Journal of Management Vol. XX No. X, Month XXXX 1–28 DOI: 10.1177/0149206314546195 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav 1 Group Ethical Voice: Influence of Ethical Leadership and Impact on Ethical Performance Lei Huang University of Nebraska-Lincoln Ted A. Paterson University of Idaho Interest in the important role that ethical leaders play in organizations has expanded in recent years because of several high-profile corporate ethical breakdowns and the increased responsibility placed upon corporate leaders as a result. In the present study, we introduce a new outcome of ethical leadership: group ethical voice. We further theorized and tested two mediating mechanisms linking ethical leadership with group ethical voice. Using two field studies and one experimental study, we found support for our assertion that ethical leadership was positively associated with group ethical voice. We also found support for most of our hypothesized mediating mechanisms (ethical culture and group ethical voice efficacy) linking ethical leadership with group ethical voice—except for the indirect effect of upper-level ethical leadership on group ethical voice via group ethical voice efficacy. We further found that group ethical voice positively influenced ethical performance (significant for the sales groups, marginally significant for the customer service groups). Contributions to both ethical leadership and voice literature are discussed along with the limitations of the current study and directions for future research. Keywords: ethical voice; ethical leadership; ethical voice efficacy; ethical performance A series of recent corporate ethical meltdowns have damaged trust in capitalism and forprofit enterprises across the globe (Yandle, 2010). The high-profile nature of the companies involved (e.g., Enron, Tyco, Lehman Brothers, etc.), coupled with the severe and widespread Acknowledgments: An earlier version of this paper was presented at the 2013 Annual Meeting of the Southern Management Association. We would like to thank our action editor, Karl Aquino, and two anonymous reviewers for their helpful and developmental feedback. Corresponding author: Lei Huang, Department of Management, College of Business Administration, University of Nebraska-Lincoln, Lincoln, NE 68588-0491, USA. E-mail: lhuanglinc@gmail.com 546195JOMXXX10.1177/0149206314546195Journal of Management / Month XXXXHuang, Paterson / Group Ethical Voice research-article2014 Downloaded from jom.sagepub.com by guest on April 3, 2015 2 Journal of Management / Month XXXX consequences of these ethical lapses, has led to a series of calls for reform. Despite regulations enforced by the government, serious concerns remain about the ethicality of business practices, thus prompting many to wonder what organizations can do to prevent such ethical meltdowns in the future. An expanding literature on ethical leadership offers some promising insights into the important impact that leaders can have in such situations (see Brown & Treviño, 2006, for a review). However, and somewhat surprisingly, with a few exceptions (e.g., Mayer, Aquino, Greenbaum, & Kuenzi, 2012; Schaubroeck et al., 2012), this literature has typically focused on outcomes that are not overtly ethical in nature, such as employees’ performance (Walumbwa, Morrison, & Christensen, 2012) and their satisfaction, motivation, and commitment (Brown & Treviño, 2006). While these prior studies have explained why ethical leadership matters, a renewed focus on meaningful ethical outcomes that can inform both ethical leadership theory and practice with effective ways to decrease unethical conduct in organizations is warranted. One such outcome that may promote the spread of ethicality from leaders throughout an organization and apprise organizational leaders of ethical lapses before they become crises is employee voice. There is evidence suggesting that ethical leadership encourages employees to speak up because their opinions are likely to be respected and valued by ethical leaders (Walumbwa & Schaubroeck, 2009). Yet these studies have focused on voice in general without specifying the content of voice. Thus, it remains unknown whether or not ethical leadership impacts employees’ ability to speak up regarding ethical issues in the workplace. To explore this question, in the current study, we introduce the concept of group ethical voice, which we define as a form of group expression that challenges, and seeks to change, the current behaviors, procedures, and policies that are not normatively appropriate. We focus on voice at the group level because prior research indicates that group voice offers protection to the individual, reflects shared concerns from a work unit, delivers greater social influence, and thus is more likely to attract attention from management (Frazier & Bowler, in press; Walumbwa et al., 2012). We ground this new concept in the existing voice literature and specify that ethical voice is concerned with change-oriented suggestions that seek to challenge existing ethical lapses so as to inform the management of potential ethical catastrophes before they occur. In addition to describing why ethical leadership influences group ethical voice, we also strive to understand how such effects happen. Recent voice research has favored the perspective that voice behavior is more likely to occur when voice agents