RELATIONAL MOTIVES SHAPING RESPONSES TO THE RECEIPT OF INTERPERSONAL HELPING AND HARMING BEHAVIORS: A DYADIC PERSPECTIVE By Brent John Lyons A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology – Doctor of Philosophy 2013 ABSTRACT RELATIONAL MOTIVES SHAPING RESPONSES TO THE RECEIPT OF INTERPERSONAL HELPING AND HARMING BEHAVIORS: A DYADIC PERSPECTIVE By Brent John Lyons Interpersonal behaviors in organizations, such as helping (e.g., organizational citizenship behaviors) and harming (e.g., aggression, counterproductive work behaviors) have important consequences for employee functioning and well-being. Traditionally, organizational research examining factors influencing responses to the receipt of help and harm has focused on individual differences or the situation in which the acts occur. Much less research, by comparison, has taken a dyadic perspective by examining the relationship between the target and initiator of the behaviors. The dyadic perspective is important because the experience of being helped or harmed may differ depending on the relationship between the target of the behaviors and the initiator of the behaviors. This dissertation investigates outcomes of interpersonal helping and harming from a dyadic relationship perspective and examines how outcomes of helping and harming vary within relationships over time. Three within-relationship processes are examined to explain how the receipt of help and harm can relate to the engagement of help and harm and relationship satisfaction: social exchange, affective events, and self-regulation. Social motives at the relationship level of analysis are hypothesized to influence the within-relationship processes. Drawing from the Hierarchical Model of Approach and Avoidance Social Motives and Goals (Gable, 2006), coworker relationships were characterized based on approach (i.e., concerned about approaching closeness) and avoidant (i.e., concerned about avoiding threat or conflict) social motives. v A sample of 55 working adults from three organizations completed daily surveys about their interactions with two coworkers over 20 work days. The majority of variance in helping, harming, and relationship satisfaction occurred within and between relationships, as opposed to between individuals. Employees who received help from a coworker engaged in help towards that coworker and employees who received harm from a coworker engaged in harm towards that coworker. Positive affective states related to receiving help, engaging in help, and relationship satisfaction, and negative affective states related to receiving harm, engaging in harm and relationship satisfaction. Positive affective states mediated the relationship between help receipt and help engagement and between help receipt and relationship satisfaction. Promotion regulatory focus related to receiving help and engaging in help and harm and prevention regulatory focus related to receiving harm and engaging in harm. Promotion regulatory focus mediated relationships between the receipt of help and the engagement in help and harm. Although relationship approach-avoidance social motives did not influence within-relationship social exchange and affective mechanisms, prevention regulatory focus more strongly evoked harm engagement in avoidance-oriented relationships compared to approach-oriented relationships. This suggests that employees who are avoidant and particularly sensitive to threat within a relationship are more likely to use harmful behaviors to avoid experiencing future threat. Overall, the results suggest that dyadic relationships are important to understanding interpersonal relations in the work context. Social exchange and affective mechanisms are robust across relationships so it is important to manage the reciprocation of help and harm and affect within relationships. The results also speak to how interpersonal goals can be framed in order to increase the occurrence of helping behaviors and decrease the occurrence of harming behaviors. vi Copyright by BRENT JOHN LYONS 2013 vii TABLE OF CONTENTS LIST OF TABLES vii LIST OF FIGURES viii INTRODUCTION The Relational Self Approach and Avoidance Motives in Interpersonal Relationships Hierarchical Model of Approach and Avoidance Social Motivation Social Exchange and the Reciprocation of Help and Harm Affective State Responses to the Receipt of Help and Harm Affective Reactions to the Receipt of Help and Harm Affect-driven Engagement in Help and Harm and Satisfaction Promotion and Prevention Regulatory Foci Exploratory Investigation: Emotional Intelligence and Interpersonal Trust Emotional Intelligence Interpersonal Trust 1 8 11 17 19 22 22 24 30 35 35 36 METHOD Participants Procedure Measure Individual Level Variables (Level 3) General Approach and Avoidance Disposition Emotional Intelligence Individual Demographics Relationship Level Variables (Level 2) Approach and Avoidance Relationship Motives Liking Trust Characteristics of Coworker Relationships Within-Relationship Level Variables/ Daily Experiences (Level 1) Receipt of and Engagement in Help Receipt of and Engagement in Harm Relationship Satisfaction Positive and Negative Affective States Regulatory Focus Manipulation Check 38 38 40 43 43 43 43 44 44 44 45 45 46 47 47 48 49 49 51 51 viii Analytical Approach 54 RESULTS Descriptive Statistics and Correlations Partitioning of Variance Within and Between Levels Tests of Hypotheses Within-Relationship Hypotheses Main Effects Mediation Cross-Level Hypotheses Moderation Moderated-Mediation Exploratory Analyses Emotional Intelligence Interpersonal Trust 56 56 60 63 63 63 69 73 73 75 77 77 82 DISCUSSION Partitioning of Variance Social Exchange Affective Events Regulatory Foci Approach-Avoidance Motives Practical Implications Limitations Future Research 84 84 86 88 90 92 94 96 100 APPENDICES APPENDIX A: Study Protocol APPENDIX B: Study Measures 103 104 108 REFERENCES 122 ix LIST OF TABLES Table 1a. Descriptive statistics and correlations between study variables at Level 1 and Level 2 58 Table 1b. Descriptive statistics and correlations between study variables at Level 3 59 Table 2. Intra-class correlation statistics for 3-level HLM model 62 Table 3. HLM results for help engagement as outcome and relationship motive as moderator 65 Table 4. HLM results for harm engagement as outcome and relationship motive as moderator 66 Table 5. HLM results for relationship satisfaction as outcome and relationship 67 motive as moderator Table 6. HLM results for mediators as outcomes and the effects of help and harm receipt and approach-avoidance relationship motives 68 Table 7. HLM results for mediators as outcomes and emotional intelligence as 79 individual-level moderator Table 8. HLM results for help engagement as outcome and emotional intelligence as 80 individual-level moderator Table 9. HLM results for harm engagement as outcome and emotional intelligence as 81 individual-level moderator x LIST OF FIGURES Figure 1. Model of proposed relationships for study 7 Figure 2. Graphical representation of study design 41 Figure 3. Harm engagement as outcome of the interaction between relationship motive and prevention focus 75 xi INTRODUCTION Interpersonal behaviors in the organizational context can take a variety of forms and have been conceptualized in a variety of ways. Positive helping behaviors have been termed interpersonal organizational citizenship behaviors (OCB; Spector & Fox, 2002), and prosocial behaviors (Brief & Motowildo, 1986) and negative harming behaviors have been termed interpersonal counterproductive work behaviors (CWB; Spector & Fox, 2002), incivility (Andersson & Pearson, 1999), aggression (Baron & Neuman, 1996), and harassment (Bowling & Beehr, 2006). Both helping and harming behaviors have important implications for organizations and their employees. Thus, researchers have devoted a great deal of effort to understand their occurrence, particularly concentrating on factors influencing engagement in such actions. Within this realm of inquiry, personality traits, employee attitudes, affective states, task characteristics, and organizational characteristics all have been found to influence the extent to which employees engage in helping and harming behaviors (for meta-analyses, see Berry, Ones, & Sackett, 2007; Bowling & Beehr, 2006; Hershcovis et al., 2007; Organ & Ryan, 1995; Podsakoff, Mackenzie, Paine, & Bachrach, 2000). Research indicates that employees who receive help from their fellow coworkers exhibit higher levels of job performance (Tsai, Chen, & Liu, 2007), and employees who are victims of harassment have higher levels of job dissatisfaction, depression, strain, and burnout (Bowling & Beehr, 2006). Further meta-analytic evidence has demonstrated that employees who are supported by the coworkers have reduced role ambiguity, conflict, and overload and are more likely to engage in OCBs whereas employees who are antagonized by their coworkers are more likely to engage in absenteeism and CWBs (Chiaburu & Harrison, 2008). Considering the important consequences of receiving help and harm, it appears critical to understand the factors 1 influencing target responses to the receipt of help and harm. Even though there is a sizable body of research demonstrating the positive effects of receiving help and the negative effects of receiving harm, the extant research has tended to focus on the receipt of help and harm as though the experience of being helped or harmed is the same regardless of the nature of the relationship between the actor and target. Although a small number of studies have utilized social network approaches that incorporate dyadic-level of analyses (e.g., Bowler & Brass, 2006; Lyons & Scott, 2012; Venkantarami & Dalal, 2007), extant research has tended to ignore the dyadic nature of relationships between employees who receive and engage in help and harm behaviors (Krasikova & LeBreton, 2012). The dyadic nature of interpersonal helping and harming is important because targets of help and harm are likely going to have different responses depending on idiosyncratic characteristics of their relationship with the actor who is enacting the behaviors toward them (Hershcovis & Barling, 2007; Aquino & Lamertz, 2004). For example, using a social network design, Bowler and Brass (2006) found that friendship and influence were associated with the exchange of helping behaviors, such that friends helped each other and received help from each other. Further, Lyons and Scott (2012) found that employees were helped by those coworkers whom they also helped and in whom they elicited positive affect states and were harmed by those coworkers whom they also harmed and in whom they elicited negative affect. However, even though these studies adopted a dyadic perspective using social network designs, they also used cross-sectional data which limits the extent to which their data are representative of interpersonal phenomena that are known to vary within individuals and over time (Dalal, Lam, Weiss, Welch, & Hulin, 2009). Such methodologies also make it difficult to ascertain the extent to which variance in interpersonal responses vary within relationships, between relationships, and between individuals. The current study builds upon this previous 2 work by considering responses to the receipt of help and harm within dyadic relationships and over time. Further, to date, most organizational research investigating how employees experience interpersonal behaviors has focused on broad traits (e.g., negative affectivity; Thau & Aquino, 2009) and the situation or context of the interaction (e.g., task interdependency; Chiarburu & Harrison, 2008). Despite the importance of these variables, consideration of relational and motivational characteristics that are more directly related to how employees relate to others offers the potential to increasing understanding of the effects of interpersonal behaviors at work. Social motivational characteristics of the relationship between the actor and target are expected to affect how interpersonal help and harm behaviors are interpreted, experienced, and responded to - a point which is elaborated on next. The workplace is filled with employee interactions that can be interpreted as positive, negative, and oftentimes ambiguous, and people may interpret the same interactions differently depending on who the actor of the behavior is. For example, one employee may interpret a coworker’s joke at his or her expense to be humorous and an expression of friendship while the same employee may interpret the same joke from another coworker as mean. Although research in the organizational sciences has been quiet regarding how the nature of relationships can affect interpretations of interpersonal behavior, insight can be drawn from social psychology, including relationship cognition and social motivation (Molden & Higgins, 2004; Strachman & Gable, 2006). Social psychologists have shed light on how social motivational processes underlying interpersonal relationships shape the affective, cognitive, and behavioral responses of targets of those behaviors. In line with this perspective, in recent years, researchers have adopted Gable’s (2006) Hierarchical Model of Approach-Avoidance Social Motives and demonstrated that 3 motives associated with approaching relationship incentives and avoiding relationship threats differentially relate to affective and cognitive outcomes (e.g., relationship satisfaction, subjective well-being; Elliot, Gable, & Mapes, 2006; Gable, 2006), and constructive and destructive behaviors (Impett, Gable, & Peplau, 2005; Impett & Gordon, 2010; Impett, Gordon, Kogan, Oveis, Gable, & Keltner, 2010; Impett, Starchman, Finkel, & Gable, 2008). Indeed, in addition to articulating how cognitive and motivational characteristics of relationships affect interpretations of behaviors, relationship models of social motivation can also speak to how relational motives influence engagement in interpersonal helping and harming behaviors within relationships under consideration. Therefore, in the current study, Gable’s (2006) model of social motivation is adapted to further understanding of how dyadic relational approach-avoidance motives shape targets’ responses to the receipt of help and harm, including targets’ relationship satisfaction and behavioral responses (i.e., help and harm engagement). Even though approach and avoidance social motives have important implications for adult relationships, the majority of the research adopting Gable’s (2006) model has been conducted with partners in intimate relationships for which participants report about their experiences with one other person, their significant other (e.g., dating partner, spouse). Therefore, extant research has not investigated approach and avoidance relational motives as they relate to employee relationships in the work context. In light of the social psychology research that has adopted Gable’s (2006) framework, it is expected that relational approachavoidance motives have the potential to shape relational functioning among coworkers in the work context because similar arguments have been made for related models that have been applied to the work context, including adult attachment (Bowlby, 1969). Most recently, Richards and Schat (2011) demonstrated that the adult attachment styles of coworkers, which have 4 underlying security or protection motives (similar to approach and avoidance social motives), related to their emotional support seeking and engagement in CWBs and OCBs towards other coworkers. In a similar vein, the current study applies Gable’s (2006) approach-avoidance model of social motivation in the work context. Further, much of the research that has applied models of social motivation has tended to examine single relationships between two people, such as married couples or dating partners, or individuals’ relationships with a general number of people. Outside of the social network research (Lyons & Scott, 2012; Bowler & Brass, 2006), little organizational research has examined interpersonal processes for multiple relationships held by the same person. Although the notion that employees can have qualitatively different relationships with different coworkers is not absent from organizational theory (Kahn, 1988), researchers have paid little attention to investigating such a possibility and have therefore been unable to account for how social cognitive mechanisms can account for how employees experience and respond to interpersonal behaviors differently in different relationships. In the realm of social cognition, relationship researchers have introduced the concept of a relational self to explain how mental representations of self can differ within the same individual but across relationship partners, and that subsequent attributes, evaluations, affect, goals, and behaviors can also differ across an individual’s relational selves (Chen, Boucher, & Tapias, 2006). Thus, people can experience and respond to the same behavior differently depending in their relationship with the person with whom they are interacting. An additional key proposition of the current study is that employees are likely to experience and respond to the receipt of help and harm differently depending on their approach and avoidance relationship motives (Gable, 2006) associated with the actor who is 5 engaging in the helping and harming behaviors towards them. The current study builds upon previous research by considering this as a possibility. In terms of employee reactions and responses to receiving help and harm behaviors within dyadic relationships, the current study integrates two theoretical perspectives that are related to dyadic interpersonal relations: social exchange theory (SET; Blau, 1964; Gouldner, 1960) and affective events theory (AET; Weiss & Cropanzano, 1996). SET predicts that individuals who are the recipients of helpful or harmful behaviors act in kind to reciprocate those behaviors (Blau, 1964; Gouldner, 1960). According to AET, affect is considered a key driver of attitudinal and behavioral responses to affective events (Weiss & Cropanzano, 1996). Specifically, AET and the emotion-centered model of voluntary work behavior (herein referred to as the emotion-centered model) that extended upon AET (Spector & Fox, 2002) describe how affective states can mediate relations between work events (e.g., help and harm receipt) and attitudinal (e.g., relationship satisfaction) and behavioral outcomes (e.g., help and harm engagement). Therefore, the mediating effect of affective states on the social exchange of help and harm behaviors as well as the effects of receipt of help and harm on relationship satisfaction are also considered in the current study. Finally, in line with research demonstrating that interpretations of events and affective responses to events are influenced by self-regulatory processes (Brockner & Higgins, 2001), relationship-level approach and avoidant social motives are expected to influence the social exchange of help and harm behaviors, affective reactions to the receipt of help and harm, attitudinal evaluations of the relationships, and how affective states mediate the association between help and harm receipt and help and harm engagement and relationship satisfaction. The model for the current study is summarized in Figure 1. 6 Figure 1. Model of proposed relationships for study. The current study contributes to extant research in a number of ways. First, the study examines affective and behavioral responses to the receipt of help and harm within dyadic relationships and over time. Second, the study applies Gable’s (2006) model of approachavoidance social motivation model to examine how dyadic relational motives shape affect-driven and behavioral responses to the receipt of help and harm. Third, by applying Gable’s model this study brings a theoretical perspective that has received fervent interest in social and relationship psychology into the realm of coworker relationships. Fourth, the study draws on theoretical underpinnings of the relational self in order to examine multiple relationships within the same employee (Chen et al., 2006). In the paragraphs that follow, the relational self is defined and its implications for studying interpersonal relations are highlighted. Next, a general review of approach-avoidance motivation is provided, followed by a discussion of approach and avoidance social motives and 7 how they are expected to influence behavioral and affective responses to the receipt of help and harm. SET, AET and the emotion-centered model are then described as theoretical mechanisms, commonly adopted by organizational researchers to understand relational phenomena, linking the receipt of help and harm to behavioral, affective, and attitudinal responses to the receipt of help and harm. Hypotheses are then made about how relationship-level approach-avoidance social motives are expected to influence processes stipulated by SET, AET and the emotioncentered model. Finally, additional exploratory questions are raised addressing how proximal promotion and prevention regulatory foci (Higgins, 1997; 1998) relate to the receipt of help and harm and engagement in help in harm and how their effects are affected by approach-avoidance relational motives; how individual differences in emotional intelligence (Wong & Law, 2002) affect how affective states drive behavioral responses to the receipt of harm; and how interpersonal trust relates to relational approach and avoidance motives and the engagement in help and harm and relationship satisfaction. The Relational Self Organizational researchers have long been interested in how employees can have qualitatively different types of relationships with different coworkers. Much of this theorizing has drawn on role theory (relational identity, Sluss & Ashforth, 2007; leader-member exchange, Graen & Scandura, 1987) and working models of relationships (attachment theory; Kahn, 1988; Richards & Schat, 2011). Recent developments in social and cognitive psychology have articulated processes underlying how relating with others can influence how peoples’ selfconcept functions, and how relationship-level self-regulatory processes can be activated to shape reactions to experiences within relationships. 8 Early work in social psychology was concerned with how peoples’ relations with others shape their sense of self. Symbolic interactionists postulated that the person and society are mutually constructed in the course of social interaction (Stryker & Stratham, 1985). For example, Cooley (1902) discussed how people infer self-meaning in relation to other people, through a process of imagining or inferring how one’s self appears to others, that is “a looking glass self.” With the emergence of relationship and social cognitive psychology, forming the basis of relational cognition (Reis & Downey, 1999), relationships began to be conceptualized as mental representations of the self and significant others. In their review of the social and cognitive psychological research of relational influences, Chen and colleagues (2006) conceptualize a relational self. The relational self reflects who a person is in relation to his or her significant others (e.g., "me when I'm with my coworker"). Significant others are defined as actual (vs. hypothetical) individuals whom one knows (vs. just met) with whom one feels some degree of closeness, and usually with whom one shares a relationship that can be normatively (e.g., friend, coworker) or idiosyncratically labeled (e.g., my closest teammate; Chen et al., 2006). According to the relational self, people store general and specific social knowledge in memory about specific individuals with whom they have interacted and larger groups of people. Thus, people possess multiple relational selves and these selves exist at varying levels of specificity. For example, the relationship-specific relational self designates the self in relation to a specific significant other (e.g., a coworker). The relational self is composed of attribute- and role-based conceptions of the self in the context of the relevant significant other. For example, the self in relation to a coworker with whom one is a mentor might include "joke-ster" and “fun-loving" and normative roles such as "authority figure" whereas the self in relation to a coworker with whom one is a subordinate might include “timid” and “amenable.” Research on relational 9 schemas has provided support for the notion that relational selves are composed of attributes, evaluations, affect, goals, and behavior that characterize the self when relating to certain significant others (Baldwin, 1992). For example, this research has demonstrated that peoples’ relational selves are affect laden: Baldwin, Carrell, and Lopez (1990) found in a lab study that Roman Catholic women who viewed sexually explicit images and are then shown a subliminal message of the pope with a disapproving expression showed more negative affect than the control group of women who viewed neutral images. The authors reasoned that affect appeared to be tied to women’s poperelevant relational selves. Relational selves also appear to influence the goals and motives that people pursue. Goals are thought to be stored in memory as a part of a relational self and when a relational self is activated, the goals that one typically pursues in relation to that significant other are set in motion (Andersen & Chen, 2002). For example, adopting a transference paradigm, in which a significant other representation is activated during an encounter with a new person, Andersen, Reznik, and Manzella (1996) found that the motivation to approach or avoid a relevant significant other is played out in relation to new others who are similar to the significant other. Further, outside the transference domain, theories conceptualizing relationships based on attachment theory view goals and motives as a core component of relationship working models. Attachment theory posits that attachment bonds are formed to fulfill some need for security or protection (Bowlby, 1969). For example, relationship working models associated with a secure relationship are generally more likely to contain intimacy goals than are avoidant working models. Thus, in terms of the relational self, the self activated in the context of secure relationships is more likely to pursue intimacy goals than the self in relation to significant others 10 with whom one shares an avoidant relationship. Overall, Chen and colleagues (2006) argue that relational selves are highly accessible sources of goals and motives, and provide self-regulatory direction; they orient the individual toward others in the world and guide the individual’s behavior in goal serving directions. Coworker relationships can therefore be considered specific contexts in which a relationship-specific portion of a focal employee’s self-concept is activated (i.e., a relationshipspecific relational self). Along with this activated sense of self, behaviors, cognitions, emotions, motives, and behaviors are also activated and such phenomena are expected to vary within the same employee across his/her different relational selves. This expectation parallels that of trait activation theory (Tett & Gutterman, 2000), which emphasizes how personality traits require trait-relevant situations for their expression. For example, an aggressive person will be more likely to behave aggressively in a context that cues aggression (e.g., competitive team), much like a relational context that cues a relational self and corresponding motives associated with that activated self. Related to the current study, different coworkers will activate different relational selves within the context of a focal employee and social motives associated with the activated relational self will also be activated. In the sections that follow, Gable’s (2006) hierarchical model of approach and avoidance social motives is used to articulate how the activation of approach and/or avoidance relational motives has implications for how employees will experience and respond to receipt of help and harm from certain coworkers. But first, a brief review of research on self-regulation is provided. Approach and Avoidance Motives in Interpersonal Relationships One important factor likely to influence interpretation of interpersonal interactions is the perceiver’s motives. Research has demonstrated that motivation can influence how social 11 information is processed (Strachman & Gable, 2006). For example, Molden and Higgins (2004) demonstrated that perceiver motivations influenced perceiver categorizations of ambiguous interpersonal behaviors. Indeed, an important implication of theories about self-regulation is that individual motives stimulate sensitivity to social information. Self-regulation refers to the process in which people seek to align themselves (i.e., their behaviors and self-conceptions) with appropriate goals or standards (Higgins, 1997). According to Carver and Scheier’s (1990) model, self-regulation involves a feedback process in that information from the environment is compared to an internal reference, an output occurs, the environment is reevaluated and compared to the internal reference and the process continues. One type of feedback process attempts to reduce the discrepancy between the input and internal reference (discrepancyreducing) and another type of feedback process attempts to increase the discrepancy between the input and internal reference (discrepancy-enhancing). These two systems have been termed approach and avoidance processes, respectively. People’ motivations to approach pleasure and avoid pain are referred to as the