VOICE RESILIENCE: UTILIZING ADEQUATE EXPLANATIONS TO MAINTAIN EXCHANGE RECIPROCITY By Danielle D. King A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology - Master of Arts 2015 ABSTRACT VOICE RESILIENCE: UTILIZING ADEQUATE EXPLANATIONS TO MAINTAIN EXCHANGE RECIPROCITY By Danielle D. King The importance of employee voice for organizational functioning and improvement makes understanding how to mitigate negative voice effects and maintain the voice process vital. This study explored the role explanations can play in assuaging negative effects following voice non-endorsement (lack of attention and resources allocated to the implementation of an idea). The current work is situated within a social exchange framework and uses the norm of reciprocity to demonstrate how reciprocal relationships can be altered yet maintained in the voice process. Explanation adequacy (specificity and sensitivity) is expected to influence voicer perceptions (efficacy and safety) and subsequent voice behavior. In addition, exploratory research examined the influence of locus of explanation attribution, absence of explanation, and monetary benefit exchange on voicer perceptions. To assess these hypotheses and research questions, 324 undergraduate students completed two questionnaires with a 3-5 day time lag between administrations. Analyses indicate voice safety is significantly predicted by explanation sensitivity and mediates the relation between sensitivity and subsequent voice, and voice efficacy and safety significantly predict presence and number of voice behaviors. Exploratory analyses indicate low resilience strengthens the relationship between sensitivity and safety, the absence of a response to voice has significant negative effects on efficacy, and efficacy offers significantly greater prediction of voice, as compared to safety. Copyright by DANIELLE D. KING 2015 I dedicate this thesis to my loving parents, Cyntrell Jones and Azemar King, my supportive stepparents, Lauren King and Cyril Jones, and my wonderful grandmother, Alma Vance. Thank you for giving me the necessary tools and serving as my source of motivation and strength. iv ACKNOWLEDGEMENTS I would like to acknowledge my peers, mentors, family members, and friends for making the completion of this thesis possible. I owe my deepest gratitude to my advisor, Dr. Ann Marie Ryan, who has provided me with the necessary training, encouragement, advice, and patience over the past three years. I greatly appreciate the opportunity to work with her and feel very fortunate to have such a wonderful role model to follow. I am tremendously grateful for the time and energy she has dedicated to my education and development thus far. I would also like to acknowledge my committee members, Dr. Richard P. DeShon and Dr. Daisy Chang, who set aside their valuable, limited time to participate in the assessment and development of this project. I consider myself extremely lucky to have received input and acceptance from these individuals. Lastly, I am ever thankful for my extraordinary family members and dearest friends who have inspired me, replenished my spirits in the toughest of times, and kept me sane. Thank you. v TABLE OF CONTENTS LIST OF TABLES ....................................................................................................................viii LIST OF FIGURES .................................................................................................................... ix INTRODUCTION .......................................................................................................................1 Employee Voice Behavior .................................................................................................5 Conceptualization of Voice......................................................................................5 Antecedents to Voice ...............................................................................................8 Leaders’ Role in the Voice Process ...................................................................................8 The Value of Voice Endorsement ............................................................................10 Voice Non-Endorsement and Potential Negative Effects ........................................11 Explanation Adequacy .......................................................................................................12 Voice Resilience ................................................................................................................14 The Social Exchange Norm of Reciprocity .......................................................................15 Hypotheses Development ..................................................................................................17 Exploratory Research .........................................................................................................21 METHOD ....................................................................................................................................25 Participants .........................................................................................................................25 Research Design.................................................................................................................26 Procedure ...........................................................................................................................27 Manipulation ......................................................................................................................28 Pilot Testing .......................................................................................................................28 Measures ............................................................................................................................29 Independent Variables .............................................................................................29 Mediation Variables .................................................................................................29 Dependent Variables ................................................................................................30 Control Variables .....................................................................................................30 RESULTS ....................................................................................................................................31 Initial Analyses ..................................................................................................................31 Hypotheses Analyses .........................................................................................................32 Hypotheses 1 and 2 ..................................................................................................32 Hypotheses 3a and 3b ..............................................................................................33 Hypotheses 4a and 4b ..............................................................................................33 Exploratory Analyses .........................................................................................................35 DISCUSSION ..............................................................................................................................38 Summary and Discussion of Results..................................................................................38 Theoretical and Practical Implications...............................................................................41 Recommendations for Future Research .............................................................................42 Limitations .........................................................................................................................44 vi Conclusion .........................................................................................................................46 APPENDICES .............................................................................................................................47 Appendix A: Consent Form ...............................................................................................48 Appendix B: Debriefing Form ...........................................................................................49 Appendix C: Manipulation Scenarios ................................................................................50 Appendix D: Measures ......................................................................................................55 Appendix E: Tables and Figures ........................................................................................57 REFERENCES ............................................................................................................................69 vii LIST OF TABLES Table 1: Means, Standard Deviations, Alphas Coefficients, and Correlations ..............................57 Table 2: Means and Standard Deviation of Focal Variables by Conditions ..................................58 Table 3: Hierarchical Linear Regression for the prediction of Voice and Voice Frequency by Efficacy and Safety ........................................................................................................................59 Table 4a: Model Coefficients for Voice Efficacy Mediation of the Relationship between Explanation Specificity and Voice.................................................................................................60 Table 4b: Model Coefficients for Voice Efficacy Mediation of the Relationship between Explanation Specificity and Voice Frequency...............................................................................61 Table 5a: Model Coefficients for Voice Safety Mediation of the Relationship between Explanation Sensitivity and Voice .................................................................................................62 Table 5b: Model Coefficients for Voice Safety Mediation of the Relationship between Explanation Sensitivity and Voice Frequency ...............................................................................63 Table 6a: Hierarchical linear Regression for the Prediction of Voice Efficacy by Resilience and Explanation Specificity ...........................................................................................................64 Table 6b: Hierarchical linear Regression for the Prediction of Voice Safety by Resilience and Explanation Sensitivity ...........................................................................................................65 Table 7: Relative Weights Analyses of the Prediction of Voice and Voice Frequency by Voice Efficacy and Safety .............................................................................................................66 viii LIST OF FIGURES Figure 1: Proposed Model ..............................................................................................................67 Figure 2: Interaction between Resilience and Sensitivity in the Prediction of Safety ...................68 ix INTRODUCTION Voice is the discretionary communication of work-related ideas and suggestions, with the intention to improve one’s organization (Morrison, 2011). The importance of voice in organizational functioning and improvement has been firmly established (e.g. Howard, 1995; LePine & Van Dyne, 1998; Morrison, 2011; Van Dyne, Cummings, & Parks 1995). Voice occurs when employees are unable to enact change without the permission or support of their leader (McClean, Burris, & Detert, 2013), which highlights the importance of managerial endorsement (attention and resources allocated to the implementation of an idea; Burris, 2012). However, at times, employee suggestions are not endorsed for various reasons (e.g. budget constraints or idea utility; Landau, 2009). Though research concerning “the dark side of voice” is sparse, both quantitative and qualitative studies have demonstrated that negative effects occur when endorsement is not granted (Landau, 2009). A quantitative study conducted by De Vries, Jehn, and Terwel (2012) demonstrated that, following voice non-endorsement, negative effects occurred (i.e. intragroup conflict) and employees interpreted their opportunity to voice as “pseudo voice” – voice opportunities given by managers who did not intend to consider employee input. Additionally, a qualitative study by Landau (2009) found significant negative effects, including decreased satisfaction with one’s leader and lowered intentions to voice in the future, when employee ideas were not endorsed. These findings, along with others, highlight the negative impact that nonendorsement can have on future voice attitudes and behaviors. Findings of negative voice effects highlight the need to uncover the causes and mitigating mechanisms of these effects. Landau (2009) took a first step in this direction by emphasizing the importance of effectively managing non-endorsement. She stated, “organizations need to train 1 supervisors how to communicate with employees whose ideas might not be feasible” (p. 12). In line with this, Morrison (2011) stated that, “perhaps one of the most important sources of cues about whether it is worthwhile and safe to voice is the behavior of one’s immediate supervisor” (p. 388) and that, “despite the growing body of work on the role of leader behavior…we still do not have a clear picture of exactly what it is that leaders do or do not do that shapes employee perceptions” (p. 391). Thus, the current thesis examines the influence of specific leader behaviors on voice attitudes and behaviors following voice non-endorsement. Some research highlights the potential for managerial behaviors to assuage negative effects following unfavorable outcomes. Offering adequate explanations has proven particularly effective in the context of unfavorable decisions (e.g. Bies, 1987; Bies & Shapiro, 1988; Gilliland & Beckstein, 1996), and the two criteria for explanation adequacy are specificity and sensitivity (Bies, Shapiro, & Cummings, 1988; Greenberg, 1993; Shapiro, Buttner, & Barry, 1994). These criteria are vital because mitigating effects have only been observed when the explanation was perceived as adequate (Folger, Rosenfield, & Robinson, 1983; Shaw, Wild, & Colquitt, 2003; Weiner, Folkes, Amirkhan, & Verette, 1987). Though the provision of an adequate explanation has been established as a useful and cost-effective tool, no research to date has examined its potential utility in the voice process. The current work applies explanation research to the organizational voice process, and situates this analysis within a social exchange framework. The social exchange process is guided by the norm of reciprocity (we owe others benefits due to the benefits we have gained from them; Gouldner, 1960). In the voice process, an employee provides suggestions, with the intention of helping the organization and themselves, and a leader endorses helpful ideas, in an attempt to improve the organization and reaffirm the employee’s contribution. The ideal voice 2 exchange process involves ideas exchanged for endorsement because these transactions help move each party toward her/his goals. Sharing an idea assists the leader with her/his goals (e.g. improving the organization), and endorsing an idea assists the voicer with his/her goals (e.g. helping the organization and self-promotion). However, in light of potential voice nonendorsement, the current thesis examines how this reciprocal relationship can be altered yet maintained. Within the current framework, voice is presented as the benefit given to one’s leader - due to its potential to assist with his/her goals - and an adequate explanation for voice nonendorsement is presented as a benefit returned - due to its potential to inform future behaviors and assuage interpersonal concerns. The current work is aligned with previous work, as social exchange theory is often used in the examination of voice effects, yet it aims to contribute a novel perspective by examining how to maintain exchange relationships in times of voice non-endorsement. Thus, the social exchange theory’s norm of reciprocity (Gouldner, 1960) is utilized in answering the question: Do adequate explanations mitigate the negative effects of voice non-endorsement and foster voice resilience (maintaining voice perceptions and continued voice behavior despite a previous unsuccessful attempt)? The current research connects and aims to contribute to social exchange, voice, and explanation literatures through the use of previously neglected perspectives and methods. Though social exchange theory is the framework most often used to understand the