have normative beliefs and control beliefs about voice behaviors (Liang, Farh, & Farh, 2012). Thus, in our framework connecting ethical leadership to group ethical voice, we introduce mediating mechanisms that address each of these two factors. First, to account for normative beliefs, we refer to ethical culture, which has been discussed as shared normative beliefs regarding ethical and unethical conduct and is believed to aid in the “transmission of ethical leadership in a manner that influences important ethics-related cognitions and behaviors of followers within and between organizational levels” (Schaubroeck et al., 2012: 1058). Due to its relationship with ethical leadership and its role in influencing ethical behaviors, we suggest that ethical culture helps transmit the influence of ethical leadership into the enactment of group voice behaviors. Second, we posit that shared efficacy beliefs among group members regarding speaking up on ethical issues can be considered a motivating factor that transmits ethical leadership to group ethical voice. According to Bandura (1982), efficacy belief refers to one’s estimation Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 3 of his or her capability of attaining desired outcomes. Efficacy beliefs have been shown to influence a number of organizational behaviors, including creativity (Tierney & Farmer, 2002), proactivity (Parker, 1998), and leadership (Hannah, Avolio, Luthans, & Harms, 2008). Thus, we propose that shared beliefs held by group members about whether the group as a whole is able to challenge the current behaviors, procedures, and policies that are not normatively appropriate, which we termed group ethical voice efficacy, will transmit the effect of ethical leadership on group ethical voice. We also demonstrate why group ethical voice should be of interest to organizational researchers and practitioners by testing its relationship with ethical performance (Gatewood & Carroll, 1991). Drawing from social cognitive theory (Bandura, 1986), we posit that the enactment of group ethical voice can produce important cues shaping the importance of ethical conduct in one’s value system. This can motivate group members to regulate their workrelated behaviors in a way that aligns with such ethical values. That is, as work groups speak up about ethical issues, there is heightened awareness of such issues, and therefore, compliance with ethical norms is likely to occur. In sum, the primary purpose of the present study is to introduce the concept of group ethical voice. In doing so, we examine how ethical leadership can influence it, both directly and indirectly via multiple mediating mechanisms, as well as whether group ethical voice influences ethical performance. Our research has a number of contributions to the ethical leadership and voice literatures. First, we heed the call from Morrison (2011) to specify the content of voice by defining an ethical dimension of group voice behaviors. We also broaden the scope of the voice construct by showing that ethical voice can positively influence ethical performance. Second, we draw on social learning theory (Bandura, 1977) to explain why ethical leadership directly facilitates group voice behaviors related to ethical issues and therefore provide an outcome of ethical leadership that transmits its influence broadly. Third, based on social cognitive theory (Bandura, 1986), we also describe how ethical leadership indirectly affects group ethical voice via ethical culture and group ethical voice efficacy. By so doing, we enhance our understanding of efficacy beliefs at the group level. We illustrate our proposed research model in Figure 1. The Foundation of Group Ethical Voice Despite the different terminologies used in the literature to refer to the construct of voice, voice is prosocial in nature, with the organization as the beneficiary in most cases (Burris, Detert, & Romney, 2013; Klaas, Olson-Buchanan, & Ward, 2012; Van Dyne, Ang, & Botero, 2003). Voice occurs when employees speak up with change-oriented comments and suggestions to improve organizational functioning (Van Dyne & LePine, 1998). Despite a growing and diverse literature on voice, a need in the current voice literature, as noted in Morrison’s (2011) review, is to extend and specify the content of voice. Due to increasing concerns about ethical issues in organizations and the ethical dilemmas that confront employees, we suggest that exploring the ethical content of voice behaviors is a worthwhile endeavor. While voice behaviors in general aim at improving performance (Morrison, 2011, 2014), ethical voice consists of suggestions aimed at improving ethical decision making and behaviors. As such, ethical voice is a subset of general voice behaviors and involves employees’ perceptions and belief regarding what is right and what is wrong. Similar to prosocial voice, ethical voice Downloaded from jom.sagepub.com by guest on April 3, 2015 4 Journal of Management / Month XXXX involves individuals’ evaluations of the risks associated with speaking up (Burris, 2012; Detert & Burris, 2007) but also requires individuals’ evaluation of their moral obligation to speak up about ethical issues in the workplace. In other words, when individuals perceive the risks associated with voice behaviors, they may still speak up because of their commitment to defending their moral standards or