hedonic principle (Higgins, 1997). According to the hedonic principle, the approach system is associated with movement toward desired, positive outcomes, whereas the avoidance system is associated with movement away from undesirable, negative outcomes (Carver, 1996; Gable, Reis, & Elliot, 2003). Approach and avoidance motivation are traditionally conceptualized as temperaments (Elliot & Thrash, 2002) and appear to map onto two distinct conceptual nervous systems. Gray (1990) labeled one nervous system the behavioral activation system (BAS), which is posited to facilitate behavior and produce positive affect. High BAS sensitivity indicates a general sensitivity to signals of non-punishment or reward (i.e., positive, desirable stimuli) and manifests as a perceptual readiness and strong emotional responsiveness to positive stimuli. Gray labeled 12 the second the behavioral inhibition system (BIS), which is posited to inhibit behavior and produce negative affect. High BIS sensitivity indicates a general sensitivity to signals of nonreward or punishment (i.e., negative, undesirable stimuli) and is manifested as a perceptual readiness and emotional reactivity to negative stimuli and the tendency to move away from such stimuli (Gray, 1990). The BAS is the neuroanatomical correlate of the approach system and the BIS is the neuroanatomical correlate of the avoidance system. The BAS/approach system is accompanied by a perceptual vigilance for, affective reactivity to, and a behavioral predisposition toward positive stimuli, whereas the BIS/avoidance system is accompanied by a perceptual vigilance for, affective reactivity to, and behavioral predisposition to negative stimuli (Elliot & Thrash, 2002). Informed by Gray’s (1990) theory, Gable and colleagues (2000) utilized an experience-sampling methodology to relate emotional reactions to daily experiences and found that higher BAS sensitivity predicts greater daily positive affect, whereas higher BIS sensitivity predicts greater daily negative affect. Linking neurobiological, affective, cognitive, and behavioral components of personality, in regards to the hedonic principle, Elliot and Thrash argue that “the distinction between approach and avoidance motivation is fundamental and integral to the study of affect, cognition, and behavior” (Elliot & Thrash, 2002, p. 804). In an effort to extend upon the hedonic principle that people are motivated to approach pleasure and avoid pain, Higgin’s (1997; 1998) regulatory focus theory posits that there are important differences in how people approach pleasure and avoid pain. That is, although approach and avoidance temperaments provide impetus for behavior, they do not offer strategies for how such motives are pursued (Elliot et al., 2006). Regulatory focus theory suggests that the hedonic principle of approaching pleasure and avoiding pain operates differently, depending on the needs that people are trying to satisfy. Growth and development needs predominate for those 13 who are promotion focused, whereas security needs drive those who are prevention focused. Promotion focus is concerned with nurturance needs and involves striving for ideals through aspirations and accomplishment; it prompts behaviors intended to move people closer to desired end-states, the ideal self. Prevention focus regulates security needs and involves fulfilling duties and obligations through vigilant and responsible behaviors; it prompts behaviors intended to avoid conditions that pull people away from desired end-states, the ought self (Higgins, 1997; 1998). Regulatory foci are differently involved in approach and avoidance (Higgins, Shah, & Friedman, 1997). Indeed, even though they are distinct constructs (Elliot & Thrash, 2010), approach and avoidance temperaments appear to influence regulatory foci at the strategic and behavioral level (Scholer & Higgins, 2008). Meta-analytic evidence suggests that high levels of approach temperament increase the likelihood that employees adopt eagerness promotion strategies focused on moving them towards desired end-states and high levels of avoidance temperament increases the likelihood that employees adopt vigilant prevention strategies focused on avoiding conditions that pull them away from desired end-states (Lanaj, Chang, & Johnson, 2012). Further, a recent study by Winterheld and Simpson (2011) demonstrated that regulatory focus orientations and approach-avoidance motivations influence interpersonal processes in similar ways. Specifically, they reasoned that in relationship-threatening situations (e.g., conflict), strongly promotion focused people, similar to those high in approach motivation, are sensitive to positive partner behaviors because they support the underlying desire for successful conflict resolution and relationship growth and advancement. Strongly prevention focused people, similar to those high in avoidance motivation, are sensitive to negative partner behaviors to avoid threatening relationship security. Therefore, promotion focus tends to be the strategic 14 means by which people high on approach motivation move towards a desired end-state and prevention focus tends to be the strategic means by which people high on avoidance motivation avoid increasing the discrepancy between themselves and the end-state (Lanaj et al., 2012). When conceptualized as individual differences, tendencies to adopt promotion and prevention focus are related to tendencies to adopt approach or avoidance motivation orientation (Amodio, Shah, Sigelman, Brazy, & Harmon-Jones, 2004). Although recent research has also identified interpersonal circumstances under which regulatory foci make predictions above and beyond approach-avoidance (see Winterheld & Simpson, 2011). The effects of regulatory foci will be discussed in exploratory fashion later in the current study. Until relatively recently, very little research on self-regulation has focused explicitly on social motives and goals. Historically speaking, social motivation has typically explored the need for belonging (Baumeister & Leary, 1995). The desire to fulfill the need to belong is thought to underlie employees’ preference to work in groups rather than alone (Alderfer, 1972), why they cooperate with others (Kramer, 1993), and why they refrain from engaging in actions that may harm their coworkers (Hollinger & Clark, 1982). Thwarted belonging has been demonstrated to lead to depression, sadness, and lowered self-esteem in the general population (for an overview, see Baumeister, Twenge, & Ciarocco, 2002), and engagement in self-defeating interpersonal behaviors in the work context (Thau, Aquino, & Poortvliet, 2007). Considering the importance of peoples’ basic motivation to belong, social psychologists have investigated the motivational processes involved in establishing and maintaining social bonds. These motivational processes are largely determined by social motives and goals. According to these researchers, interpersonal relationships consistently present both incentives (social support and affiliation) and threats (conflict and incivility) and thus there is utility in addressing how social motives and goals 15 associated with approaching incentives and avoiding threats within relationships leads to behavioral and affective outcomes (Gable, 2006). Early research on social motivation defined the need for affiliation as motives stemming from insecurity, rejection, and social isolation (Atkinson, Heyns, & Veroff, 1954). However, in light of contradictory findings (need for affiliation was negatively correlated with popularity and positively correlated with self-confidence; Atkinson et al., 1954) the focus shifted to two submotives – need for affiliation and fear of rejection – posited based on expectations of positive and negative reinforcement in interpersonal relationships, respectively (Mehrabian, 1976). Mehrabian (1976) found that people high in need for affiliation were less anxious, elicited more positive affect from others, and were more self-confident than people low in need for affiliation whereas people high in sensitivity to rejection were less confident, more anxious, and were judged less positively by others than people low on sensitivity to rejection. This early work on social motivation, however, did not distinguish between approach social motives from avoidance social motives (Gable, 2006). In order to further understanding of how approach and avoidance social motives differentially influence interpersonal outcomes, Gable (Gable, 2006; Gable & Strachman, 2006) adapted Elliot and Church’s (1997) Hierarchical Model of Approach and Avoidance Motivation, which was initially applied to work on achievement motivation, to the social domain. Gable’s (2006) Hierarchical Model of Approach-Avoidance Social Motivation has received fervent interest by social psychologists in recent years and has been applied to the study of dispositional social motives (Gable, 2006), and friendship relationships (Elliot et al., 2006), romantic relationships (Impett, Gable, & Peplau, 2005; Impett & Gordon, 2010; Impett, Gordon, Kogan, Oveis, Gable, & Keltner, 2010), and sexual relationships (Impett, Starchman, Finkel, & Gable, 2008). 16 Hierarchical Model of Approach-Avoidance Social Motivation. The hierarchical model (Gable, 2006) outlines how dispositional individual differences in distal social motives, environmental factors, and short-term proximal goals influence social cognition, affect, and behaviors. In this model, a motive is an affectively-based tendency that orients individuals toward positive or negative social stimuli (which is functionally similar to but distinct from a motivation temperament which reflects a neurobiological sensitivity to positive [approach] or negative [avoidance] stimuli; Elliot & Thrash, 2002; Gray, 1990). The model is hierarchical because “dispositional approach social motives and incentives in the social environment are hypothesized to predispose people to adopt short-term approach social goals; and dispositional avoidance social motives and threats in the social environment are hypothesized to predispose people to adopt avoidance social goals” (p. 177; Gable, 2006). For example, in a discussion about a task to complete an employee inventory, an employee who has a high approach social motive will be more likely to adopt approach goals, such as “I want to us to have a pleasant discussion so that I can become closer to my coworker”; whereas an employee who has a high avoidance motive will be more likely to adopt an avoidance goal, such as “I don’t want to embarrass myself or start an argument with my coworker and for neither of us to be resentful of the outcome.” The approach and avoidance motives are also posited to be associated with different social outcomes. Approach social motives lead more strongly to positive relationship outcomes, such as affiliation and companionship, whereas the avoidance social motives lead more strongly to negative relationship outcomes, such as rejection conflict (Gable, 2006). People with strong approach motives define successful interactions and relationships as those which provide desired outcomes (e.g., affiliation and social support); and painful relationships are defined as those that do not 17 provide these desired outcomes. Individuals with strong avoidance motives define positive interactions and relationships as those that lack threats to the desired outcome (e.g., disagreements and anxiety); and painful relationships are defined as those that possess these threats to the desired outcome. Recent research in social psychology has found general support for the hierarchical model of approach-avoidance social motivation. Gable (2006) found that general approach motives was associated with more positive social attitudes, satisfaction with social bonds, and less loneliness, whereas general avoidance social motives was associated with more negative social attitudes, relationship insecurity, and loneliness. Elliot et al. (2006) found that general approach motives were associated with greater subjective well-being, whereas avoidance social goals were associated with more reports of physical health symptoms 3.5 months later. In a study of romantic relationships, Impett and colleagues (2008) found that people with strong approach motives in their relationship had higher maintained sexual desire on a daily basis over six months than individuals lower in approach motives in their relationship. Further, in an event sampling study of romantic relationships, Impett and colleagues (2010) found that relationship approach motives were associated with increased relationship satisfaction on a daily basis and over time, and relationship avoidance motives were associated with decreased relationship satisfaction overtime. In an experimental investigation, Strachman and Gable (2006) found that avoidance social goals were associated with negatively biased interpretation of ambiguous social cues and more pessimistic evaluations of social actors, and that strangers remember more negative information (adjectives such as aggressive, selfish, out-spoken) about their interacting partner when avoidance social goals are primed compared to when approach social goals are primed. 18 Considering that approach motives reflect a general sensitivity and affective responsiveness to positive stimuli and avoidance motives reflect a general sensitivity and affective responsiveness to negative stimuli (Elliott & Trash, 2002; Gray, 1990), behavioral and affect-driven reactions to the receipt of help and harm are likely to vary depending on the activation of approach or avoidance social motives. Below, SET and AET theoretical perspectives are described highlighting the how help and harm can be reciprocated within relationships and the central role of affect in influencing engagement in help and harm and relationship satisfaction within relationships. Social Exchange and the Reciprocation of Help and Harm According to SET, individuals’ voluntary behaviors are driven by obligations that arise through a series of social transactions (Blau, 1964; Emerson, 1976; for a review, see Cropanzano & Mitchell, 2005). The transactions are typically considered to be interdependent and contingent on the actions of another person (Blau, 1964). When two people interact they become involved in a continuing exchange that results in the reciprocation of inputs and outputs within the relationship. This reciprocation is guided by the norm of reciprocity (Gouldner, 1960) such that individuals respond to others’ actions in kind. Norms of reciprocity can be both positive and negative: Positive norms of reciprocity motivate individuals to respond to positive treatment with positive treatment and negative norms of reciprocity motivate individuals to respond to negative treatment with negative treatment (Gallucci & Perugini, 2003; Gouldner, 1960; Perugini, Gallucci, Presaghi, & Ercolani, 2003). It thus follows that the direction of help by an actor employee toward a target coworker should trigger a positive reciprocity norm, motivating that target coworker to respond in kind by directing help toward the actor employee. Likewise, the direction of harm by an actor employee 19 toward a target coworker should trigger a negative reciprocity norm, motivating that target coworker to respond in kind by directing harm toward the actor employee. These reciprocal transactions can be self-reinforcing and cycles of social exchange can be difficult to break (Cropanzano & Mitchell, 2005). Indeed, in their theoretical model of incivility, Andersson and Pearson (1999) argued that the desire for reciprocation may eventually spark an escalating spiral of negative behavior. Empirical evidence has demonstrated that the reciprocation of help and harm is dyadic and employees develop different social exchange relationships with different coworkers (Lyons & Scott, 2012). In a social network study (dyadic level of analysis), Lyons and Scott (2012) found that the receipt of help and harm by a given employee from a certain coworker was associated with the extent to which that employee engaged in help and harm towards that coworker. Lyons and Scott (2012) also found that the behaviors exchanged between a focal employee and a given coworker were equivalent, demonstrating support for the homeomorphic reciprocity principle of SET (Gouldner, 1960). That is, engaging in help, but not harm, was associated with receiving help, and engaging in harm, but not help, was associated with receiving harm. That is, they found that employees exchanged “tat for tat” as opposed to “tit for tat” (Gouldner, 1960). However, the Lyons and Scott (2012) study utilized a cross-sectional design which limited the generalizability of the inferences to help and harm that vary within individuals and over time (Dalal et al., 2009). The current study builds upon this work by using an experience-sampling methodology. It is hypothesized that: Hypothesis 1: Receipt of help will be positively associated with engagement in help. Hypothesis 2: Receipt of harm will be positively associated with engagement in harm. The extent that these interdependent exchanges have the potential to generate highquality relationships may in fact depend upon social motives associated with the specific 20 relationships. As noted earlier, relationship outcomes tend to be more positive in approachoriented relationships as opposed to avoidant-oriented relationships (Gable, 2006) and social exchange may be one mechanism explaining such differences. SET (Blau, 1964, Gouldner, 1960) describes the reciprocation of help in harm as being driven by a cognitive rational calculation of who owes what to whom. Sensitivity to such debits and credits would therefore affect the social exchange. According to Gable’s (2006) approach-avoidance model, approachavoidance social motives sensitize individuals toward positive or negative social stimuli. This is because approach-oriented motives individuals are particularly attuned to achieving positive aspirations for the relationship, such as growing closer to someone. In avoidance-oriented relationships, individuals are particularly attuned to avoiding threats for the relationship, such as conflict. As a result, approach social motives sensitize individuals to positive interactions and tend to be related more strongly to positive relationship outcomes, such as stronger affiliation and companionship, whereas the avoidance social motives, that sensitize individuals to negative interactions, tend to be more strongly related to negative relationship outcomes, such as rejection (Gable, 2006). It is thus expected that employees in approach-oriented relationships will be more sensitive to receiving help, thereby amplifying calculations of positive reciprocal debits and credits, and enhancing positive social exchange and reciprocation of help engagement. However, in avoidance-oriented relationships employees will be more sensitive to receiving harm, thereby amplifying calculations of negative reciprocal debits and credits, and enhancing negative social exchange and reciprocation of harm engagement. Hypothesis 3: The association between help receipt and help engagement will be stronger in approach-oriented relationships compared to avoidance-oriented relationships. 21 Hypothesis 4: The association between harm receipt and harm engagement will be stronger in avoidance-oriented relationships compared to approach-oriented relationships. Affective State Responses to the Receipt of Help and Harm Affective Reactions to Receipt of Help and Harm. According to AET (Weiss & Cropanzano, 1996), fluctuations in affect are predictable and influence workplace behaviors and attitudes. Affective states are highly variable and fluctuate within persons over time (Dalal et al., 2009). Affective states are a functional mechanism that organizes ongoing activity to force attention on pressing events that are relevant to physiological needs or that induce disturbing cognitive associations, such as threats to well-being, goal attainment, or esteem and physiologically energize individuals to induce appropriate action (Spector & Fox, 2002). Affective-events – such as receiving help from a coworker or being criticized by a coworker – disrupt existing endogenous affective cycles. In turn, affect (anger, joy) induced by these events subsequently influences work attitudes (e.g., job satisfaction) and work behaviors, such as helping behaviors, confrontational or harming behaviors, and/or work withdrawal. Positive events, such as receiving help, elicit positive affective states because they are appraised as enhancing well-being, esteem or bringing one closer to goal achievement (Weiss & Cropanzano, 1996; Spector & Fox, 2002). Negative events, such as receiving harm, elicit negative affective states because they are appraised as being threatening to well-being, esteem or goal attainment (Weiss & Cropanzano, 1996). Hypothesis 5: Receipt of help will be positively associated with experiences of positive affective states. 22 Hypothesis 6: Receipt of harm will be positively associated with experiences of negative affective states. An additional component of AET suggests that individual dispositions influence the impact that work events have on affective reactions (Weiss & Cropanzano, 1996), a feature similar to what Bolger and Zuckerman (1995) call differential reactivity. That is, people high in a certain disposition should react more severely (i.e., have a stronger affective reaction) than people low in that disposition. The majority of studies on differential activation have investigated the impact of two personality traits, extraversion and neuroticism, on affective reactions to stressful events (Bolger & Zuckerman, 1995) Extraversion relates to the BAS that is more sensitive to signals of reward (Gray, 1990), and extraverts are predisposed to react more positively to events that elicit positive affective states. Neuroticism relates to the BIS that is more sensitive to signals of punishment (Gray, 1990), and neurotic individuals are predisposed to react more negatively to events that elicit negative affective states. In a daily-diary study, Rodell and Judge (2009) demonstrated that hindrance stressors related more strongly to anger for people high on neuroticism (Rodell & Judge, 2009). Additionally, in another daily-diary study, Judge, Scott, and Ilies (2006) demonstrated that perceived injustice related more strongly to state hostility for people higher in trait hostility. Self-regulation theories have also examined how motives sensitize individuals to experience specific affective states (Brockner & Higgins, 2001). Activation of approach and avoidance motivational systems are associated with different affective outcomes (Gable et al., 2000). For example, Gable and colleagues (2000) found that high BAS sensitivity was associated with more daily positive affect and high BIS sensitivity was predictive of increased daily negative affect. High BAS sensitivity reflects an increased attentiveness to potential reward and 23 high BIS sensitivity reflects an increased attentiveness to punishment. Approach motivation is associated with feelings of eagerness, excitement, and elation whereas avoidance motivation is associated with feelings of anxiety and frustration (Carver & White, 1994; Gray, 1990). In this model, the approach and avoidance motivation systems are viewed as responsible for positive and negative feelings, respectively, but approach is not associated with negative feelings and avoidance is not associated with positive feelings. It is thus expected that when a focal employee interacts with a specific coworker, social motives associated with the activation of the relationship-specific relational self will function similarly to the differential reactivity effects of motivational dispositions. Approach social motives will intensify positive affective reactions to positive events, whereas avoidance social motives will intensify negative affective reactions to negative events. Hypothesis 7: The association between receipt of help (a positive event) and positive affective states will be stronger in approach-oriented relationships compared to avoidance-oriented relationships. Hypothesis 8: The association between receipt of harm (a negative event) and negative affective states will be stronger in avoidance-oriented relationships compared to approach-oriented relationships. Affect-Driven Engagement in Help and Harm and Satisfaction. According to AET, people engage in behaviors that are intended to deal with their particular affective reactions (Weiss & Cropanzano, 1996). Affective states induce action tendencies that tend to elicit certain behaviors, quite often motivating behavior that will reduce negative feelings and enhance positive feelings (Spector & Fox, 2002). Behaviors most directly influenced by affective events are what Weiss and Cropanzano call “affect-driven behaviors” and help and harm are considered 24 affect-driven behaviors. Dalal, and colleagues (2009) demonstrated, via two experience-sampling studies, that OCB and CWB are affect-driven phenomenon and exhibit considerable withinperson variation. Another relevant theory is Spector and Fox’s (2002) emotion-centered model. According to the emotion-centered model, situations induce affective states which influence the likelihood that individuals will engage in either help or harm. In line with AET, affect focuses attention on the events that caused it. Action tendencies accompanying affective states can immediately and automatically prompt behavior, or they can facilitate intentions to engage in a given behavior when the opportunity arises (Sepctor & Fox, 2002; Weiss & Cropanzano, 1996). In general, action tendencies accompanying positive affective states (joy, happiness) are prosocial, motivating individuals to draw in or engage the perceived source of the pleasant feelings (resulting in behaviors that include friendliness and cooperation; help), whereas action tendencies accompanying negative affective states (anger, hostility) are antisocial, motivating individuals to repel the perceived source of the unpleasant feelings (resulting in behaviors that include aggression and antagonism; harm). Importantly, negative affective states interfere with cognitive processes involved in moral judgments, lower inhibitions, and prime aggressive thoughts and scripts (Berkowitz, 2003; Tedeschi & Felson, 1994), increasing the likelihood that the states of action readiness associated with negative affective states will be manifested. The target for the help or harm is predicated on the perceived agent of the situation that induced the affective state. Thus, an individual who feels mistreated by a coworker will most likely direct harm toward that individual rather than people in general (Spector & Fox, 2002). Recently, empirical research adopting experience-sampling methodologies and examining the behavioral consequences of affective states provided evidence supporting 25 propositions of AET and the emotion-centered model. For example, Rodell and Judge (2009) demonstrated that daily experiences of challenge stressors (i.e., stressors associated with gains and growth) related to daily engagement in OCB and this effect was mediated by attentiveness (positive affective state), and daily experiences of hindrance stressors (i.e., stressors viewed as obstacles to growth) related to daily engagement in CWB and this effect was mediated by anger (negative affective state). Ilies, Scott, and Judge (2006) demonstrated that employees who experience positive affective states over a period of time (e.g., day, week) were more likely to engage in OCBs during that time, and Judge, Scott, and Ilies (2006) found that daily perceptions of interpersonal mistreatment related to higher daily engagement in CWB and this effect was mediated by state hostility. The empirical evidence presented here also supports the two-factor approach to affect, on which the emotion-centered model is based (Spector & Fox, 2002). According to the two-factor approach, positive and negative affect are seemingly controlled by different areas of the brain – positive affect is controlled by the BAS and negative affect by the BIS (Spector & Fox, 2002). Thus, high (low) positive affective states are expected to lead to high (low) engagement in helping, but is not expected to be related to engagement in harming, whereas high (low) negative affective states are expected to lead to high (low) engagement in harming but is not expected to be related to engagement in helping. In a meta-analysis, Dalal (2005) found that at the betweenperson level, OCB and CWB engagement were at best moderately negatively related and, thus, thought of as relatively distinct constructs. Dalal and colleagues (2009) replicated the betweenperson effects at the within-person level in an experience-sampling study and found that employees engaged in more OCB during experiences of positive affective states (but not 26 negative affective states) and more CWB during experiences of negative affective states (but not positive affective states). Avoidance motivation is associated with anxiety emotions and the action tendency associated with anxiety is inhibition and withdrawal behaviors (Gray, 1990; Lazarus, 1991). Therefore, it is expected that individuals in avoidance relationships will engage in different forms of harm in response to the receipt of harm than is typically examined adopting AET and the emotion-centered model. That is, unlike anger and hostility that are associated with more active aggressive action tendencies (Lazarus, 1991), anxiety and frustration may be associated with more covert and passive forms of aggression, even though both forms of aggression may serve a similar underlying function of repelling the perceived source of the negative affective states. In their theoretical conceptualization of workplace aggression, Baron and Neuman (1996) distinguished between active and passive aggressive behaviors. Active aggression produces harm through the performance of some behavior (e.g., insulting or criticizing a coworker) whereas passive aggression is consistent with withdrawal and delivers harm through the withholding of some action (e.g., avoiding or ignoring a coworker). Therefore, in addition to traditional conceptualizations of interpersonal CWBs and aggression, in the current study, engagement in harm in avoidance relationships may also expected to be associated with passive harming behaviors. Hypothesis 9: Positive affective states will be positively associated with engagement in help (but not harm). Hypothesis 10: Negative affective states will be positively associated with engagement in harm (but not help). 