voice process, the current research takes a different perspective in examining how exchange reciprocity can be maintained in times of non-endorsement. This perspective may help uncover the psychological processes that mitigate negative voice effects, which is important because although we know that explanations mitigate negative reactions to decisions, we still do not 3 know what makes them work (Frey & Cobb, 2010). This work also aims to contribute a new perspective to explanation literature, as novel dependent variables will be assessed. Explanations have been linked to various justice perceptions (e.g. distributive and procedural; Bobocel & Farrell, 1996; Konovosky & Cropanzano, 1991). However, the current research highlights the importance of explanations beyond the justice domain by examining their potential influence on social exchange relationships and subsequent organizational citizenship behaviors (Van Dyne et al., 1995). Lastly, voice effects are examined in the current work using the previously segmented explanation and voice literatures. A meta-analysis of explanations (Shaw et al., 2003) demonstrated the value of adequately explaining many unfavorable outcomes (e.g. receiving a poor performance evaluation, firing an employee, an unwanted relocation), yet no studies examined explanations for voice decisions (e.g. non-endorsement). More specifically, the influence of explanation adequacy on voice attitudes and decision has not been previously explored. The current thesis examines these effects. The current thesis also seeks to contribute to four previously neglected areas of research. First, Chiaburu, Lorinkova, and Van Dyne (2013) asserted that, unlike other affiliative citizenship behaviors (e.g. helping), change-oriented citizenship behaviors (e.g. voice) have received less research attention. Second, although voice behavior is not a one-time experience and involves relational aspects (e.g. its dependence on managers’ endorsement; Landau, 2009; Withey & Cooper, 1989), few studies to date have investigated the role of responses to voice (Burris, 2012). Third, a recent review of voice research by Morrison (2011) proposed that both contextual (e.g. organizational culture) and individual (e.g. personality) factors influence voice efficacy and safety, yet there was no mention of experiential antecedents linked to these judgments. Fourth, previous studies often examine the aspects of explanations (e.g. specificity 4 and sensitivity) combined (Greenberg, 1993, 1994) or as interactions (Shapiro et al., 1994), yet few studies have heeded the suggestion of Bies and colleagues (1988) to assess whether specificity and sensitivity are independently associated with different subordinate thoughts and reactions. The current work seeks to ameliorate such disparities by focusing on change-oriented voice behavior, investigating the influence of responses to voice, examining leader behaviors as antecedents to voice efficacy and safety, and considering the effects of specificity and sensitivity on separate voice antecedents. This work begins with the conceptualization of and antecedents to voice behavior and continues with a discussion of the influential role leaders play in the voice process. Following this discussion, the value of voice endorsement and potential negative effects of voice nonendorsement are explained in greater detail. Next, the newly introduced role of explanation adequacy and voice resilience in the voice process is explained. The social exchange framework’s norm of reciprocity is then presented to provide the background of, and lead-in to, the development of current hypotheses. The proposed study methods, analyses, and results are then presented. Finally, a discussion of findings, implications, and conclusions is presented. Employee Voice Behavior Conceptualization of Voice. Hirschman (1970) presented the term voice to denote one of three potential responses to dissatisfaction (along with exit or loyalty). Voice has since been expanded to include more than just a response to dissatisfaction and is now considered, in essence, the discretionary communication of work related ideas, suggestions, or opinions with the intent to improve organizational functioning (Burris, Detert, & Chiaburu, 2008; Detert & Burris, 2007; Detert & Trevino, 2010; LePine & Van Dyne, 1998; Morrison, 2011; Van Dyne & LePine, 1998). Most of Hirschman’s work focused on clients of organizations deciding to share 5 or withhold voice, yet this framework has subsequently been applied to organizational behavior research. Since that time, a clearer picture of what constitutes voice behavior and different types of voice has emerged. In the nomological network of extra-role behaviors (behaviors that go beyond role expectations to benefit the organization) presented by Van Dyne et al. (1995), voice is placed in the challenging/promotive category. It is promotive in that it aims to cause action and it is challenging because its goal is to alter the status quo. Voice is also distinct from affiliative extra-role behaviors (e.g. helping) because it is associated with risks (e.g. potentially damaging relationships or reputations). Most recently, Maynes and Podsakoff (2013) conceptualized different types of voice along two intersecting dimensions: preservation vs. challenge oriented (voice that either preserves or challenges the status quo) and promotive vs. prohibitive (voice that either advocates for or stops something). This circumplex resulted in the four types of voice: supportive voice (preservation/promotive), constructive voice (challenge/promotive), defensive voice (preservation/prohibitive), and destructive voice (challenge/prohibitive). Along with the refinement of the voice construct over time, different research approaches and agendas have been enacted. Early research aimed to determine the individual difference antecedents to voice (e.g. personality and attitudes; LePine & Van Dyne, 2001; Withey & Cooper, 1989). This stream of research was followed by an examination of organizational factors and their influence on voice behavior (e.g. organizational culture and size; Edmonson, 2003; Milliken, Morrison, & Hewlin, 2003). Finally, recent work on employee voice involves identifying the psychological mechanisms that facilitate voice behavior (Liang, Farh, & Farh 2012; Ng & Feldman 2011). Contributing to the overall utility of voice research, Van Dyne and LePine (1998) conducted a longitudinal study that provided support for its convergent, 6 discriminant, and predictive validity across multiple rating sources, and also demonstrated a link between extra-role behavior and performance. In an effort to answer the call of researchers to begin focusing on the different types of voice to more clearly understand the distinct antecedents and outcomes of each (Morrison, 2011), this thesis focuses on constructive voice. Voice is here conceptualized as a changeoriented citizenship behavior (proactive behavior aimed at improving work processes and outcomes; Chiaburu et al., 2013). Though voice may be expressed to one’s leader, peers, or subordinates, the current thesis defines voice as the sharing of constructive ideas or suggestions with someone in the organization that holds the necessary power and resources to bring about the proposed change (for similar definitions see Detert & Burris, 2007; LePine & Van Dyne, 1998). Implicit within this definition are non-anonymous, constructive, and challenging elements. For example, suggesting a new solution to a current budget problem or sharing an innovative idea meant to improve employee efficiency are voice behaviors. Placing anonymous thoughts in a suggestion box, reporting unethical dealings to an outside source (whistle-blowing), or stating complaints without suggestions are not considered voice. Though driven by good-intentions, the risks associated with engaging in voice behavior often make employees uncomfortable with voicing (Dutton, Ashford, O’Neill, Hayes, & Wierba, 1997; Milliken et al., 2003). Many organizations view nonconforming behavior negatively (Nemeth & Staw, 1989), and voice could be misinterpreted as bossiness in an effort to undermine leadership (Tepper, Duffy, Hobbler, & Ensley, 2004). Argyris and Schon (1978) found that voice to leaders, that is intended to spark change, may be seen as a form of defiance and intentional challenging of authorities. Fears associated with voice range from existence losses (e.g. termination) to relatedness losses (e.g. humiliation; Alderfer, 1969; Maslow, 1943). 7 Discomfort with voicing may also be due to the potential direct costs (e.g. time and energy expended) or indirect costs (e.g. confronting people with power; Withey & Cooper, 1989). Although individual differences have been linked to voice behavior (e.g. proactive personality and self-esteem) employees usually engage in voice only after cognitively assessing the potential costs and benefits of their actions (Dutton et al., 1997; Kish-Gephart, Detert, Trevino, & Edmonson, 2009; Milliken et al., 2003). A closer examination of previously established antecedents to employee voice behaviors sheds light on the processes that lead to voice behavior. Antecedents to Voice. The two key antecedents to voice behavior are voice efficacy and voice safety, and research has shown that these judgments are influenced by both individual (e.g. personality) and contextual (e.g. organizational culture) factors. Voice is more likely to occur when individuals believe that improvement is likely (efficacy) and when the costs of voice seem low (safety; Withey & Cooper, 1989). Efficacy involves the assessment of whether desired changes are likely to follow from engaging in voice, and Withey and Cooper (1989) showed that people who believe improvement is possible will be more likely to voice. Safety is also a major concern because the act of voicing has the potential to damage interpersonal relationships if others in the organization disagree with or feel threatened by one’s suggestions (LePine & Van Dyne, 1998; Van Dyne & LePine, 1998). Consequently, employee voice behavior more often occurs when the anticipated benefits outweigh the potential costs, and leaders often play a significant role in this decision-making process. The role of leaders in establishing voice efficacy and safety will be discussed in the following sections. Leaders’ Role in the Voice Process Leaders influence subordinates indirectly due to their control over organizational policies, structures, and climate, and leaders also play an influential role in organizational voice 8 processes. Voice is often called “speaking up” because subordinates typically share ideas with those who have the necessary resources, power, or authority to make changes (Detert & Burris, 2007). However, since voicing to organizational leaders can be particularly risky for interpersonal relationships (Detert & Burris, 2007), most employees will not offer challenges to supervisors that have signaled disinterest in input from below (Hornstein, 1986). Research has also shown that the act of voicing can be facilitated by high quality leader-member relationships and certain leader characteristics (e.g. openness; Detert & Burris, 2007; Edmonson, 2003; Morrison & Phelps, 1999; Van Dyne, Kamdar, & Joireman, 2008; Walunbwa & Schaubroeck, 2009). The salient power differential between leaders and subordinates makes leaders an important part of the voice process and may influence whether employee voice is facilitated, enhanced, or silenced. Leaders are most often the targets at which voice is directed, and leaders generally decide whether employee voice is listened to and/or endorsed. Leaders first decide whether voice opportunities are made available and the way in which these processes are maintained. Implications from the work of Liang and colleagues (2013) on the psychological mechanisms that lead to voice behavior showed that, in order to maximize innovative suggestions and idea sharing, leaders can increase subordinates’ felt obligation for constructive change by emphasizing the value of voice as a way to give back to the organization and by making the voicing process a positive experience for employees. Leaders’ decision of whether ideas proposed are endorsed is an important element in the voice process because employees often do not feel it is appropriate to bypass their boss with suggestions (Detert & Edmonson, 2011). Thus, direct leaders may serve as the “gate-keepers” for employee input and subsequent implementation. Finally, many of the beneficial outcomes associated with voice are contingent 9 on leader responsiveness following voice behavior (Edmonson, 1999; Stamper & Van Dyne, 2001; Van Dyne et al., 1995). Overall, leaders influence employee perceptions of viable voice opportunities, voice instrumentality (whether voice has influence), and expected outcomes of voice behaviors. The Value of Voice Endorsement. As said by Senge (1990), “it is just not possible any longer to ‘figure it out’ from the top” (p. 4). Valuable information is gained when employees speak up (Glauser, 1984; Morrison & Milliken, 2003), and this competitive advantage tool has become increasingly important as the emphasis on flexibility, innovation, and improvement has grown (Howard, 1995). Employee voice plays a critical role in both individual and organizational outcomes. For the individual, endorsed voice behavior can lead to a sense of contributing to the organization, increased visibility, team learning, improved work processes, and the receipt of positive feedback (Edmonson, 1999; Stamper & Van Dyne, 2001; Van Dyne et al., 1995). Benefits to the organization include crisis prevention, higher levels of overall effectiveness, innovation, and improved adaptation to changes in the external environment (Nemeth, 1986; Nemeth & Staw, 1989; Van Dyne et al., 1995). Recent meta-analytic work also showed that voice is related to performance dimensions including in-role performance, creativity, and implementation of new ideas (Ng & Feldman, 2011). Important to note is that the act of voicing in itself does not bring about change - the utility of voice rests in action taken by those in power following their receipt of voice (McClean et al., 2013). The idea that the value of voice is linked to its impact on outcomes was initially suggested by Hirschman (1970). He argued that individuals judge the value of voice in organizations by whether it produces change. Such ideas are in line with the instrumental perspective (voicing is not enough, people must also perceive that their viewpoints are a part of 10 the decision-making process; Shapiro, 1993). This perspective argues that outcomes will only be perceived as fair when individuals perceive that they had influence on the decision-making process (Thibault & Walker, 1975). Contrary to this is the value-expressive model, which posits that, regardless of its influence on the outcome, the importance of voice rests in the perception that one has the opportunity to express his/her voice (McFarlin & Sweeney, 1996). Interestingly, both the value-expressive and instrumental perspectives have been supported (Lind & Tyler, 1988). However, empirical findings of negative voice effects show that not all voice opportunities and behaviors lead to positive outcomes. Voice Non-Endorsement and Potential Negative Effects. Though sparse attention has been paid to the potential that providing voice opportunities leads to negative outcomes, there have been both quantitative and qualitative findings in support of this claim. Research has supported the existence of a frustration effect (when participants are frustrated that their input did not affect the decision made; Folger, 1977; Harlos, 2001), and this negative voice effect is shown to lower fairness judgments (Folger, Rosenfield, Grove & Corkran, 1979; Lind & Lissak, 1985). For example, a study conducted by Avery and Quinones (2002) found a significant interaction effect between voice behavior and voice instrumentality, such that voice behavior had a negative effect on perceived fairness when instrumentality was low and voice behavior had no effect on perceived fairness when instrumentality was high. Later, a qualitative study conducted by Landau (2009) compared the experiences of three groups of employees (those whose voice was endorsed, those whose voice attempt was unsuccessful, and those who chose not to voice). The outcome variables included procedural justice, voice propensity (future intent to voice), and ratings of supervisors as voice managers. Results showed that those who spoke up and were not endorsed perceived the lowest presence of procedural justice and had the least favorable attitudes 11 toward their supervisor as a