values (Gentile, 2010). Although it has been investigated primarily at the individual level, voice behavior has recently received research attention at the group level (e.g., Frazier & Bowler, in press; Walumbwa et al., 2012). Group voice refers to the proposal of constructive suggestions, new ideas, and advice by a workgroup (Frazier & Bowler, in press). Prior research suggests that group voice may be more effective than individual voice because it reflects shared concerns from the focal work environment, delivers greater social influence, and thus is more likely to attract attention from the management team (Frazier & Bowler, in press; Walumbwa et al., 2012). We suggest that each of these reasons is even more important within the context of ethical issues due to their highly sensitive nature. Indeed, acting as a sole voice in speaking out against perceived ethical lapses involves substantial personal risk for a single individual (Treviño & Brown, 2004). In an interview study conducted by Milliken, Morrison, and Hewlin (2003), 20% of the interviewees stated that they would not speak up to their commanding superiors about ethical issues primarily due to the fear of being labeled negatively, of damaging interpersonal relationships, and of being the target of retaliation or punishment. In this regard, speaking up as a group to challenge ethical issues is much more likely to protect individuals from these adverse consequences of speaking up, because it prevents a single group member from being the primary target of negative responses. Therefore, the group may be a “safer vehicle” for representing and eventually communicating concerns about unethical conduct than the individual. We define group ethical voice as a form of group expression that challenges, and seeks to change, the current behaviors, procedures, and policies that are not normatively appropriate. We posit that the work group functions as an arena for raising and discussing ethical issues Figure 1 Proposed Research Model Upper-level Manager’s Ethical Leadership Lower-level Manager’s Ethical Leadership Ethical Culture Group Ethical Voice Group Ethical Voice Efficacy Level 1 Level 2 Group Ethical Performance H1 H1 H2 H3 H4 Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 5 because it is a social platform through which employees can exchange ideas about what constitutes an ethical issue, which ethical issues are of interest to the group, and how, when, and to whom to speak up about these issues. Indeed, social interactions in a group setting can help to develop shared perceptions about speaking up (Klaas et al., 2012; Morrison, WheelerSmith, & Kamdar, 2011). Employees may make sense about what behaviors are more acceptable and whether in this particular group speaking up is encouraged or discouraged (Morrison & Milliken, 2000). Such shared perceptions, particularly in situations where employees are confronted with ethical dilemmas or issues, may help them make decisions about speaking up. Having established the basic concepts regarding the group ethical voice construct, we now proceed to explore its relationship with other constructs. In the following sections, we will explain why and how ethical leadership can influence the ethical voice of a group, including two mediating mechanisms, and how ethical voice contributes to ethical performance. Ethical Leadership and Group Ethical Voice The most widely cited definition of ethical leadership is provided by Brown, Treviño, and Harrison (2005) as “the demonstration of normatively appropriate conduct through personal actions and interpersonal relationships, and the promotion of such conduct to followers through two-way communication, reinforcement, and decision-making” (p. 120). This definition and formulation of ethical leadership draws upon Bandura’s (1977) social learning theory to explain that followers’ ethical behavior is learned, in part, by their observation, retention, and reproduction of their leaders’ ethical behaviors. In the present study, we argue that the social learning processes help to activate group ethical voice. In organizational contexts, employees tend to evaluate the favorableness of the context (i.e., whether it promotes or hinders speaking up) primarily through observations of and interactions with their leaders (Burris, Detert, & Chiaburu, 2008; Van Dyne, Kamdar, & Joireman, 2008). Therefore, leadership behaviors are considered key factors influencing employees’ evaluations of the risk inherent in speaking up, the specific message they desire to communicate, and their decisions to speak up (Detert & Burris, 2007). Ethical voice involves employees making judgments about the nature of the events they observe in the workplace and determining if there is a need to correct normatively inappropriate behaviors by speaking up. In this regard, ethical leaders provide important clues to their subordinates about normatively appropriate behaviors through their personal ethical conduct and promotion of ethical conduct to others (Brown et al., 2005). Moreover, because of their commitment to two-way communication and listening to their subordinates (Brown et al., 2005), they also empower work groups to speak up. As a result, ethical leaders facilitate not only an environment in which voice behaviors are accepted but one in which agreement on ethical issues is likely to be reached. Moreover, prior research