barsade, S. G. (2002). The ripple effect: Emotional contagion and its influence on group behavior. Administrative Science Quarterly, 47(4), 644-675. Bandura, A., & Cervone, D. (1983). Self-evaluative and self-efficacy mechanisms governing the motivational effects of goal systems. Journal of personality and social psychology, 45(5), 1017. Bolin, J. H. (2014). Hayes, Andrew F.(2013). Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression‐Based Approach. New York, NY: The Guilford Press. Journal of Educational Measurement, 51(3), 335-337. Brotheridge, C. M., & Grandey, A. A. (2002). Emotional labor and burnout: Comparing two perspectives of “people work”. Journal of vocational behavior, 60(1), 17-39. Byron, K. (2008). Carrying too heavy a load? The communication and miscommunication of emotion by email. Academy of Management Review, 33, 309–327 Carlson, D., Ferguson, M., Hunter, E., & Whitten, D. (2012). Abusive supervision and work– family conflict: The path through emotional labor and burnout. The Leadership Quarterly, 23(5), 849-859. Carver, C. S., & Scheier, M. F. (1999). Themes and issues in the self-regulation of behavior. Advances in social cognition, 12(1), 1. Chau, S. L., Dahling, J. J., Levy, P. E., & Diefendorff, J. M. (2009). A predictive study of emotional labor and turnover. Journal of Organizational Behavior, 30(8), 1151-1163. Clark, M. S., & Taraban, C. (1991). Reactions to and willingness to express emotion in communal and exchange relationships. Journal of Experimental Social Psychology, 27(4), 324-336. Cohen, T. R., & Thompson, L. (2011). Chapter 1 When are Teams an Asset in Negotiations and when are they a Liability?. In Negotiation and groups (pp. 3-34). Emerald Group Publishing Limited. Coˆte´, S. (2005). A social interaction model of the effects of emotion regulation on work strain. Academy of Management Review, 30, 509–530 Cote, S., & Morgan, L. M. (2002). A longitudinal analysis of the association between emotion regulation, job satisfaction, and intentions to quit. Journal of Organizational Behavior: 112 The International Journal of Industrial, Occupational and Organizational Psychology and Behavior, 23(8), 947-962. De Dreu, C. K. (2003). Time pressure and closing of the mind in negotiation. Organizational Behavior and Human Decision Processes, 91(2), 280-295. De Dreu, C. K., Giebels, E., & Van de Vliert, E. (1998). Social motives and trust in integrative negotiation: The disruptive effects of punitive capability. Journal of Applied Psychology, 83(3), 408. Diefendorff, J. M., & Gosserand, R. H. (2003). Understanding the emo- tional labor process: A control theory perspective. Journal of Organizational Behavior, 24, 945–959. http://dx.doi.org/10.1002/job.230 Diefendorff, J. M., & Greguras, G. J. (2009). Contextualizing emotional display rules: Examining the roles of targets and discrete emotions in shaping display rule perceptions. Journal of Management, 35(4), 880-898. Diefendorff, J., Morehart, J., & Gabriel, A. (2010). The influence of power and solidarity on emotional display rules at work. Motivation and Emotion, 34(2), 120-132. Diefendorff, J. M., & Richard, E. M. (2003). Antecedents and consequences of emotional display rule perceptions. Journal of Applied Psychology, 88(2), 284. Drach-Zahavy, A., Yagil, D., & Cohen, I. (2017). Social model of emotional labour and client satisfaction: exploring inter-and intrapersonal characteristics of the client− provider encounter. Work & Stress, 31(2), 182-208. Ekman, P. (1993). Facial expression and emotion. American psychologist, 48(4), 384. Ekman, P., & O’Sullivan, M. (1991). Who can catch a liar? American Psychologist, 46, 913–920 Elfenbein, H. A., & Ambady, N. (2002). On the universality and cultural specificity of emotion recognition: A meta-analysis. Psychological Bulletin, 128, 203–235. Erez, M. (1977). Feedback: A necessary condition for the goal setting-performance relationship. Journal of Applied psychology, 62(5), 624. Fischer, A. H., & Manstead, A. S. R. (2008). Social functions of emotion. In M. Lewis, J. Haviland & L. Feldman Barrett (Eds.), Handbook of Emotion (3rd edn.) (pp. 456–569). New York: Guilford. French, J. R., Raven, B., & Cartwright, D. (1959). The bases of social power. Classics of organization theory, 7, 311-320. Fridlund, A. J. (1994). Human Facial Expression: An Evolutionary View. San Diego, CA: Academic Press. 