27 It thus follows that affective states may serve as a partial mediator of the social exchange associations between help and harm receipt and help and harm engagement. Hypothesis 11: Positive affective states will mediate the association between receipt of help and engagement in helping (but not harming). Hypothesis 12: Negative affective states mediate the association between receipt of harm and engagement in harm (but not help). AET also makes direct predictions about linking affective states to attitudinal outcomes. According to AET, an attitude is an evaluative judgment about some foci that is influenced by affective experiences (Weiss & Cropanzano, 1996). The affective components of attitudes reflect the recall of affective experiences associated with the evaluative foci and such attitudes are expected to vary with affect. Social psychologists have demonstrated that affective states are predictive of relationship satisfaction over time (Gonzagos, Campos, & Bradbury, 2007; Impett et al., 2010). For example, Impett and colleagues (2010) drew from Frederickson’s (2001) broaden-and-build theory which suggests that positive affective states broaden people’s attention and thinking, and these broadened outlooks help people to discover and build consequential personal resources such as social support and enhanced feelings of satisfaction. Therefore, it is expected that: Hypothesis 13: Positive affective states will be positively associated with relationship satisfaction. Hypothesis 14: Negative affective states will be negatively associated with relationship satisfaction.. It thus follows that affective states may serve as a partial mediator of the relationship between receipt of help and harm and relationship satisfaction. This expectation is in line with 28 meta-analytic evidence demonstrating that employees who experience support from their coworkers have higher work attitudes, including job satisfaction (Chiaburu & Harrison, 2008), and employees who experience antagonism or harassment from their coworker have lower work attitudes (Bowling & Beehr, 2006; Chiaburu & Harrison, 2008). Hypothesis 15: Positive affective states will mediate the association between receipt of help and relationship satisfaction. Hypothesis 16: Negative affective states mediate the association between receipt of harm and relationship satisfaction. Finally, in line with arguments made for hypotheses 3, 4, 7, 8, 13, 14, 15 and 16 approach-avoidance relationship motives may moderate the strength of the mediating effects of affective states on the social exchange of help and harm and the effects of receipt of help and harm on relationship satisfaction. Hypothesis 17: The mediating effect of positive affective states on the association between help receipt and help engagement will be stronger in approach-oriented relationships compared to avoidance-oriented relationships. Hypothesis 18: The mediating effect of negative affective states on the association between harm receipt and harm engagement will be stronger in avoidance-oriented relationships compared to approach-oriented relationships. Hypothesis 19: The mediating effect of positive affective states on the association between help receipt and relationship satisfaction will be stronger in approach-oriented relationships compared to avoidance-oriented relationships. 29 Hypothesis 20: The mediating effect of negative affective states on the association between harm receipt and relationship satisfaction will be stronger in avoidance-oriented relationships compared to approach-oriented relationships. Promotion and Prevention Regulatory Foci Although the majority of research on motivational variables in interpersonal contexts has drawn on approach-avoidance motivations, a handful of recent studies have drawn from regulatory focus theory. In particular this research has explored how interpersonal experiences evoke promotion and prevention regulatory foci that then in turn induce strategic behavioral responses (Crowe & Higgins, 1997; Molden, Lucas, Gardner, Dean, & Knowles 2009; Osyerman, Uskul, Yoder, Nesse, & Williams, 2007). Independent of approach-avoidance relationship motives, interpersonal behaviors from specific coworkers may make prevention or promotion concerns temporarily salient. Research on romantic relationships has demonstrated that although most people should feel good (bad) about positive (negative) partner behaviors, depending on the meaning of the positive (negative) behaviors to them, promotion or prevention concerns may be made salient and uniquely influence subsequent behavioral responses to the receipt of such behaviors (Higgins & Schoeler, 2008; Winterheld & Simpson, 2011). Evocation of promotion and prevention foci may have unique effects on engagement in help and harm in dyadic relationships compared to that which would be expected by approach-avoidance motives alone. In fact, promotion and prevention regulatory foci may act more proximally to influence the engagement help and harm than approach-avoidance relationship motives. Therefore, the current study examined how receipt of help and harm evoke promotion and prevention regulatory foci which in turn affect engagement in help and harm. The effects of proximal promotion and prevention regulatory foci were be considered in the context of approach and 30 avoidance relationship motives that are expected to affect how the receipt of help and harm relate to the adoption of such foci. As was briefly summarized earlier, the two motivational systems proposed by regulatory focus theory (Higgins, 1997; 1998) include: (1) promotion focus, which facilitates the fulfillment of people’s nurturance needs through the pursuit of hopes and aspirations and is concerned with personal growth and advancement, and (2) prevention focus, which facilitates the achievement of security needs through the fulfillment of duties and obligations and is concerned with safety and protection. When pursuing promotion concerns people strive toward rewarding outcomes (i.e., social gains), and they try to avert the absence of positive outcomes (i.e., non-gains, or missed opportunities). When pursuing prevention concerns people work to avert negative outcomes (i.e., social losses) and strive toward the absence of negative outcomes (i.e., non-losses, or absence of threats). Some theorists do contend that neither regulatory foci are identical to the approach (concerned with approaching positive outcome) or avoidance (concerned with avoiding negative outcome) systems, such that both regulatory foci are concerned with obtaining positive outcomes (i.e., prevention focus is concerned with security and safety and promotion focus in concerned with growth and nurturance; Crowe & Higgins, 1997). However, regulatory foci and approachavoidance have been found to influence interpersonal outcomes in similar ways (Higgins & Schoeler, 2008; Winterheld & Simpson, 2011). Indeed, approach motives increase people’s tendency to adopt promotion strategies, and avoidance motives increase people’s tendency to adopt prevention strategies (Lanaj et al., 2012). When activated by some relationship-threatening event (e.g., an argument) a prevention focus leads people to be more sensitive to negative partner behaviors and less willing to engage in risky conflict-resolution behaviors, similar to individuals high in avoidance motivation (Osyerman et al., 2007). Similarly, when promotion focus is 31 primed (after experiencing a positive event), people are more sensitive to positive partner behaviors and are more willing to engage in risky creative conflict-resolution behaviors, similar to people high in approach motivation (Winterheld & Simpson, 2011). However, recent evidence suggests that regulatory foci can account for behavioral responses in ways that are unique to expectations of approach-avoidance motivation. For example, Scholer, Zou, Fujita, Stroessner, and Higgins (2010) found that when people viewed a relationship threat (e.g., argument) has having long-term detrimental effects on their relationship, people with a prevention focus people did not adopt the typical avoidance behaviors (e.g., avoiding the situation), but instead adopted approach oriented behaviors (e.g., actively working to resolve the conflict) in order to restore safety and security in the relationship. Likewise, Winterheld and Simpson (2011) found that in the face of relationship conflict, preventionfocused people engaged in relationship restoration behaviors (e.g., discussion to resolve conflict). Further, Molden and colleagues (2009) found that, people who perceived they had been excluded by being ignored evoked a promotion focus that lead them to reestablish social contact, whereas people who perceived they had been excluded by being rejected evoked a prevention focus and withdrew from social contact. As such, previous research has demonstrated that positive interpersonal events (e.g., receipt of help) tend to evoke promotion focus where negative interpersonal events (e.g., receipt of harm) tend to evoke prevention foci. However, the effects of promotion and prevention regulatory foci on engagement in interpersonal behaviors (e.g., engagement in help and harm) tend to differ depending on the individuals’ perception and interpretation of the inciting event. As relationship approach-avoidance social motives may affect individuals’ interpretations of help and harm receipt, approach and avoidance motivation may also affect how individuals 32 interpret the receipt of help and harm in terms of gains or losses. For example, in an approach relationship, help receipt may be more likely to be perceived as a social gain and evoke promotion regulatory focus whereas in an avoidance relationship, help receipt may not be perceived as a social gain and be less likely to evoke promotion regulatory focus. Likewise, in an avoidance relationship, harm receipt may be more likely to be perceived as a social loss and evoke prevention regulatory focus whereas in an approach relationship, harm receipt may not be perceived as a social loss (but an opportunity for gain) and be less likely to evoke prevention regulatory focus. However, an alternative possibility could be that promotion and prevention regulatory foci are not relevant to avoidant relationships (which is concerned about avoiding threat), and they are only evoked in response to help and harm in approach relationships in which positive aspirations are of primary concern (Crowe & Higgins, 1997). Therefore, one question of interest is if approach and avoidance relationship motives influence the evocation of promotion and prevention regulatory foci in response to the receipt of help and harm. Research Question 1: Do help and harm receipt evoke promotion and prevention regulatory foci within relationships? Does the strength of the association between help and harm receipt and promotion and prevention regulatory foci differ across approach and avoidance relationships? Further, given recent research on the unique effects of regulatory foci in predicting behavioral responses to positive and negative behaviors (Crowe & Higgins, 1997; Molden et al., 2009; Scholer et al., 2010; Winterheld & Simpson, 2011), another question of interest is how promotion and prevention foci relate to the engagement of help and harm towards the coworker with whom a focal employee either has an approach or avoidant relationship motive. 33 Research Question 2: Do promotion and prevention regulatory foci affect engagement in help and harm within relationships? Does the strength of the association between prevention and promotion regulatory foci and help and harm engagement differ across approach and avoidance relationships? It thus follows that promotion and prevention regulatory foci may also partially mediate the association between help and harm receipt and help and harm engagement. Such reasoning is consistent with SET (Blau, 1964, Gouldner, 1960). According to SET, outcomes of the exchange of socioemotional resources (e.g., interpersonal helping and harming behaviors) address social needs such as esteem and affiliation. Socioemotional outcomes are symbolic and represent how a person is valued by the exchange partner (Cropanzano & Mitchell, 2005). Depending on the interpretation of an exchange relationship – whether individuals are perceived to be valued or devalued – will influence the decisions individuals may about how to move forward in the exchange relationship (Cropanzano & Mitchell, 2005). This decision process involves a rational calculation of the consequences and benefits of strategies that can be adopted to attain the desired outcome (Meeker, 1971). Therefore, in a relationship of interdependent exchanges individuals will evaluate their exchanges as they relate to desired outcomes, including belonging. When individuals receive help they may interpret such actions as a social gain bringing them closer to their desired belonging state and focus their regulation on achieving the desired outcome (i.e., promotion focus). When individuals receive harm they may interpret such actions as a social loss distancing them from their desired affiliation state and focus their regulation on preventing further distancing from that state (i.e., prevention focus). Thus, promotion and prevention regulatory foci may also serve as a mediating mechanism between the receipt of help and harm and engagement in help and harm in dyadic relationships. 34 Research Question 3: Do promotion and prevention regulatory foci mediate associations between help and harm receipt and help and harm engagement? Does the relationship approach and avoidance motive moderate this mediation? Exploratory Investigation: Emotional Intelligence and Interpersonal Trust Emotional intelligence. Although separate from the primary focus of the current study, emotional intelligence was also examined as an individual difference a moderator of the withinrelationship associations between help and harm receipt, positive and negative affective states, and engagement in help and harm. During the last two decades, emotional intelligence has become an increasingly popular in organizational research (Joseph & Newman, 2010). This is likely because research has suggested that emotional intelligence is a consistent predictor of job performance (Law, Wong, & Song, 2006; Wong & Long, 2002; for a meta-analytic review, see Joseph & Newman, 2010). Additional evidence suggests that emotional intelligence also predicts leadership effectiveness (Kafetsios, Nezlek, & Vassiou, 2011). Emotional intelligence has its roots in social intelligence, and early on it was defined as the ability of a person to deal with his or her emotions (Salovey & Mayer, 1990). More recently emotional intelligence as an ability has been defined in terms of four dimensions (Davies, Stankov, & Roberts, 1998): (1) understanding of one’s own emotions, (2) understanding of others’ emotions, (3) regulation of one’s own emotions (i.e., in terms of rapid recovery from psychological distress and being able to control one’s temper), and (4) emotional expression that is beneficial for performance (i.e., directing emotions towards constructive activities). Therefore, individuals high in EI should be able to recognize their emotions, regulate their emotions, and manage their emotions in a way that is beneficial to performance (Wong & Long, 2002). Effective management of emotions is an important factor affecting interpersonal relations as 35 individuals who are not sensitive to their own and others’ emotions will have trouble regulating and managing their emotions and will have problems interacting with others (Law et al., 2006). For example, in the current study individuals high in emotional intelligence may be better at not letting their emotions affect their interpersonal behaviors. This may be particularly valuable in response to receiving harm. Interpersonal trust. In addition to addressing the above research questions, the current study also assessed how dyadic approach-avoidance relational motives shape dyadic interpersonal trust and how interpersonal trust affected engagement in helping and harming behaviors and relationship satisfaction. Even though interpersonal trust has typically been conceptualized as dyadic in nature (i.e., trustor and trustee; Mayer, Davis, & Schoorman, 1995), the majority of extant research has examined interpersonal trust as a generalized experience of the individual without considering the specific relationship between the turstor and trustee. The approach of the current study follows a small number of previous studies that have examined interpersonal trust from a dyadic perspective (e.g., Chua, Ingram, & Morris, 2008; Yakovleva, Reilly, & Werko, 2010). Trust is defined by Mayer and colleagues (1995) as the willingness of a trustor to be vulnerable to a trustee based on positive expectations about the trustee’s actions. This definition of trust captures three components: (a) benevolence which reflects expectations about caring or supportive motives, (b) ability, which reflects expectations about competence and skills, and (c) integrity which reflects expectations about a consistent adherence to sound principles. Trust has also been conceptualized to include two factors by McAllister (1995): (a) affect-based trust which reflects expectations of reciprocal care and concern and (b) cognition-based trust which reflects expectations of reliability and dependability. Recently, researchers have highlighted the 36 similarities in both Mayer and colleagues (1995) and McAllister’s (1995) conceptualizations of trust, with benevolence having much in common with affect-based trust and ability and integrity having much in common with cognition-based trust (for example, see Colquitt, LePine, Piccolo, Zapata, & Rich 2012). In the current study McAllister’s (1995) two-factor conceptualization of trust is drawn upon to explore how approach-avoidance dyadic relationship motives affect dyadic affect- and cognition-based interpersonal trust, and how interpersonal trust affects engagement in helping and harming behaviors and relationship satisfaction. 37 METHOD Participants Participants were recruited from three organizations: a large administrative department of a mid-Western governmental agency, two grocery stores of a mid-Western food services retailer, and administrative departments of a large mid-Western University. In order to recruit participants, the researcher gained permission from senior-level management at each organization (i.e., staff supervisor at the government agency, HR manager at the food retailer, and department chairperson and Dean at the university). The organization representatives at the government agency and food retailer provided the researcher with a list of employees who were eligible to participate. Research has demonstrated that employees of different hierarchical positions within an organization (e.g., supervisors and subordinates) exert unique influences on employee interpersonal experiences (Chiaburu & Harrison, 2008). As such, in order to avoid confounding formal status differences within employee relationships with the focal relational constructs of interest, eligible participants were limited to one hierarchical position within each organization. That is, in the government agency eligible participants were limited to nonsupervisory support staff; non-supervisory cashiers and food handlers in the grocery stores; and administrative assistants in the university departments. After they were identified, all eligible prospective participants received an email from the organization representative requesting their participation (for an example of the email sent to university administrative staff see Appendix A). Although it is difficult to ascertain exactly how many individuals received the recruitment email, the email was sent to 195 individuals (100 from the government agency, 81 from the grocery stores, and 14 from the university departments). Two of the university administrative staff recruited an additional one participant each, resulting in a final recruitment total of 197 individuals. Of those who the email was sent to, 78 (44 from the government agency, 9 from the 38 grocery stores, and 16 from the university departments) signed up to participate. At the government agency and grocery stores, participants met with the researcher to review the study logistics and address any questions. They were also provided with an information sheet (in the form of Question & Answers, see Appendix A) outlining logistical issues, the timeframe of their participation, assurance of confidentiality, and method of payment. The administrative staff at the university received this information sheet via email. On a pre-specified date (typically on a Monday one week after participants received the information sheet) participants received an email containing a link to the pre-survey. Of the 78 individuals who initially signed-up to participate, 69 completed the pre-survey. One week following the pre-survey, the 69 participants received the Day 1 survey on a Monday – all 69 participants completed this survey. After data collection had been completed, an additional 14 participants were removed from the sample because their survey responses evidenced that they did not engage in the survey efforts or did not understand the instructions. Specifically, these participants either accessed the survey and only completed a small proportion of questions, only completed a small number of the daily surveys (five or less of 20), and/or responded to questions about the wrong relationship partner (e.g., one person alternated between different coworkers for each of the 20 survey days). The final sample consisted of 55 participants which represented a 28% response rate (of the 197 initially contacted). Of those 55 participants, 83% were female (17% male), 87% White (13 % Black/African-American), with a mean age of 39 years (SD=15.52, ranging from 20 to 66 years), and had an average organizational tenure of 8 years and one month (SD=10 years 6 months). Of the final 55 participants, 34 were from the government agency (76% female; 85% white, 15% black/African-American; mean age=38 years, SD=10.62; mean organizational tenure = 4 years 10 months, SD=5 years one month), six were from the grocery stores (83% female; 39 83% white, 17% black/African-American; mean age=27 years, SD=10.24; mean organizational tenure = 2 years 4 months, SD=10 years), and 15 from the university administration (100% female; 93% white, 7% black/African-American; mean age=47 years, SD=12.91; mean organizational tenure = 17 years 6 months, SD=14 years 8 months). Although participants from each organization were mostly white and female, participants from the grocery stores tended to be younger and had lower organizational tenure than participants from the government agency and university administration. In fact, the average age and organizational tenure of university administration participants was highest relative to participants from the government agency or grocery stores. Compensation for participation varied across organization as supervisors and managers at each organization had different expectations regarding acceptable levels of remuneration. At the government agency, participants were entered into a draw for eight prizes of $25.00. Each day a participants completed a daily survey, their names were entered into a lottery. At the grocery stores, every participant was paid $20.00 regardless of the amount of surveys they completed and they were also entered into a lottery system for two prizes of $60.00. University staff participants were each paid $40.00 regardless of the number of surveys they completed and they were entered into a lottery for four prizes of $60.00. Procedure Data collection took place over 5 weeks. At the beginning of week 1, participants completed the pre-survey. Participants received the pre-survey via email. The pre-survey contained the consent form (for an example of the email sent to university administrative staff see Appendix A). The pre-survey contained the relationship motive manipulation (see instructions below) for which each participant was asked to nominate two coworkers, one with 40 whom they have an approach relationship (Coworker A) and one with whom they have an avoidance relationship (Coworker B). This is to ensure variability on the relationship motive variable. See Figure 2 for a graphical representation of this design. Coworker B Coworker A Approach Motive Avoidance Motive Focal employee Figure 2. Graphical representation of study design. Each participant reported about two relationships: A coworker with whom they have high approach relationships motives (Coworker A), and a coworker with whom they have high avoidance relationships motives (Coworker B). To ensure confidentiality of individuals, as opposed to indicating the nominated coworkers’ actual names, participants were asked to indicate an alias for each nominated coworker. The pre-survey also contained questions assessing additional relationship characteristics for each nominated coworker, including their hierarchical status relative to each coworker, their requirement to work with each coworker, their liking of each coworker, their trusting of each coworker, and their length of relationship with each coworker. Finally, the presurvey assessed individual difference variables, including approach-avoidance disposition, emotional intelligence, and demographics (i.e., age, gender, race/ethnicity, organizational tenure). One week following the pre-survey, participants received the first of the daily surveys via email. Each day for 20 work days, participants received an email at the end of each work day between 2pm and 4pm that contained a link to each day’s survey. Concerns of contrast effects 41 made it troublesome for participants to report on their experiences of both nominated coworkers at the same time. To address this concern, prior to data collection, participants were randomly assigned to either complete the daily measures about one coworker (e.g., Coworker A; approach relationship) for the first half of data collection (10 work days) and then complete the daily measures about their experiences with the second nominated coworker (e.g., Coworker B; avoidance relationship) for the last half of data collection (10 days), or vice versa. In order to increase the likelihood that participants would attend to the appropriate coworker they nominated in the pre-survey, the daily emails also indicated the alias of the nominated coworker for which the respective survey would be completed about. The daily surveys were designed to be short (approximately 5 minutes in length) and assess the variables expected to vary within the relationships over time. These variables included: receipt of help and harm, positive and negative affective states, promotion and prevention regulatory foci, engagement in help and harm, and relationship satisfaction. In order to decrease the possibility that the ordering of survey measures would affect participants’ responding, the order of measures in each daily survey were randomly counterbalanced, alternating each day. The survey was divided into two halves and each half began with relatively easy and innocuous questions (i.e., affective states, promotion and prevention regulatory foci) to earn the respondent’s trust before moving on to more sensitive questions (e.g., receipt/engagement in help/harm). For example, the first half of the daily surveys assessed positive and negative affective states, receipt of help and harm, and the second half assessed promotion and prevention regulatory foci, engagement in help and harm and relationship satisfaction. 