has suggested that employees tend to remain silent to avoid unfavorable reactions from management if they do not know whether speaking up to challenge the status quo is acceptable in the work environment (Morrison, 2014; Morrison & Milliken, 2003). This is understandable since it is often the manager in an organization—particularly, those at the top—who sets the overall tone in the organization with respect to the standard of ethical conduct and people’s perceptions about whether speaking up to challenge the status quo is likely to be accepted (Detert & Treviño, 2010). Quite often these managers also serve as role models indicating a moral standard of appropriate work behaviors in Downloaded from jom.sagepub.com by guest on April 3, 2015 6 Journal of Management / Month XXXX organizations (Andrews, 1989; Falkenberg & Herremans, 1995; Mayer et al., 2012). Therefore, leaders’ behaviors and communication are factored in when groups develop shared beliefs about whether speaking up is safe or acceptable (i.e., group voice safety beliefs, Morrison et al., 2011). By observing and modeling their managers’ ethical leadership, group members should be motivated to discuss and internalize the meaningfulness of promoting ethical values in work practices (Mayer et al., 2012; Walumbwa et al., 2012). Thus under the supervision of ethical leaders, groups are more likely to speak up to challenge and seek to correct ethical problems. We thus hypothesize the following: Hypothesis 1: Ethical leadership is positively associated with group ethical voice. The Mediating Role of Ethical Culture While ethical leadership is expected to activate group ethical voice directly through rolemodeling processes, we also suggest that ethical leadership may indirectly influence group ethical voice. Specifically, prior voice research has suggested that the normative beliefs held by employees regarding whether speaking up is encouraged and favored (i.e., voice climate) can determine the occurrence of voice behaviors (Morrison et al., 2011). Other studies have also shown that ethical leadership plays a role in forming the ethical culture of work group or unit (e.g., Schaubroeck et al., 2012). Combining these two streams of thought, we propose that ethical culture transmits the influence of ethical leadership to motivate ethical voice in the workplace. Ethical culture is defined as “a subset of organizational culture, representing a multidimensional interplay among various ‘formal’ and ‘informal’ systems of behavioral control that are capable of promoting either ethical or unethical behavior” (Treviño, Butterfield, & McCabe, 1998: 451-452). The “formal” component of the ethical culture may include organizational policies as well as work procedures and practices that are enforced under the code of ethics, leadership structures that are designed for minimizing the occurrence of unethical conduct, reward systems that recognize the value of ethical behaviors, and specialized training programs that help to raise organizational members’ awareness of ethical issues (Schaubroeck et al., 2012; Treviño, 1990). The “informal” component is conveyed through personal stories that indicate the occurrence of ethical or unethical conduct or through peers’ influence via the use of language or behaviors reflective of their attitudes and values toward ethical conduct (Ardichvili, Mitchell, & Jondle, 2009; Falkenberg & Herremans, 1995). Both of these components refer to the surface-level shared cultural elements (Schein, 1985) that are observable and interpretable by individuals in the work environment. These shared cultural elements can inform organizational members of conduct that is desirable or undesirable in the work environment (Treviño et al., 1998). An important characteristic of ethical culture is that the shared understandings regarding ethical norms in a work environment can be shaped by external factors that affect organizational members’ cognitions about ethical or unethical conduct (Treviño et al., 1998). Prior research has found a positive impact of ethical leadership on ethical culture (e.g., Schaubroeck et al., 2012) because leadership behaviors, through social learning processes (Bandura, 1977), affect how followers evaluate what is right and what is wrong in certain situations (Mayer, Kuenzi, Greenbaum, Bardes, & Salvador, 2009; Treviño & Brown, 2004). Leaders also tend to embed their values and implicit beliefs into the shared assumptions held at the Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 7 level of work unit through their handling of critical incidents, which is informative of leaders’ attitudes and values toward ethical conduct, or through sharing personal stories or past experiences, which can reflect leaders’ moral judgments about exemplary ethical or unethical behaviors (Schaubroeck et al., 2012; Schein, 1985). Through these mechanisms, ethical leadership helps to create shared understanding among group members regarding the normatively appropriate or inappropriate conduct. Therefore, consistent with prior research, we expect a positive relationship between ethical leadership and ethical culture. We also expect that ethical culture mediates the relationship between ethical leadership and group ethical voice. Specifically, prior research has found that a stronger ethical culture helped to increase work groups’ intentions to report unethical behaviors