voice manager. This group of unsuccessful voicers also expressed the lowest intentions to voice in the future. Most recently, a quantitative study conducted by De Vries and colleagues (2012) demonstrated that negative effects (reduced future voice behavior and intra-group conflict) occurred when employees’ suggestions were not endorsed and voice opportunities were assumed to be “pseudo voice” – voice opportunity given by managers who never intended to consider employee input. Feelings of experienced deceit and unfairness following perceived pseudo voice opportunities led to decreased future voice behaviors and increased group conflict. This work demonstrated that whether offering voice opportunities led to positive or negative effects was dependent upon how employees perceived the motives of their manager in providing these opportunities. These findings, along with others, highlight the impact that voice endorsement can have on future voice attitudes and behaviors. Unfortunately, a reality within the voice process is that, at times, managers must withhold endorsement of employee voice for various reasons (e.g. time/budget constraints, ideas given that have been previously attempted, or suggestions that would likely lead to negative outcomes for the organization; Landau, 2009). Subordinates are not often privy to all information that managers have access to, and this may limit subordinates’ ability to objectively assess the feasibility and potential outcomes of their suggested ideas. Thus, findings concerning negative effects following voice non-endorsement do not imply that managers should endorse all suggestions – such an implication would be highly difficult to uphold and unreasonable to expect. However, the demonstrated importance of leader behaviors and consideration of employee voice highlight the important role of managerial decision communication in times of voice non-endorsement. Explanation Adequacy 12 In most times of delivering bad news, organizational agents fail to give explanations or they provide very vague explanations due to concerns about emotional distress or potential litigation (Folger & Skarlicki, 2001; Smeltzer & Zener, 1992). However, research shows that explanations lead to enhanced fairness perceptions (Bies & Shapiro, 1987, 1988), and subsequently reduce both employee turnover (Brockner, DeWitt, Grover, & Reed, 1990) and destructive behavior (e.g. theft; Greenberg, 1990). A meta-analytic review of the effects of explanations by Shaw and colleagues (2003) demonstrated the significant, positive effect of explanations on employee attitudes. Results showed that employees were 43% less likely to retaliate after an unfavorable decision if an adequate explanation was provided. Bies et al. (1988) examined the effect of explanations, following the refusal of a subordinate’s request, on anger and conflict inducing behaviors. Results indicated that it was not the mere provision of an explanation that influenced outcomes, but the adequacy of the explanation that explained a significant portion of variance in subordinate reactions. The mitigating effects of explanations are contingent on their perceived adequacy (specificity and sensitivity; Bies et al., 1988; Shaw et al., 2003). To elucidate this claim, Shapiro and colleagues (1994) examined the factors that enhance the perceived adequacy of explanations. Results indicated that adequate explanations are those in which the reasons or criteria for making a decision were specific, rather than vague, and the explainer exhibited sensitivity and concern for the receiver. Specificity involves whether an explanation is tailored to the individual and provides detailed information about what lead to the decision (Ployhart, Ryan, & Bennett, 1999; Shapiro et al., 1994). Sensitivity refers to the manner in which the explanation is conveyed and highlights whether the communicator exhibits sincere concern for the recipient (Ployhart et al., 1999; Shapiro et al., 1994). As an adequate explanation is a useful tool in the communication of 13 unfavorable outcomes, it may also assuage the negative effects that follow voice nonendorsement and foster voice resilience. Voice Resilience Resilience is the “capacity to rebound or bounce back from adversity, conflict, [or] failure” (Luthans, 2002; p. 702). Although resilience has been traditionally portrayed as trait-like and relatively fixed (Block, 1963), there is increasing evidence that it is developable (Bonanno, 2004; Masten, 1994, 2001; Masten & Reed, 2002; Youssef & Luthans, 2005). Empirical evidence shows that there are multiple methods for building resilience, like facilitating positive emotions (Tugade & Fredrickson, 2004), altering the levels of risk and personal assets (Masten, 2001), and fostering self-enhancement (Greenwald, 1980). Resilience may be useful in the voice process because, although an individual’s suggestion is not endorsed, future voice attempts may successfully contribute to the organization. For example, imagine that a medical student unknowingly suggests a change to a surgeon that would likely harm a patient. Although the surgeon will not endorse the suggestion, it would be unfounded for the surgeon to then assume that the student is incapable of offering helpful suggestions. Additionally, a response to the student that shows disinterest in future suggestions and contempt may be harmful to both the student and the leader, because this may decrease the student’s likelihood of sharing future potentially helpful suggestions. Thus, in line with the current work’s hypotheses, it may be most beneficial for the surgeon to give a specific and sensitively delivered explanation, which would facilitating learning on the part of the student and assuage interpersonal tension. Overall, leaders play a significant role in the voice process and engaging in future voice behaviors despite the adversity of non-endorsement (voice resilience) may benefit both individuals and others in the organization. The process of maintaining engagement in voice behavior despite a previous 14 unsuccessful attempt is termed here as voice resilience, and the ways this may be accomplished are detailed in the following sections. The Social Exchange Norm of Reciprocity A social exchange is a two-sided, mutually contingent, and mutually rewarding process. This process involves “transactions” or “exchanges,” and is limited to actions that are contingent on rewarding reactions from the other party involved (Blau, 1964). Thus, a social exchange emerges as actions generate obligations on the part of others (Emerson, 1976). According to Blau (1964), the principle underlying social exchange is that by supplying a reward or service to another, the receiving party is then obligated to return a benefit. A social exchange is distinguished from an economic exchange by its unspecified obligations, returns that cannot be bargained, and its ability to engender feelings of personal obligation, gratitude, and trust (Blau, 1964). One of the major tenets of social exchange theory is that, in order to develop trust and commitment to social relationships, parties must abide by certain rules of exchange (Cropanzano & Mitchell, 2005). The rules and norms of exchange form the guidelines for the exchange process. One of the best-known exchange rules is reciprocal interdependence (Gouldner, 1960). This rule emphasizes a contingent, interpersonal process whereby an action by one party leads to a response by another. A reciprocal exchange is characterized by the absence of explicit bargaining, yet one party’s actions are contingent on the other’s behavior (Molm, 2003). In the enactment of reciprocal interdependence, if one party supplies a benefit the receiving party would respond in kind (Gergen, 1969). This process is said to begin as “one participant makes a ‘move,’ and if the other reciprocates, new rounds of exchange initiate” (Cropanzano & Mitchell, 2005, p.876) 15 Reciprocity is defined as “a mutually gratifying pattern of exchanging goods and services,” governed by the “moral norm: You should give benefits to those who give you benefits” (Gouldner, 1960, p. 170). Reciprocity connotes that each party involved in a social exchange has rights to receive benefits after benefits have been given, and duties to return benefits once they have been received. In addition, the norm of reciprocity is a kind of obligation based on prior experiences and actions. This norm asserts that people should help those who have helped them and those whom you have helped have an obligation to return assistance. Overall, it is said that when one party assists another, an obligation is generated and is present until some benefit is returned – referred to by Gouldner as “the shadow of indebtedness” (p.174). Parsons (1951) states that reciprocity is “inherent in the nature of social interaction” (p. 445). Importance has been attributed to the norm of reciprocity as a vital element of maintaining social equilibrium and a governing rule for interactions. The principal of reciprocity is said to be a foundational basis on which much of primitive social and ethical life rests (Gouldner, 1960; Thurnwald, 1932). Simmel (1950) further asserts that social equilibrium is made possible by “the reciprocity of service and return service” and that “contacts among men rest on the schema of giving and returning the equivalence” (p. 387). In sum, scholars attest that “to require a benefit, or to be grateful to him who bestows it, is probably everywhere, at least under certain circumstances, regarded as a duty” (Westermarck, 1908, p.154). The norm of reciprocity is an important framework for the voice process. As Gouldner (1960) purports, it is morally improper to break off relations with someone to whom you are still indebted. Thus, it is likely to cause rifts in the voice process if employee voice behavior does not garner an appropriate response. Gouldner (1960) further asserted that reciprocity involves two obligations: helping the person who has helped you and not injuring that person. Adequate 16 explanations address the two needs of reciprocity as they entail sharing specific information and delivering one’s message in a sensitive manner. As a social exchange is characterized by unspecified obligations (Blau, 1964) and is governed by the norm of reciprocity (Gouldner, 1960), the current work hypothesizes that the extra-role sharing of specific and sensitive explanations is an adequate reward to the voicer such that, despite an unsuccessful voice attempt, they will maintain positive perceptions and continue investing effort into the extra-role voice process (See Figure 1). Hypotheses Development Voice efficacy and voice safety are integral judgments in the voice process because a lowering of these perceptions leads employees to believe that their voice behavior will not foster change and that their suggestions will lead to negative outcomes. However, leaders’ communication of specific explanations, especially in times of non-endorsement, may reassure voicers that their suggestions are taken seriously (voice efficacy), and an interpersonally sensitive presentation may demonstrate that they will not incur additional costs (voice safety). In line with this, Edmonson (2003) found that leader behaviors that indicate openness to employee input may decrease the salience of the hierarchical relationship with subordinates and lead to fewer perceived potential risks and costs for subordinates in deciding whether to speak up. Bad news victims typically want to understand the causes and/or reasons behind a negative event (e.g. Louis, 1980; Wong & Weiner, 1981), and researchers have shown that feelings of anger and resentment are significantly reduced when adequate explanations are presented (Folger et al., 1983; Weiner et al., 1987). As self-efficacy is the extent to which a person believes that they are capable of successfully performing a specific behavior (Prussia & Kinicki, 1996), more specific insight into the causes of non-endorsement may help individuals 17 learn from and improve upon previous voice behaviors. Thus, increased explanation specificity may lead to higher voice efficacy, compared to vague explanations. The way decisions are explained also affects the recipient’s perception of the decision process and decision maker (for review see Whitener, Brodt, Korsgaard, & Werner, 1998). Factors such as offender’s perceived repentance and boss’ sincerity are key factors in mitigating negative reactions (Blumstein, 1974; Rubin, Brockner, Eckenrode, Enright, & Johnson-George, 1980). Greater sensitivity in non-endorsement explanations may foster trust between the voicer and his/her leader, as work has shown that demonstrating personal interest, listening to subordinate input, and taking appropriate action following input reduces perceived risks in honest communication (Bass & Riggio, 2006; Edmoson, 2003). Explanation adequacy has also been linked to perceived trustworthiness and fairness (Bies et al., 1988). Though previous research concerning the unique effects of specificity and sensitivity is sparse, previous findings were used to distinguish these potential effects. The work of Frey and Cobb (2010) demonstrated that there was no direct effect between explanation specificity and interactional perceptions. In this study, the authors noted that, “judgments of how we are treated are formed independently of the specificity of information provided in messages given to us” (p. 1226). Thus, the provision of information (specificity) is hypothesized to influence attitudes concerning voice efficacy, while the interpersonal component of adequacy (sensitivity) is here linked to safety perceptions. Therefore, it is hypothesized that: Hypothesis 1: Increased explanation specificity will lead to higher subsequent voice efficacy. Hypothesis 2: Increased explanation sensitivity will lead to higher subsequent voice safety. 18 The link between voice efficacy and safety with voice is well established (Milliken et al., 2003; Morrison, 2011; Withey & Cooper, 1989), and is consistent with motivation theories that link expectancy beliefs and effort (Vroom, 1964). Indeed, terms such as “quiescence silence” and “defensive silence” have been used to refer to situations in which concerns about negative interpersonal repercussions and futility inhibit voice behavior (Pinder & Harlos, 2001; Van Dyne, Ang, & Botero, 2003). Research on constructive voice has also shown that withholding of suggestions for improvement may be due to fears of posing a threat to authority figures or reluctance to request a change in the practices or behaviors of leaders (Detert & Burris, 2007; Kish-Gepert et al., 2009; Milliken et al., 2003). The assumption that one will be unsuccessful or experience negative consequences after voicing, understandably, will lead to lower subsequent behavior. Therefore, the following is hypothesized: Hypothesis 3a: Higher voice efficacy will lead to higher subsequent voice. Hypothesis 3b: Higher voice safety will lead to higher subsequent voice. Parallels between in-role explanation research (e.g. performance rating explanations) and extra-role feedback (e.g. responses to voice) may supplement the current hypotheses. For example, performance feedback literature shows that those who receive specific, rather than vague, explanations tend to perform better and are more satisfied with the appraisal process (Hammer & Hammer, 1976; Henderson, 1984; Rice, 1987). This highlights the potential positive impact of adequate explanations on overall perceptions of and future responses to an organizational process. Further, linkages between explanations and future voice have been alluded to in previous research. For example, Cheung (2013) posited that, “employee[s] may use the provision of timely information and adequate explanations from the supervisor to make appropriate 19 suggestions on improving the effective functioning of the organization” (p. 557). Kouzes and Posner (1987) also argued that, "without information, you can be certain that people will not extend themselves to take responsibility or vent their creative energies" (p. 157), and Kanter (1983) stated that access to information helps people to feel capable of taking initiative. Additionally, empirical work has suggested that sharing specific information about voice decisions may be a valuable input for subordinate psychological processes and future voice. De Vries and colleagues (2012), in their work concerning pseudo voice, implied that employees most often perceive pseudo voice when leaders do not provide information about how their input was used, and this perception leads to lowered voice. Fuller, Marler and Hester (2006) also examined access to company’s strategy information and found that such access fostered individual felt responsibility for constructive change and future voice. Results implied that employees often need information that enables them to align their behaviors with organizational objectives before they engage in voice. In total, adequate explanations may help foster voice resilience (engaging in future voice behaviors despite a previous unsuccessful attempt), because offering adequate explanations helps restore equilibrium to social structures and social relationships (Goffman, 1971; Scott & Lyman, 1968; Stokes & Hewitt, 1976). One example of this is the work of Gilliland and Beckstein (1996) which demonstrated that explanation adequacy predicted receiver perceptions and future intentions to invest in the system to which they were previously denied (i.e. submitting work to a journal following a rejection). The aforementioned linkages lead to the following hypotheses: Hypothesis 4a: The relationship between explanation specificity and subsequent voice will be mediated by voice efficacy. 