42 Finally, at the end of the last survey (Day 20) participants also completed a manipulation check (described in detail below). For the manipulation check, participants rated their approach and avoidance social motivation for both of the coworkers they nominated. Measures Full scale items for all measures are listed in Appendix B. Individual Level Variables (Level 3) General Approach and Avoidance Disposition. Individual differences in general motivational tendencies were measured using the BAS and BIS scales (Carver & White, 1994). BAS was measured using 13 items (e.g., “When I see an opportunity for something I like, I get excited right away”) and the BIS was measured with seven items (e.g., “If I think something unpleasant is going to happen I usually get pretty ‘worked up’”). Participants responded to the items on a scale ranging from 1 (strongly disagree) to 5 (strongly agree) scale, and their responses were averaged to form the BAS and BIS indices. Cronbach’s alpha reliability for the BAS was 0.80 and 0.79 for the BIS. Individual differences in BAS and BIS are independent and individuals can be both high and low on both the BAS and BIS (Carver & White, 1994). Emotional Intelligence. Individual differences in emotional intelligence were assessed with the Wong Law Emotional Intelligence Scale (WLEIS; Wong & Law, 2002). The 16-item scale consists of four dimensions measured with four items each: The Self-Emotion Appraisal (SEA) dimension assesses an individual’s self-perceived ability to understand their emotions (e.g., “I have a good understanding of my own emotions”); the Others’ Emotion Appraisal (OEA) dimension assesses a person’s tendency to be able to perceive other peoples’ emotions (e.g., “I am sensitive to the emotions and feelings of others”); the Use of Emotion (UOE) dimension concerns the self-perceived tendency to motivate oneself to enhance performance 43 (e.g., “I am a self-motivating person”); and the Regulation of Emotion (ROE) concerns the selfperceived ability to regulate their own emotions (e.g., “I have good control of my own emotions”). Participants indicated their agreement with each item on a scale ranging from strongly disagree (1) to strongly agree (5). Means, standard deviations, and Cronbach’s alpha reliabilities for the emotional intelligence sub-dimensions were: SEA (M=4.30, SD=0.53, α=0.88), OEA (M=4.03, SD=0.51, α=0.78), UOE (M=4.24, SD=0.60, α=0.78), and ROE (M=3.90, SD=0.82, α=0.82). The WLEIS elicits a global emotional intelligence scale with higher scores indicating greater emotional intelligence (Law, Wong, & Song, 2006; Wong & Law, 2002; Wong, Wong, & Law, 2005). Using Mplus version 6.11, a confirmatory factor analysis was conducted to evaluate the factorial structure of the WLEIS. The model reflected the four correlated dimensions and a second-order factor behind all of the dimensions. This model fit the data well: χ² (df=100) =118.74, p=.09, RMSEA=.06 (90% confidence interval: 0.00, 0.09), CFI=0.95, SRMR=0.07. This suggests that the 16 items represent an estimate of global emotional intelligence. In line with previous research (Law, Wong, & Song, 2006; Wong & Law, 2002; Wong, Wong, & Law, 2005), the 16 items were averaged to represent an overall emotional intelligence score for each individual (M=4.12, SD=0.43, α=0.86). Individual Demographics. Participants provided information about their age, gender, race/ethnicity, and organizational tenure. Relationship Level Variables (Level 2) Approach and Avoidance Relationship Motives. In the pre-survey, participants were instructed to select two coworkers with whom they interact on a daily basis and who is not in a supervisor position relative to them (see Appendix B for instructions). Participants were provided with a description of approach and avoidance relationships and then asked to select one 44 coworker with whom they have a relationship characterized by an approach social motive (Coworker A) and a second with whom they have a relationship characterized by an avoidance social motive (Coworker B). Relationship descriptions were based on Elliot and colleagues’ (2006) Approach and Avoidance Social Goals Scale which was designed to measure relationships about friendships and close relationships. The wording was adapted to apply to coworkers for this study. Participants were asked to record an alias for each nominated coworker that they could use to remember the person throughout the duration of the study. For the approach relationship, the description read: “This coworker is someone with whom who you try to deepen, grow and develop your relationship with by sharing fun and meaningful experiences.” In terms of the avoidance relationship, the description read: “This coworker is someone whom you try to make sure that nothing bad happens in your relationship by avoiding conflicts and situations that could cause harm to your relationship. You try to avoid getting embarrassed, betrayed, or hurt in your relationship with this coworker.” The relationship motive variable was dummy-coded, the approach relationship was coded as 1 and the avoidance relationship was coded as 0 for each participant. Liking. Participants were asked to indicate the extent to which they like each of their nominated coworkers on a one-item scale ranging from I don’t like Coworker A (or B) at all (1) to I like Coworker A (or B) very much (5). Trust. Trust in coworkers was assessed using McAllister’s (1995) trust scale that assesses two dimensions of trust: affect-based and cognition-based. Affect-based trust reflects expectations of emotional investment, reciprocal care and concern, and cognition-based trust reflects expectations of professionalism, reliability and dependability. Participants indicated their trust for both the coworkers they nominated, Coworker A and Coworker B. McAllister’s (1995) 45 trust scales, originally designed to reference a generic individual, were adapted to reference Coworker A and Coworker B (see Appendix B).The 11-item scale consisted of five items assessing affect-based trust (e.g., “I freely share my ideas and feelings with Coworker A”) and six items assessing cognition-based trust (e.g., “I can rely on Coworker A (or Coworker B) to not make my job more difficult with careless work.”). Participants were asked to indicate their agreement with each item on a scale ranging from 1 (Strongly disagree) to 5 (Strongly agree). Using Mplus version 6.11, a confirmatory factor analysis was conducted to evaluate the factorial structure of the trust scales. A model reflecting the two affect- and cognition-based dimensions of trust was compared to a model reflecting a single trust dimension. The two-factor model fit the data reasonably well (χ² [df=104] =220.69, p<.001, RMSEA=.08 [90% confidence interval: 0.04, 0.11], CFI=0.97, SRMR=0.10) although it fit the data significantly better than the single factor structure (χ² [df=108] =347.74, p<.001, RMSEA=0.20 [90% confidence interval: 0.18, 0.23], CFI=0.74, SRMR=0.14), Δ χ² (4)=127.06, p<.05. Cronbach’s alpha reliability for affectbased trust for Coworker A relationship was 0.86 and 0.89 for the Coworker B relationship. Cronbach’s alpha reliability for cognition-based trust for Coworker A relationship was 0.90 and 0.93 for Coworker B relationship. Characteristics of Coworker Relationships. In light of concerns that difference in organizational hierarchical status, task interdependency (Harrison & Chiarubu, 2008) and the length of relationship between participants and their nominated coworkers would affect reported experiences, coworker status, whether or not the coworkers are required to work with each other to complete work tasks, and length of relationship were also assessed to be included as controls in subsequent analyses. Specifically, participants were asked if both Coworker A and Coworker 46 B were a higher status than them in the organization (Yes=1, No=0), if they were required to work with Coworker A and Coworker B to complete work tasks (Yes=1, No=0), and the tenure of their relationships (see Appendix B). Of those individuals nominated as Coworker A, participants were required to work with 60% of them and 55% of those nominated to work for Coworker B. Although participants were asked to nominate coworkers who were of the same status as them, 26% of Coworker A (approach motive) coworkers were a higher status and 22% of Coworker B (avoid motive) coworkers were a higher status. Within-Relationship Level Variables/ Daily Experiences (Level 1) Receipt of and Engagement in Help. For each day of data collection, participants indicated if Coworker A or Coworker B helped them (help receipt) and if they helped Coworker A or Coworker B during the work day (help engaged), for 10 days each. Similar to an approach taken in an experience-sampling study by Dalal and colleagues (2009), the receipt of and engagement in help was assessed with six item interpersonal-OCB items. The help scale was adapted from items from existing sources including Smith, Organ, and Near’s (1983) measure of OCB and Borman and Motowildo (1997) measure of prosocial behaviors. For example, one item included: “tried to help me” (“tried to help this coworker”). The receipt and engagement of help was not assessed using extant scales because such scales include some items that are not likely to occur frequently enough throughout the day to be assessed within person on a daily basis, and because large number of items will create excessive demands on participants as they complete measures on a daily basis (Dalal et al., 2009). Items that were expected to only occur rarely were omitted and the assessment of help in the current study contained relatively high frequency behaviors (see Appendix B). Each day participants indicated the extent they received the helpful behaviors from the coworker and engaged in the helpful behaviors towards the coworker on a 47 scale ranging from 1 (never) to 5 (very often). Participants’ daily help receipt and engagement scores were computed by averaging across all items for each day. Average Cronbach’s alpha scores for the help receipt scale was 0.96 (ranged from 0.92 to 0.98) and 0.94 (ranged from 0.88 to 0.98) for the help engagement scale. Receipt of and Engagement in Harm. Similar to the helping variables, for each day of data collection, participants indicated if Coworker A or Coworker B harmed them (harm receipt) and if they harmed Coworker A or Coworker B during the work day (harm engaged), for 10 days each. Similar to an approach taken in the experience-sampling study by Dalal and colleagues (2009), receipt of and engagement in harm was assessed with behaviors that are relatively high frequency and low severity. Passive and active forms of harm were also assessed, each with four items. These items were drawn from Dalal and colleagues’ (2009) assessment of CWB and Spector, Fox, Penney, Bruursema, Goh, and Kessler’s (2006) 32-item CWB checklist. Items were selected based on their resemblance to Neuman and Baron’s (1996) distinction between active and passive aggression. For example, an active harming behavior included “insulted or made fun of me” (“insulted or made fun of this coworker”) and a passive item included “ignored me” (“ignored this coworker”). Each day participants indicated the extent they received the harmful behaviors from the coworker and engaged in the harmful behaviors towards the coworker on a scale ranging from 1 (never) to 5 (very often). However, the low frequency of active behaviors made it difficult to analyze the data for the active and passive harm behaviors separately as for some days there was no variance for the active items. As such, in order to facilitate data analyses, the active and passive helping behaviors were combined to form an eight-item inventory of harming. Participants’ daily harm receipt and engagement scores were computed by averaging across all items for each day. Average Cronbach’s alpha reliability 48 scores for harm receipt was 0.86 (ranged from 0.72 to 0.96) and 0.83 for harm engagement (ranged from 0.73 to 0.96). Relationship Satisfaction. Relationship satisfaction was assessed with items adapted from previous studies assessing daily satisfaction in romantic relationships (Campbell, Simpson, Boldry, & Kashy, 2005; Impett et al., 2010; Neff & Karney, 2009; see Appendix B). Participants indicated their agreement with three items (e.g., “I am satisfied with my relationship with this coworker today”). Responses ranged from 1 (strongly disagree) to 5 (strongly agree). Participants’ daily relationship satisfaction scores were computed by averaging across all items for each day. Average Cronbach’s alpha reliability scores for relationship satisfaction was 0.94 (ranged from 0.88 to 0.97). Positive and Negative Affective States. Affective states were assessed using an adaptation of the Positive and Negative Affect Schedule (PANAS; Watson et al., 1988; Tellegen, Watson, & Clark, 1999). Affective states assessed reflected feelings associated with the twofactor conceptualization of approach and avoidance motivation (Carver & White, 1994; Gray, 1990). Research has suggested that affective states associated with approach and avoidance motivation are bipolar, in that approach motivation is associated with certain positive (i.e., cheerfulness) and negative (i.e., dejection) affective states and avoidance motivation is associated with certain positive (i.e., quiescence) and negative (i.e., agitation) affective states (Carver & Sheier, 1998). Similar approaches have been adopted in research assessing approach motivation system influences on affective reactions to positive and negative events (Carver, 2004), and with research assessing affective reactions to successes and failures in regulatory foci (Higgins et al., 1997). Participants were asked to indicate how intensely they felt a series of adjectives during the day: 1 (very slightly) to 5 (extremely). Positive affective adjectives 49 included: “discouraged,” “downhearted,” and “sad” (representing dejection), and “excited,” “happy,” and “joyful,” (representing cheerfulness). Negative affective adjectives included: “distressed,” “nervous,” and “jittery” (representing agitation), and “at ease,” “calm,” and “relaxed” (representing quiescence). In light of research demonstrating that hostility affective states relate to engagement in deviance behaviors (e.g., Judge et al., 2006), hostility-related adjectives were also included as a part of the negative affective states scale in order to more fully capture the broader content domain of negative affective states that may be related to harm receipt and harm engagement. In terms of positive affective states, due to an almost perfect correlation between “calm” and “relaxed” at several time points, “relaxed” was removed from the scale so that model fit could be assessed. With this resulting five item scale, the single factor conceptualization of positive affective states (χ² [252]=318.61, p<.001, RMSEA=0.10, CFI=0.98, SRMR=0.08) fit the data significantly better than the two factor conceptualization (χ² [194]=1140.79, p<.001, RMSEA=0.32, CFI=0.82, SRMR=0.14). Δ χ² (58)=822.19, p<.05. The single factor measure of positive affective states was used in all subsequent analyses. Participants were assigned a single score based on the mean aggregate of the positive affective states items for each day. Average Cronbach’s alpha reliability for positive affective states was 0.96 (ranged from 0.92 to 0.98). In terms of negative affective states, several items - mainly the agitation- (e.g., “jittery”) and hostility-related (e.g., “hostile”) items - were infrequently endorsed by participants across time which resulted in a lack of variance at several time points making it difficult to assess model fit using confirmatory factor analysis. Removing scale items to increase the acceptability of the model fit would have greatly sacrificed the content of the scale. Therefore, in order to 50 maintain the integrity of the scale, and to move ahead with the data analyses, all of the negative affective states items were mean aggregated into a single score for each participant per day. Average Cronbach’s alpha reliability scores for negative affective states was 0.89 (ranged from 0.80 to 0.98). Regulatory Focus. Although Molden and Finkel (2010) assessed daily promotion and prevention foci, they only used one item. In the current study, this approach was expanded upon and daily promotion and prevention foci were assessed with three items each drawing on Lockwood, Jordan, and Kunda’s (2002) measure of individual differences in promotion and prevention foci. The items were adapted to reflect daily experiences about a relationship with a coworker (see Appendix B). For example, a promotion focus item included: “I thought about my hopes and aspirations for my relationship with this coworker.” An example prevention focus item included: “I was oriented towards preventing losses when interacting with this coworker.” Participants indicated their agreement with each item, ranging from 1 (strongly disagree) to 5 (strongly agree). Average Cronbach’s alpha reliability scores for promotion was 0.88 (ranged from 0.74 to 0.97) and 0.87 for prevention (ranged from 0.77 to 0.93). Manipulation Check Following the completion of the final survey on Day 20, participants completed the manipulation check. The manipulation check was intended to verify that the coworkers participants nominated as having approach- and avoidance-oriented social motivation did in fact reflect this distinction. In order to so, participants evaluated their approach and avoidance social motivation with each nominated coworker. It was expected that participants would report having a higher approach-oriented relationship motives with the coworker whom they nominated as their approach relationship (i.e., Coworker A) compared to the coworker whom they nominated 51 as their avoidant relationship (i.e., Coworker B) and that they would report having higher avoidant-oriented relationship motives with the coworker whom they nominated as their avoidant relationship compared to the coworker whom they nominated as their approach relationship. Participants completed Elliot and colleagues (2006) measure of approach-avoidance social motivation for both Coworker A and Coworker B by indicating their agreement to items on an eight-item scale with item responses ranging from 1 (strongly disagree) to 5 (strongly agree): four items for approach (e.g., “I try to deepen my relationship with Coworker A [or Coworker B]”) and four items for avoidance (e.g., “I try to avoid disagreements and conflict with Coworker A [or Coworker B]”). Of the 55 participants, 38 completed the manipulation check (69% response rate). There was no significant difference between the 38 participants who completed the manipulation check and the 17 who did not in terms of: proportion of gender, race, and organizational membership, and mean differences in age, approach and avoidance disposition, emotional intelligence, and average levels of receipt of help and harm, positive and negative affective states, promotion and prevention regulatory foci, and engagement in help and harm. However, individuals who completed the manipulation check had significantly less organizational tenure (M=75.21, SD=83.78) than individuals who did not complete the manipulation check (M=147.88, SD=183.12), t(53)=2.04, p<.05. Cronbach’s alpha for the approach social motivation scale was 0.95 for Coworker A and 0.93 for Coworker B. Cronbach’s alpha for the avoidance social motivation scale was 0.87 for Coworker A and 0.78 for Coworker B. There was no significant difference in approach social motivation between Coworker A (M=3.81, SD=1.03) and Coworker B (M=3.57, SD=1.08) (t [37]=1.00, n.s.) although the effect was in the expected direction such that the mean for Coworker A was slightly higher than the mean for Coworker B. Further, there was no significant 52 difference in avoidance social motivation between Coworker A (M=3.68, SD=0.88) and Coworker B (M=3.85, SD=0.75) (t [37]=1.22, n.s.) although the effect was also in the expected direction such that the mean for Coworker B was slightly higher than the mean for Coworker A. Therefore, the manipulation check was generally not supportive of the manipulation. There are a number of possible explanations as to why the current manipulation check was not supportive of the manipulation (which will be elaborated on in the discussion) but alternative evidence from the data provided validity evidence for the manipulation. For example, this evidence can be found from inspection of the associations between the relationship motive (at Level 2) and average levels of positive and negative affective states and promotion and prevention regulatory foci (at Level 1). Using relationship motive (approach=1, avoid=0) as a predictor of Level 1 intercepts, relationship motive was significantly and positively related positive affective states (b10=0.72, p<.001), significantly and negatively related to negative affective states (b10=-0.30, p<.05), significantly and positively related to promotion focus (b10=0.22, p<.05), and significantly and negatively related to prevention focus (b10=-0.51, p<.001). These results are in line with previous research demonstrating the positive emotions are more common in approach-oriented relationships, negative emotions are more common in avoidance-oriented relationships (Gable et al., 2000), and meta-analytic evidence demonstrating that high levels of approach temperament increase the likelihood that individuals adopt promotion strategies and high levels of avoidance temperament increase the likelihood that individuals adopt prevention strategies (Lanaj et al., 2012). These results suggest that the more proximal daily experiences of affective states and regulatory foci (relative to the distal relationship motive) that are representative of broader approach and avoidance orientations tended to map onto the relationship manipulation. 53 Analytical Approach The data for the current study is hierarchically nested such that days were nested within relationships and relationships were nested within individuals. In order to analyze the hierarchically nested data hierarchical linear modeling version 6.01 (HLM; Raudenbush & Byrk, 2002) was used. The data consisted of three levels. Level 3 was the individual level consisting of individual differences in general approach and avoidance disposition, emotional intelligence, gender, age, and race/ethnicity. Level 2 was the relationship level and consisted of the relationship motive manipulation, trust, liking, relationship tenure, and hierarchical status difference. Level 1 was the within-relationship (daily experiences) level consisting of receipt of and engagement in help, receipt of and engagement in harm, relationship satisfaction, positive and negative affective states, and promotion and prevention regulatory foci. Thus, the Level 1 data could vary within the relationships at Level 2, which were nested within individuals at Level 3. To test the hypothesized within-relationship effects among help and harm receipt, affective states, regulatory foci, help and harm engagement, and relationship satisfaction (Hypotheses 1, 2, 5, 6, 9, 10, 11, 12, 13, 14, 15, 16 and research questions 1, 2, and 3), each dependent variable (help and harm engagement, relationship satisfaction) was regressed onto each of the mediator (affective states, regulatory foci) and independent (help and harm receipt) predictor variables with all of the Level 1 predictors centered at the individuals’ mean. To test the hypothesized cross-level moderation and moderated-mediation effects of the Level 2 relationship motive on the within-relationship effects (Hypotheses 3, 4, 7, 8, 17, 18, 19, 20 and research questions 1, 2, 3), relationship motive was added to the above equations as a Level 2 predictor of the intercept and slope of each Level 1 effect of interest (e.g., the within-relationship effect between help receipt and positive affective states). Continuous Level 2 variables (i.e., liking, trust) were 54 centered at the individual. The Level 3 individual difference variables (gender, race/ethnicity, age, approach disposition, avoidance disposition), which were centered at the grand mean, were included as a predictor of the Level 2 intercept as controls in all analyses. To test the moderated mediation hypotheses, the indirect effects across approach and avoidance relationship motives were compared following techniques proposed by Edwards and Lambert (2007) and Preacher, Rucker, and Hayes (2007). According to this approach, mediation is framed in terms of a path model and the relationships between variables are modeled using regressions. The process involves several steps to assess moderated mediation: (1) the dependent variables (help and herm engagement, relationship satisfaction) are regressed onto the independent variables (help and harm receipt) for an estimate of the direct effect, (2) the mediator variables (positive and negative affective states, promotion and prevention regulatory foci) are regressed onto the independent variables for an estimate of path a, (3) the dependent variables are regressed onto the mediator variables, controlling for the direct effect, for an estimate of path b, and (4) the indirect effects (product of path a and path b estimates) are compared across different levels of the moderator (approach [1] and avoidance [0] relationship motive). Similar approaches have recently been utilized by organizational researchers testing moderated-mediation with multi-level data using HLM (e.g., Bacharach & Bamberger, 2007; Chowdhury & Endres, 2010; Duffy, Ganster, Shaw, Johnson, & Pagon, 2006; Liu, Chen, & Yao, 2011; Scott & Barnes, 2011). In all analyses, participant gender, race, age, approach and avoidance dispositions, relationship-level status difference, liking, and relationship tenure were included as controls. Data analyses were based on 576 reported interactions. This averages to 5.24 interactions per relationship and 10.47 interactions per person. 55 RESULTS Descriptive Statistics and Correlations Means, standard deviations, and correlations are shown in Table 1a and 1b. Level 1 within-relationship correlations are above the diagonal of Table 1a, Level 2 betweenrelationships are below the diagonal of Table 1a, and Level 3 individual correlations are in Table 1b. For the relationship-level correlations, Level 1 variables are mean aggregated over the 10 day period into single scores for each relationship, and for the individual level correlations Level 1 variables are mean aggregated over the 20 day period into single scores for each participant and Level 2 variables were mean aggregated across both relationships into a single score for each participant. As can be seen in Table 1a, approach-avoidance relationship motive (approach=1, avoidance=0) did not significantly correlate with the means of relationship-level help or harm receipt, positive or negative affective states, help or harm engagement, relationship satisfaction and promotion and prevention regulatory foci. Further, preliminary analyses were conducted to examine how the Level 2 relationship variables (liking, status, relationship tenure, interdependency, affect-based and cognition-based trust) related to approach and avoidance-oriented relationship motives. Results indicated that there was no significant difference in approach or avoidance-oriented relationships in the likelihood that individuals nominated someone who was of a higher status than them in their organization (χ² [1]=0.20, n.s.) or that they nominated someone who they were required to work with in order to complete work tasks (χ² [1]=0.33, n.s.). However, liking was significantly higher for the coworker they nominated as the approach relationship (M=4.69, SD=0.77) than the 56 coworker they nominated as the avoidance relationship (M=3.81, SD=1.18) (t[54]=4.90, p<.05). Relationship tenure did not differ significantly across approach (M=45.03, SD=44.71) and avoidance (M=38.90, SD=36.02) relationships, t(54)=1.10, n.s. In terms of trust, both affect-based and cognition-based trust were significantly higher for the coworker nominated as the approach relationship (affect-based: M=4.06, SD=0.86; cognition-based: M=4.29, SD=0.77) than the coworker nominated as the avoidance relationship (affect-based: M=3.18, SD=1.04; cognition-based: M=3.37, SD=1.16), t (54)>4.61, p<.05. As participants were recruited from three different organizations (i.e., government agency, grocery stores, university administration), means of Level 1 variables were compared across organizations. There were no significant differences. Overall, results also did not change with organization included as a control. Further, whether or not participants were required to work with the nominated coworkers also had no effect on the means of the Level 1 variables and overall results were also not changed by including the interdependency variable as a control. Therefore, for purposes of simplicity, organization and interdependency were not included in the analyses for the testing of hypotheses. 