at work as they see eye to eye regarding how they should act in situations involving ethical issues (Schaubroeck et al., 2012). Therefore, ethical culture should encourage work groups to actively inform management of ethical issues that are in need of their attention. As described in prior research (Falkenberg & Herremans, 1995; Treviño, 1990), the power of organizations as a formal system can encourage individuals and work units within the system to comply with behaviors that are expected by the system. Thus, we argue that in a strong ethical culture, work groups will challenge unethical conduct as they see doing so as a form of compliance with organizational expectations. These above explanations are consistent with social learning theory (Bandura, 1977). Specifically, ethical culture can be a motivating factor for work groups to retain and reproduce ethical conduct they observe from their managers’ ethical leadership. As described in prior research, the effect of role modeling and vicarious learning processes can be reinforced when leadership behaviors produce stimuli that help observers to internalize their observation and reflect it in their own regulation of self-directed behaviors (Manz & Sims, 1981). In this regard, ethical leadership influences ethical voice via the activation of ethical culture in the workplace. We thus expect that ethical leadership positively and indirectly affects group ethical voice by enabling an ethical culture in the organization. Hypothesis 2: Ethical culture mediates the positive relationship between ethical leadership and group ethical voice. The Mediating Role of Ethical Voice Efficacy While ethical culture indicates the favorableness of the environment for ethical voice, we argue that work groups’ judgment regarding their ability to deliver their concerns about ethical issues also influences whether or not they speak up. As described in Ajzen (1991), behaviors are determined not only by normative beliefs (e.g. ethical culture) held by individuals in a given environment but also by control beliefs, which reflect the extent to which one believes that such behavior will result in desirable outcomes. This control belief is conceptually related to the concept of self-efficacy, which refers to one’s judgment about “how well one can execute courses of action required to deal with prospective situations” (Bandura, 1982: 122). At the group level, self-efficacy has been specifically applied to the domain of voice behavior such that it refers to a “shared belief about whether group members are able to voice effectively” (Morrison et al., 2011). In the present study, we extend the group-level voice efficacy construct by specifying it in the domain of ethics. We posit that group ethical voice efficacy indicates a group’s collective confidence about speaking up on ethical issues. Downloaded from jom.sagepub.com by guest on April 3, 2015 8 Journal of Management / Month XXXX We propose that ethical leadership will have a positive impact on group ethical voice efficacy. Specifically, management typically represents the standards of ethical values deemed appropriate by the organization (Mayer et al., 2009). According to social cognitive theory (Bandura, 1986), this representation of ethical values from management should affect how group members develop efficacy beliefs regarding their behaviors in alignment with those ethical values. This takes place through interactions with their managers wherein work groups perceive ethical leaders as ones who respect the values of integrity and honesty, recognize work inputs reflecting consistent moral standards, and encourage conversations promoting ethical conduct (Brown & Treviño, 2006). Therefore, the observational learning process can help to verify groups’ judgments regarding whether the reproduction of behaviors promoting ethical values in the workplace will be accepted and encouraged by their leaders. As a result of that, work groups should be more likely to develop stronger efficacy beliefs regarding their capability to engage in behaviors that challenge unethicality when they are supervised by ethical leaders. Following prior literature that found a positive relationship between efficacy and domainspecific behaviors (e.g., Tierney & Farmer, 2002), we also propose that groups are more likely to speak up regarding ethical issues when they have stronger ethical voice beliefs. Prior research has suggested that elevated efficacy beliefs often lead to greater confidence in regulating one’s behaviors in a way that can maximize the chance of attaining goals (Bandura, 1989). Stronger efficacy beliefs have also been shown to help people sustain their efforts to accomplish their goals, a self-regulated process through which people focus on resolving identified problems instead of avoiding doing so due to obstacles or difficulties (Wood & Bandura, 1989). Therefore, higher group ethical voice efficacy reflects shared beliefs among group members regarding their ability to speak up about unethical conduct and can be a source of motivation for work groups as they consider raising ethical issues with management. These groups are also more likely to demonstrate sustaining efforts to challenge ethical problems they observe in the workplace. Therefore, we expect that ethical leadership indirectly influences group ethical voice via group ethical voice efficacy. We thus