20 Hypothesis 4b: The relationship between explanation sensitivity and subsequent voice will be mediated by voice safety. Exploratory Research Three exploratory research questions are examined. First, the current research aims to assess the role of resilience in the voice process. In addition to previously discussed considerations of resilient behavior (i.e. voicing despite the adversity of non-endorsement) in the voice process, the role of dispositional trait resilience “the capacity to rebound or bounce back from adversity” (Luthans, 2002; p. 702) - which makes resilient behaviors more likely for some - will be examined. Dispositional resilience will be assessed as a potential moderating factor that alters the influence of specificity and sensitivity on voice perceptions (i.e. efficacy and safety); since the negative effects of non-endorsement are expected to present themselves in the initial link to voice perceptions (efficacy and safety) it is here examined whether resilience can ameliorate these effects. This buffering effect is expected due to the “bouncing back” nature of those considered high in resilience, which may lead to negative experiences not exhibiting great influence on subsequent perceptions and behaviors. Previous research has linked dispositional resilience to a host of desired outcomes such as decreased stress and anxiety (Davydov et al., 2010), fewer post-traumatic stress symptoms following traumatic experiences (Bensimon, 2012), and improved cardiovascular recovery from stress (Tugade & Fredrickson, 2004). Thus, it is here expected that those low in resilience (i.e. those without the dispositional protective factor) will subsequently be more sensitive to explanation components and will experience worse outcomes when presented with inadequate explanations following non-endorsement. Those high on resilience will have this individual difference to draw from in times of adversity (i.e. nonendorsement paired with low specificity and sensitivity), and thus may be less sensitive to the 21 effects of explanation adequacy. The following moderating effects of resilience are here explored: Research Question 1: Does resilience interact with explanation adequacy (specificity and sensitivity) to predict efficacy and safety perceptions in such a way that the relationship between adequacy and perception (efficacy and safety) is weaker for highly resilient individuals? As a second exploratory area, three different elements of exchange will be assessed to determine their influence on reciprocity between leaders and voicers. First, explanations for nonendorsement may contain information that conveys an internal or external attribution. Internally attributed non-endorsement is one that places responsibility on the individual (e.g. idea utility), while external is non-endorsement attributed to factors external to the voicer (e.g. organizational constraints). In terms of locus of non-endorsement attribution, both internally and externally attributed explanations offer potential costs and benefits. Internal explanations may offer a learning opportunity for voicers to better understand what was wrong with their idea and how it can be improved in the future. On the other hand, the risk of damaging voicer efficacy and safety perceptions in the interpersonally uncomfortable internal attribution explanation may be ameliorated in the use of an external attribution for non-endorsement. As no previous work has linked explanations and voice literatures, the role of explanation attribution is assessed here as an exploratory research question. A second exploratory factor that may influence voicer perceptions and experiences is the provision of an explanation for non-endorsement. Previous work has hinted at that importance of explanations by suggesting that the reason employees are often frustrated when their ideas are not used may be because they did not receive a response (De Vries et al., 2012). However, such 22 an assumption should be empirically assessed. The current work empirically tests this claim. By linking explanation prevision and subsequent efficacy and safety perceptions this work seeks to shed light on whether interpersonal and instrumental concerns and ameliorated (or heightened) when employee simply receive (or do not receive) a response from their leader. Third, as the current work discusses the potential for exchanged explanations to maintain reciprocity in the voice process, the claim may be raised that other exchanged benefits (e.g. a monetary bonus) may also help maintain reciprocity and ameliorate negative effects following voice non-endorsement. This exploratory question aims to help clarify the boundary conditions of voice process reciprocity maintenance and voice resilience by assessing whether various exchanged benefits are adequate, or the exchange of information (i.e. voice) for information (i.e. explanation) is alone considered an exchange “in kind.” Gouldner (1960) asserts that “whether in fact there is a reciprocity norm specifically requiring that returns for benefits received be equivalent is an empirical question” (p.171). Thus, the current work will assess this claim empirically. Overall, the three above-mentioned exchange elements will be explored in the following research question: Research Question 2: Are voice efficacy and safety predicted by exchange factors beyond explanation adequacy (i.e. locus: internal vs. external; presence: response vs. no response; reward type: explanation vs. bonus)? Lastly, the current work seeks to provide additional support for the influence of efficacy and safety, while also parsing apart their relative, unique importance. This research will examine the comparative variable importance information provided by each perceptual construct in the prediction of voice. Thus, the following research question will be examined: 23 Research Question 3: Does voice efficacy or safety have a larger relative weight in the prediction of voice? Is each weight significant in the prediction of voice? 24 METHOD Participants Participants included 532 undergraduate students. Participants were given the opportunity to offer suggestions at time 1 and, as current variables of interest involved responses received concerning their previous suggestions, 139 participants were removed from analyses because they chose not to give a suggestion at time 1. Independent samples t-tests comparing those who completed both measures to those who only completed measures at time 1 demonstrated significant differences on voice efficacy, t(461) = -8.93, p < .01, with those who gave suggestions exhibiting higher mean efficacy levels, voice safety, t(461) = -5.61, p < .01, with those who gave suggestions exhibiting higher mean safety levels, self-esteem, t(461) = -2.17, p < .05, with those who gave suggestions exhibiting higher mean self-esteem levels, and resilience, t(461) = -2.30, p < .05, with those who gave suggestions exhibiting higher mean resilience levels. Frequencies also demonstrated a change is the distribution of participants’ gender representation (no suggestions = 68% female; suggestions given= 82% female) and ethnic group membership (no suggestions = 54.7% majority members; suggestions given = 75.8% majority members) between conditions. In addition, as the study relied upon participation at two time points, 69 students were removed from analyses because they provided suggestions at time one but did not complete time 2 measures. Independent samples t-tests comparing those who completed both measures to those who only completed measures at time 1 demonstrated significant differences on voice efficacy, t(391) = -3.58, p < .01, with those who completed time 1 and time 2 measures demonstrating higher efficacy, voice safety, t(391) = -2.23, p < .05,with with those who completed time 1 and time 2 measures demonstrating higher safety perceptions, self-esteem, t(391) = -2.21, p < .05, with those who completed time 1 and time 2 measures 25 demonstrating greater self-esteem evaluations, and resilience, t(391) = -2.07, p < .05, with those who completed time 1 and time 2 measures demonstrating higher resilience levels. Frequencies also demonstrated differences between groups based on gender (30% of those who completed only time 1 measures were male, while those who completed both time 1 and time 2 measures were only 17.3% male) and ethnicity (44% of the sample completing only time 1 were minority ethnicity group members and only 24% of the sample completing both time points belonged to a minority ethnicity group). Thus, 324 undergraduate students (83% female; 76% majority ethnic group members; Mage= 19, SD = 1.91) provided usable data for current analyses. Only participants above the age of 18 were permitted to participate. Participants also indicated their year in undergraduate education (about 25% from each class category), 91.2% had previous work experience, and 53% were currently employed. Individuals in the current sample were randomly assigned to one condition and all conditions contained between 25 and 35 participants. Research Design The current design is a 2 (high specificity, low specificity) x 2 (sensitive, insensitive) x 2 (locus of non-endorsement: internal, external) between-subjects study. In addition, two other exploratory manipulation conditions were assessed in the current work: (1) no response provided to the voicer and (2) a monetary bonus given in exchange for sharing ideas. Written responses were used to recreate an organizational voice process, because, although employee voice is often verbal, it is not limited to verbal expressions (Hirschman, 1970) and may include emails and written communication (Withey & Cooper, 1989). Explanation research also typically delivers messages in writing (see Shapiro et al., 1994); this may be due to the increased likelihood that participants will miss or misunderstand a verbally delivered explanation. Thus, in an effort to 26 simulate a written voice process in an organization, a written format was used for the current study. Procedure Participants were contacted and recruited via a University online Human Participation in Research (HPR) website. All participation was voluntary and participants received two credits for their participation (See Appendix A and B for Consent and Debriefing Forms). Data collection involved two 30-minute online sessions per participant. Time points were separated by a 3-5 day period. This time frame was chosen to simulate responding to voiced suggestion in timely manner, while also ensuring that adequate time was allotted for the perceived consideration of ideas. At time one, participants first read background information about a fictional marketing organization and were then asked to place themselves in the perspective of a current employee. After reading company background information, participants were told that their supervisor is working on a new marketing campaign. Participants then reviewed the current marketing strategies (marketing plan and flyer) that the supervisor created. Study participants were then given an opportunity to provide suggestions about how the current marketing strategy can be improved. Once suggestions were provided, participants completed all time 1 and control measure scales and provided demographic information. At time two, participants received a non-endorsement response from their supervisor. Within the non-endorsement response explanation focal independent variables were manipulated, and participants were randomly assigned to one condition. After reading the supervisor’s explanation, participants were then shown marketing materials created for their supervisor’s newest organizational client. After reviewing these materials, participants were again given a 27 chance to provide input. Participants then completed scales measuring time 2 mediator variables, dependent variables and manipulation checks. Manipulation Manipulation scenarios (See Appendix C) were crafted following the example of previous manipulations of explanation specificity and sensitivity (Frey & Cobb, 2010; Ployhart et al., 1999; Sharpiro et al., 1994). To manipulate explanation specificity and sensitivity, students received explanations for non-endorsement. Specificity was manipulated by the presence (or absence) of personalized information and the provision of specific (or vague) details about the reasons for non-endorsement. Sensitivity was manipulated by the presence (or absence) of expressed concern for the recipient and whether the explanation was presented in a polite (or blunt) manner. In addition, these specificity and sensitivity manipulations were used in conjunction with the internal versus external attribution of non-endorsement. In the internal attribution conditions, non-endorsement was attributed to the quality of one’s ideas. In the external condition, non-endorsement was attributed to a lack of organizational resources available for idea implementation. This provided eight conditions. Two additional conditions were included based on exploratory research questions: some participants were randomly assigned to a condition in which they were made aware that their suggestions would not be used but received no response from their supervisor concerning their voiced ideas and suggestions, and some participants were randomly assigned to a condition in which they were notified that their ideas would not be used but that they would receive a monetary bonus in exchange for sharing their ideas. Pilot Testing 28 Pilot testing of explanation adequacy manipulations was conducted. Voice nonendorsement explanations were presented to 26 student participants. These participants completed scales measuring explanation specificity (Shapiro et al., 1994; α = .87) and sensitivity (Shapiro et al., 1994; α = .91). Six participants were removed from analyses due to failed attention checks that instructed participants to leave certain questions unanswered, leaving a usable sample of 20 participants. Comparison of mean differences demonstrated that high specificity conditions received higher specificity ratings (M = 3.65, SD = .82) as compared to low specificity conditions (M = 2.18, SD = .78). The same pattern of results was observed in comparing high sensitivity conditions (M = 3.65, SD = 1.00) to the low sensitivity conditions (M = 2.92, SD = .94). T-tests demonstrated that high and low specificity conditions differed significantly, t(19) = 4.95, p < .01, along with high and low sensitivity conditions t(19) = 3.79, p < .01. Overall, the level of each explanation adequacy was perceived as different and in the direction intended between conditions. Measures All measures (See Appendix D) were completed using a five-point (strongly disagree to strongly agree) scale unless otherwise noted. Independent Variables. As the explanations presented to participants were intended to manipulate specificity and sensitivity, a four-question measure of explanation specificity (Shapiro et al., 1994; α = .87) and a five-question measure of sensitivity (Shapiro et al., 1994; α = .91) was completed by each participant. Mediation Variables. The voice efficacy scale was adapted from the Burris et al. (2008) measure. This scale assesses futility perceptions concerning voice behavior. The three questions (α = .91) ask whether the respondent views speaking up as a waste of time, would say that it is 29 useless for them to try to speak up, or believes that nothing will change even if they speak up. The current measure of voice safety was adapted from research conducted by Edmonson (1999) concerning psychological safety. Psychological safety is defined as the belief that it is safe to engage in interpersonal risk taking. This three-question measure (α = .89) was adapted to fit the voice context. An example question reads, “It is safe for me to speak up here.” Dependent Variables. Voice (the provision of ideas/suggestions intended to help the organization; Van Dyne & LePine, 1998; α = .95) was assessed with a 6-item measure and actual voice behaviors presented by participants. For the voice measure, an example question reads, “I would speak up in this group with ideas for new projects or changes in procedures.” Actual voice behavior (voice frequency) was gathered as an additive variable of the number of ideas/suggestions provided by participants. The primary investigator counted and subsequently coded the number of suggestions provided by each participant to determine the voice frequency score. Control Variables. Previous research has found that self-esteem (LePine & Van Dyne, 1998) and proactive personality (Crant, 2003) significantly predict voice. Therefore these measures were examined as potential controls in the current analyses. The ten-question selfesteem scale (Rosenberg, 1965; α = .75) assessed individuals’ perception of personal worth. The ten-question proactive personality scale (Siebert, Crant, & Kraimer, 1999; α = .78) measured the extent to which individuals are prone to bringing about change and affecting their surroundings. In addition, based on current hypotheses that resilience may influence voice resilience, a tenquestion measure of resilience (Campbell-Sills & Stein, 2007; α = .85) was administered to assess this dispositional individual difference variable. One example question asks whether participants “tend to bounce back after illness or hardship.” 