57 Table 1a 1 2 .07 3 4 .51** .53** -.22** -.17** -0.2 .87** -.16 .89** -0.27 .72** .75** 5 .53** -.25** .72** .67** 7 -.33** .18** -.39** -.30** -.38** 8 -.38** .16* -.40** -.36** -.43** 9 .56** -.26** .72** .67** .71** .47** 6 -.15* .26** -.28** -.21** -.37** .55** .74** -.45** -.46** -.50** 10 11 12 13 14 15 1. Promote 2. Prevent 3. Helpr 4. Helpe 5. Pos af 0.08 .60** .62** .61* 6. Neg af 7. Harmr 8. Harme 9. Satisfac -.22* 0.25* -.28* 0.22* -.34* 0.24 .61** -.29* -.31* -.39* -.36* .74** -.27* -.33* -.39* .72** -.50** -.39** -.47** .81** .57** .61** -.56** .79** -.50** -.53** 10. Motive -0.05 -0.14 -0.05 0.00 0.03 0.06 0.12 -0.01 .04 11. Trusta 12. Trustc 0.00 0.08 -0.1 -0.09 -0.04 -0.03 -0.01 0.00 0.03 0.05 0.05 0.08 0.08 0.09 0.00 -0.02 .06 .10 .53** .50** .77** 13. Status 14. Liking 0.05 .28* 0.15 -0.08 -0.05 .30* -0.05 .30* -0.02 .33* 0.05 -0.05 0.09 -0.11 0.06 -.20* 0.04 .41** 0.1 .63** 0.06 .52** .15* 15. Tenure 0.00 0.06 0.04 0.03 -0.1 -0.04 0.06 0.05 -.03 .34** 0.00 0.08 0.08 0.06 0.16 0.12 M SD 3.43 0.77 2.65 0.83 3.33 1.09 3.44 1.04 3.3 1.15 1.28 0.47 1.25 0.44 1.22 0.37 3.89 0.80 0.5 0.5 3.62 0.84 3.83 0.83 0.24 4.25 41.95 0.43 1.08 40.53 Descriptive statistics and correlations between study variables at Level 1 and Level 2; N3=55, N2=86, N1=576, Level 1 correlations are in the top diagonal and level 2 correlations are in the bottom diagonal; Level 2 correlations were calculating by mean aggregating Level 1 variables for each Level 2 (relationship); means (M) and standard deviations (SD) are for Level 2; helpr/harmr=help/harm receipt, helpe/harme= help/harm engagement, af=affect, satisfac=satisfaction, trusta=affect trust, trustc=cognitive trust, relationship status coded as different status=1, same status=0, relationship motive coded as approach motive=1, avoidance motive=0, **p<.001, *p<.05 58 Table 1b 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Promote .24 2. Prevent .56* -.08 3. Helpr .59* -.05 .91** 4. Helpe .52* -.16 .70** .71* 5. Pos af -.04 .26 -.20 -.15 -.27* 6. Neg af -.10 .18 -.21 -.16 -.23 .44* 7. Harmr -.18 .14 -.21 -.22 -.30* .43* .83** 8. Harme .61** -.10 .79** .76** .81** -.39* -.28* -.32* 9. Satisfac 10. Motive 11. Trusta .48* .11 .58* .55* .51* -.11 -.24 -.36* .56** 12. Trustc .44* .10 .40* .36* .61** -.41* -.43* -.48* .61** .61* 13. Status 14. Liking .40* .08 .46* .41* .42* -.30* -.31* -.29* .58** .64* .58* 15. Tenure .08 -.01 .22 .17 -.03 -.17 -.04 -.04 .10 -.06 .12 16. Gender .00 .07 .00 -.05 -.07 -.14 -.19 -.16 -.03 -.04 -.09 17. Age -.10 -.21 .06 .01 -.18 -.06 -.01 -.06 -.16 -.09 -.07 18. Race -.27* -.13 -.19 -.25 -.23 .00 -.17 -.12 -.08 -.16 -.19 19. Avoidance .21 -.07 .05 .06 -.03 -.04 .10 .06 .18 .12 .03 20. Approach .20 -.12 -.08 -.04 .12 .14 .07 .06 .11 -.04 .00 21. EI .12 .05 .11 .11 .30 -.05 -.20 -.16 .24 .25a .10 M 3.43 2.66 3.33 3.44 3.35 1.32 1.26 1.21 3.89 3.63 3.83 SD 0.64 0.68 0.91 0.89 0.83 0.33 0.32 0.27 0.53 0.64 0.74 Descriptive statistics and correlations between study variables at Level 3; N3=55, N2=86, N1=576, Level 1 correlations are in the top diagonal and level 2 correlations are in the bottom diagonal; Level 2 correlations were calculating by mean aggregating Level 1 variables for each Level 2 (relationship); means (M) and standard deviations (SD) are for Level 2; helpr/harmr=help/harm receipt, helpe/harme= help/harm engagement, af=affect, satisfac=satisfaction, trusta=affect trust, trustc=cognitive trust, EI=emotional intelligence, relationship status coded as different status=1, same status=0, relationship motive coded as approach motive=1, avoidance motive=0, **p<.001, *p<.05 59 Table 1b (cont’d) Variable 14. Liking 15. Tenure 16. Gender 17. Age 18. Race 19. Avoidance 20. Approach 21. EI 14 15 16 17 18 19 20 21 .14 .00 .33 -.14 .40* .26 -.20 .20 .27* .18 .17 .24 .23 .19 .17 -.03 -.22 -.10 -.12 .09 .15 .11 -.25 .05 -.25 -.01 -.28* .26 M 4.27 40.01 .84 39.00 .87 3.42 3.82 4.12 SD 0.71 34.13 0.37 12.52 0.34 0.67 0.46 0.43 N3=55, correlations are between Level 3 variables; means (M) and standard deviations (SD) are for Level 3; EI=emotional intelligence, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded as different status=1, same status=0, relationship motive coded as approach motive=1, avoidance motive=0, **p<.001, *p<.05 Partitioning of Variance Within and Between Levels Before testing the hypotheses, the null models (regressions with no Level 1, Level 2, or Level 3 predictors) in HLM for each Level 1 endogenous variable were inspected to determine the extent of variance in the Level 1 variables accounted for at each level. Null models separate the variance in a given level 1 variable within relationships and between relationships (Level 2) and individuals (Level 3). The intercept represents the mean level of the variable for all the days of data collection. If there is a lack of variance at the within-relationship level, and most of the variance is explained at the between-relationship or individual level then a 3-level model is not appropriate for the data as there is only two levels of data (relationship and individual). If there is a lack of variance at the relationship level, and most of the variance is explained within relationships or between individuals, then a 3-level model is not appropriate and there is only two levels (within-relationship and individual). Likewise, if there is a lack of variance at the individual level, there is only two levels of data (within relationship and relationship) a 3-level model is not appropriate. Finally, if all the variance is explained at either the relationship or 60 individual level (and not at any other level), HLM is inappropriate as there is only one level of variance. The results of the null models are presented in Table 2. The results indicated that for each endogenous Level 1 variable, there were differences in the proportion of variance at each level. Specifically, 19.08% of the variance in positive affective states was at the within-relationship level, 59.87% at the relationship-level, and 21.05% at the individual level. For negative affective states, 42.86% was at the within-relationship level, 57.14% at the relationship level, and 0% at the individual level. Similarly, harm engagement also had 0% variance at the individual level but 25% variance at the within-relationship level and 75% variance at the relationship level. For help engagement, 23.58% of the variance was explained at the within-relationship level, 30.89% at the relationship level, and 45.53% at the individual level. For promotion and prevention regulatory foci, 33.8% and 37.86% of variance was at the within-relationship level, respectively, 45.07% and 36.89% at the relationship-level, respectively, and 21.13% and 25.24% at the individual level, respectively. For relationship satisfaction, 36% of variance was explained at the within-relationship level and 64% was explained at the between relationship level; almost no variance in relationship satisfaction was explained between persons. Overall, the proportion of explained variance tended to be greatest at the relationship level (except for prevention regulatory foci) and smallest at the individual level (except for positive affective states and help engagement) with very little variance in negative affective states and harm engagement being explained at the individual level. The above results suggested that 3-level HLM was appropriate and that there was within-relationship and between-relationship variance to be explained. 61 Table 2 Level-1 Level-2 Level-3 variance variance variance component component component Variable Intercept (b000) (e2) ICC1 (r2) ICC2 (u2) Promotion 3.36 0.24 33.80% 0.32** 45.07% 0.15* Prevention 2.67 0.39 37.86% 0.38** 36.89% 0.26** Positive affective state 3.25 0.29 19.08% 0.91** 59.87% 0.32* Negative affective state 1.33 0.15 42.86% 0.20** 57.14% 0.00 Help engagement 3.35 0.29 23.58% 0.38** 30.89% 0.56** Harm engagement 1.23 0.04 25.00% 0.12** 75.00% 0.00 Relationship satisfaction 3.81 0.35 36.01% 0.62** 63.92% 0.00 Intra-class correlation statistics for 3-level HLM model; N3=55, N2=86, N1=576, g000 is the pooled intercept Representing average level of variable across individuals, ICC1= e2/(e2+r2+u2), ICC2=r2/(e2+r2+u2), ICC3=u02/(e2+r2+u2), *p<.05, **p<.001 62 ICC3 21.13% 25.24% 21.05% 0.00% 45.53% 0.00% 0.00% Tests of Hypotheses Within-Relationship Hypotheses Main Effects. Results for the within-individual hypotheses can be found in Table 3, Table 4, Table 5, and Table 6. As is shown in Table 3, Table 4, and Table 5 the dependent variables (help and harm engagement, relationship satisfaction) were regressed onto the independent variables (help and harm receipt) and mediator variables (positive and negative affective states, promotion and prevention regulatory foci). Table 3 displays the results for tests of hypotheses 1, 3, 7, 9, 11, 17 and research question 2 and 3 regressing the dependent variables help engagement onto the independent variables (receipt of help and harm) and mediators (positive and negative affective states, promotion and prevention regulatory foci). Table 4 displays the results for tests of hypotheses 2, 4, 10, 12, 18 and research question 2 and 3, regressing the dependent variables harm engagement onto the independent variables and mediators. Table 5 displays the results for tests of hypotheses 13, 14, 15, 16, 19, and 20 regressing the dependent variables relationship satisfaction onto the independent variables and mediators. Model 1 on the left side of Table 3, Table 4, and Table 5 displays the results relating the independent variables (help and harm receipt) directly to the dependent variable (hypotheses 1, 2, 3 and 4), Model 2 displays the results with positive and negative affective states as mediators added to the model (hypotheses 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20), and Model 3 of Table 3 and Table 4 displays the results with promotion and prevention regulatory foci as mediators added to the model (research questions 2 and 3). Model 4 of Table 3 and Table 4 displays the results with both affective states and regulatory foci mediators included in the model. 63 Hypothesis 1 predicted that help receipt would be positively associated with help engagement. Hypothesis 2 predicted that harm receipt would be positively associated with harm engagement. As shown in Model 1 of Table 3 and 4, results suggest that help receipt was significantly and positively associated with help engagement (b100=0.52, p<.001) and was significantly and negatively related to harm engagement (b100=-0.05, p<.05). Harm receipt was significantly and positively associated with harm engagement (b200=0.25, p<.05) and, surprisingly, significantly and positively related to help engagement (b200=0.16, p<.05). Hypothesis 1 and 2 were therefore supported. Table 6 displays the results in which the mediator variables were regressed onto the independent variables (hypotheses 5, 6, 7, 8 and research question 1). Hypothesis 5 predicted that help receipt would be positively associated with positive affective states. Hypothesis 6 predicted that harm receipt would be positively associated with negative affective states. Evidence for research question 1 can also be found in Table 6, relating help and harm receipt to promotion and prevention regulatory foci. Results suggest that help receipt was positively and significantly related to positive affective states (b100=0.37, p<.05) but not to negative affective states (b100=0.05, n.s.). Receipt of harm was significantly and positively associated with negative affective states (b200=0.34, p<.05) but was not significantly related to positive affective states (b200=0.18, n.s.). Hypothesis 5 and 6 were therefore fully supported. Regarding research question 1, help receipt was significantly and positively related to promotion regulatory foci (b100=0.22, p<.05) but was not significantly related to prevention regulatory foci (b100=0.05, n.s.) and harm receipt was significantly and negatively related to promotion regulatory foci (b200=-0.16, p<.05) and was significantly and positively related to prevention regulatory foci (b200=0.23, p<.05). 64 Table 3 Variable Intercept (b000) Gender (b001) Race (b002) Age (b003) Avoidance disposition (b004) Approach disposition (b005) Status difference (b020) Relationship tenure (b030) Liking (b040) Relationship motive (b010) Help received (b100) Harm received (b200) Positive affective state (b300) Negative affective state (b400) Promotion focus (b500) Prevention focus (b600) Cross-level moderation Help receipt* Relationship (b110) Harm receipt* Relationship (b210) Positive affect *Relationship (b310) Negative affect*Relationship (b410) Promotion focus* Relationship (b510) Prevention focus* Relationship (b610) Model 1 b s.e. 4.05** 0.37 -0.04 0.32 -0.73* 0.31 0.00 0.01 0.08 0.22 -0.03 0.29 0.13 0.17 0.00 0.00 0.20** 0.04 0.08 0.06 0.52** 0.06 0.16* 0.07 Model 2 b s.e. 4.03** 0.37 -0.04 0.31 -0.67* 0.31 0.00 0.01 0.07 0.22 -0.04 0.29 0.03 0.10 0.00 0.00 0.17* 0.04 0.02 0.04 0.48** 0.07 0.16* 0.08 0.14* 0.05 0.04 0.06 Model 3 b s.e. 4.09** 0.37 -0.08 0.31 -0.77* 0.32 0.00 0.01 0.10 0.22 -0.04 0.29 0.13 0.16 0.00 0.00 0.17* 0.04 0.11 0.07 0.49** 0.07 0.18* 0.08 Model 4 b s.e. 4.07** 0.37 -0.07 0.30 -0.71* 0.32 0.00 0.01 0.09 0.22 -0.06 0.29 0.02 0.09 0.00 0.00 0.14* 0.03 0.04 0.05 0.46** 0.07 0.20* 0.09 0.14* 0.05 0.04 0.06 0.09 0.06 0.06 0.04 0.13* 0.05 0.08 -0.08 0.10 0.17 0.07 0.02 0.09 0.15 0.10 0.14 0.06 0.17 0.06 0.04 0.05 -0.04 0.10 0.16 0.07 0.03 0.08 0.15 0.10 0.15 0.06 0.17 0.07 -0.08 0.07 0.06 0.07 -0.13 0.09 0.06 HLM results for help engagement as outcome and relationship motive as moderator; N3=55, N2=86, N1=576, b=unstandardized coefficient, s.e.=standard error, b000 is the pooled intercept representing average level of variable across individuals, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded as different status=1, same status=0, relationship motive coded as approach motive=1, avoidance motive=0, **p<.001, *p<.05 65 Table 4 Variable Intercept (b000) Gender (b001) Race (b002) Age (b003) Avoidance disposition (b004) Approach disposition (b005) Status difference (b020) Relationship tenure (b030) Liking (b040) Relationship motive (b010) Help received (b100) Harm received (b200) Positive affective state (b300) Negative affective state (b400) Promotion focus (b500) Prevention focus (b600) Cross-level moderation Help received*Relationship (b110) Harm received*Relationship (b210) Positive affect *Relationship (b310) Negative affect*Relationship (b410) Promotion focus*Relationship (b510) Prevention focus*Relationship (b610) Model 1 b s.e. 1.36* 0.23 -0.08 0.12 -0.06 0.16 0.00 0.00 0.07 0.06 0.04 0.08 -0.10 0.09 0.00 0.00 -0.04 0.03 -0.08 0.06 0.02 0.25 0.05 0.05* Model 2 b s.e. 1.44** 0.21 -0.19 0.11 -0.08 0.13 0.00 0.00 0.07 0.05 0.10 0.07 -0.04 0.07 -0.03 0.05 -0.05 0.03 0.00 0.00 -0.03 0.02 0.20* 0.05 -0.02 0.02 0.15* 0.04 Model 3 b s.e. 1.35** 0.14 -0.07 0.08 -0.11 0.10 0.00 0.00 0.04 0.04 0.04 0.06 -0.03 0.05 0.00* 0.00 -0.03 0.03 -0.07 0.05 -0.03 0.02 0.19* 0.06 -0.08* 0.06 0.05 0.01 0.03 0.09 0.03 0.10 0.03 0.08 0.02 0.02 0.03 0.10 0.08 -0.09* 0.05 0.03 0.02 -0.04 0.03 0.02 0.04 0.02 Model 4 b s.e. 1.41** 0.13 -0.18* 0.09 -0.07 0.09 0.00 0.00 0.07 0.04 0.07 0.06 -0.03 0.06 0.00* 0.00 -0.05 0.03 -0.04 0.04 -0.02 0.02 0.16* 0.06 -0.02 0.03 0.15 0.04 -0.05 0.03 0.02 0.02 0.04 0.01 0.01 -0.05 0.05 -0.06* 0.03 0.10 0.03 0.07 0.04 0.02 HLM results for harm engagement as outcome and relationship motive as moderator; N3=55, N2=86, N1=576, b=unstandardized coefficient, s.e.=standard error, b000 is the pooled intercept representing average level of variable across individuals, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded Table 5. HLM results for relationship relationship motive coded asrelationship motive as moderator as different status=1, same status=0, satisfaction as outcome and approach motive=1, avoidance motive=0, **p<.001, *p<.05 66 Table 5 Variable Intercept (b000) Gender (b001) Race (b002) Age (b003) Avoidance disposition (b004) Approach disposition (b005) Status difference (b020) Relationship tenure (b030) Liking (b040) Relationship motive (b010) Help received (b100) Harm received (b200) Positive affective state (b300) Negative affective state (b400) Cross-level moderation Help received*Relationship (b110) Harm received*Relationship (b210) Positive affect *Relationship (b310) Negative affect*Relationship (b410) Model 1 b s.e. 4.07** 0.19 -0.03 0.18 -0.28* 0.20 -0.01 0.00 0.08 0.11 -0.02 0.19 0.04 0.18 0.00 0.00 0.04 0.18 0.32* 0.11 0.42* 0.06 -0.38* 0.12 -0.05 -0.24 0.08 0.22 Model 2 b s.e. 4.11** 0.22 -0.04 0.21 -0.28* 0.21 -0.01 0.01 0.11 0.12 0.00 0.19 -0.03 0.15 0.00 0.00 0.11 0.09 0.20* 0.08 0.31* 0.06 -0.22* 0.11 0.27* 0.06 -0.32* 0.08 -0.23* 0.11 -0.02 0.10 -0.16 0.27 0.02 0.11 0.04 0.32 HLM results for relationship satisfaction as outcome and relationship motive as moderator; N3=55, N2=86, N1=576, b=unstandardized coefficient, s.e.=standard error, b000 is the pooled intercept representing average level of variable across individuals, gender coded as male=0, female=1, race coded as white=1, nonwhite=0, relationship status coded as different status=1, same status=0, relationship motive coded as approach motive=1, avoidance motive=0, **p<.001, *p<.05 67 Table 6 Variable Intercept (b000) Gender (b001) Race (b002) Age (b003) Avoidance disposition (b004) Approach disposition (b005) Status difference (b020) Relationship tenure (b030) Liking (b040) Relationship motive (b010) Help received (b100) Harm received (b200) Cross-level moderation Help received *Relationship (b110) Harm received *Relationship (b210) Positive affective b states s.e. Negative affective states b s.e. Promotion focus b s.e. Prevention focus b s.e. 3.82** -0.08 -0.58 -0.01 0.25 0.24 0.3 0.01 1.41** -0.08 -0.01 0 0.12 0.1 0.1 0 3.67** 0.19 -0.54 0 0.19 0.25 0.17 0.01 2.41** 0.32 -0.06 -0.01 0.39 0.27 0.37 0.01 0.09 0.1 -0.05 0 0.34** 0.29* 0.37** 0.19 0.29 0.21 0 0.09 0.14 0.06 -0.04 0.16 0.04 0 0.01 -0.21* -0.05 0.05 0.08 0.08 0 0.06 0.07 0.03 0.11 0.27 0.27* 0 0.12 -0.03 0.22* 0.14 0.19 0.13 0 0.09 0.11 0.05 -0.18 -0.1 0.3 0 -0.35* -0.18 0.05 0.16 0.21 0.18 0 0.1 0.1 0.06 -0.18 0.11 0.34* 0.18 -0.16* 0.07 0.23* 0.09 -0.02 0.11 -0.03 0.05 0.1 0.1 -0.08 0.13 -0.26 0.18 0 0.21 -0.24 0.19 -0.33 0.22 HLM results for mediators as outcomes and the effects of help and harm receipt and approach-avoidance relationship motives; N3=55, N2=86, N1=576, b=unstandardized coefficient, s.e.=standard error, b000 is the pooled intercept representing average level of variable across individuals, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded as different status=1, same status=0, relationship motive coded as approach motive=1, avoidance motive=0, **p<.001, *p<.05 68 Hypothesis 9 predicted that positive affective states would be positively related to help engagement. Hypothesis 10 predicted that negative affective states would be positively related to harm engagement. Hypothesis 13 predicted that positive affective states would be positively related to relationship satisfaction and hypothesis 14 predicted that negative affective states would be negative related to relationship satisfaction. Research questions 2 questioned how promotion and prevention regulatory foci relate to engagement in help and harm. As shown in Model 2 of Table 3, positive affective states was significantly and positively related to help engagement (b300=0.14, p<.04) but not to harm engagement (see Model 2 of Table 4; b300=0.02, n.s.). As shown in Model 2 of Table 4, negative affective states was significantly and positively related to harm engagement (b400=0.15, p<.05) but not to help engagement (see Model 2 of Table 3; b400=0.04, n.s.). Hypothesis 9 and 10 and were therefore supported. In terms of hypothesis 13, as shown in Model 2 of Table 5, positive affective stats was signiticantly and positively related to relationship satisfaction (b300=0.27, p<.05) and negative affective states was significantly and negatively related to relationship satisfaction (b400=-0.32, p<.05). Hypothesis 13 and 14 were therefore supported. Regarding research question 2, as shown in Model 3 of Table 3, promotion regulatory focus was significantly and positively related to help engagement (b500=0.13, p<.05) and was, as shown in Model 3 of Table 4, significantly and negatively related to harm engagement (b500=-0.08, p<.05). Also, as shown in Model 3 of Table 3, prevention regulatory focus was not significantly related to help engagement (b600=0.05, n.s.) and was, as shown in Model 3 of Table 4, significantly and positively related to harm engagement (b600=0.06, p<.05). Mediation. Hypothesis 11 predicted that positive affective states would mediate the association between help receipt and help engagement and hypothesis 12 predicted that negative 69 affective states would mediate the association between harm receipt and harm engagement. Hypothesis 15 predicted that positive affective states would mediate the association between help receipt and relationship satisfaction and hypothesis 16 predicted that negative affective states would mediate the association between harm receipt and relationship satisfaction. Research question 3 questioned how promotion and prevention regulatory foci would mediate the associations between help and harm receipt and help and harm engagement. In order assess the significance of the indirect mediating effects, the PRODCLIN program was used (MacKinnon, Fritz, Williams, & Lockwood, 2007; Tofighi & MacKinnon, 2011). The PRODCLIN program calculates confidence intervals around an indirect effect based on the distribution of the product approach introduced by MacKinnon, Lockwood, Hoffman, West, and Sheets (2002). This approach is beneficial relative to Sobel’s test (1982) because it does not assume the distribution of the product of the two regression coefficients that make up the indirect effect is normal (MacKinnon et al., 2002). However, the Sobel test, for which the standard error was calculated with the first-order Taylor series expansion which is recommended for multi-level data (Krull & MacKinnon, 1999), was also used in calculating the indirect effects to cross-verify the results. Initial evidence of mediation can be inferred from how the coefficients of the direct effect of the dependent variables (help and harm engagement, relationship satisfaction) regressed onto the independent variables (help and harm receipt) change with the inclusion of the mediator variables (positive and negative affective states, promotion and prevention regulatory foci). Regarding hypothesis 11, as shown in Model 1 of Table 3, help receipt was significantly and positively related to help engagement (b100=0.52, p<.001). With the addition of positive affective states (see Model 2), this coefficient was slightly reduced, although it remained significant (b100=0.48, p<.001). Based on the Sobel test (z=2.54, p<.05) and the PRODCLIN 70 program 95% confidence intervals (CI=0.01, 0.09) the indirect effect was significant. This suggests that positive affective states partially mediated the association between help receipt and help engagement, supporting hypothesis 11. Regarding hypothesis 12, as shown in Model 1 of Table 4, harm receipt was significantly and positively related to harm engagement (b200=0.25, p<.001) and this coefficient was slightly decreased with the addition of negative affective states (see Model 2 of Table 4; b200=0.20, p<.05), although it remained significant. Based on the Sobel test, the indirect effect was not significant (z=1.66, n.s.) and neither were the results of the PRODCLIN program (95% CI=-0.002, 0.12). Therefore, negative affective states did not mediate the association between harm receipt and harm engagement and hypothesis 12 was not supported. Regarding hypothesis 15, as shown in Model 1 of Table 5, help receipt was significantly and positively related to relationship satisfaction (b100=0.42, p<.001). With the addition of positive affective states (see Model 2), this coefficient was reduced, although it remained significant (b100=0.31 p<.001). Based on the Sobel test (z=1.73, p=.08) the indirect effect was not significant but the PRODCLIN program 95% confidence intervals (CI=0.05, 0.16) indicated that the indirect effect was significant. This suggests that positive affective states partially mediated the association between help receipt and relationship satisfaction, supporting hypothesis 15. Regarding hypothesis 16, as shown in Model 1 of Table 5, harm receipt was significantly and negatively related to relationship satisfaction (b200=-0.38, p<.001). With the addition of negative affective states (see Model 2), this coefficient was reduced, although it remained significant (b200=-0.22 p<.05). Based on the Sobel test (z=-1.71, p=.09) the indirect effect was not significant and the PRODCLIN program 95% confidence intervals (CI=-0.25, 0.00) also indicated that the indirect was not significant. This suggests that negative affective 71 states did not mediate the association between harm receipt and relationship satisfaction. Hypothesis 16 was not supported. In terms of research question 3, with the addition of promotion regulatory focus to the model, the relationship between help receipt and help engagement decreased from (b100=0.52, p<.001; see Model 1 of Table 3) to (b100=0.49, p<.001; see Model 3 of Table 3), although the coefficient remained significant. Based on the Sobel test (z=1.94, p=0.05) and the PRODCLIN program (95% CI=.003, .06) the indirect effect was significant. This suggests that that promotion regulatory focus partially mediated the association between help receipt and help engagement. Further, the coefficient of the relationship between help receipt and harm engagement (b100=0.05, p<.05; see Model 1 of Table 4) also decreased to non-significance with the inclusion of promotion regulatory focus into the model (b100=-0.03, n.s. see Model 3 of Table 4). Indeed, based on the Sobel test (z=-2.28, p<.05) and PRODLCLIN program (95% CI=.004, .04) the indirect effect was significant. This suggests that promotion regulatory focus fully mediated the association between help receipt and harm engagement. Further, the coefficient of the relationship between harm receipt and harm engagement (b200=0.25, p<.001; see Model 1 of Table 4) decreased slightly with the inclusion of promotion and prevention regulatory foci into the model (b200=0.19, p<.05; see Model 3 of Table 4). However, neither the indirect effects of promotion (z=1.74, n.s.; PRODCLIN program 95% CI=-.001, .03) or prevention (z=1.83, n.s.; PRODCLIN program 95% CI=-.001, .03) regulatory focus were significant. Therefore, promotion and prevention regulatory foci did not mediate the association between harm receipt and harm engagement. Finally, the coefficient of the relationship between harm receipt and help engagement (b200=0.16, p<.001; see Model 1 of Table 4) actually increased slightly with the inclusion of promotion and prevention regulatory foci into the model (b200=0.18, p<.05; see 72 Model 3 of Table 4). The indirect effect of promotion regulatory focus (z=-1.57, n.s.; PRODCLIN program 95% CI=-0.05, 0.00) was not significant. Therefore, promotion and prevention regulatory foci did not mediate the effect of harm receipt on help engagement. Cross-Level Hypotheses Moderation. Hypotheses 3 predicted that the association between help receipt and help engagement would be stronger in approach relationships compared to avoidance relationships. Hypothesis 4 predicted that the association between harm receipt and harm engagement would be stronger in avoidance relationships compared to approach relationships. Results of slopes-asoutcomes are shown in the “cross-level moderation” section in Model 1 of Tables 3 and 4. Results revealed that relationship motive was not significantly related to the within-relationship slope between help receipt and help engagement (b110=0.08, n.s.) or between help receipt and harm engagement (b110=0.05, n.s.). Further, relationship motive was not significantly related to the within-relationship slope between harm receipt and harm engagement (b210=0.01, n.s.) or between harm receipt and help engagement (b210=-0.08, n.s.). Hypotheses 3 and 4 were therefore not supported. Hypothesis 7 predicted that the relationship between help receipt and positive affective states would be stronger for individuals in approach-oriented relationships compared to individuals in avoidance-oriented relationships. Hypothesis 8 predicted that the association between harm receipt and negative affective states would be stronger in avoidance-oriented relationships compared to approach-oriented relationships (see Table 6). Results of the slopes-asoutcomes regressions revealed that relationship motive was not significantly related to the within-relationship slope between help receipt and positive affective states (b110=-0.02, n.s.) or negative affective states (b110=-0.03, n.s.) nor was it related to the within-relationship slope 73 between harm receipt and negative affective states (b210=0.00, n.s.) or positive affective states (b210=-0.26, n.s.). Hypotheses 7 and 8 were therefore not supported. Further inspection of the data revealed that relationship motive did not significantly relate to any of the Level 1 slopes of the dependent variables (help engagement, harm engagement, relationship satisfaction) regressed on to positive and negative affective states (Model 2 of Tables 3, 4, and 5) and receipt of help and harm (Model 1 of Tables 3, 4, and 5). However, relationship motive did significantly relate to the Level 1 slopes of the association between prevention regulatory focus and harm engagement (b610=-0.09, p<.05; see Model 3 of Table 4). Simple slopes analysis (Aiken & West, 1991) indicated that the association between prevention regulatory focus and harm engagement was not significant in approachoriented relationships (b20=-0.01, n.s.) and significant and positive in avoidance-oriented relationships (b20=0.06, p<.05). Refer to Figure 3 for a graphical representation of this interaction. Overall, the results suggest that approach and avoidance relationship motives did not affect how within-relationship positive and negative affective states or help and harm receipt relate to help and harm engagement. However, the association between prevention regulatory foci and harm engagement was affected by relationship-level motives. 