hypothesize the following: Hypothesis 3: Group ethical voice efficacy mediates the positive relationship between ethical leadership and group ethical voice. Ethical Performance as Outcome of Ethical Voice So far, we have explained why and how ethical leadership can encourage group ethical voice in the workplace. We further contend that group ethical voice also has meaningful impact on organizational outcomes. While voice behavior is often seen as a desirable outcome in and of itself, this is largely because it is assumed that voice leads to change in workplace behaviors that can positively contribute to organizational functioning (Morrison, 2011). In the case of group ethical voice, we posit that this particular brand of voice behaviors will have a positive impact on the ethical performance of employee groups. Both corporate and social interests in ethical performance have grown substantially in the past 20-plus years because organizations have realized that behaviors of low credibility (e.g., misrepresentation of information, breach of integrity) from their employees damage the whole value system of an organization in the long run (Becker, 1998; Brooks, 1989). This practical concern has led the assessment of ethical performance to go beyond the Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 9 organizational level—where ethical performance refers to performance achievements in line with corporate code of ethics (Brooks, 1989)—toward one at lower levels in an organization. As Gatewood and Carroll (1991) suggested, ethical performance can be conceptualized and measured at the group level, where it reflects compliance with certain regulatory, professional, or organizational standards while one’s obligation to moral standards are also met. While prior studies have relied on different normative frameworks to develop the ethical performance construct, in this paper, we adopt a deontological approach to conceptualize ethical performance wherein universal duties and rights are of primary concern (Treviño & Nelson, 2014). This approach aligns with the “rights” approach described in DeGeorge (1990) and Gatewood and Carroll (1991) such that behavioral outcomes can be considered ethical when such outcomes protect certain rights of the affected parties (e.g., a customer’s right to be treated with respect). Consistent with Kant’s theory of ethics, the deontological approach to ethical performance assumes that there are universal laws that should govern human behavior and that the intent of the individual in carrying out their actions is the primary determinant of ethicality. This is in contrast to the consequentialist approach, which focuses on the consequences of behavior to determine if it is ethical or not, and virtue ethics, which emphasizes one’s character in determining ethical behavior (Treviño & Nelson, 2014). Thus, in this study, we conceptualize the ethical performance construct in a way that is consistent with a deontological approach, placing primary emphasis on the work groups’ effectiveness in respecting and protecting the rights of those they come in contact with and their ability to comply with the duties expected of them. While it seems logical that group ethical voice and ethical performance would be related in some way, the nature of the relationship could be contested. One perspective argues that group ethical voice behaviors are an indication of an unethical work environment (because there are ethical lapses that people are speaking up about), and therefore, one would expect a negative relationship. However, another perspective argues that because group ethical voice behaviors are occurring, unethical conduct is being rooted out of the organization, and therefore, we would expect a positive relationship between group ethical voice and ethical performance. We adopt the latter perspective and ground our theorizing in the social cognitive theory. Per this theory, human behavior is governed by intentionality and self-control (Bandura, 2001), which can be regulated through either self-influence or what Wood and Bandura (1989) called “emergent interactive agency”—a system of interactional influence coming from proximal agents. Consistent with the latter, when a group speaks up to challenge ethical issues, this group-level behavior influences intentions to engage in subsequent behaviors that are consistent with the ethical beliefs defended by speaking up. That is, group ethical voice behaviors can function as a group-level self-regulating mechanism that motivates group members to reflect on the consensus or the majority view of the group and therefore to align their own behaviors with the moral standards represented in their voice behaviors. Therefore, we propose that group ethical voice positively influences group ethical performance. Hypothesis 4: Group ethical voice is positively associated with ethical performance. Overview of Methods So far, we have proposed a direct relationship between ethical leadership and group ethical voice, two mediating mechanisms linking this relationship, and a positive relationship between Downloaded from jom.sagepub.com by guest on April 3, 2015 10 Journal of Management / Month XXXX ethical voice and ethical performance. In the following sections, we conducted two field studies and an experiment to test the proposed relationships. Specifically, given that ethical voice is a proposed