30 RESULTS Initial Analyses Initial analyses were conducted to assess scale properties, manipulation strength, and control variable inclusion (See Appendix E for Tables and Figures). First, predictor (specificity and sensitivity) and mediator (efficacy and safety) scales were assessed using confirmatory factor analysis (CFA). The chi-square statistic, comparative fit index (CFI) and root mean square error of approximation (RMSEA) were used to evaluate model fit. CFAs were conducted in the statistical program R using listwise deletion and maximum likelihood estimation. Hu and Bentler (1999) have suggested values to assess adequacy of comparative fit indices (RMSEA < .06; NNFI and CFI > .95). For both sets of variables (independent variables = specificity and sensitivity; mediators = efficacy and safety) the one factor model demonstrated poorer fit (independents: Χ2= 321.45, df = 27, p < .001, RMSEA = .18, and CFI = .77; mediators: Χ2= 330,60, df = 9, p < .001, RMSEA = .33, and CFI = .70) than the two factor model (independent variables: Χ2= 142.2, df = 26, p < .001, RMSEA = .12, and CFI = .91; mediators: Χ2= 88.91, df = 8, p < .001, RMSEA = .02, and CFI = .99). Thus specificity and sensitivity, as well as efficacy and safety were analyzed as separate and distinct variables in all subsequent analyses. Second, manipulation check analyses were conducted to ensure that those in the low explanation specificity condition perceived the explanation to be less specific than those in the high specificity condition and that those in the sensitive condition perceived the explanation to be more sensitive than those in the insensitive condition. Independent samples t-tests indicated that specificity significantly differed between conditions (low M = 1.87, SD = .69; high M = 2.37, SD = .06), t(258) = 5.68, p < .01. The same pattern of expected results was also demonstrated between sensitivity manipulation conditions (low M = 2.42, SD = .75; high M = 31 2.99, SD = .69), t(258) = 6.31, p < .01. These findings supported current manipulations as effective for intended purposes. Lastly, correlations were examined to determine which of the proposed control variables should be included as controls in subsequent analyses. See Table 1 for means, standard deviations, alpha coefficients, and correlations. Correlational analyses demonstrated that selfesteem and resilience should be controlled for in the prediction of efficacy and safety and that self-esteem, proactive personality, and resilience should be controlled for in the prediction of voice. See Table 2 for means by condition for focal variables. Hypotheses Analyses Hypotheses 1 and 2. A one-way between-subjects analysis of covariance (ANCOVA) was conducted to determine whether there were significant mean group differences on voice efficacy for participants in each specificity condition, and whether there were significant mean group differences on voice safety of participants in each sensitivity condition. Control variables (i.e. self-esteem and resilience) I and time 1 measures were controlled for in the corresponding analyses (e.g. time 1 voice efficacy was controlled for in the prediction of time 2 voice efficacy)II. Results demonstrated that explanation specificity did not have a significant effect on voice efficacy, F(4, 260) = 2.45, p > .05, ns. A second ANCOVA demonstrated that sensitivity I Control variables were included due to findings that individual differences were significantly correlated with condition and thus may not have been distributed completely at random throughout the current sample. When analyses were conducted without control variables, hypothesis tests fell below generally accepted significance levels. II As the current work assessed change in efficacy and safety following explanation presented, time 1 efficacy and safety were controlled for in analyses. 32 of explanation significantly predicted perceived voice safety, F(4, 260) = 4.77, p < .05, R2 =.17. Thus, Hypothesis one was not supported and Hypothesis two was supported (See Table 2). III Hypotheses 3a and 3b. Hierarchical linear regression analyses were conducted to assess the prediction of voice by voice efficacy and voice safety (See Table 3). Variables that demonstrated a significant relationship with voice were controlled for (self-esteem, proactive personality, and resilience were entered in step 1). In addition, all variables were grand mean centered prior to analyses. Results demonstrated that voice efficacy significantly predicted voice score, β = .39, SE = .04, p < .01, and voice frequency score, β = .23, SE = .12, p < .01. In addition, multiple linear regression revealed that voice safety perceptions have a significant effect on voice score, β = .31, SE = .04, p < .01, but not voice frequency score, β = .06, SE = .15, p > .05, ns. These findings provided support for Hypotheses 3a and partial support for Hypothesis 3b. Hypotheses 4a and 4b. Mediation analyses were conducted based on the methods described by Andrew Hayes (2013). Hypotheses were analyzed using the program PROCESS (Hayes, 2013) to determine whether the conditions for mediation were met. In applying these conditions to the current model, the following criteria must be met: variation in levels of explanation specificity significantly account for variation in voice efficacy and variation in levels of explanation sensitivity significantly account for variation in voice safety; variation in voice efficacy and voice safety significantly account for variation in voice, and, when controlling for voice efficacy and voice safety, the previously significant relationships between explanation specificity and explanation sensitivity with voice should be reduced or no longer significant. To III Regression analyses were also conducted to test hypothesis one and two using the participant rated specificity and sensitivity measures. The same pattern of results was observed. 33 test these conditions, a series of separate regression models are estimated (Judd & Kenny, 1981). First, the mediators are regressed onto the independent variables. Second, the dependent variable is regressed onto the independent variables. Third, the dependent variable is regressed onto both the independent variables and the mediators. Separate coefficients for each equation are presented. To establish mediation, the first two equations must be significant, and the mediators must affect the dependent variable in the third equation. Finally, the effect of the independent variable on the dependent variable should be less in the third equation than in the second. Sobel tests (Sobel, 1982) and bootstrap confidence intervals were used to determine significance of indirect effects. PROCESS results did not support the hypothesis that the relationship between explanation specificity and voice was mediated by efficacy (See Table 4a). The total effect (the summation of the direct and indirect effects) was not statistically different from zero, t(322) = 1.74, p > .05, ns. In addition, the indirect effect was also not statistically different from zero, as revealed by a 95% bootstrap confidence interval that included zero (lower limit = -.05 to upper limit = .05). Normal theory-based Sobel test also demonstrated non-significant mediation results, Z = -.20, p > .05, ns. Similarly, simple mediation analyses demonstrated that the relationship between explanation specificity and voice frequency score was not mediated by voice efficacy. Indirect effect bootstrap confidence intervals included zero (lower limit = -.07 to upper limit = .05) and the Sobel test result was not significant, Z = -.12, p > .05, ns (See Table 4b). PROCESS results supported the hypothesized mediation of the relationship between explanation sensitivity and voice by voice safety (See Table 5a). Both bootstrap confidence intervals (lower limit = .01 to upper limit = .12) and Sobel test results, Z = 2.76, p < .01, supported the expectation that the indirect effect was significantly different from zero. 34 PROCESS results also demonstrated that the indirect of effect of explanation sensitivity on voice frequency score, via explanation sensitivity, was marginally significant. Bootstrap confidence intervals (lower limit = .01 to upper limit = .12) and Sobel test results, Z = 1.92, p = .05, provided marginal support for this expectation. Thus, Hypothesis 4a was not supported and Hypothesis 4b was supported (See Table 5b). Exploratory Analyses Three exploratory analyses were conducted. First, hierarchical linear regression using interaction terms was used to assess the interactions between resilience and explanation specificity and sensitivity in the prediction of voice efficacy and safety, respectively. Significant results were followed-up with simple slopes analyses. Resilience was assessed as a moderator of the relationship between explanation specificity and voice efficacy as well as the relationship between explanation sensitivity and voice safety. Control and corresponding time one variables were entered into step one of the hierarchical linear regression model, main effects of specificity (or sensitivity) and resilience were entered in step two and the interaction terms were entered in step three. A significant interaction was not observed between resilience and specificity in the prediction of voice efficacy, β = -.08, SE = .09, p > .05, ns (See Table 6a). However, a significant interaction was observed between resilience and explanation sensitivity in the prediction of voice safety, β = -.16, SE = .07, p < .01 (See Table 6b). The significant interaction was further probed using simple slopes analyses (Preacher, Curran, & Bauer, 2006). These analyses demonstrated that the slope coefficient was not significant at high levels of resilience (+1 standard deviation above the mean), z = .20, p > .05, ns, but was significant at low levels of resilience (-1 standard deviation below the mean), z = 4.89, p <.01 (See Figure 2). Thus, simple slopes analyses demonstrated that those high on dispositional resilience were unaffected by explanation 35 sensitivity, while those low on resilience demonstrated higher mean levels of voice safety in times of high explanation sensitivity and significantly lower levels of safety when explanation sensitivity was low. Second, separate multivariate analyses of covariance (MANCOVA) were used to predict voice efficacy and safety based on locus of non-endorsement attribution, response presence, and reward type exchanged (See Table 2). Planed comparisons were conducted by dummy coding conditions and carrying out analyses that compared the focal group to the most relevant comparison groups of interest. For locus of attribution, internal was compared to external explanations. In the case of assessing explanation presence effects, absence of an explanation was compared to the provision of an adequate as well as an inadequate explanation. Monetary bonus provision was also compared to the provision of an adequate as well as an inadequate explanation. In the prediction of voice efficacy, MANCOVA results demonstrated that attribution locus (internal vs. external), F (1, 260) = .16, p >.05, ns; absence of response (vs. inadequate response), F (1, 94) = .26, p > .05, ns; monetary bonus provision (vs. inadequate explanation), F(1, 94) = .37, p > .05, ns; and monetary bonus (vs. adequate explanation), F(1,98) = .41, p > .05, ns, did not demonstrate significant differences in prediction. However, absence of response (vs. adequate response) did significantly predict voice efficacy, F(1,98) = 5.30, p < .05. In terms of voice safety, locus of explanation attribution (internal vs. external), F(1, 260) = .01, p > .05, ns; absence of response (vs. inadequate response), F(1,94) =.16, p > .05, ns; absence of response (vs. adequate response), F(1,98) = 1.98, p > .05, ns; monetary bonus provision (vs. inadequate explanation), F (1, 94) = .49, p > .05, ns; and monetary bonus 36 provision (vs. adequate explanation), F(1,98) = 3.03, p > .05, ns, did not significantly predict voice safety. Lastly, relative weights analyses using confidence interval tests of significance was conducted to assess the relative influence of voice efficacy and voice safety perceptions on voice (Tonidandel et al., 2009). Relative importance analysis was conducted in the statistical program R. Relative weights analysis is often used to examine the proportionate contribution a predictor makes to the R2 (total variance explained by predictors; LeBrenton, Hargis, Griepentrog, Oswald, & Ployhart, 2007; Johnson & LeBrenton, 2004). In the prediction of voice score, both efficacy and safety demonstrated significant weight. However, in the prediction of voice frequency, only efficacy demonstrated a significant relative weight in prediction (See Table 7). Efficacy contributed greater weight in the prediction of both voice and voice frequency. 37 DISCUSSION Summary and Discussion of Results The current work assessed the role of explanation adequacy in influencing voice perceptions and subsequent voice behavior. Findings demonstrated that explanation sensitivity significantly predicted voice safety and that the relationship between sensitivity and subsequent voice was mediated by voice safety. This highlights the important role of sensitive explanation delivery in times of voice non-endorsement, and also sheds light on safety perceptions as the meditational mechanism that forms this linkage. This finding was in line with current hypotheses and furthers previous research concerning the utility of explanation sensitivity in its application to voice safety. In addition, current findings provided support for the assertion that efficacy and safety significantly predict voice, and also provided novel information affirming the significant and greater relative influence of efficacy, as compared to safety, on subsequent voice. This work is in line with previous voice research (for review see Morrison, 2011) and also contributes a more nuanced understanding of these effects. Though voice efficacy and safety are often presented as the two fundamental voice perceptions that influence voice behavior, understanding their relative contribution and significance may offer valuable information to guide future research and implementation. Also, by assessing these predictors’ influence on a predominantly used voice scale (Van Dyne & LePine, 1998) this work bolsters their expected value. Results further showed that the effect of explanation sensitivity on voice safety is moderated by dispositional resilience - those lower on resilience are more sensitive to the effects of sensitivity, and that providing an adequate explanation, as compared to no response, makes a significant, positive difference for voice efficacy. These finding demonstrate the importance of 38 dispositional resilience in buffering the negative effects of voice non-endorsement and the value of providing responses to voice behaviors. Non-significant findings also provide potential insight into the voice process. Current findings demonstrated five non-significant effects: (1) explanation specificity does not significantly impact voice efficacy, (2) safety perceptions do not predict voice frequency ratings, and there were no observed differences on voice efficacy or safety when comparing (3) locus of explanation attribution, (4) absence of an explanation vs. presentation of an inadequate explanation, and (5) an explanation (adequate or inadequate) as compared to the provision of a monetary reward. First, the unsupported prediction of voice efficacy by explanation specificity may be due to context characteristics and may demonstrate a boundary condition for the utility of adequate explanations in voice processes. Aspects of the scenario may have limited the influence of specificity on efficacy. Perhaps, as this was a role-playing task, limits to realism may have made participants’ efficacy more dependent on personal traits or views (e.g. general expectations and experiences concerning intern efficacy), as opposed to the one time supervisor explanation provided. On the other hand, these results may depict the limited utility of disclosing more specific information in times of voice non-endorsement. Second, the finding that safety perceptions predicted voice but not voice frequency provides potentially useful information for the utility and limits of voice safety influence. This finding may not be surprising, as the number of suggestions can be a function of other factors (e.g. having ideas to share). As shown by current findings, it may be the case that employees who do not feel safe in the voice process will likely not voice at all, while those who do feel safe will voice – with varied frequency. 