74 Figure 3 Harm engagement as outcome of the interaction between relationship motive and prevention focus; relationship motive coded as avoidance = 0 and approach = 1. Moderated Mediation. Hypothesis 17 predicted that relationship motive would moderate the mediating effect of positive affective states on the association between help receipt and help engagement. Hypothesis 18 predicted that relationship motive would moderate the mediating effect of negative affective states on the association between harm receipt and harm engaged. Hypothesis 19 predicted that relationship motive would moderate the mediating effect of positive affective states on the association between help receipt and relationship satisfaction and hypothesis 20 predicted that relationship motive would moderate the mediating effect of negative affective states on the association between harm receipt and relationship satisfaction. Research question 3 explored how relationship motives would moderate the mediating effects of promotion and prevention regulatory foci on the associations between help and harm receipt and 75 help and harm engagement. Edwards and Lambert’s (2007) moderated path analysis approach was followed by comparing the indirect effects across levels of moderator variable, relationship motive (approach=1, avoidance=0). Regarding hypothesis 17, positive affective states significantly mediated the association between help receipt and help engagement in approach (z=2.74, p<.05; PRODLCLIN program 95% CI=.02, .13) and avoidance relationships (z=2.43, p<.05; PRODLCLIN program 95% CI=.01, .08) but the indirect effect did not differ significantly across relationships (z=0.03, n.s.). Therefore, hypothesis 17 was not supported. In terms of hypothesis 18, negative affective states did not significantly mediate the association between harm receipt and harm engagement in approach (z=1.46, n.s.; PRODLCLIN program 95% CI=.00, .08) or avoidance relationships (z=0.92, n.s.; PRODLCLIN program 95% CI=-.04, .13) and the indirect effect also did not differ significantly across relationships (z=0.01, n.s.). Therefore, hypothesis 18 was also not supported. Regarding hypothesis 19, positive affective states significantly mediated the association between help receipt and relationship satisfaction in approach (z=2.30, p<.05; PRODLCLIN program 95% CI=.02, .18) and avoidance relationships (z=4.13, p<.05; PRODLCLIN program 95% CI=.05, .13) but the indirect effect did not differ significantly across relationships (z=-0.01, n.s.). Therefore, hypothesis 19 was not supported. In terms of hypothesis 20, negative affective states did not significantly mediate the association between harm receipt and relationship satisfaction in approach (z=-0.44, n.s.; PRODLCLIN program 95% CI=-.12, .07) or avoidance relationships (although the Sobel test was significant, the PRODCLIN analysis was not: z=-2.19, p<.05; PRODLCLIN program 95% CI=-.15, .02) and the indirect effect also did not differ significantly across relationships (z=-1.29, n.s.). Therefore, hypothesis 20 was also not supported. 76 Finally, regarding research question 3, promotion regulatory focus mediated the association between help receipt and help engagement in approach (z=2.99, p<.05; PRODLCLIN program 95% CI=.02, .08) and avoidance relationships (z=2.03, p<.05; PRODLCLIN program 95% CI=.002, .005) but the indirect effect did not differ significantly across relationships (z=0.10, n.s.). Promotion regulatory focus also mediated the association between help receipt and harm engagement in both approach (z=-1.97, p<.05; PRODLCLIN program 95% CI=-.005, -.15) and avoidance relationships (z=-4.45, p<.05; PRODLCLIN program 95% CI=-.03, -.01) although the indirect effect did not differ significantly across relationships (z=0.08, n.s.). Prevention regulatory focus did not significantly mediate the association between harm receipt and harm engagement in approach relationships (z=0.07, n.s.; PRODLCLIN program 95% CI=.003, .003) but it did have a significant mediating effect in avoidance relationships (z=2.40, p<.05; PRODLCLIN program 95% CI=.01, .03). However, the indirect effects in the approach and avoidance relationships were not significantly different (z=-0.12, n.s.). All other mediating models where positive and negative affective states and promotion and prevention regulatory foci were mediators between associations of help and harm receipt and help and harm engagement and relationship satisfaction had no significant mediating effects in either approach or avoidance relationships. Therefore, approach-avoidance relationship motives did not moderate the mediation of the specified within relationship mechanisms. Exploratory Analyses Emotional Intelligence. The effects of individual differences in emotional intelligence on the within-relationship associations between help and harm receipt, positive and negative affective states, and help and harm engagement were also examined using Edwards and Lambert’s (2007) moderated path analysis approach. In order to follow this approach, the 3-level 77 data were divided into two separate data sets consisting of two levels each based on the Level-2 relationship motive manipulation. One data set consisted of approach-oriented relationships and another of avoidance-oriented relationships. These two new data sets thus had two levels where Level 2 was the individual level and Level 1 was the daily experiences level nested within individuals. Participant gender, race, age, hierarchical status difference, liking, and relationship tenure were included as controls in these analyses and set as predictors of Level 1 intercepts. Emotional intelligence was included as a predictor of Level 1 intercepts and as well as a predictor of Level 1 slopes of the association between the independent variables and the mediators. In the first step of the analyses the mediators (positive and negative affective states) were regressed onto the independent variables (receipt of help and harm) and the interaction between emotional intelligence and the independent variables. The mediators-as-outcomes results for both approach and avoidance data sets are shown in Table 7. Emotional intelligence was not significantly related to positive affect states in the approach relationship data (b04=0.56, n.s.) or the avoidance relationship data (b04=0.23, n.s.). Emotional intelligence was not significantly related to overall levels of negative affective states in either the approach (b04=-0.19, n.s.) or avoidance (b04=-0.04, n.s.) data. Further, emotional intelligence did not affect the slopes of the associations between help and harm receipt and positive affective states in either the approach (help receipt: b11=0.07, n.s.; harm receipt: b21=0.04, n.s.) or avoidance (help receipt: b11=-0.04, n.s.; harm receipt: b21=-0.32, n.s.) data. Emotional intelligence also had no significant effect on the slopes of the association between help receipt and negative affective states in the approach (b11=-0.07, n.s.) and avoidance (b11=0.04, n.s.) or harm receipt and negative affective states in the approach (b21=-0.49, n.s.) or avoidance data (b21=-0.40, n.s.). 78 Table 7 Variable Positive affective states Approach Avoidance b s.e. b s.e. 3.97** 0.32 3.71** 0.47 Negative affective states Approach Avoidance b s.e. b s.e. 1.51** 0.15 1.28** 0.14 Intercept (b00) -0.27 0.34 0.00 0.41 -0.14 0.12 -0.06 0.12 -0.49 0.33 -0.46 0.41 -0.17 0.11 0.13 0.12 0.02 0.01 -0.02 0.01 0.00 0.00 -0.01 0.00 0.56 0.30 0.23 0.37 -0.19 0.11 -0.04 0.10 0.45* 0.20 -0.42 0.34 -0.07 0.07 -0.09 0.12 0.26 0.17 0.26 0.21 -0.01 0.05 -0.01 0.10 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 -0.22* 0.10 -0.18 0.11 0.25* 0.11 0.27 0.18 0.36* 0.07 0.38** 0.06 -0.05 0.03 -0.06 0.03 0.04 0.19 -0.32 0.24 -0.49 0.28 -0.40 0.34 0.07 0.11 -0.04 0.13 -0.04 0.07 -0.07 0.05 Gender (b01) Race (b02) Age (b03) Emotional intelligence (b04) Status difference (b05) Liking (b06) Relationship tenure (b07) Harm received (b10) Help received (b20) Cross-level moderation Harm received* emotional intelligence (b11) Help received*emotional intelligence (b21) HLM results for mediators as outcomes and emotional intelligence as individual-level moderator; N2=55, approach N1=358, avoid N1=362, approach relationships are on left of each column and avoidance relationships are italicized on right, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded as different status=1, same status=0, **p<.001, *p<.05 79 Table 8 Model 1 Variable Intercept b00 Gender (b01) Race (b02) Age (b03) Emotional intelligence (b04) Status difference (b05) Liking (b06) Relationship tenure (b07) Harm received (b10) Help received (b20) Negative affective states (b30) Positive affective states (b40) Cross-level moderation Harm received* emotional intelligence (b11) Help received*emotional intelligence (b21) Negative affective states* emotional intelligence (b31) Positive affective states* emotional intelligence (b41) Model 2 Approach b s.e. 3.92** 0.42 -0.08 0.42 Avoidance b s.e. 3.95** 0.39 0.11 0.38 Approach b s.e. 3.70** 0.43 0.18 0.45 Avoidance b s.e. 4.00** 0.38 0.12 0.37 -0.51 0.01 0.27 0.32 0.10 0.00 -0.12 0.52** -0.64* 0.00 0.19 -0.37 0.25 0.00 0.15 0.48** -0.58 0.01 0.33 0.29 0.18 0.00 -0.11 0.52** 0.35 0.01 0.32 0.26 0.20 0.00 0.11 0.06 -0.66* 0.00 0.31 -0.33 0.24 -0.01 0.10 0.49** 0.27 0.01 0.32 0.33 0.17 0.00 0.09 0.07 0.20* 0.13* 0.07 0.04 0.08 0.12* 0.05 0.06 0.33 0.01 0.34 0.23 0.18 0.00 0.10 0.07 0.26 0.01 0.34 0.32 0.19 0.01 0.08 0.07 -0.28 0.22 -0.11 0.15 -0.33 0.26 -0.10 0.18 0.08 0.16 -0.06 0.13 0.17 0.14 -0.01 0.14 0.15 0.20 -0.21 0.12 -0.09 0.09 -0.07 0.12 HLM results for help engagement as outcome and emotional intelligence as individual-level moderator; N2=55, approach N1=358, avoid N1=362, approach relationships are on left of each column and avoidance relationships are italicized on right, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded as different status=1, same status=0, **p<.001, *p<.05 80 Table 9 Model 1 Variable Intercept (b00) Gender (b01) Race (b02) Age (b03) Emotional intelligence (b04) Status difference (b05) Liking (b06) Relationship tenure (b07) Harm received (b10) Help received (b20) Negative affective states (b30) Positive affective states (b40) Cross-level moderation Harm received* emotional intelligence (b11) Help received*emotional intelligence (b21) Negative affective states* emotional intelligence (b31) Positive affective states* emotional intelligence (b41) Approach b s.e. 1.41** 0.21 -0.13 0.10 -0.12 0.15 0.00 0.00 -0.12 0.13 -0.05 0.06 -0.10 0.05 0.00 0.00 0.21* 0.08 0.01 0.01 Model 2 Avoidance b s.e. 1.33** 0.21 -0.05 0.11 -0.03 0.15 0.00 0.00 -0.15 0.13 -0.13 0.10 0.00 0.05 0.00 0.00 0.27* 0.05 -0.06* 0.02 Approach b s.e. 1.21** 0.08 -0.07 0.06 0.00 0.04 0.00 0.00 -0.07 0.09 -0.04 0.04 -0.06 0.04 0.00 0.00 0.09 0.06 0.00 0.02 0.14* 0.05 0.00 0.01 Avoidance b s.e. 1.43** 0.13 -0.19 0.10 -0.04 0.09 0.00 0.00 -0.15 0.12 -0.12 0.09 -0.01 0.05 0.00 0.00 0.21 0.05 -0.05* 0.02 0.14* 0.04 -0.02 0.02 0.11 0.17 0.08 0.10 0.09 0.14 0.18 0.10 0.01 0.01 -0.04 0.04 0.03 0.07 -0.03 0.03 0.11 0.13 -0.06 0.11 0.06 0.04 -0.03 0.05 HLM results for harm engagement as outcome and emotional intelligence as individual-level moderator; N2=55, approach N1=358, avoid N1=362, approach relationships are on left of each column and avoidance relationships are italicized on right, gender coded as male=0, female=1, race coded as white=1, non-white=0, relationship status coded as different status=1, same status=0, **p<.001, *p<.05 81 Next, the dependent variables (help and harm engagement) were regressed onto the independent variables (help and harm receipt; Model 1) and then the independent variables with the addition of the mediators (positive and negative affective states; Model 2). The results for both approach and avoidance data with help engagement as the outcome are displayed in Table 8 and the results with harm engagement as the outcome are displayed in Table 9. Emotional intelligence had no direct effect on levels of help and harm engagement. Further, emotional intelligence did not affect the slope-as-outcomes associations between help and harm receipt or positive and negative affective states and help and harm engagement. Overall emotional intelligence did not affect associations between help and harm receipt, positive and negative affective states, and help and harm engagement. Emotional intelligence did also not moderate any mediating effect of positive or negative affective states on the association between the receipt of help and harm and engagement in help and harm. Interpersonal Trust. How affect- and cognition-based trust were related to approachavoidance relationship motives and engagement in help and harm and relationship satisfaction was examined next. Affect- and cognition-based trust (Level 2) were tested as mediators of the association between approach-avoidance relationship motives (Level 2) and engagement in help and harm and relationship satisfaction (Level 1). Multi-level mediation procedures were followed based on the recommendations of Bauer, Preacher, and Gil (2006). Individual differences in gender, age, race, and approach and avoidance disposition were included as Level 3 controls, and liking, relationship tenure, and status difference were included as Level 2 controls in the analyses. First, affect- and cognition-based trust were regressed onto relationship motives (approach=1, avoidance=0). Relationship motive was not significantly related to affect-based trust (b=0.09, s.e.=0.10, n.s.) or cognition-based trust (b=0.23, s.e.=0.14, n.s.). Therefore, 82 mediation was not supported. Affect-based trust did not significantly relate to help engagement (b010= 0.00, s.e.=0.11, T=0.04, n.s.) but cognition-based trust did significantly and positively relate to help engagement (b020= 0.19, s.e.=0.07, T=2.86,p<.05.). Neither affect- (b010= -0.10, s.e.=0.10, T=-1.05, n.s.) or cognition-based trust (b020= -0.09, s.e.=0.07, T=1.34, n.s.) significantly related to harm engagement. Further, affect-based trust did not significantly relate to relationship satisfaction (b010= 0.10, s.e.=0.12, T=0.84, n.s.) but cognition-based trust did significantly and positively relate to relationship satisfaction (b010= 0.53, s.e.=0.10, T=5.23,p<.001.). Therefore, although trust is not related to approach-avoidance relationship motives, cognition-based trust related to help engagement and relationship satisfaction. This suggests that individuals were more likely to help someone whom they viewed as reliable and dependable and they may do so because they are motivated to foster future obligations or are certain that they can assist each other in their future performance (McAllister, 1995). At least within work relationships, such concern for future obligations relevant to performance may override concerns that the other persons’ needs are met and thereby affect-based trust was not related to helping. Similarly, within the work context, perceptions that a coworker is reliable and dependable may be particularly important for relationship satisfaction. 83 DISCUSSION The focus of the current study was to examine how relational approach-avoidance motives affect within-relationship processes underlying the dyadic exchange of interpersonal helping and harming behaviors and relationship satisfaction. Compared to the between-person, within-person, or cross-sectional social network-approaches, the within-relationship approach represents a shift in terms of acknowledging that behavior varies in predictable ways within relationships over time. In the current study, employees from three organizations reported about their interpersonal experiences with two coworkers for ten days each. The results were generally supportive of SET (Blau, 1964; Gouldner, 1960), AET (Weiss & Cropanzano, 1995) and the emotion-centered model (Spector & Fox, 2002) as theoretical explanations for social exchange and affective responses to the receipt of help and harm. The results also shed light on how regulatory focus theory (Higgins, 1997; 1998) is useful in explaining engagement in helping and harming behaviors and how promotion focus mediates the exchange of helping and harming behaviors. Relationship-level approach-avoidance social motives (Gable, 2006) did not exert influence on behavioral, affective, or attitudinal responses to the receipt of help and harm but relationship motives did affect how prevention regulatory focus related to engagement in harming behaviors. Below, the theoretical implications of the results are discussed starting with the within-relationship processes followed by the cross-level effects of relationship approachavoidance motives. Partitioning of Variance. Results of the current study revealed that engagement in help and harm, relationship satisfaction, positive and negative affective states, and promotion and prevention regulatory foci exhibited substantial variability within and across relationships. Almost one-quarter of the variance in help engagement was due to within-relationship variance; 84 almost one-third of the variance in help engagement was due to between-relationship variance; while under half was due to individual differences. This suggests that the majority of variance in help engagement was due to within and between-relationship effects. Variability in harm engagement was even less dependent on the individual compared to help engagement; most of the variance in harm engagement was either within relationships or between relationships. Similar to the results of Dalal and colleagues (2009), albeit a small difference withinrelationships, harm engagement varied within individuals more so than help engagement. One explanation underlying this difference is the fact that employees are more likely to regulate their engagement in harm than their engagement in help because organizations are more likely to discourage harm (Dalal et al., 2009). Further, relative to engagement in help, engagement in harm appeared to vary more across relationships. This is consistent with theorizing related to workplace aggression suggesting that perpetration of aggression is largely influenced by the nature of the relationship between the perpetrator and target (Hershcovis & Barling, 2007; Aquino & Lamertz, 2004). This evidence also demonstrates the importance of considering within relationship processes, including SET, AET, and regulatory foci that explain dyadic engagement in helping and harming behaviors. Variance partitioning of relationship satisfaction indicated that individuals did not generally have high or low relationship satisfaction over time, but relationship satisfaction varied across relationships and within relationships over time. Such results is consistent with AET (Weiss & Cropanzano, 1996) suggesting that attitudes are constructed by integrating affect and information relevant to specific situations but also stored information relative to the attitude object. Therefore, relationship satisfaction is driven not only be fluctuations in affect and 85 situational information over time, but also by stored information relevant to the coworker that is less dynamic. The majority of variance in positive and negative affective states and promotion regulatory focus was explained between-relationships compared to within-relationships, although similar amounts of variance in prevention focus was explained within relationships and between relationships. This suggests that the relationship may serve as an important context that evokes affective and regulatory reactions and these reactions tend to be stable within the relationship but vary across relationships. These results are consistent with the notion of relational self for which cognitions, affect, and goals tend to be activated in relation to a significant other (Baldwin, 1992). Further, a large proportion of the variance in negative affective states and promotion and prevention regulatory foci also varied within-relationships over-time. Greater within-relationship variation in negative affective states, relative to positive affective states, may be due to how individuals tend to structure their environments in ways to increase their tendency to have positive experiences and decrease their tendency to have negative experiences (e.g., avoiding situations that cause distress; Carstensen, Isaacowitz, & Charles, 1999). As such, negative stimuli within relationships that would evoke negative affective states are likely to not only be less common over time but more sporadic relative to positive affective states. Further, withinrelationship fluctuations in regulatory foci may be a function of self-regulatory reactions to daily situations that evoke promotion and prevention foci. Overall, these results justify the use of within-relationship and between-relationship designs in the study of these interpersonal phenomena. Social Exchange. The results of the current study were generally supportive of SET and positive and negative norms of reciprocity that underlie social exchanges (Blau, 1964, Gouldner, 86 1960). Similar to findings of a recent social network study by Lyons and Scott (2012), the extent to which a focal employee received help and harm from a given coworker significantly related to the amount of help and harm they engaged in towards that coworker, respectively. Receipt of help was more strongly associated with engagement in help than engagement in harm, supporting the positive norm of reciprocity, and receipt of harm was more strongly associated with engagement in harm than engagement in help, supporting the negative norm of reciprocity. However, inconsistent with expectations receipt of help was negatively related to engagement in harm and receipt of harm was positively related to engagement in help. These results are counter to previous research demonstrating that the positive norm of reciprocity is independent of the negative norm of reciprocity (Lyons & Scott, 2012), that positive and negative social exchanges may in fact not be independent. The current study utilized an experience-sampling longitudinal design which has benefits over the cross-sectional design utilized by Lyons and Scott (2012). Experience-sampling designs are thought to be less susceptible to biases associated with cross-sectional retrospective accounts, including consistency motifs or illusory correlations (Podsakoff, MacKenzie, Podsakoff, & Lee, 2003). For example, by relying on retrospective accounts in the Lyons and Scott (2012) study, participants may have artificially inflated their associations between similar behaviors (e.g., help receipt and help engagement) and deflated their associations between dissimilar behaviors (e.g., help receipt and harm engagement). The results of the current study suggest that receipt of help may make individuals less likely to engage in harming behaviors perhaps because help receipt fosters positive evaluations of the exchange relationship reducing the likelihood that harm would engaged in. This result is consistent with research on abusive supervision that has demonstrated that subordinates are less likely to engage in deviance towards their supervisor or organization 87 when they evaluate their relationship with their supervisor as having high-quality exchange (Tepper, 2007). However, promotion regulatory focus also mediated the association between help receipt and harm engagement and its effects may serve as an alternative explanation. The results of regulatory foci are elaborated upon below. Although harm receipt was strongly related to engagement in harm, to lesser degree it was also positively related to engagement in help. That is, in response to harm employees engaged in both harmful and helping behaviors. Helping behaviors in response to harm may in fact reflect the manifestation of relationship maintenance and reconciliation strategies (Aquino, Tripp, & Bies, 2006) related to forgiveness (Fehr & Gelfand, 2012). In the context of interdependent working relationships (participants in the current study were asked to nominate coworkers with whom they worked with on a daily basis), targets of harm may be less motivated for revenge or avoidance and instead desire for rapport, initiating benevolence and prosocial behaviors to achieve that means. Future research will benefit from examining the motivational underpinnings of help responses to the receipt of harm in dyadic relationships. Overall, these unexpected social exchange results provide evidence counter to previous research demonstrating that social exchange is “tat for tat” (Lyons & Scott, 2012), and in fact, employees may also exchange “tit for tat” (Gouldner, 1960). Affective Events. In addition to the effects of social exchange, the results of the current study were also consistent with AET (Weiss & Cropanzano, 1996) and the valence symmetry arguments of the emotion-centered model (Spector & Fox, 2002). That is, the receipt of help and harm from a certain coworker were associated with the amount of positive and negative affective states elicited from that coworker, respectively. Help receipt was not related to negative affective states and harm receipt was not related to positive affective states supporting arguments of 88 valence symmetry (Spector & Fox, 2002). Additionally, positive and negative affective states were also related to the amount of help and harm engaged in towards certain coworkers, respectively. Positive affective states were not associated with harm engagement and negative affective states were not associated with help engagement also supporting valence symmetry arguments (Spector & Fox, 2002). As expected, action tendencies accompanying positive affective states were prosocial, motivating individuals to draw in or engage the perceived source of the positive affect whereas action tendencies accompanying negative affective states were antisocial, motivating individuals to repel the perceived source of the negative affect (Spector & Fox, 2002). Indeed, positive affective states partially mediated the association between receipt of help and help engagement. This suggests that positive norms of reciprocity may partially be explained by the management of positive affect. Receiving help may evoke positive emotions, and in order to draw the source of the positive emotions closer, individual reciprocate with help. However, unexpectedly, negative affective states did not mediate the association between the receipt of harm and harm engagement. One possible explanation could be that the expression of negative affective states and harming behaviors are less desirable in organizations and are less common compared to expressions of positive affective states and helping behaviors (Dalal et al., 2009). For example, opportunities to engage in harm, relative to help, are also less common and may depend on opportunity, including the presence of a supervisor or coworkers. As such, reacting to positive affective states in response to helping is easier to do than reacting to negative affective states in response to harming. The mediating effects of negative affective states on the reciprocation of harming behaviors may depend on the extent to which the context permits the enactment of harm. Research has demonstrated that workplace aggression is more common in organizational contexts that have climates permitting aggression (Bowling & Beehr, 2006). 