extension of the voice construct, we first used Study 1a to test the construct validity of group ethical voice and to test the direct relationship between ethical leadership and group ethical voice. Following Study 1a, Study 1b involved an experiment wherein we attempted to demonstrate the causality of the ethical leadership–ethical voice relationship, which cannot be verified by the cross-sectional design used in Study 1a. In the next step, we used Study 2 to test the complete proposed model with all hypothesized relationships, including the direct relationships between ethical leadership and group ethical voice, the two proposed mediating mechanisms, and ethical performance as the outcome of group ethical voice. In both Study 1a and Study 2, while hypothesizing about ethical leadership and its relationship with other constructs in general, we operationalized ethical leadership at two levels, upper level (i.e., store managers) and lower level (i.e., group managers), due to the nature of our samples. We suggest that this is also consistent with the nature of the setting where store managers and group managers are both references of ethical leadership behaviors. Study 1a Sample and Procedures We invited 907 employees from 42 chain stores of a consumer electronics company in China to participate in this study. Each store is relatively small in size so that store managers (hereafter referred to as upper-level managers) often have direct interactions with group members. This setting enables employees to observe and interact with their upper-level manager and functional group manager (hereafter referred to as lower-level manager) in their daily jobs. The invited participants work in different areas, such as sales, customer service, and logistics. We collected two waves of survey data. Ten percent of all the respondents received a supermarket promotional card valued at 100 RMB (approximately $15) as a lottery bonus for participating. At Time 1, we asked 155 lower-level managers to rate their upper-level managers’ ethical leadership and collected 139 responses (89.7%). Nine hundred seven employee participants were also asked to rate their lower-level managers’ ethical leadership. Seven hundred forty-five responses were collected (82.1%). At Time 2, which was about 6 weeks after we collected employee responses, we asked one assistant store manager from each of the 42 stores to provide ratings of group ethical voice. Forty of them returned their completed surveys (95.2%). We were then able to match up these 40 responses with those from 137 lower-level managers and 696 employees. After omitting groups with unsatisfactory response rates (i.e., 60%; Roth & BeVier, 1998; Timmerman, 2005), we obtained a final sample of 689 employees nested in 134 groups from 40 stores. Group size ranged from 4 to 9 members, with an average of 5.1 members. Among the 689 employee respondents, 424 were female (61.5%). The average age and organizational tenure were 26.2 and 4.6 years, respectively. Among the 134 lower-level managers included in our final sample, 21 were female (15.7%). The average age and organizational tenure were 33.3 and 5.0 years, respectively. Among the 40 upperlevel managers, 17 were female (43%). The average age and organizational tenure were 38.5 and 9.3 years, respectively. Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 11 Measures One of the authors, a bilingual in Mandarin and English, translated the survey material and measures from English into Chinese. We then followed the standard back-translation procedure (Brislin, Lonner, & Thorndike, 1973) to ensure that the English and Chinese versions of the survey material and scales were appropriately translated. We used a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) to measure our study variables. Ethical leadership. We assessed ethical leadership using the 10-item Ethical Leadership Scale (ELS) developed and validated by Brown et al. (2005). Following previous ethical leadership research, we aggregated lower-level managers’ evaluations of their upper-level managers to obtain an upper-level manager’s ethical leadership measure (Mayer et al., 2009, 2012). Similarly, we aggregated employees’ responses to obtain the ratings of lower-level ethical leadership, but we changed the referent from “my store manager” to “my group manager.” A sample item was “My store (group) manager discusses business ethics or values with employees.” The Cronbach’s alpha was .93 for upper-level ethical leadership and .92 for lower-level ethical leadership. Group ethical voice. We measured group ethical voice by adapting the six-item voice behavior scale developed by Van Dyne & LePine (1998). Specifically, we asked the assistant store managers to assess each group’s voice behavior with a particular focus on issues of workplace ethics. We changed the reference from “this employee” to “this group.” The items were “This group develops and makes recommendations concerning ethical issues that affect their work,” “This group speaks up and encourages other units in the organization to get involved in issues that affect the business ethics of the work environment,” “This group communicates its opinions about ethical issues to others even if its opinion is different and others disagree with it,” “This group keeps well informed about ethical issues where the group’s opinion might be useful,” “This group