39 The remaining non-significant finding demonstrated an absence of significant mean differences on voice efficacy and safety based upon locus of explanation attribution (external vs. internal), the absence of an explanation (as compared to an inadequate explanation), and the provision of a monetary bonus (as compared to both an adequate or inadequate explanation). These findings provide detailed comparisons of potential managerial responses to voice and, if not due to limitations in the current study, demonstrate that efforts directed towards choosing between these exchanges may be futile. The demonstrated similarity in subsequent voice perceptions following externally or internally attributed non-endorsement explanations was a surprising and interesting finding. This exploratory research question was undertaken based on the expectation that different perceptions would result from placing blame on the voicer or the organization. Findings demonstrated that perhaps managers or researchers need not be concerning about non-endorsement attributions in the voice process as much as was expected. However, such non-significant differences should be interpreted with caution, as these effects may also depend on unexplored moderation factors (e.g. whether the voicer believes the explanation reason presented, how invested the voicer was in their idea being endorsed). Results also demonstrated that the absence of an explanation was not significantly different from the provision of an inadequate explanation. This demonstrates that the adage “saying anything is better than saying nothing” may not apply to voice non-endorsement. Participants responded well to receiving adequate explanations, as compared to no explanation, but supervisors who delivered an inadequate explanation did not improve voice perceptions, even as compared to those who do not explain their decision at all. This finding may further bolsters the value of an adequate explanation and reaffirm the harm in ignoring voice behavior or providing non-specific and non-sensitive responses. 40 Lastly, offering monetary rewards to voicers, as compared to the provision of an explanation, did not produce significant differences for subsequent voice efficacy and safety perceptions. This finding is in line with the assertion that reciprocity exchange can be maintained in times of non-endorsement, and demonstrates that managerial benefits exchanged may not have to be information (e.g. voice behavior) exchanged for information (e.g. an explanation). One consideration in interpreting this finding involves the type and quantity of benefit exchanged and how effects may be altered based on these factors. This highlights that what some voicers value and accept may be different from others. For example, although the current study did not specify a dollar amount in the monetary bonus condition, certain values exchanged may be perceived as appropriate and may be differentially effective in voice exchanges. Theoretical and Practical Implications Current findings provide implications for theoretical and practical decisions. In terms of theoretical implications, the demonstrated role of explanations, the norm of reciprocity, and resilience in the voice process provide implications for future theory development. First, by parsing apart the unique effects of explanation adequacy elements (specificity and sensitivity) on perceptual elements (efficacy and safety) and finding unique patterns and relationships, this work demonstrates that future research should further uncover the psychological processes that fuel and mediate voice effects. Second, general expectations based the social exchange framework were supported by current findings. In uncovering the significant difference that an adequate explanation can produce and the absence of a difference between explanations and monetary benefits, this work demonstrates the important role of explanations, but also shows that the most important element if this process is the maintenance of reciprocity; which may not be limited to mutual information sharing. This provides additional support for the consideration of the voice 41 process as a social exchange and may help guide future scholars in the voice theory development. Lastly, this work demonstrated the importance of resilience as a behavioral outcome (voice resilience despite non-endorsement) as well as its role as a dispositional trait in influencing voice effects. This provides guidance for future developments to build upon and expand. For example, future research should continue to consider the potential for voice resilience and role of disposition resilience in the voice process, as well as examine the effects of other perseverance characteristics (e.g. general affect, locus of control, hardiness) on voice effects. Many practical decisions may be guided by the current work. As it is well understood that all voiced ideas and suggestions may not be endorsed, it is useful to understand how leaders should manage such instances. Results indicate the importance of conveying non-endorsement messages in a sensitive manner, the value in ensuring that employees receive a benefit in exchange for their voice behavior – whether that is an explanation or some tangible reward - and the especially important role that explanation sensitivity plays in the voice process for those who do not possess dispositional resilience. Such findings may be used to train managers in how to properly respond to non-endorsed voice, to reap the many benefits that subsequent useful suggestions may provide. In addition, this work may be used to inform employees how and why negative voice effects occur. It may be beneficial to alert employees to the influence nonendorsement and managerial response may exert on their future voice perceptions and behaviors so that they too may be on guard against voice non-resilience. Recommendations for Future Research In addition to the theoretical advancements mentioned that future work may expound upon, future empirical work should pursue the application of additional theories to the voice 42 process, examine current and additional effects using differing types of voice behavior, further investigate the determinants of voice efficacy, and explore additional moderators that may alter explanation adequacy effects. First, the application of additional theories in investigating and understanding the voice process would further current understandings. For example, work that links fairness theories to understanding which elements exchanged by leaders maintain the reciprocal exchange may prove useful. Feedback literature may also provide a theoretical lens to understand parallel effects in the voice process. In addition, future work should consider the role of leadership theories in informing our understanding of the role of leaders in voice exchanges. Second, future work should consider different types of voice to help clarify previous findings and delineate potentially unique effects (e.g. explanations may be useful with non-endorsement voice intended to change the status quo vs. voice intended to maintain it). Such research would also help situate findings in the general voice literature relative to different voice types. Third, this work demonstrated the importance of voice efficacy as a significant predictor of voice and voice frequency, demonstrating relatively greater influence than voice safety; however, the explanation adequacy component examined (specificity) was not a significant predictor of voice efficacy. Thus, future work should further examine factors that may influence voice efficacy (e.g. whether previous ideas have generally been endorsed; relationship with current leader; time and effort spent in idea generation). Such assessments would offer greater potential to influence subsequent voice behavior. Fourth, as mentioned above, despite the demonstrated importance of trait resilience in current effects, future work should uncover other moderators that may alter voice effects (e.g. general affect, locus of control, hardiness). A fifth consideration for future work involves the use of various measures of voice and capturing actual voice behavior. Although intentions often translate into actual behaviors (Ajzen & Fishbein 1977), current results 43 showed some differential findings between voice scale scores and voice frequency and a small relationship between these forms of voice measurement (r = .13). This highlights the potential value in future voice research not only assessing self-reported voice scores, but also measuring actual voice behavior - furthering our knowledge about how these measurement approaches relate. Lastly, current research assessed the presence and frequency of voice but did not delve into the utility or quality of suggestions given. As the value of voice often depends upon such characteristics, future research would do well to assess whether and why the quality of future suggestions might change following non-endorsement (e.g. individuals may put less effort in generating thoughtful suggestions following rejection of previous ideas or they may be motivated to put in more effort to avoid repeated non-endorsement) and how explanation aspects (e.g. provision and adequacy) may alter these effects. Limitations The current work and findings are strengthened by the time-lagged assessment of the voice process, the delineation of unique predictions between the elements of explanation adequacy and voice perception, the use of a specific and consistent form of voice throughout the study, and the measurement of voice based on a previously validated scale as well as the assessment of actual voice behaviors. However, current findings should also be interpreted in light of current limitations. First, the use of a student sample, though potentially representative of interns, is not representative of most working adults and may limit current research generalizability. Second, as the marketing materials presented at time one and time two were different, differences between these materials (e.g. quality and number of perceived flaws) may have affected participants’ likelihood of voicing suggestions. Despite the materials being vetted by 4 SMEs for similarity on appeal, individual participants’ differing views concerning the 44 quality of the two sets of materials may have influenced their voicing behaviors. Third, observed differences between those who completed both time points and provided suggestions at time 1 and those who did not demonstrate that missing data may not be completely at random. Although these differences do not impact conclusions, because the independent variables involving explanation elements could only be presented to those who chose to voice and completed both time points, this may limit the interpretation and generalizability of results garnered from the final sample. However, there is not enough theoretical or empirical evidence currently available to explain potential gender or ethnicity effects on motivations to voice. Fourth, the ratings of explanation specificity and sensitivity, despite being significantly different across conditions in the expected direction, fell at or near the midpoint of the scale and may not be completely independent. The relatively low ratings, especially in the high specificity and sensitivity conditions, may be due to manipulation strength - as it was important to balance manipulation strength with realism (e.g. not making something so insensitive as it would fall outside the bounds of realistic managerial behavior). However, it is noteworthy that effects were present despite the absence of extreme scale scores, and thus, even larger effects may be expected if these manipulations further demonstrated higher or lower sensitivity levels. There was also a significant correlation between explanation specificity and sensitivity rating. A post-hoc MANCOVA further demonstrated that specificity condition predicted specificity (F(1, 260)= 32.47, p< .01) as well as sensitivity (F(1, 260)= 12.64, p < .01) rating; while sensitivity condition only predicted sensitivity (F(1, 260)= 21.22, p < .01) rating. These associations, despite the demonstration of two separate factors in the confirmatory factor analyses, are intuitive - as an explanation given with more detail would seem more sensitive due to time and effort exerted to provide information to the voicer; while sensitivity in explanation presentation would not likely 45 influence perceived specificity. Thus, perceptions may not be parallel to their independence in manipulations and, thus, interpretations of specificity manipulations should be interpreted in light of this association. Lastly, though the marketing firm and scenario information was intended to portray a realistic scenario, it cannot be assumed that participants’ actions in the current study are a direct reflection of how they would behave as an actual organizational member. Current concerns of generalizability, company material equivalence, and scenario realism should be considered in interpretation of this research and improved upon in future work. Conclusion In the present study, the norm of reciprocity was used to understand and predict the role of explanation adequacy in mitigating the potential negative effects of voice non-endorsement. Explanation sensitivity showed connections to perceptions of voice safety, efficacy and safety both predicted voice, and safety was shown to mediate the relationship between sensitivity and subsequent voice. In addition, low resilience strengthened the relationship between sensitivity and voice safety, the importance of providing adequate explanation (vs. no response) for voice efficacy was shown, and voice efficacy demonstrated significantly greater weight than in the prediction of voice. These effects demonstrate the important role that leaders play in the voice process, especially in times of voice non-endorsement, and help to further the voice, social exchange, and explanation literatures. 