89 Future research will benefit from examining opportune conditions under which negative affective states mediate harm reciprocation. Results of the current study were also generally supportive of AET (Weiss & Cropanzano, 1996) arguments surrounding affective components of attitudes. Positive affective states positively related to relationship satisfaction and negative affective states negatively related to relationship satisfaction. Positive affective states also mediated the association between harm receipt and relationship satisfaction. In line with AET, receiving help, a positive event, may evoke positive affective states which then inform attitudinal evaluations of the attitudinal object (i.e., relationship satisfaction), the interacting coworker. However, contrary to AET expectations negative affective states did not mediate the association between harm receipt and relationship satisfaction. This may be because negative affective states are not frequent enough within relationships to counter the broader positive (within-relationship and relationship-level) information that individuals incorporate into their construction of a relationship attitude. Such an explanation is supportive of results demonstrating that relationship satisfaction tended to be high across relationships. Regulatory Foci. Promotion and prevention regulatory foci played an important role in the reciprocation of help and harm. The effects that promotion and prevention regulatory foci had on the reciprocation of help and harm may have come about via the symbolic representation of how help and harm affects individuals’ sense of belonging within a relationship. Momentary situations that provide feedback to individuals about their belongingness are known to be related to social-oriented promotion and prevention regulatory foci (Crowe & Higgins, 1997; Molden et al., 2009; Oyserman et al., 2007; Winterheld et al., 2011). Research has demonstrated that interpersonal feedback oriented towards one’s progress in achieving belonging (helping) is 90 related to promotion regulatory focus and feedback oriented towards losses associated with belonging (e.g., rejection, loneliness; harming) is related to prevention regulatory focus (Crowe & Higgins, 1997). Consistent with previous research (e.g., Molden et al., 2009; Oyserman et al., 2007; Winterheld et al., 2011) in the current study, receipt of help positively related to promotion regulatory focus, but not to prevention regulatory focus, and the receipt of harm positively related to prevention regulatory focus and negatively related to promotion regulatory focus. The social loss associated with receiving harm may have in fact reduce a sense of social gain (“not belonging”) thereby reducing promotion regulation. Receiving help, on the other hand, is perhaps a less salient indicator of the presence or absence of social losses (e.g., rejection vs. nonrejection) and may therefore be less influential in inducing prevention regulatory focus. Promotion regulatory focus also positively related to help engagement and negatively related to harm engagement whereas prevention focus positively related to harm engagement but was not related to help engagement. Such findings are consistent with previous research demonstrating that promotion focus can inspire more risky behaviors that are strategically inclined to achieve relationship gains, including engaging in helpful behaviors (Molden et al., 2009; Winterheld et al., 2011). This strategy also means that behaviors that would impede the desired progress (harm) would be less desirable. Prevention-focused strategies, on the other hand, tend to be less risky and are more concerned with safety and security. Behaviorial outcomes of prevention focus are thus thought to be intended to repel the source of social loss. Such behaviors include withdrawing from social contact (Molden et al., 2009) and confronting the instigator to stop the harmful treatment (Oyserman et al., 2007). Prevention focus may not relate to help engagement because engagement in helping would be too risky for the precautious 91 and safety-oriented prevention strategy (Winterheld et al., 2011). Promotion focus in particular appears to play an influential role in directing both help and harm engagement. Indeed, promotion focus partially mediated the association between help receipt and help engagement and fully mediated the association between help receipt and harm engagement. Social gain associated with receiving help may have induced promotion focus that then may have lead to prosocial behaviors and the avoidance of detrimental behaviors (harm) in order to improve belonging. The mediating effect of promotion focus on the association between help receipt and harm engagement is particularly interesting because it provides evidence for a motivational mechanism that connects positive and negative social exchanges (Gouldner, 1960). The results of the current study suggest that receiving help is not only associated with engaging in help via positive norms of reciprocity, positive affective states (AET, the emotion-centered model), and promotion regulatory focus, but receiving help also reduces engagement in harm through the effects of promotion regulatory focus. As will be discussed as practical implications below, this has important implications for how organizations can manage the occurrence of harmful interpersonal behaviors in organizations. Approach-Avoidance Relationship Motives. Contrary to expectations, relationshiplevel approach and avoidance social motivation did not affect social exchange and affect-driven processes relating help and harm receipt to help and harm engagement and relationship satisfaction. Although, approach-avoidance relationship motives did affect the association between prevention focus and harm engagement. As previous research has demonstrated, social exchange and affective states accounts for the reciprocation of help and harm are robust and may not differ across contexts (Lyons & Scott, 2012). Similarly, it appears as though sensitivity to affiliation-related stimuli in approach-oriented relationships and sensitivity to rejection-stimuli in 92 avoidance-oriented relationships may not exert a contextual effect on the social exchange and affect-driven processes. Although previous research drawing from the approach-avoidance model of social motivation (Gable, 2006) has demonstrated that individuals exhibit more positive affect, attitudes, and behaviors in approach relationships and more negative affect, attitudes and behaviors in avoidance relationships (Elliot et al., 2006; Gable, 2006; Impett et al., 005; Impett & Gordon, 2010; Impett et al., 2010; Impett, Starchman, Finkel, & Gable, 2008), that research examined affect, attitudes and behaviors as singular outcomes of approach-avoidance relationships with one relationship partner. Evidence of the current study, on the other hand, suggests that social exchange and affective events theories may generalize across relationship contexts within and between individuals regardless of the approach or avoidance orientation of the relationship. Another possible explanation surrounding the null findings of relationship motives is that the manipulation did not work or that it was not sensitive to capture changes in relational self over time. Such explanations will be discussed in more detail in the limitations section of the discussion. However, it is worth noting here that validity evidence did support the approachavoidance relationship manipulation (see Manipulation Check) and the relationship motives did influence the association between prevention regulatory focus and harm engagement. That is, prevention regulatory focus was more strongly related to harm engagement in avoidance relationships compared to approach relationships. Accordingly, the avoidance relationship may serve as a context in which individuals are particularly attuned to potential relational threat. With its concern for security and safety, prevention regulatory focus in an avoidance relationship will be particularly inclined to induce behavior to repel potential sources of threat (Molden, Lee, & Higgins, 2008). Therefore, in an 93 avoidance relationship where threats are perceived to be more likely, prevention regulatory foci will be more strongly related to harm engagement than in approach-oriented relationships in which individuals are less sensitive to relational threats. Avoidance relationships may be so concerned with threat that social loss is a more prevalent interpretation of behavior than social gain. Within an avoidance relationship, without a perceived opportunity for social gain, promotion regulatory focus may be unlikely to take effect on harming behaviors. Approach relationships did not affect the associations between promotion and prevention regulatory foci and help and harm engagement. This is could be because indicators of social gain or social loss are less pertinent to one’s sense of belonging in approach relationships in which the interpretation of interactions as social gain is commonplace. Overall, these results are counter to some traditional conceptualizations of promotion and prevention regulatory foci that argue regulatory foci pertain mainly to the approach domain of the hedonic principle (Crowe & Higgins, 1997; Higgins, 1997; 1998). Instead, these results provide evidence that prevention regulatory focus is also affected by the avoidance domain (Lanaj et al., 2012). The current study therefore provides support to the notion that both regulatory foci can be strategic inclinations towards achieving both approach and avoidance outcomes (Molden et al., 2008). Practical Implications In addition to the theoretical implications, the current study also has a number of practical implications for the management of interpersonal behaviors in organizations. The substantial within- and between-relationship variance in the engagement in help and harm suggest that there are limits to the effectiveness of staffing strategies – such as selecting for “helpers” or selecting against or terminating “harmers” – to foster helping and deter harming among employees. The 94 results of the current study suggest that managers would do well to train and develop employees to foster high quality relationships. In line with results supporting SET, managers could simply encourage employees to engage in more help (and thereby receive more help) and engage in less harm (and thereby receive less harm; Lyons & Scott, 2012). Further, in line with AET (Weiss & Cropanzano, 1996) positive affective states positively related to help engagement and negative affective states positively related to harm engagement. Therefore, managers should work to engender employee relationships that are ripe with positive affective states and deter circumstances inciting negative affective states. To do so, managers may target relational stressors that lead to negative affective states. For example, managers could restructure interpersonal and job circumstances to decrease ambiguity, conflict and stress by clearly communicating responsibilities and expectations among coworkers (Litzky, Eddleston, & Kidder, 2006). Conflict management training could also help employees manage stressful communication within their relationships thereby reducing incidents of negative affective states (Judge et al., 2006). Such strategies may also be effective in improving relationship satisfaction among coworkers. It is important to note that conflict is commonplace in the workplace and it is unrealistic to expect that managers could reduce harm entirely and foster universal relationship satisfaction. However, managers can do much to cultivate an environment that is open and tolerant of diverse viewpoints and encourage cooperative norms among coworkers preventing disagreements and conflict between individuals from being misinterpreted as harmful personal attacks (De Dreau & Weingart, 2003). Managers can also encourage their employees to use more collaborative and less contentious communication when experiencing stress or expressing disagreements (Lovelace, Shapiro, & Weigart, 2001). 95 Finally, results also indicated that promotion regulatory focus was related to increased engagement in helping behaviors and decreased engagement in harmful behaviors, whereas prevention focus was related to increased engagement in harmful behaviors. It thus follows that managers can perhaps encourage engagement in helpful behaviors and curb engagement in harmful behaviors by fostering promotion regulatory foci within relationships and by decreasing the extent to which employees draw upon prevention regulatory focus strategies within relationships. Research has demonstrated that promotion strategies can be primed by framing end-states in terms of achievements, accomplishments, and aspirations, as opposed to what individuals “ought” to do (Crowe & Higgins, 1997). For example, managers could train employees to interpret behaviors within relationships in terms of fulfilling hopes and aspirations for meaningful, cooperative, and productive collaborations. In such cases, employees may be more likely to interpret both helpful and harmful behaviors as potential for social gain and evoke promotion regulatory focus which would then lead to engagement in helpful behaviors and less harmful behaviors. By avoiding framing relationships in terms of “oughts” (e.g., “don’t be rude”, “follow the rules”, “mind your manners”), managers may also be able to reduce the extent to which prevention regulatory focus are evoked from harmful behaviors within relationships. Such framing may be particularly useful and adaptive in relationships where coworkers are vigilant to rejection (i.e., avoidant relationships) and security and precautionary needs are more likely to drive harmful behaviors. Managers could identify such relationships by noting which coworkers have a history of conflict or mistreatment and working with those coworkers to frame their interactions in a positive light. Limitations 96 As with all research, the current study also had some limitations. First, all constructs were measured via self-reports. One limitation of self-report data is the potential for common method bias to inflate relationships between variables. However, help and harm receipt and help and harm engagement were only moderately correlated with one another and correlations between other variables (within relationships, between relationships, and between individuals) had low correlations suggesting that same source bias cannot account for the effects alone. Further, significant interactions between relationship motives and prevention focus are also difficult to explain with same-source bias alone (Podsakoff et al., 2003). An additional concern with the use of self-report data is the concern of socially desirable responding, particularly regarding the sensitive nature of the harm variables. However, recent empirical evidence suggests that self-reports of help and harm engagement may in some cases be as accurate as “other reports” of the same information (Berry, Carpenter & Barratt, 2012; Dalal et al., 2009). Indeed, in a recent meta-analysis Berry and colleagues (2012) demonstrated that self-reports of CWB are comparable to other-reports of CWB and they were similarly related to antecedents and outcomes of CWB. Dalal and colleagues (2009) also demonstrated that withinindividual relationships between positive and negative affective states and OCB and CWB were comparable for self- and other-reported data. Further, as some forms of help and harm are emitted in private or in an unobservable fashion, other-reports may also be less accurate than self-reports (Dalal et al., 2009). Concerns regarding common method variance were also reduced for the Level 1 and Level 2 variables because the scores were centered at the individuals’ mean. This means that variable relations were compared relative to the individual means which reduces the potential effects of individual differences in responses biases – such as general responses biases or 97 affectivity – affecting the results (Podsakoff et al., 2003). Mood is also a concern for common method bias (Podsakoff et al., 2003) but affective states were included as predictors in the analyses essentially controlling for the effects of mood. Nevertheless, future research that overcomes the methodological limitations associated with self-report is recommended. For example, trained observers could be used to code video tapes of coworkers interacting. A second limitation of the design of the current study is the lack of control that is present in experience-sampling studies. Even though relational approach-avoidance motives were manipulated, participants still nominated their own choices for each coworker. Further, the causal associations between the within-relationship variables are difficult to ascertain given the current design. Indeed, it is possible that alternative, unmeasured variables could explain the associations between variables in the current study. It is also possible that the model tested in the current study could also be represented by alternative causal pathways. For example, the exchange of help and harm may lead to emotional evaluations of the exchange and subsequent experiences of positive and negative affective states (Lawler, 2001). Although the proposed model is one possible representation of the data, other alternative representations are possible. Accordingly, future research would benefit from adopting experimental methods or time-lagged research designs to test the proposed relations ( for an example see Dalal et al., 2009) and to increase confidence surrounding inferences of causality. Third, the sample was predominately female and white, which limits generalizability of the findings to men and persons of color. Gender and race were included as controls in the analyses. The results revealed that persons of color engaged in more help than Whites and that men engaged in more harm than women. Caution should be made in interpreting these findings 98 as they are based on small sample sizes. Future researchers should investigate gender and racial differences more thoroughly. Fourth, the analytical framework (Edwards & Lambert, 2007) relied on regression analyses and path analyses, which assumed that study variables are measured without error. Although internal consistency for study variables tended to be high, violations of this assumption can be addressed by using structural equation modeling of latent variables, which takes measurement error into account. Fifth, although validity evidence supported the approach-avoidance relationship distinction of the two coworkers nominated by each participant, the manipulation check of the current study failed to find significant differences in approach and avoidance social motivation between the two relationships. The manipulation check was assessed five weeks after the manipulation raising the question of the stability of the approach-avoidance relational self over the five week period of the study. Perhaps relational approach-avoidance motives changed over time? Theory of the relational self suggests that it has both transient and chronic influences (Andersen & Chen, 2002). The relational self is represented as knowledge structures that develop over time and are activated based on relevant situational cues. Variability in relational selves emerges on the basis of transient cues of knowledge structure accessibility. Knowledge structures are activated when relevant situational cues are salient, such as the presence of the nominated coworker. Stability of the relational self is derived from the chronic accessibility of the knowledge structures. Therefore, although the relational self is stable in its structure over time, the relational self is activated depending on the situation (e.g., whether the nominated coworker is present). Considering that participants in the current study were asked to nominate coworkers with whom they work on a daily basis and not complete surveys on days the 99 nominated coworker was not present, the approach-avoidance relational selves were expected to be chronically activated throughout the duration of the study. However, the assessment of approach-avoidance relational motives within the current study was unable to assess potential variations in relational selves over time. Such variation may in part explain the failure of the manipulation check. Future research will benefit from assessing the relationship construct of interest on a more frequent basis in order to capture potential variability of the relational self. Future Research In addition to the above recommendations, the results of the current study offer insight into possible directions for future research. First, compared to engagement in help, engagement in harm varied more across relationships. Such findings provide impetus for further exploration of alternative relationship-level variables that may influence outcomes of harm across relationships. For example, Hershcovis and Barling (2007) highlight relational power as an important relationship variable that could affect responses to workplace aggression. Targets who experience harm from perpetrators in positions of power may have more negative consequences because the perpetrator has more control over the employees’ social and/or work-related resources. Status difference was included as a covariate in the current study but it did not have effect on the outcome variables. However, status and other forms of power (e.g., social capital; Aquino & Lamertz, 2004) may affect behavioral, affective and attitudinal responses to aggression, including variables not measured in the current study. For example, research on abusive supervision has demonstrated that subordinates may not reciprocate abusive supervisor behaviors towards their supervisor, but may redirect that aggression towards less powerful targets or by engaging in deviance towards the organization (Tepper, 2007). 100 Further, although the approach-avoidance social motives assessed in the current study did not tend to affect engagement in help or harm within relationships, additional research has highlighted other possible social motives that may influence help and harm engagement. For example, Rioux and Penner (2001) identified prosocial values, organizational concern, and impression management as three motivational variables that are influential in OCB engagement. It could be that prosocial, impression management and organizational concern motives could influence the exchange of interpersonal helping and harming differently than the motivational variables assessed in the current study. For example, individuals who have high prosocial motives who concerned about the well-being of others may be more likely to reciprocate help with help, but also be more likely to reciprocate harm with helping behaviors, and less likely to reciprocate harm with harm. Research on workplace aggression also highlights additional motivational variables that could influence engagement in harm within dyadic relationships. For example, individuals motivated to maintain personal esteem may aggress against those who threaten that esteem. Lam, Van der Vegt, Walter, and Huang (2011) demonstrated that esteem threat may explain why high performers can experience higher levels of victimization from coworkers relative to poorer performers. Within the context of a dyadic relationship, an esteem-threatening coworker may motivate higher levels of harming behaviors. Future research can examine such a possibility. Another potential area for future research could be relationship maintenance and reconciliation strategies that follow harm. One possible explanation of the result in the current study where receipt of harm was related to engagement in help could be that help engagement is a part of reconciliation strategy. Although, promotion and prevention regulatory foci did not explain how receipt of harm related to help engagement, previous research has examined how 101 self-regulation can influence relationship maintenance and forgiveness processes. For example, Molden and Finkel (2010) demonstrated that both promotion and prevention regulatory foci related to forgiveness following a relationship offense. They argued that promotion focus motivates forgiveness through perceived benefits to repairing the relationship (e.g., perceptions that the individual will provide future benefits) and prevention focus motivates repair through perceived costs of further relationship deterioration. Although types of harming behavior were not compared in the current study, harming behaviors that symbolize possible social gains (e.g., passive exclusion) versus harming behaviors that symbolize social losses (e.g., active rejection) may differentially affect how promotion and prevention regulatory foci relate to forgiveness and reconciliation behaviors (Molden et al., 2009). Therefore, future research will benefit from differentiating between different types of harming behaviors in order to more clearly delineate how motives for reconciliation can offset the exchange of harmful behaviors. 102 APPENDICES 103 APPENDIX A Study Protocol 104 Recruitment Email Greetings, I am a researcher at Michigan State University who is conducting a research project. The study is titled “Goals and Relationships” and the purpose of this study is to examine the role of personal goals in the quality of coworker relationships. The study will last five weeks, with the projected start date of end of January (exact date TBD). I would like to invite you to participate in this interesting study. In return for your participation, you will be paid $40 ($10 for each week) and entered into a raffle for an additional four prizes of $60. You will have the opportunity to increase your odds of winning. Currently there are 15 positions available; sign-up is done on a first-come-firstserved basis. The requirements for participants are the following: 1. Email me (see instructions below) to register as a participant for the study. 2. Read the Question and Answer sheet to address any of your questions about the study. Please feel free to email or call me with additional questions. 3. Complete the pre-survey you will receive through email. You will receive this survey one week before you commence the daily surveys (for four weeks). 4. One week after the pre-survey, for four weeks (20 working days), one time each day after work hours you will complete a brief (about 5-10 minutes) survey online. You will receive an email each day reminding you to complete the survey. In return for your participation, you will be paid $40. Further, you will also be entered into a random drawing for one of two prizes of $60. Each daily survey completed will be an entry into the lottery, such that if all 20 daily surveys are completed, you will receive 20 entries into the lottery. If you are willing to participate, please email Brent Lyons at lyonsbr3@msu.edu indicating your name and email address. Brent will then contact you by email to provide instructions for participating in the study. If you prefer not to participate, you can simply not reply to this message. Please also email or phone Brent if you have any questions about the study. Thank you for considering my request. Brent John Lyons, M.A. Department of Psychology Michigan State University East Lansing, MI 48824 | Phone: (517) 281-4518 | Email: lyonsbr3@msu.edu 105 Consent Form This study is designed to investigate how relationship goals impact workplace interpersonal experiences. Interpersonal relationships are an important aspect of employees' work environment, but researchers currently have very little understanding of how relationship goals influence how people experience interpersonal interactions. Research that investigates this topic will help build knowledge that will aid understanding of interpersonal aspects of the workplace. If you choose to participate in this study, you will be asked to fill out a series of brief surveys during a 5 week period. You will be asked to spend approximately 5-10 minutes each afternoon. If you choose to participate in this study, you authorize the researchers to have access the questionnaires that you complete. If you agree to participate in this study you will be compensated $40 and entered into a random drawing for one of two prizes of $60. Each daily survey completed will be an entry into the lottery, such that if all 20 daily surveys are completed, you will receive 20 entries into the lottery. Some of the questions in the survey address sensitive topics. We recommend you complete the surveys in a private place. Your participation in this research is completely voluntary. You are free to terminate your participation at any time without penalty. Your participation in this study will be kept confidential to the maximum extent allowable by law. Your data will be included in a summary report along with the data from others. The report will not include any information that will allow anyone to identify any of your individual responses. If participants have any questions in regards to this study, they may contact Dr. Ryan, Department of Psychology, 333 Psychology Building, Michigan State University, East Lansing, MI 48824, Phone: (517) 353-8855, E-mail: ryanan@msu.edu. If you have questions or concerns about your role and rights as a research participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University Human Research Protection Program at (517) 3552180, Fax (517) 432-4503, or e-mail irb@msu.edu. Also feel free to send them a letter at 207 Olds Hall, MSU, East Lansing, MI 48824. 106 Information Sheet “MSU Research - Goals and Relationships” – Questions and Answers What is the purpose of this study? The purpose of this study is to examine how personal goals employees have with certain coworkers impacts how they interact with those coworkers, and in turn, impacts important organizational outcomes such as job satisfaction and intentions to leave the job. What will I be doing as a participant? You will be completing one pre-survey followed by 20 daily surveys over a period of five weeks (one survey a day for 20 work days). Each survey will take 5 to 10 minutes to complete except the pre-survey and last survey will be 10 to 15 minutes long. One survey will be emailed to you at the end of each work day. You can complete the survey after your shift has been completed. The survey will ask questions about the types of interactions you had with your coworkers and the emotions you experienced that day. The first day of the survey will be TBD and end on TBD. Who will be able to see my data? Your responses will be completely confidential. Data will be presented in aggregate form and your data will not be identifiable. Only Brent Lyons will have access to your data. Your responses to the surveys will have no implication for your standing in your job. What happens if I miss a day of work and can’t complete the survey for that day (e.g., I am sick and do not show up to work, I go on vacation…)? Only complete the survey for the day the survey is emailed to you. If you miss a day of work do not complete a survey for that missed day. You can continue participating in the study on the days you return to work. How will I be compensated? How do I win the cash prize? You will be paid $40 ($10 for each week). There are also two prizes of $60 each. Every day you complete a survey your name is entered into the draw for one prize. So the more days you complete the survey, the greater chance you have of winning. Prize drawings will occur in the beginning of TBD. Will I have access to the results? Yes. Brent Lyons will distribute an executive summary of aggregate findings among all participants. How can I contact Brent if I am having technical difficulties or have questions about my participation? You can call or email Brent at any time. Phone: 517-281-4518; Email: brent.j.lyons@gmail.com Thank you for your participation in my research! 107 APPENDIX B Study measures 108 Motivational disposition (Carver & White, 1994) BIS/Avoidance 1. If I think something unpleasant is going to happen I usually get pretty "worked up" 2. I worry about making mistakes 3. Criticism or scolding hurts me quite a bit 4. I feel pretty worried or upset when I think or know somebody is angry at me 5. Even if something bad is about to happen to me, I rarely experience fear or nervousness ® 6. I feel worried when I think I have done poorly at something 7. I have very few fears compared to my friends ® BAS/Approach 8. When I get something I want, I feel excited and energized 9. When I'm doing well at something, I love to keep at it 10. When good things happen to me, it affects me strongly 11. It would excite me to win a contest 12. When I see an opportunity for something I like, I get excited right away 13. When I want something, I usually go all-out to get it 14. I go out of my way to get things I want 15. If I see a chance to get something I want, I move on it right away 16. When I go after something I use a "no holds barred" approach 17. 