113 Friedman, R., Anderson, C., Brett, J., Olekalns, M., Goates, N., & Lisco, C. C. (2004). The positive and negative effects of anger on dispute resolution: evidence from electronically mediated disputes. Journal of Applied Psychology, 89(2), 369 Frijda, N. H. (1986). The emotions: Studies in emotion and social interaction. Paris: Maison de Sciences de l'Homme. Gabriel, A. S., Daniels, M. A., Diefendorff, J. M., & Greguras, G. J. (2015). Emotional labor actors: A latent profile analysis of emotional labor strategies. Journal of Applied Psychology, 100(3), 863. Gabriel, A. S., & Diefendorff, J. M. (2015). Emotional labor dynamics: A momentary approach. Academy of Management Journal, 58(6), 1804-1825 Goldberg, L. S., & Grandey, A. A. (2007). Display rules versus display autonomy: emotion regulation, emotional exhaustion, and task performance in a call center simulation. Journal of occupational health psychology, 12(3), 301. Grandey, A. A. (2000). Emotional regulation in the workplace: A new way to conceptualize emotional labor. Journal of occupational health psychology, 5(1), 95. Grandey, Alicia A., and Allison S. Gabriel. "Emotional Labor at a Crossroads: Where do we Go from here?" Annual Review of Organizational Psychology and Organizational Behavior 2 (2015): 323-49. ProQuest. Web. 26 Nov. 2018 Grandey, A. A., Dickter, D. N., & Sin, H. P. (2004). The customer is not always right: Customer aggression and emotion regulation of service employees. Journal of Organizational Behavior: The International Journal of Industrial, Occupational and Organizational Psychology and Behavior, 25(3), 397-418. Grandey, A., Diefendorff, J., & Rupp, D. E. (Eds.). (2013). Emotional labor in the 21st century: Diverse perspectives on emotion regulation at work. Routledge. Grandey, A. A., & Melloy, R. C. (2017). The state of the heart: Emotional labor as emotion regulation reviewed and revised. Journal of occupational health psychology, 22(3), 407. Grandey, A., Tam, A., & Brauburger, A. (2002). Affective states and traits of young workers: A diary study. Motivation and Emotion, 26, 31–55. http://dx.doi.org/10.1023/A:1015142124306 Grant, A. M. (2013). Rocking the boat but keeping it steady: The role of emotion regulation in employee voice. Academy of Management Journal, 56(6), 1703-1723. Gross, J. J. (1998). Antecedent-and response-focused emotion regulation: divergent consequences for experience, expression, and physiology. Journal of personality and social psychology, 74(1), 224. 114 Gross, J. J. (1998). The emerging field of emotion regulation: an integrative review. Review of general psychology, 2(3), 271Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. Berkeley: University of California Press. Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: implications for affect, relationships, and well-being. Journal of personality and social psychology, 85(2), 348. Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion: Studies in emotion and social interaction. Editions de la Maison des sciences de l’homme. Hess, U., Blairy, S., & Kleck, R. E. (2000). The influence of facial emotion displays, gender, and ethnicity on judgments of dominance and affiliation. Journal of Nonverbal behavior, 24(4), 265-283. Hülsheger, U. R., & Schewe, A. F. (2011). On the costs and benefits of emotional labor: A meta- analysis of three decades of research. Journal of occupational health psychology, 16(3), 361. Jones, R. G., & Best, R. G. (1995). A further examination of the nature and impact of emotional work requirements on individuals and organizations. Paper presented at the annual meeting of the Academy of Management, Vancouver, British Columbia. Kalokerinos, E. K., Greenaway, K. H., Pedder, D. J., & Margetts, E. A. (2014). Don’t grin when you win: The social costs of positive emotion expression in performance situations. Emotion, 14(1), 180. Kotsou, I., Nelis, D., Grégoire, J., & Mikolajczak, M. (2011). Emotional plasticity: Conditions and effects of improving emotional competence in adulthood. Journal of Applied Psychology, 96(4), 827. Kammeyer‐Mueller, J. D., Rubenstein, A. L., Long, D. M., Odio, M. A., Buckman, B. R., Zhang, Y., & Halvorsen‐Ganepola, M. D. (2013). A meta‐analytic structural model of dispositonal affectivity and emotional labor. Personnel Psychology, 66(1), 47-90. Klein, H. J. (1989). An integrated control theory model of work motivation. Academy of Management Review, 14(2), 150-172. Lazarus, R. S. (1991). Progress on a cognitive-motivational-relational theory of emotion. American psychologist, 46(8), 819. Lee, F., & Tiedens, L. Z. (2001). 2. Is it lonely at the top?: The independence and interdependence of power holders. Research in organizational behavior, 23, 43-91. Locke, E. A., & Latham, G. P. (1990). A theory of goal setting & task performance. Prentice- Hall, Inc. 115 Lopez, S. H. (2010). Workers, managers, and customers: Triangles of power in work communities. Work and Occupations, 37(3), 251-271. Manstead, A. S., & Fischer, A. H. (2001). Social appraisal. Appraisal processes in emotion: Theory, methods, research, 221-232. Matsui, T., Okada, A., & İnoshita, O. (1983). Mechanism of feedback affecting task performance. Organizational Behavior and Human Performance, 31(1), 114-122. Patton, C., & Balakrishnan, P. S. (2012). Negotiating when outnumbered: agenda strategies for bargaining with buying teams. International Journal of Research in Marketing, 29(3), 280-291. Pinkley, R. L. (1995). Impact of knowledge regarding alternatives to settlement in dyadic negotiations: Whose knowledge counts?. Journal of Applied Psychology, 80(3), 403. Richards, J. M., & Gross, J. J. (2000). Emotion regulation and memory: the cognitive costs of keeping one's cool. Journal of personality and social psychology, 79(3), 410. Rupp, D. E., & Spencer, S. (2006). When customers lash out: the effects of customer interactional injustice on emotional labor and the mediating role of discrete emotions. Journal of Applied Psychology, 91(4), 971. Sally, D., & O'Connor, K. (2003). Team negotiations. Marq. L. Rev., 87, 883. Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. (1987). Emotion knowledge: Further exploration of a prototype approach. Journal of Personality and Social Psychology, 52, 1061–1086. Sinaceur, M., & Tiedens, L. Z. (2006). Get mad and get more than even: When and why anger expression is effective in negotiations. Journal of Experimental Social Psychology, 42(3), 314-322. Sliter, M., Jex, S., Wolford, K., & McInnerney, J. (2010). How rude! Emotional labor as a mediator between customer incivility and employee outcomes. Journal of occupational health psychology, 15(4), 468. Sy, T., Côté, S., & Saavedra, R. (2005). The contagious leader: impact of the leader's mood on the mood of group members, group affective tone, and group processes. Journal of applied psychology, 90(2), 295. Thompson, L., Peterson, E., & Brodt, S. E. (1996). Team negotiation: An examination of integrative and distributive bargaining. Journal of Personality and Social Psychology, 70(1), 66. 116 Totterdell, P., & Holman, D. (2003). Emotion regulation in customer service roles: Testing a model of emotional labor. Journal of occupational health psychology, 8(1), 55. Tschan, F., Rochat, S., & Zapf, D. (2005). It's not only clients: Studying emotion work with clients and co‐workers with an event‐sampling approach. Journal of Occupational and Organizational Psychology, 78(2), 195-220. Vahle-Hinz, T., Mauno, S., de Bloom, J., & Kinnunen, U. (2017). Rumination for innovation? Analysing the longitudinal effects of work-related rumination on creativity at work and off-job recovery. Work & Stress, 31(4), 315-337. Van Dijk, E., Van Kleef, G. A., Steinel, W., & Van Beest, I. (2008). A social functional approach to emotions in bargaining: When communicating anger pays and when it backfires. Journal of Personality and social Psychology, 94(4), 600. Van Kleef, G. A., Anastasopoulou, C., & Nijstad, B. A. (2010). Can expressions of anger enhance creativity? A test of the emotions as social information (EASI) model. Journal of Experimental Social Psychology, 46(6), 1042-1048. Van Kleef, G. A., & Côté, S. (2007). Expressing anger in conflict: When it helps and when it hurts. Journal of Applied Psychology, 92(6), 155. Van Kleef, G. A., De Dreu, C. K., & Manstead, A. S. (2004). The interpersonal effects of anger and happiness in negotiations. Journal of personality and social psychology, 86(1), 57. Van Kleef, G. A., De Dreu, C. K., & Manstead, A. S. (2004). The interpersonal effects of emotions in negotiations: a motivated information processing approach. Journal of personality and social psychology, 87(4), 510. Van Kleef, G. A., De Dreu, C. K., & Manstead, A. S. (2006). Supplication and appeasement in conflict and negotiation: The interpersonal effects of disappointment, worry, guilt, and regret. Journal of personality and social psychology, 91(1), 124. Van Kleef, G. A., De Dreu, C. K., Pietroni, D., & Manstead, A. S. (2006). Power and emotion in negotiation: Power moderates the interpersonal effects of anger and happiness on concession making. European Journal of Social Psychology, 36(4), 557-581. Van Kleef, G. A., Homan, A. C., Beersma, B., Van Knippenberg, D., Van Knippenberg, B., & Damen, F. (2009). Searing sentiment or cold calculation? The effects of leader emotional displays on team performance depend on follower epistemic motivation. Academy of Management Journal, 52(3), 562-580. Van Kleef, G. A. (2010). The emerging view of emotion as social information. Social and Personality Psychology Compass, 4(5), 331-343. Van Kleef, G. A. (2009). How emotions regulate social life: The emotions as social information (EASI) model. Current directions in psychological science, 18(3), 184-188. 117 Wagner, D. T., Barnes, C. M., & Scott, B. A. (2014). Driving it home: How workplace emotional labor harms employee home life. Personnel Psychology, 67(2), 487-516. Webster, D. M., & Kruglanski, A. W. (1994). Individual differences in need for cognitive closure. Journal of personality and social psychology, 67(6), 1049 Weiss, H. M., Suckow, K., & Cropanzano, R. (1999). Effects of justice conditions on discrete emotions. Journal of applied psychology, 84(5), 786. Yukl, G., & Falbe, C. M. (1991). Importance of different power sources in downward and lateral relations. Journal of applied psychology, 76(3), 416. 118