46 APPENDICES 47 Appendix A: Consent Form Marketing Study The purpose of this research study is to understand how different situations influence behavior. You will read a scenario and be asked to role-play and complete a series of questionnaires that assess your thoughts and decisions in the situation explained. This research study will take about 60 minutes (Two 30-minute online sessions) to complete, and participants will be compensated with course credits in accordance with the guidelines set forth by the HPR system - only those that participate in both sessions will receive full credit for this experiment. Students will receive 1 course credit for each session they participate in. Participation is completely voluntary, you may choose not to participate at all, to not answer certain questions, or to discontinue your participation at any time without consequences (e.g., will not affect treatment you will receive, will not affect your grade or evaluation, etc.). Also, you have the right to request that your responses not be used in the data analyses. Participation in this research study does not involve any foreseeable risks. The benefit of participating in this research study, however, is that you have the opportunity to learn more about the research process and also help contribute to scientific advancement. This experiment is confidential and your consent form will be stored in a separate place from your responses. To help us protect your confidentiality, only your last name and email address will be requested. Your confidentiality will be protected to the maximum extent allowed by law. All data will be stored on the hard drive of a secure computer, and will only be accessed by trained experimenters. Data will be stored for five years after the publication of research stemming from this project---as specified by the American Psychological Association. At the conclusion of this research the experimenter will provide a thorough explanation of the study. It is our goal that you learn about the research you participated in today. The experimenter will be happy to answer any questions you have about the research. Dr. Ann Marie Ryan, a professor in the Department of Psychology is conducting this scientific study. If you have questions about the study, contact Ann Marie Ryan, Ph.D., Department of Psychology, 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, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University's Human Research Protection Program at 517-3552180, Fax 517-432-4503, or e-mail irb@msu.edu or regular mail at 408 West Circle Drive. Room 207 Olds Hall, MSU, East Lansing, MI 48824. Checking the box below indicates your voluntary consent to participate in this research: 48 Appendix B: Debriefing Form Marketing Study Thank you for your participation in this study. Below you will find more information about the purpose of this study as well as a list of counseling and information resources. The purpose of this study was to examine the influence of explanations (specific or sensitive; Sharpiro, Buttner, & Barry, 1994) on social exchanges (Blau, 1964) and voice efficacy, safety, and voice (discretionary communication of ideas, suggestions, or opinions about work-related issues with the intent to improve things; Morrison, 2011). Companies are interested in gaining important and innovative information from their employees. Yet managers must withhold endorsement of employee ideas for various reasons (e.g. time/budget constraints; Landau, 2009). Therefore, investigating the impact explanations have on future voice behavior, following an idea not being endorsed, has the potential to inform our literature and human resource practices aimed at encouraging voice behavior from all employees. Voice and explanation research have not been previously linked to one another. The aim of this study is to examine the potential influence of explanations on future voice behavior – in an effort to uncover the underlying processes that increase or decrease employee voice behavior. The company and marketing strategies that you read about are fictional and all information within the stories was created for the purpose of this study. If answering any of the survey questions led you to feel distressed and you would like to speak to someone about your thoughts, please take advantage of the free resources listed below: • Listening Ear Crisis Intervention Center 1017 East Grand River East Lansing, MI, 44423 24-Hour Crisis Hotline: 517-337-1717 Business Phone: 517-337-1728 Participants who are interested in learning more about the results of this study may send the researchers a request for a summary of the findings via email at ryanan@msu.edu. They may also send any comments, questions or concerns regarding the study to the principal investigator, Dr. Ann Marie Ryan at: Department of Psychology, Michigan State University, East Lansing, MI 48824, E-mail: ryanan@msu.edu. We would like to thank you again for your participation 49 Appendix C: Manipulation Scenarios Condition 1: High Sensitivity / High Specificity / External Attribution Condition 2: Low Sensitivity / High Specificity / External Attribution 50 Condition 3: High Sensitivity / Low Specificity / External Attribution Condition 4: Low Sensitivity / Low Specificity / External Attribution 51 Condition 5: High Sensitivity / High Specificity / Internal Attribution Condition 6: Low Sensitivity / High Specificity / Internal Attribution 52 Condition 7: High Sensitivity / Low Specificity / Internal Attribution Condition 8: Low Sensitivity / Low Specificity / Internal Attribution 53 Condition 9: Monetary Bonus Condition 10: No Response to Voice Behavior 54 Appendix D: Measures Manipulation Check Explanation Specificity (scale = 1 strongly disagree to 5 strongly agree; Shapiro et al., 1994; α = .87) 1. My supervisor’s explanation seemed generic 2. My supervisor gave specific reasons for not using my idea 3. My supervisor gave vague reasons for not using my idea 4. My supervisor gave reasons for their decision that were specific to me Explanation Sensitivity (scale = 1 strongly disagree to 5 strongly agree; Shapiro et al., 1994; α = .87) 1. My supervisor communicated their decision in a sincere manner 2. My supervisor communicated their decision is a friendly manner 3. My supervisor seemed sensitive in their communication 4. My supervisor seemed concerned about my understanding of why my idea was not used 5. My supervisor seemed concerned about my feelings Mediation Variables Voice Efficacy (scale = 1 strongly disagree to 5 strongly agree; Burris et al., 2008; α= .91) Scale = 1 (strongly disagree) to 5 (strongly agree) 1. Trying to improve things by speaking up to my supervisor is a waste of time 2. It is useless for me to suggest new ways of doing things to my supervisor 3. Nothing will change even if I share my ideas with my supervisor Voice Safety (scale = 1 strongly disagree to 5 strongly agree; Edmonson, 1999; α = .89) Scale = 1 (strongly disagree) to 5 (strongly agree) 1. It is safe for me to make suggestions to my supervisor 2. It is safe to give my opinions to my supervisor 3. It is safe for me to speak up to my supervisor Dependent Variables Voice (Van Dyne&LePine, 1998; alpha = .89) Scale = 1(strongly disagree) to 7 (strongly agree) 1. I would develops and makes recommendations concerning issues that affect this organization 2. I would speaks up and encourages others to get involved in issues that affect the company 3. I would communicates opinions about work issues to others in this organization even if my opinion is different and others in the group disagree with me 4. I would keeps well informed about issues where my opinion might be useful to this company 5. I would gets involved in issues that affect the quality of work life here in this organization 55 6. I would speak up in this organization with ideas for new projects or changes in procedures Control Variables Rosenberg Self Esteem (scale = 1 strongly disagree to 5 strongly agree; Rosenberg 1965; α = .75) 1. On the whole, I am satisfied with myself 2. At times, I think I am no good at all 3. I feel that I have a number of good qualities 4. I am able to do things as well as most other people 5. I feel I do not have much to be proud of 6. I certainly feel useless at times 7. I feel that I am a person of worth, at least on an equal plane with others 8. I wish I could have more respect for myself 9. All in all, I am inclined to feel that I am a failure 10. I take a positive attitude toward myself Proactive Personality (scale = 1 strongly disagree to 5 strongly agree; Siebert et. al., 1999; α = 78) 1. Wherever I have been, I have been a powerful force for constructive change 2. I am constantly on the lookout for new ways to improve my life 3. If I see something I don’t like, I fix it 4. I am always looking for better ways to do things 5. No matter what the odds, if I believe in something I will make it happen 6. Nothing is more exciting than seeing my ideas turn into reality 7. I love being a champion for my ideas, even against opposition 8. I excel at identifying opportunities 9. If I believe in an idea, no obstacle will prevent me from making it happen 10. I can spot a good opportunity long before others can Connor- Davidson Resilience Scale (scale = 1 not true at all to 5 nearly all the time; based on how participants have felt over the past month; Campbell-Sills & Stein.2007; α = .85) 1. Able to adapt to change 2. Can deal with whatever comes 3. Tries to see humorous side of problems 4. Coping with stress can strengthen me 5. Tend to bounce back after illness or hardship 6. Can Achieve goals despite obstacles 7. Can stay focused under pressure 8. Not easily discouraged by failure 9. Thinks of self as strong person 10. Can handle unpleasant feelings 56 Appendix E: Tables and Figures Table 1: Means, Standard Deviations, Alphas Coefficients, and Correlations M(SD) (1) (2) (3) (4) (5) (6) (7) (8) (9) 1. Specificity 2.14(.76) (.82) 2. Sensitivity 2.70(.79) .53** (.85) 3. Efficacy 3.59(.80) -.01 .13* (.87) 4. Safety 3.70(.66) -.03 .16** .45** (.88) 5. Voice 3.72(.57) -.10 .08 .55** .51** (.88) 6. Voice Freq. 2.03(1.58) -.13* -.06 .24** .14* .13* 7. Self-Esteem 3.80(.65) -.09 -.10 .21** .20** .26** -.06 (.90) 8. Proactivity 3.77(.48) .06 .02 .08 .10 .21** -.08 .40** (.83) 9. Resilience 3.82(.57) -.10 -.01 .24** .21** .27** .01 .53** .50** (.87) Note. * p < .05; ** p < .01. N = 324. Reliability coefficients are on the diagonal. Scales ranged from 1(strongly disagree) to 5(strongly agree). 57 Table 2: Means and Standard Deviation of Focal Variables by Condition Condition Non-Specific Specific Non-Sensitive N 126 134 130 Efficacy M (SD) 3.56 (.87) 3.67 (.69) 3.61 (.83) Sensitive 130 3.62 (.74) Safety M (SD) 3.72 (.68) 3.72 (.65) 3.68a (.71) 3.76a (.62) Voice M (SD) 3.72 (.59) 3.75 (.55) 3.73 (.62) 3.74 (.53) V. Freq. M (SD) 1.92 (1.66) 2.02 (1.46) 2.02 (1.54) 1.92 (1.58) Internal Attrib. 130 3.68 (.80) 3.76 (.63) 3.81b (.55) 1.83 (1.63) External Attrib. 130 3.55 (.76) 3.68 (.70) 3.65b (.58) 2.12 (1.48) Inadequate Explanation 62 3.54 (.93) 3.71 (.71) 3.74 (.63) 1.97 (1.66) Adequate Explanation 66 3.66c (.67) 3.77 (.60) 3.77d (.50) 1.97 (1.50) Bonus 32 3.63 (.89) 3.59 (.56) 3.81 (.47) 2.38 (1.86) 3.55d (.67) 2.16 (1.37) No Response 32 3.39c (.83) 3.63 (.66) Note. * p < .05; ** p < .01. Means that share a subscript differ significantly (p < .05). The adequate explanation was composed of both internal and external attribution explanations that provided both specific and sensitive information. The inadequate explanations included internal and external explanations with neither specific nor sensitive information. Bonus and no response conditions were compared to both inadequate and adequate explanation conditions. 58 Table 3: Hierarchical Linear Regression for the prediction of Voice and Voice Frequency by Efficacy and Safety Voice Step 1 Step 2 Variable b SE β b SE β Constant 3.25 .03 5.52 .03 Self-Esteem .13 .06 .15* .06 .05 .06 Proactivity .10 .07 .08 .14 .06 .12* Resilience .15 .07 .15* .03 .06 .02 Efficacy .28 .04 .39** Safety .27 .04 .31** R2 Adjusted R2 ΔR2 .10 .09 Voice Frequency Step 1 Step 2 b SE β b SE β 2.03 .09 2.03 .09 -.19 .16 -.08 .28 .16 -.11 -.34 .21 -.10 -.27 .21 -.08 .30 .20 .11 .13 .19 .05 .45 .12 .23** .14 .15 .06 .42 .41 .02 .01 .32** .08 .06 .06** Note. * p < .05; ** p < .01. All variables were grand mean centered prior to analyses. 59 Table 4a: Model Coefficients for Voice Efficacy Mediation of the Relationship between Explanation Specificity and Voice Outcome Antecedent X (Specificity) M (Efficacy) Constant M (Efficacy) Coeff. SE p -.01 .06 >.05 3.62 .13 <.01 2 R = .00 F(1, 322) =.04 , p > .05 Y (Voice) Coeff. SE p -.07 .03 =.05 .40 .03 <.01 2.45 .14 <.01 2 R = .31 F(2, 321) =72.58, p < .01 60 Table 4b: Model Coefficients for Voice Efficacy Mediation of the Relationship between Explanation Specificity and Voice Frequency Outcome Antecedent X (Specificity) M (Efficacy) Constant M (Efficacy) Coeff. SE p -.01 .06 >.05 3.62 .13 <.01 R2 = .00 F(1, 322) = .04, p > .05 61 Y (Voice Freq.) Coeff. SE p -.26 .11 <.05 .46 .11 <.01 .92 .46 <.05 R2 = .07 F(2, 321) = 12.35, p < .01 Table 5a: Model Coefficients for Voice Safety Mediation of the Relationship between Explanation Sensitivity and Voice Outcome M (Safety) Antecedent Coeff. SE p X (Sensitivity) .13 .05 <.01 M (Safety) Constant 3.34 .13 <.01 2 R = .03 F(1, 322) =8.29 , p < .01 Y (Voice) Coeff. SE p .00 .04 >.05 .44 .04 <.01 2.08 .17 <.01 2 R = .26 F(2, 321) =55.39 , p <.01 62 Table 5b: Model Coefficients for Voice Safety Mediation of the Relationship between Explanation Sensitivity and Voice Frequency Outcome Antecedent X (Sensitivity) M (Safety) Constant M (Safety) Coeff. SE p .13 .05 <.01 3.34 .13 <.01 2 R = .03 F(1, 322) = 8.29, p < .01 Y (Voice Freq.) Coeff. SE p -.16 .11 >.05 .37 .13 <.01 1.11 .54 <.05 2 R = .03 F(2, 321) = 4.28, p < .05 63 Table 6a: Hierarchical linear Regression for the Prediction of Voice Efficacy by Resilience and Explanation Specificity Variable Constant Self-Esteem Efficacy- Time 1 Resilience Specificity Resil. X Spec. R2 Adjusted R2 ΔR2 Step 1 b SE -1.33 .04 .16 .07 .38 .08 β .13* .28** Step 2 b SE -2.41 .04 .08 .08 .36 .08 .19 .09 .06 .06 .11 .11 .04** .13 .12 .03* 64 β .07 .26** .14* .05 b -.01 .08 .36 .20 .05 -.14 Step 3 SE β .04 .08 .06 .08 .26** .09 .14* .06 .05 .09 -.08 .13 .12 .01 Table 6b: Hierarchical linear Regression for the Prediction of Voice Safety by Resilience and Explanation Sensitivity Variable b Constant 3.70 Self-Esteem .09 Safety- Time 1 .32 Resilience Sensitivity Resil. X Sens. Step 1 SE β .03 .06 .09 .06 R2 .12 2 Adjusted R .11 2 ΔR .12** Note. * p < .05; ** p < .01. .30** b 3.23 .06 .34 .10 .17 Step 2 SE β .12 .06 .06 .06 .07 .04 .17 .16 .05** 65 .31** .09 .21** b 3.21 .06 Step 3 SE .12 .06 .06 .34 .09 .18 -.21 .06 .07 .04 .07 .32** .08 .22** -.16** .19 .18 .02** β Table 7: Relative Weights Analyses of the Prediction of Voice and Voice Frequency by Voice Efficacy and Safety Voice Voice Frequency Raw Rescaled CI CI Relative Relative Sig. Sig. Variables Weight Weight Low High Efficacy .22 55.94* .13 .30 Safety .17 44.05* .09 .26 Note. * = significant relative weight in prediction. 66 Rescaled Raw Relative Relative CI Sig. CI Sig. Weight Weight Low High .05 81.54* .00 .10 .01 18.46 -.01 .03 Figure 1: Proposed Model 67 Figure 2: Interaction between Resilience and Sensitivity in the Prediction of Safety 4 3.9 Voice Safety 3.8 3.7 Low Resilience 3.6 3.5 3.4 3.3 High Resilience 3.2 3.1 3 Low Sensitivity High Sensitivity Explanation Sensitivity 68 REFERENCES 69 REFERENCES Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84, 888-918. Alderfer, C. P. (1969). An empirical test of a new theory of human needs. Organizational Behavior and Human Performance, 4(2), 142-175. Argyris, C. & Schon, D. A. (1978). Organizational Learning. Reading, MA: Addison-Wesley. Avery, D. R., & Quinones, M. A. (2002). Disentangling the effects of voice: The incremental roles of opportunity, behavior, and instrumentality in predicting procedural fairness. Journal of Applied Psychology, 87(1), 81-86. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 6. Bass, B. M., & Riggio, R. E. (2006). Transformational leadership. Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Bensimon, M. (2012). Elaboration on the association between trauma, PTSD, and posttraumatic growth: The role of trait resilience. Personality and Individual Differences, 52(7), 782787. Bies, R. J. (1987). Beyond “voice”: The influence of decision maker justification and sincerity on procedural fairness judgments. Representative Research in Social Psychology, 17, 314. Bies, R. J., & Shapiro, D. L. (1987). Interactional fairness judgments: The influence of causal accounts. Social Justice Research, 1, 199-218. Bies, R. J., & Shapiro, D. L. (1988). Voice and justification: Their influence on procedural fairness judgments. Academy of Management Journal, 31(3), 676-685. Bies, R. J., Shapiro, D. L., & Cummings, L. L. (1988). Causal Accounts and Managing Organizational Conflict: Is it enough to say it’s not my fault? Communication Research, 15, 381-399. Blau, P. (1964). Exchange and power in social life. New York: John Wiley and Sons. Block, J., & Turula, E. (1963). Identification, ego control, and adjustment. Child Development, 34(4) 945-953. Blumstein, P. W. (1974). The honoring of accounts. American Sociological Review, 39, 551-566. 