1 will often do things for no other reason than that they might be fun 18. I crave excitement and new sensations 109 Emotional intelligence (Wong & Law, 2002) The Wong Law Emotional Intelligence Scale (WLEIS) Self-Emotions Appraisal (SEA) 1. I have a good sense of why I have certain feelings most of the time. 2. I have good understanding of my own emotions. 3. I really understand what I feel. 4. I always know whether or not I am happy. Others-Emotions Appraisal (OEA) 5. I always know my friends’ emotions from their behavior. 6. I am a good observer of others’ emotions. 7. I am sensitive to the feelings and emotions of others. 8. I have good understanding of the emotions of people around me. Use of Emotion (UOE) 9. I always set goals for myself and then try my best to achieve them. 10. I always tell myself I am a competent person. 11. I am a self-motivating person. 12. I would always encourage myself to try my best. Regulation of Emotion (ROE) 13. I am able to control my temper so that I can handle difficulties rationally. 14. I am quite capable of controlling my own emotions. 15. I can always calm down quickly when I am very angry. 16. I have good control of my own emotions. 110 Demographics 1. What is your gender? Male ____ Female ____ 2. What is your age (in years)? _______ 3a. Is your ethnicity Hispanic/Latino of any race? Yes___ No___ 3b. Please select your race/ethnicity (choose one or more) from the following groups. a. Aboriginal/Native Canadian b. Asian c. Black or of African descent d. White e. Two or more races 4. How long have you been working at the organization that currently employs you? _____years ____months 111 Relationship Motives (adapted from Elliot, Gable, & Mapes, 2006). Instructions: You will report on your interactions with two coworkers throughout the duration of this study. Carefully read the descriptions below in order to determine which two coworkers you will report about. You will select two coworkers whom you work with on a daily basis and have at least several interactions with throughout the day, and who are NOT of a different rank (i.e., do not choose your supervisor) than you in your organization. Even though you will be asked to think of two persons, you will be asked to provide an alias (or alternative name) for these two individuals in order to protect their privacy. Make sure you provide an alias/alternative name that you will remember for each person as you’ll need to remember each person throughout the duration of the study (e.g., if your coworker, John, has red hair, his alias could be Red Hair). In order to maintain confidentiality, it is important that you do NOT tell the coworkers you are rating them. Neither the coworkers you rate nor your organization will see your ratings, so please respond as honestly as possible. Coworker A: This coworker is someone with whom who you try to deepen, grow and develop your relationship with by sharing fun and meaningful experiences. Coworker A alias: John Smith Coworker B: This coworker is someone whom you try to make sure that nothing bad happens in your relationship by avoiding conflicts and situations that could cause harm to your relationship. You try to avoid getting embarrassed, betrayed, or hurt in your relationship with this coworker. Coworker B alias: Jane Doe 112 Liking How much do you like coworker A?* 1. 2. 3. 4. 5. I don’t like coworker A at all I dislike coworker A a little bit I neither like or dislike coworker A I like coworker A a little bit I like coworker A very much *Note: An identical scale was used to asses liking for Coworker B . 113 Trust (adapted from McAllister, 1995) Affect-based Trust 1. 2. 3. 4. 5. I freely share my ideas and feelings with Coworker A I can talk freely to Coworker A about difficulties I am having at work With Coworker A, we would both feel a sense of loss if one of us is transferred Coworker A responds caringly when I share my problems Coworker A and I have invested a lot in our working relationship Cognition-based Trust 6. Coworker A approaches his/her job with dedication 7. I see no reason to doubt Coworker A's competence for the job 8. I can rely on Coworker A not to make my job more difficult 9. Most people trust and respect Coworker A as a coworker 10. My peers consider Coworker A to be trustworthy 11. If people knew more about Coworker A, they would be more concerned and monitor his/her performance more closely (reverse coded) *Note: an identical scale was used to assess trust in Coworker B. 114 Characteristics of Coworkers 1. Is Coworker A a higher status than you in the organization? YES NO 2. How long have you been working with Coworker A? _____years ____months 3. Are you required to work with Coworker A to complete your work tasks? YES NO *Note: identical items were to assess characteristics of Coworker B. 115 Receipt and Engagement of Help (Dalal et al., 2009) Receipt of Help 1. 2. 3. 4. 5. 6. Went out of his/her way to be nice to me Tried to help me Defended my opinion or suggestion Went out of his/her way to include me in a conversation Tried to be available to me Spoke highly about me to others Engagement in Help 1. 2. 3. 4. 5. 6. Went out of my way to be nice to this coworker Tried to help this coworker Defended this coworker’s opinion or suggestion Went out of my way to include this coworker in a conversation Tried to be available to this coworker Spoke highly about this coworker to others 116 Receipt and Engagement in Harm (Dalal et al., 2009; Spector, Fox, & Penny, 1996) Receipt of harm Passive 1. Ignored me 2. Tried to avoid interacting with me 3. Spoke poorly about me to others 4. Did something to make me look bad Active 5. Behaved in an unpleasant manner toward me 6. Tried to harm me 7. Criticized my opinion or suggestion 8. Insulted or made fun of me Engagement in harm Passive 1. Ignored this coworker 2. Tried to avoid interacting with this coworker 3. Spoke poorly about this coworker to others 4. Did something to make this coworker look bad Active 5. Behaved in an unpleasant manner toward this coworker 6. Tried to harm this coworker 7. Criticized this coworker’s opinion or suggestion 8. Insulted or made fun of this coworker 117 Relationship Satisfaction (Impett et al., 2010; Neff & Karney, 2009; Campbell, Simpson, Bowlby, & Kashy, 2005) 1. 2. 3. I am satisfied with this coworker today I am satisfied with my relationship with this coworker today I felt close with this coworker today 118 Positive/Negative Affective States (Watson et al., 1988; Tellegen et al., 1999) Positive Affective States 1. Happy 2. Joyful 3. Excited 4. At ease 5. Calm 6. Relaxed Negative Affective States 1. Discouraged 2. Downhearted 3. Sad 4. Distressed 5. Nervous 6. Jittery 7. Angry 8. Hostile 9. Irritated 10. Frustrated 119 Regulatory Focus (adapted from Lockwood, Jordan, & Kunda, 2002). Promotion focus 1. I imagined experiencing good things that I hope might happen in my relationship with this coworker 2. I was oriented towards achieving successes when interacting with this coworker 3. I thought about my hopes and aspirations for my relationship with this coworker Prevention focus 4. I imagined experiencing bad things that I fear might happen in my relationship with this coworker 5. I thought about how I can prevent failures when interacting with this coworker 6. I was oriented towards preventing losses when interacting with this coworker 120 Manipulation Check – Social Motivation (Elliot, Gable, & Mapes, 2006) Approach social motivation 1. I try to deepen my relationship with Coworker A 2. I try to move toward growth and development in my relationship with Coworker A 3. I am trying to enhance the bonding and intimacy in my relationship with Coworker A 4. I am trying to share many fun and meaningful experiences with Coworker A Avoidance social motivation 5. I try to avoid disagreements and conflicts with Coworker A 6. I try to stay away from situations that could harm my relationship with Coworker A 7. I try to avoid getting embarrassed, betrayed, or hurt by Coworker A 8. I try to make sure that nothing bad happens to my relationship with Coworker A 121 REFERENCES 122 REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park: Sage. Alderfer, C. P. (1972). Existence, relatedness, and growth: Human needs in organizational settings. New York: Academic Press. Amodio, D. M., Shah, J. Y., Sigelman, J., Brazy, P. C., & Harmon-Jones, E. (2004). Implicit regulatory focus associated with resting frontal cortical asymmetry. Journal of Experimental Social Psychology, 40, 225-232. Andersson, L. M., & Pearson, C. M. (1999). Tit for tat? The spiraling effect of incivility in the workplace. Academy of Management Review, 24, 452-471. Andersen, S. M., Reznik, I., & Manzella, L. M. (1996). Eliciting transient affect, motivation, and expectancies in transference: Significant-other representations and the self in social relations. Journal of Personality and Social Psychology, 71, 1108–1129. Aquino, K. & Lamertz, K. (2004). A relational model of workplace victimization: Social roles and patterns of victimization in dyadic relationships. Journal of Applied Psychology, 89, 1022-1034. Aquino, K. & Thau, S. (2009). Workplace victimization: Aggression from the target’s perspective. Annual Review of Psychology, 60, 717 – 741. Aquino, K., Tripp, T.M., Bies, R.J. (2006). Getting even or moving on? Power, procedural justice, and types of offense as predictors of revenge, forgiveness, reconciliation, and avoidance in organizations. Journal of Applied Psychology, 91, 653-658. Atkinson, J. W., Heyns, R. W., & Veroff, J. (1954). The effects of experimental arousal of the affiliation motive on thematic apperception. Journal of Abnormal and Social Psychology, 49, 405–410. Baldwin, M. W. (1992). Relational schemas and the processing of information. Psychological Bulletin, 112, 461–484. Baldwin, M. W., Carrell, S. E., & Lopez, D. F. (1990). Priming relationship schemas: My advisor and the Pope are watching me from the back of my mind. Journal of Experimental Social Psychology, 26, 435–454. Bacharach, S. & Bamberger, P. (2007). Organizational context and post-event distress: 9/11 and the New York City firefighters. Academy of Management Journal, 50, 849 – 868. Baron, R. A., & Neuman, J. H. (1996). Workplace violence and workplace aggression: Evidence on their relative frequency and potential causes. Aggressive Behavior, 23, 161-173. 123 Bauer, D.J., Preacher, K.J. & Gil, K.M. (2006). Conceptualizing and testing random indirect effects and moderated mediation in multilevel models: new procedures and recommendations. Psychological Methods, 11, 142-163. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as fundamental human motivation. Psychological Bulletin, 117, 497–529. Baumeister, R. F., Twenge, J. M., & Ciarocco, N. (2002). The inner world of rejection: Effects of social exclusion on emotion, cognition, and self-regulation. In J. P. Forgas & K. D. Williams (Eds.), The social self: Cognitive, interpersonal, and intergroup perspectives (pp. 161–174). New York: Psychology Press. Berkowitz, L. (2003). Affect, aggression and antisocial behavior. In R. J. Davidson, K. Scherer, & H. H. Goldsmith (Eds.), Handbook of affective sciences (pp. 804-823). New York: Oxford University Press. Berry, C. M., Carpenter, N. C., Barratt, C. L. (2012). Do other-reports of counterproductive work behavior provide an incremental contribution over self-reports? A meta-analytic comparison. Journal of Applied Psychology, 97, 613-636. Berry, C. M., Ones, D. S., & Sackett, P. R. (2007). Interpersonal deviance, organizational deviance, and their common correlates: A review and meta-analysis. Journal of Applied Psychology, 92, 410-424. Blau, P. M. (1964). Exchange and power in social life. New York: John Wiley & Sons. Bolger, N., & Zuckerman, A. (1995). A framework for studying personality in the stress process. Journal of Personality and Social Psychology, 69, 890–902. Borman, W. C., & Motowidlo, S. J. (1997). Task performance and contextual performance: The meaning for personnel selection research. Human Performance, 10, 99–109. Bowler, W. M., & Brass, D. J. (2006). Relational correlates of interpersonal citizenship behavior: A social network perspective. Journal of Applied Psychology, 91, 70-82. Bowling, N. A., & Beehr, T. A. (2006). Workplace harassment from the victim’s perspective: A theoretical model and meta-analysis. Journal of Applied Psychology, 91, 998-1012. Brief, A. P., & Motowidlo, S. J. (1986). Prosocial organizational behavior. Academy of Management Review, 11, 710-725. Brockner, J., & Higgins, E.T. (2001). Regulatory focus theory: Implications for the study of emotions at work. Organizational behavior and Human Decision Processes, 86, 35-66. Campbell, L., Simpson, J. A., Boldry, J. G., & Kashy, D. A. (2005). Perceptions of conflict and support in romantic relationships: The role of attachment anxiety. Journal of Personality and Social Psychology, 88, 510-531. Carver, C. S. (2004). Negative affects deriving from the behavioral approach system. Emotion, 4, 3-22. 124 Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319–333. Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. New York: Cambridge University Press. Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory of socioemotional selectivity. American Psychologist, 54, 165–181 Chen, S., Boucher, H.C., & Tapias, M.P. (2006). The relational self revealed: Integrative conceptualization and implications for interpersonal life. Psychological Bulletin, 132, 151-179. Chen, P.Y, & Spector, P.E. (1991). Negative affectivity as the underlying cause of correlations between stress and strain. Journal of Applied Psychology, 76, 398 – 407. Chiaburu, D. S., & Harrison, D. A. (2008). Do coworkers make the place? Conceptual synthesis and meta-analysis of lateral social influences in organizations. Journal of Applied Psychology, 93, 1082-1103. Chowdhury, S. & Endres, M. L. (2010). The impact of client variabilty in nursing on occupational strain and injury: The cross-level moderating effect of safety climate. Academy of Management Journal, 52, 182-198. Chua, R.Y.J., Ingham, P., & Morris, M.W. (2008). From the head to the heart: Locating cognition and affect-based trust in managers’ professional networks. Academy of Management Journal, 51, 436-452. Colquitt, J. A., LePine, J. A., Piccolo, R. F., Zapata, C. P., & Rich, B. L. (2012). Explaining the justice performance relationship: Trust as exchange deepener or trust as uncertainty reducer? Journal of Applied Psychology, 97, 1-15. Cooley, C. H. (1902). Human nature and social order. New York: Scribner. Cropanzano, R., & Mitchell, M. S. (2005). Social exchange theory: An interdisciplinary review. Journal of Management, 31, 874–900. Cropanzano, R., Paddock, J., Rupp, D. E., Bagger, J., & Baldwin, A. (2008). How regulatory focus impacts the process-by-outcome interaction for perceived fairness and emotions. Organizational Behavior and Human Decision Processes, 105, 36-51. Crowe, E., & Higgins, E. T. (1997). Regulatory focus and strategic inclinations: Promotion and prevention in decision making. Organizational Behavior and Human Decision Processes, 69, 117-132. Dalal, R. S. (2005). A meta-analysis of the relationship between organizational citizenship behavior and counterproductive work behavior. Journal of Applied Psychology, 90, 12411255. 125 Dalal, R.S., Lam, H., Weiss, H.M., Welch, E.R., & Hulin, C.L. (2009). A within person approach to work behavior and performance: Concurrent and lagged citizenship - counter productivity associations, and dynamic relationships with affect and overall job performance. Academy of Management Journal, 52, 1051-1066. Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: In search of an elusive construct. Journal of Personality and Social Psychology, 75, 989–1015. De Dreu, C. K. W. & Weingart, L.R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: a meta-analysis. Journal of Applied Psychology, 88, 741-749. Duffy, M.K., Ganster, D.C., Shaw, J.D., Johnson, J.L., & Pagon, M., 2006. The social context of undermining behavior at work. Organizational Behavior and Human Decision Processes, 101, 105-121. Elliot, A.J., & Church, M.A. (1997). A hierarchical model of approach and avoidance achievement motivation. Journal of Personality and Social Psychology, 72, 218232. Elliot, A. J., Gable, S. L., & Mapes, R. R. (2006). Approach and avoidance motivation in the social domain. Personality and Social Psychology Bulletin, 32, 378-391. Elliot, A. J., & Thrash, T. M. (2002). Approach–avoidance motivation in personality: Approach and avoidance temperaments and goals. Journal of Personality and Social Psychology, 82, 804–818. Elliot, A.J., & Thrash, T.M. (2010). Approach and avoidance temperament as basic dimensions of personality, Journal of Personality, 78, 865-906. Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broadenand-build theory of positive emotions. American Psychologist, 56, 218-226. Gable, S. L. (2006). Approach and avoidance social motives and goals. Journal of Personality, 71, 175-222. Gable, S. L., Reis, H. T., & Elliot, A. J. (2000). Behavioral activation and inhibition in everyday life. Journal of Personality and Social Psychology, 78, 1135-1149. Gallucci, M., & Perugini, M. (2003). Information seeking and reciprocity: A transformational analysis. European Journal of Social Psychology, 33, 473–495 Fehr, R. & Gelfand, M. J. (2012). The forgiving organization: A multilevel model of forgiveness at work. Academy of Management Review. 37, 664-688. Gonzagos, G.C., Campos, B., & Bradbury, T. (2007). Similarity, convergence, and relationship satisfaction in dating and married couples. Journal of Personality and Social Psychology, 94, 34-38. 126 Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161–178. Graen, G. B., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (Vol. 9, pp. 175208). Greenwich, CT: JAI Press. Gray, J. A. (1990). Brain systems that mediate both emotion and cognition. Cognition and Emotion, 4, 269–288. Hershcovis, M.S., & Barling, J. (2007). Towards A relational model of workplace aggression. In J. Langan-Fox, C.L. Cooper, & R. Klimoski (Eds.), Dysfunctional Workplace: Management Challenges and Symptoms, (p. 268-284). Cheltenham, UK: Edward Elgar Publishing Ltd. Hershcovis, M. S., Turner, N., Barling, J., Arnold, K. A., Dupre´, K. E., Inness, M., et al. (2007). Predicting workplace aggression: A meta-analysis. Journal of Applied Psychology, 92, 228–238. Higgins, E.T. (1997). Beyond Pleasure and Pain. American psychologist, 52, 1280-1300. Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. Advances in Experimental Social Psychology, 30, 1–46. Higgins, E. T., Shah, J., & Friedman, R. (1997). Emotional responses to goal attainment: Strength of regulatory focus as moderator. Journal of Personality and Social Psychology, 72, 515-525. Ilies, R., Scott, B. A., & Judge, T. A. (2006). The interactive effects of personal traits and experienced states on intraindividual patterns of citizenship behavior. Academy of Management Journal, 49, 561-575. Impett, E. A., & Gordon, A. M. (2010). Why do people sacrifice to approach rewards versus to avoid costs? Insights from attachment theory. Personal Relationships, 17, 299-315. [PDF] Impett, E. A., Gordon, A. M., Kogan, A., Oveis, C., Gable, S. L., & Keltner, D. (2010). Moving toward more perfect unions: Daily and long-term consequences of approach and avoidance goals in romantic relationships. Journal of Personality and Social Psychology, 99, 948-963. Impett, E., A., Strachman, A., Finkel, E. J., & Gable, S. L. (2008). Maintaining sexual desire in intimate relationships: The importance of approach goals. Journal of Personality and Social Psychology, 94, 808-823. Impett, E. A., Peplau, L. A., & Gable, S. L. (2005). Approach and avoidance sexual motivation: Implications for personal and interpersonal well-being. Personal Relationships, 12, 465– 482. Hollinger, R. C., & Clark, J. P. (1982). Formal and informal social controls of employee deviance. Sociological Quarterly, 23, 333–343. 127 Joseph, D. L., & Newman, D. A. (2010). Emotional intelligence: An integrative meta-analysis and cascading model. Journal of Applied Psychology, 95, 54-78. Judge, T. A., Scott, B. A., & Ilies, R. (2006). Hostility, job attitudes, and workplace deviance: Test of a multilevel model. Journal of Applied Psychology, 91, 126-138. Kahn, W. (1998). Relational systems at work. Research in Organizational Behavior, 20, 39-76 Kafetsios, K., Nezlek, J.B., & Vassiou, K. (2011). A Multilevel analysis of relationships between leaders’ and subordinates’ emotional intelligence and emotional outcomes. Journal of Applied Social Psychology, 41, 5, 119-1142. Kramer, R. M. (1993). Cooperation and organizational identification. In J. K. Murnighan (Ed.), Social psychology in organizations: Advances in theory and research (pp. 244–268). Englewood Cliffs, NJ: Prentice Hall. Krasikova, D., & LeBreton, J. M (2012). Just the two of us: Misalignment of theory and methods in examining dyadic phenomena. Journal of Applied Psychology, 97, 739-757. Krull, J. J., & MacKinnon, D. P. (1999). Multilevel mediation modeling in group-based intervention studies. Evaluation Review, 23: 418–444. Lanaj, K., Chang, C.-H., & Johnson, R. E. (in press). Regulatory focus and work-related outcomes: A meta-analysis. Psychological Bulletin. Lam, C., Van der Vegt, G. S., Walter, F. & Huang, X. (2011). Harming high performers: A social comparison perspective on interpersonal harming in work teams. Journal of Applied Psychology, 96, 588-601. Lawler, E.J. (2001). An affect theory of social exchange. American Journal of Sociology, 107, 321-52. Liu, D., Chen, X. P., & Yao, X. (2011). From autonomy to creativity: A multilevel investigation of the mediating role of harmonious passion. Journal of Applied Psychology, 96, 294309. Lockwood, P., Jordan, C., & Kunda, Z. (2002). Motivation by positive or negative role models: Regulatory focus determines who will best inspire us. Journal of Personality and Social Psychology, 83, 854-864. Lyons, B.J. & Scott, B.A. (2012). Integrating social exchange and emotion centered explanations for the receipt of help and harm: A social network approach. Organizational Behavior and Human Decision Processes, 117, 66-79. Lovelace, K., Shapiro, D. L., & Weingart, L. R. (2001). Maximizing cross-functional new product teams’ innovativeness and constraint adherence: A conflict communications perspective. Academy of Management Journal, 44, 779–783. 128 MacKinnon, D. P., Fritz, M. S., Williams, J., & Lockwood, C. M. (2007). Distribution of the product confidence limits for the indirect effect: Program PRODCLIN. Behavior Research Methods, 39, 384–389. Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20, 709-734. McAllister, D. J. (1995). Affect- and cognition-based trust as foundations for interpersoanl cooperation in organizations, Academy of Management Journal, 38, 24-59. Meeker, B. F. (1971). Decisions and exchange. American Sociological Review, 36, 485-495. Mehrabian, A. (1976). Questionnaire measures of affiliative tendency and sensitivity to rejection. Psychological Reports, 38, 199–209. Molden, D. M., & Finkel, E. J. (2010). Motivations for promotion and prevention and the role of trust and commitment in interpersonal forgiveness. Journal of Experimental Social Psychology, 46, 255-268. Molden, D. C., & Higgins, E. T. (2004) Categorization under uncertainty: Resolving vagueness and ambiguity with eager versus vigilant strategies. Social Cognition, 22, 248-277. Molden, D. C., Lee, A. Y., & Higgins, E. T. (2008). Motivations for promotion and prevention. In J. Shah & W. Gardner (Eds.) Handbook of motivation science (pp. 169-187). New York: Guilford Press. Molden, D. C., Lucas, G. M., Gardner, W. L., Dean, K., & Knowles, M. (2009). Motivations for prevention or promotion following social exclusion. Being rejected versus being ignored. Journal of Personality and Social Psychology, 96, 415-431. Neff, L. A. & Karney, B. R. (2009). Stress and reactivity to daily relationship experiences: How stress hinders adaptive processes in marriage. Journal of Personality and Social Psychology, 97, 435-450. Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775-802. Osyerman, D. Uskul, A.K., Yoder, N., Nesse, R.M., and Williams, D.R. (2007). Unfair treatment and self-regulatory focus. Journal of Experimental Social Psychology, 43, 505-512. Perugini, M., Gallucci, M., Presaghi, F., & Ercolani, A. P. (2003). The personal norm of reciprocity. European Journal of Personality, 17, 251–283. Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26, 513-563. Reis, H. T., & Downey, G. (1999). Social cognition in relationships: Building essential bridges between two literatures. Social Cognition, 17, 97–117. 129 Richards, D.A., & Schat, A. (2011). Attachment at (not to) work: Applying attachment theory to explain individual behavior in organizations. Journal of Applied Psychology, 96, 169182. Rioux, S. M., & Penner, L. A. (2001). The causes of organizational citizenship behavior: A motivational analysis. Journal of Applied Psychology, 86, 1306–1314. Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace behaviors: A multidimensional scaling study. Academy of Management Journal, 38, 555-572. Rodell, J. B., & Judge, T. A. (2009). Can “good” stressors spark “bad” behaviors?: The mediating role of emotions in links of challenge and hindrance stressors with citizenship and counterproductive behaviors. Journal of Applied Psychology, 94, 1438-1451. Sackett, P. R., & DeVore, C. J. (2001). Counterproductive behaviors at work. In N. Anderson, D. Ones, H. Sinangil, & C. Viswesvaran (Eds.), Handbook of industrial, work, and organizational psychology, vol. 1: 145–164. London: Sage. Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination, Cognition and Personality, 9, 185–211. Scholer, A. A., & Higgins, E. T. (2008). Distinguishing levels of approach and avoidance: An analysis using regulatory focus theory. In A. J. Elliot (Ed.), Handbook of approach and avoidance motivation (pp. 489–504). New York, NY: Psychology Press. Scholer, A. A., Zou, X., Fujita, K., Stroessner, S. J., & Higgins, E. T. (2010). When risk seeking becomes a motivational necessity. Journal of Personality and Social Psychology, 99, 215-231. Sluss, D.M., & Ashforth, B.E. (2007). Relational identity and identification: Defining ourselves through work relationships. Academy of Management Review, 32, 9-32. Scott, B. A., & Barnes, C. M. (2011). A multilevel field investigation of emotional labor, affect, work withdrawal, and gender. Academy of Management Journal, 54, 116-136. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological Methodology, (pp. 290-312). Washington DC: American Sociological Association. Spector, P. E., & Fox, S. (2002). An emotion-centered model of voluntary work behavior: Some parallels between counterproductive behavior and organizational citizenship behavior. Human Resource Management Review, 12, 269-292. Spector, P. E., Fox, S., Penney, L. M., Bruursema, K., Goh, A., & Kessler, S. (2006). The dimensionality of counterproductivity: Are all counterproductive behaviors created equal? Journal of Vocational Behavior, 68, 446-460. Strachman, A. & Gable, S.L. (2006). Approach and avoidance relationship commitment. Motivation and Emotion, 30, 117-126. 130 Stryker, S., & Statham, A. (1985). Symbolic interaction and role theory. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed.), vol. 1: 311–378. New York: Random House. Tedeschi, J. T., & Felson, R. B. (1994). Violence, aggression, and coercive actions. Washington, DC: American Psychological Association. Tepper, B. J. (2007). Abusive supervision in work organizations: Review, synthesis and research agenda. Journal of Management, 33, 261-289. Tett, R. P., & Guterman, H. A. (2000). Situation trait relevance, trait expression, and crosssituational consistency: Testing a principle of trait activation. Journal of Research in Personality, 34, 397–423. Thau, S., Aquino, K., & Poortvliet, P. M. (2007). Self-defeating behaviors in organizations: The relationship between thwarted belonging and interpersonal work behaviors. Journal of Applied Psychology, 92, 840–847. Tofighi, D. & MacKinnon, D. P. (2011). RMediation: An R package for mediation analysis confidence intervals. Behavior Research Methods, 43, 692-700. Tsai, W., Chen, C., & Liu, H. (2007). Test of a model linking employee positive moods and task performance. Journal of Applied Psychology, 92, 1570-1583. Weiss, H. M., & Cropanzano, R. (1996). An affective events approach to job satisfaction. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 18, pp. 174). Greenwich, CT: JAI Press. Winterheld, H. A., & Simpson, J. A. (2011). Seeking security or growth: A regulatory focus perspective of motivations in romantic relationships. Journal of Personality and Social Psychology, 101, 935-954. Wong, C. S., & Law, K. S. (2002). The effects of leader and follower emotional intelligence on performance and attitude: An exploratory study. The Leadership Quarterly, 13, 243–274. Wong, C-S., Wong, P., & Law, K. S. (2005). The interaction effect of emotional intelligence and emotional labor on job satisfaction: A test of Holland’s classification of occupations. In C. E. J. Hartel, W. J. Zerbe & N. M. Ashkanasy (Eds.), Emotions in organizational behavior(pp. 235–250). Mahwah, NJ: Erlbaum. Yakovleva, M., Reilly, R.R., & Werko, R. (2010). Why do we trust? Moving beyond individual to dyadic perceptions. Journal of Applied Psychology, 79-91. 131