70 Bobocel, D. R., & Farrell, A. C. (1996). Sex-based promotion decisions and interactional fairness: Investigating the influence of managerial accounts. Journal of Applied Psychology, 81, 22-35. Bonanno, G. A. (2004). Loss, trauma, and human resilience: Have we underestimated the human capacity to thrive after extremely aversive events? American Psychologist, 59(1), 20-28. Brockner, J., DeWitt, R., Grover, S., & Reed, T. (1990). When it is especially important to explain why: Factors affecting the relationship between managers’ explanations of a layoff and survivors’ reactions to the layoff. Journal of Experimental Social Psychology, 26(5), 389-407. Burris, E. R. (2012). The risks and rewards of speaking up: Managerial responses to employee voice. Academy of Management Journal, 55(4), 851-875. Burris, E. R., Detert, J. R., & Chiaburu, D. S. (2008). Quitting before leaving: The mediating effects of psychological attachment and detachment on voice: Journal of Applied Psychology, 93, 912-922. Campbell-Sills, L., & Stein, M. B. (2007). Psychometric analysis and refinement of the connordavidson resilience scale (CD-RISC): Validation of a 10-item measure of resilience. Journal of Traumatic Stress, 20(6), 1019-1028. Cheung, M. F. (2013). The meditational role of perceived organizational support in the effects of interpersonal and informational justice on organizational citizenship behavior. Leadership and Organization Development Journal, 34(6), 551-572. Chiaburu, D. S., Lorinkova, N. M., & Van Dyne (2013). Employees’ social context and changeoriented citizenship: A meta-analysis of leader, coworker, and organizational influences. Group & Organizational Management, 38(3), 291-333. Crant, M. J. (2003). Speaking up when encouraged: Predicting voice behavior in a naturallyoccurring setting. Paper presented at the annual meeting of the Academy of Management, Seattle. Cropanzano, R. & Mitchell, M. S. (2005). Social exchange theory: An interdisciplinary review. Journal of Management, 31, 874-900. Davdov, D. M., Stewart, R., Ritchie, K., & Chaudieu, I. (2010). Resilience and mental health. Clinical Psychology Review, 30(5), 479-495. De Vries, G., Jehn, K. A., & Terwel, B. W. (2012). When employees stop talking and start fighting: The detrimental effects of pseudo voice in organizations. Journal of Business Ethics, 105(2). 221-230. 71 Detert, J. R., & Burris, E. R. (2007). Leadership behavior and employee voice: Is the door really open? Academy of Management Journal, 50(4), 869-884. Detert, J. R., & Edmonson, A. (2011). Implicit voice theories: Taken-for-granted rules of selfcensorship at work. Academy of Management Journal, 54(3), 461-488. Detert, J. R., & Trevino, L. K. (2010). Speaking up to higher-ups: How supervisors and skiplevel leaders influence employee voice. Organizational Science, 21, 249-270. Dutton, J. E., Ashford, S. J., O’Neill, R. M., Hayes, E., & Wierba, E. E. (1997). Reading the wind: How middle managers assess the context for selling issues to top managers. Strategic Management Journal, 18, 407-425. Edmonson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44, 350-383. Edmonson, A. C. (2003). Speaking up in the operating room: How team leaders promote learning in interdisciplinary action teams. Journal of Management Studies, 40, 14191452. Emerson, R. (1976). Social exchange theory. Annual Review of Sociology, 2, 335-362. Faul, F., Erdfelder, E., Lang, A. & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175 – 191. Folger, R. (1977). Distributive and procedural justice: Combined impact of “voice” and improvement on experienced inequality. Journal of Personality and Social Psychology, 35, 108-119. Folger, R., Rosenfield, D., Grove, J., & Corkan, L. (1979). Effects of “voice” and peer opinions on response to inequality. Journal of Personality and Social Psychology, 37, 2253-2261. Folger, R., Rosenfield, D., & Robinson, T. (1983). Relative deprivation and procedural justifications. Journal of Personality and Social Psychology, 45(2), 268-273. Folger, R. & Skarlicki, D. P. (2001). Fairness as a dependent variable: Why tough times can lead to bad management. In R. Cropanzano (Ed.), Justice in the workplace: From theory to practice, Mahwah, NJ: Erlbaum. Frese, M., Teng, E., & Wijnen, C. J. (1999). Helping to improve suggestion systems: Predictors of making suggestions in companies. Journal of Organizational Behavior, 20(7), 11391155. 72 Frey, F. M., & Cobb, A. T. (2010). What matters in social accounts? The roles of account specificity, source expertise, and outcome loss on acceptance. Journal of Applied Social Psychology, 40(5), 1203-1234. Fuller, J. B., Marler, L. E., & Hester, K. (2006). Promoting felt responsibility for constructive change and proactive behavior: Exploring aspects of an elaborated model of work design. Journal of Organizational Behavior, 27, 1089-1120. Gergen, K. (1969). The Psychology of Behavior Exchange. Reading, MS: Addison-Wesley. Gilliland, S. W., & Beckstein, B. A. (1996). Procedural and distributive justice in the editorial review process. Personnel Psychology, 49, 669-691. Glauser, M. J. (1984). Upward information flow in organizations: Review and conceptual analysis. Human Relations, 37, 613-643. Goffman, E. (1971). Relations in public: Micro studies of the public order. New York: Basic Books. Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161-178. Greenberg, J. (1990). Employee theft as a reaction to underpayment inequality: The hidden cost of pay cuts. Journal of Applied Psychology, 75(5), 561-568. Greenberg, J. (1993). The social side of fairness: Interpersonal and informational classes of organizational justice. In R. Cropanzano (Ed.), Justice in the workplace: Approaching fairness in human resource management (pp. 79-103). Hillside, NJ: Lawrence Erlbaum Associates. Greenwald, A. G. (1980). The totalitarian ego: Fabrication and revision of personal history. American Psychologist, 35(7), 603-618. Hammer, C. & Hammer, E. (1976). Behavior modification on the bottom line. Organizational Dynamics, 4(4), 3-10. Harlos, K. P. (2001). When organizational voice systems fail: More on the deaf-ear syndrome and frustration effects. The Journal of Applied Behavioral Science, 37(3), 324-342. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. New York: Guilford Press. Henderson, R. (1984). Performance appraisal. Reston, VA: Reston Publishing. Hirschman, A. O. (1970). Exit, voice, and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard University Press. 73 Hornstein, H. A. (1986). Managerial courage: Revitalizing your company without sacrificing your job. New York: Wiley. Howard, A. (1995). The changing nature of work. San Francisco: Jossey-Bass. Johnson, J. W., & LeBrenton, J. M. (2004). History and use of relative importance indices in organizational research. Organizational Research Methods, 7, 238-257. Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating mediation in evaluation research. Evaluation Research, 5, 602-619. Kanter, R. M. (1983). Change matters: Innovation for productivity in the American workplace. New York: Simon & Schuster. Kassing, J. W. (2000). Investigating the relationship between superior-subordinates relationship quality and employee dissent. Communication Research Reports, 17(1), 58-69. Kish-Gephart, J. J., Detert, J. R., Trevino, L. K., & Edmonson, A. C. (2009). Silenced by fear: The nature, sources and consequences of fear at work. Research in Organizational Behavior, 29, 163-193. Konovosky, M. A., & Cropanzano, R. (1991). Perceived fairness of employee drug testing as a predictor of employee attitudes and job performance. Journal of Applied Psychology, 76, 698-707. Kouzes, J. M. & Posner, B. Z. (1987). The leadership challenge. San Francisco: Jossey-Bass. Landau, J. (2009). When employee voice is met by deaf ears. S.A.M. Advanced Management Journal, 74(1), 4-28. LeBrenton, J. M., Hargis. M. B., Griepentog, B., Oswald, F. L., & Ployhart, R. E. (2007). A multidimensional approach for evaluating variables in organizational research and practice. Personnel Psychology, 60, 475 – 498. LePine, J. A., & Van Dyne, L. (1998). Predicting voice behavior in work groups. Journal of Applied Psychology, 83(6), 853-868. LePine, J. A., & Van Dyne, L. (2001) Voice and cooperative behavior as contrasting forms of contextual performance: Evidence of differential relationships with Big Five personality characteristics and cognitive ability. Journal of Applied Psychology, 86, 326-336. Liang, J., Farh, C. I. C., & Farh, J. L. (2012). Psychological antecedents of promotive and prohibitive voice: A two-wave examination. Academy of Management Journal, 55(1), 7192. 74 Lind, E. & Lissak, R. (1985). Apparent impropriety and procedural judgments. Journal of Experimental Social Psychology, 21, 19-29. Lind, E. & Tyler, T. (1988). The Social Psychology of Procedural Justice. New York: Plenum Press. Louis, M. (1980). Surprise and sense making: What newcomers experience in entering unfamiliar organizational settings. Administrative Science Quarterly, 25, 226-251. Luthans, F. (2002). The need for and meaning of positive organizational behavior. Journal of Organizational Behavior, 23 (6), 695-706. Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50, 370-396. Masten, A. S. (1994). Resilience in individual development: Successful adaptation despite risk and adversity. In. M. C. Wang (Ed.), Educational resilience in inner-city America: Challenges and prospects (pp. 3-25). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Masten, A. S. (2001). Ordinary magic: Resilience processes in development. American Psychologist, 56(3), 227. Masten, A. S., & Reed, M. G. (2002). Resilience in development. Handbook of Positive Psychology, 74-88. McClean, E. J., Burris, E. R., & Detert, J. R. (2013). When does voice lead to exit? It depends on leadership. Academy of Management Journal, 58(2), 525-548. McFarlin, D. B. & Sweeny, P. D. (1996). Does having a say matter only if you get your way? Instrumental and value-expressive effects of employee voice. Basic and Applied Social Psychology, 18(3), 289-303. Milliken, F. J., Morrison, E. W., & Hewlin, P. F. (2003). An exploratory study of employee silence: Issues that employees don’t communicate upward and why. Journal of Management Studies, 40, 1453-1476. Molm, L. D. (2003). Theoretical comparisons of forms of exchange. Sociological Theory, 21(1), 1-17. Morrison, E. W. (2011). Employee voice behavior: Integration and directions for future research. The Academy of Management Annals, 5(1), 373-412. Morrison, E. W. & Milliken, F. J. (2000). Organizational silence: A barrier to change and development in a pluralistic world. Academy of Management Review, 25, 706-725. Morrison, E. W. & Phelps, C. C. (1999). Taking charge at work: Extra-role efforts to initiate workplace change. Academy of Management Journal, 42, 403-419. 75 Mynes, T. D., & Podsakoff, P. M. (2013). Speaking more broadly: An examination of the nature, antecedents, and consequences of an expanded set of employee voice behaviors. Journal of Applied Psychology. Advance online publication. doi: 10.1037/a0034284 Nemeth, C. J. (1986). Differential contributions of majority and minority influence. Psychological Review, 93, 23-32. Nemeth, C. J., & Staw, B. M. (1989). The tradeoffs of social control and innovation in groups and organizations. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (pp. 175-210). New York: Academic Press. Ng, T. W., & Feldman, D. C. (2011). Employee voice behavior: A meta-analytic test of the conservation of resources framework. Journal of Organizational Behavior, 33, 216-234. Parsons, T. (1951). The Social System. Free Press. Pinder, C. C., & Harlos, K. P. (2001). Employee silence: Quiescence and acquiescence as responses to perceived injustice. Research in Personnel and Human Resources Management, 20, 331-369. Ployhart, R. E., Ryan, A. M., & Bennett, M. (1999). Explanations for selection decisions: Applicants’ reactions to informational and sensitivity features of explanations. Journal of Applied Psychology, 84(1), 87-106. Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437-448. Prussia, G. E., & Kinicki, A. J. (1996). A motivational investigation of group effectiveness using social-cognitive theory. Journal of Applied Psychology, 81(2), 187-198. Rice, B. (1987). Performance review: The job nobody likes. In D. Organ (Ed.), The applied psychology of work behavior. Plano, TX: BPI. Robin, J. Z., Brockner, J., Eckenrode, J., Enright, M. A., & Johnson-George, C. (1980). Weakness as strength: Test of a “my hands are tied” ploy in bargaining. Personality and Social Psychology Bulletin, 6, 216-221. Rosenberg, M. (1965). Rosenberg self esteem scale. doi: http://dx.doi.org/10.1037/t01038-000 Scott, M. B., & Lymann, S. M. (1968). Accounts. American Sociological Review, 33, 46-62. Senge, P. M. (1990). The fifth discipline: The art and practice of the learning organization. New York: Doubleday. 76 Shapiro, D. (1993). Reconciling theoretical differences among procedural justice researchers by reevaluating what it means to have one’s view “considered”: Implications for third-party managers. In R. Cropanzano (Ed.), Justice in the workplace (pp. 51-78). Hillsdale, NJ: Lawrence Erlbaum. Sharpiro, D. L., Buttner, E. H., & Barry, B. (1994). Explanations: What factors enhance their perceived adequacy? Organizational Behavior and Human Decision Processes, 58, 346368. Shaw, J. C., Wild, E., & Colquitt, J. A. (2003). To justify or excuse?: A meta-analytic review of the effects of explanations. Journal of Applied Psychology, 88, 444-458. Siebert, S. E., Crant, J. M., & Kraimer, M. L. (1999). Proactive personality and career success. Journal of Applied Psychology, 84, 416-427. Simmel, G. (1950). The Sociology of Georg Simmel. Free Press. Smeltzer, L. R., & Zener, M. F. (1992). Development of a model for announcing major layoffs. Group and Organizational Management, 17, 446-472. Sobel, M. E. (1982). Asymptomatic confidence intervals for indirect effects in structural equation models. In S. Leinhart (Ed.), Sociological methodology 1982 (pp. 290-312). San Francisco: Jossey-Bass. Stamper, C., & Van Dyne, (2001). Work status and organizational citizenship behavior: A field study of restaurant employees. Journal of Organizational Behavior, 22, 517-536. Stokes, R., & Hewitt, J. P. (1976). Aligning actions. American Sociological Review, 838-849. Tepper, B. J., Duffy, M. K., Hoobler, J., & Ensley, M. D. (2004). Moderators of the relationship between coworkers’ organizational citizenship behavior and fellow employees’ attitudes. Journal of Applied Psychology, 89, 455-465. Thibault, J. & Walker, L. (1975). Procedural Justice: A psychological analysis. Hillsdale, NJ: Elrbaum. Thurnwald, R. (1932). Economics in Primitive Communities. London: Oxford University Press. Tonidandel, S., LeBrenton, J. M., & Johnson, J. W. (2009). Determining the statistical significance of relative weights. Psychological Methods, 14(4), 387-399. Tugade, M. M., & Frederickson, B. L. (2004). Resilient individuals use positive emotions to bounce back from negative emotional experiences. Journal of Personality and Social Psychology, 86(2), 320-333. 77 Van Dyne, L., Ang, S., & Botero, I. C. (2003). Conceptualizing Employee Silence and Employee Voice as Multidimensional Constructs. Journal Of Management Studies, 40, 1359-1392. Van Dyne, L., Cummings, L. L., & Parks, J. M. (1995). Extra-role behaviors: In pursuit of construct and definitional clarity (A bridge over muddied water). Research in Organizational Behavior, 17, 215-285. Van Dyne, L., Kamdar, D., & Joireman, J. (2008). In-role perceptions buffer the negative impact of low LMX on helping and enhance the positive impact of high LMX on voice. Journal of Applied Psychology, 93, 1195-1207. Van Dyne, L., & LePine, J. A. (1998). Helping and voice extra-role behaviors: Evidence of construct and predictive validity. Academy of Management Journal, 41, 108-119. Vroom, V. H. (1964). Work and Motivation. New York: Wiley. Walumbwa, F. O. & Schaubroeck, J. (2009). Leader personality traits and employee voice behavior: Mediating roles of ethical leadership and work group psychological safety. Journal of Applied Psychology, 94, 1275-1286. Weiner, B., Folkes, V., Amirkhan, J., & Verette, J. (1987). An attributional analysis of excuse giving: Studies of a naive theory of emotion. Journal of Personality and Social Psychology, 52, 316-324. Westermarck, E. (1908). The Origin and Development of the moral Ideas. London: Macmillan. Whitener, E. M., Brodt, S. E., Korsgaard, M. A., & Werner, J. M. (1998). Managers as initiators of trust: An exchange relationship framework for understanding managerial trustworthy behavior. Academy of Management Review, 23(3), 513-530. Withey, M. J., & Cooper, W. H. (1989). Predicting exit, voice, loyalty, and neglect. Administrative Science Quarterly, 34, 521-539. Wong, C. & Weiner, B (1981). When people ask why questions. Journal of Personality and Social Psychology, 40, 650-663. Youssef, C. M., & Luthans, F. (2007). Positive organizational behavior in the workplace: The impact of hope, optimism, and resilience. Journal of Management, 33(5), 774-800. 78