gets involved in ethical issues that affect the quality of work life,” and “This group speaks up with ideas for initiatives or changes in procedures to improve ethics in the organization.” The Cronbach’s alpha was .93. Control variables. Following prior research investigating multilevel ethical leadership (Mayer et al., 2009), we controlled for age, gender, and organizational tenure of both upperlevel managers and lower-level managers. We also controlled for group size because it is considered a factor affecting leaders’ ability to influence follower behaviors (Walumbwa & Schaubroeck, 2009) and group-level voice (Walumbwa et al., 2012). In addition, we controlled for the type of group (e.g., sales, customer service, and logistics). Measurement and Analytic Procedures Data aggregation. Following suggestions for conducting multilevel research (Kozlowski & Klein, 2000), we first evaluated the appropriateness of data aggregation for ethical leadership rated by both lower-level managers and employee respondents. Specifically, we first calculated within-group interrater agreement (rwg[j] ; James, Demaree, & Wolf, 1984) to justify aggregating employee ratings of ethical leadership to the lower level and lower-level Downloaded from jom.sagepub.com by guest on April 3, 2015 12 Journal of Management / Month XXXX managers’ ratings of ethical leadership to the upper level. The mean rwg(j) values were .87 (ranged from .81 to .95) for upper-level ethical leadership (rated by lower-level managers) and .84 (ranged from .72 to .96) for lower-level ethical leadership (rated by employee respondents), both exceeding the suggested cutoff value of .70 (Chen, Mathieu, & Bliese, 2004). This indicated that respondents in our sample had sufficient agreement on their leaders’ ethical leadership. We further calculated the intraclass correlation coefficient (ICC; Bliese, 2000) to determine if there was sufficient between-group variance for the ethical leadership measures. The ICC values were as follows: lower-level ethical leadership (ICC1 = .26, ICC2 = .64) and upper-level ethical leadership (ICC1 = .33, ICC2 = .63). These ICC values fell into an acceptable range for aggregation (James, 1982). Therefore, we found support for data aggregation of ethical leadership measures. Analytic strategy. Given the multilevel nature of our data, we used multilevel modeling techniques (Hofmann, 1997) through HLM 7 (Raudenbush, Bryk, Cheong, Congdon, & du Toit, 2011), which accounts for potential biased estimates of parameters and errors due to nonindependence of observations (Bliese & Hanges, 2004; Klein & Kozlowski, 2000). Prior to testing the relationship between ethical leadership and ethical voice, using Mplus 6.1 (Muthén & Muthén, 2010), we performed a series of confirmatory factor analyses (CFAs) to estimate the fit of the measurement model and validate our proposed construct of ethical voice. Results Measurement model tests. We performed a set of CFAs to examine the factor structure of all study variables. We assessed the fit of the measurement model using chi-square statistics and multiple fit indices (Hu & Bentler, 1999), including comparative fit index (CFI), TuckerLewis index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR). The CFA results suggested that our proposed construct of group ethical voice produced acceptable fit: χ2 = 13.01, df = 9, CFI = 1.00, TLI = .99, RMSEA = .06, SRMR = .02. In addition, we also estimated the fit of our three-factor baseline measurement model in which ethical leadership at upper and lower levels and group ethical voice were each loaded on a unique factor. The CFA results showed that this baseline measurement model also produced acceptable fit: χ2 = 618.20, df = 296, CFI = .96, TLI = .95, RMSEA = .04, SRMR = .03. We compared this baseline model with an alternative model in which we loaded items of upper- and lower-level ethical leadership on one single latent factor. The chi-square difference test showed that our baseline model produced significantly better fit with the data: ∆χ2(2) = 670.69, p < .01. Preliminary tests. Means, standard deviations, and correlations of study variables are reported in Table 1. As expected, both upper-level (r = .41, p < .01) and lower-level ethical leadership (r = .26, p < .01) correlated significantly with group ethical voice. All the control variables except for upper-level managers’ organizational tenure and lower-level managers’ age were significantly correlated with one or more primary study variables. We thus excluded these two control variables when testing the model in hierarchical linear modeling (HLM). Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 13 As reported in Table 2, the results of our preliminary tests showed that both upper-level (γ = .34, p < .01) and lower-level ethical leadership (γ = .32, p < .01) were positively associated with group ethical voice. These results thus rendered preliminary support for Hypothesis 1. In Study 1a, we have verified the construct validity of group ethical voice and provided initial evidence for a positive relationship between ethical leadership and group ethical voice. In the next study, we use an experimental design to demonstrate the causal impact of ethical leadership on ethical voice. Study 1b Sample and Procedures Participants consisted of 305 undergraduate students in a large, midwestern U.S. university who were invited to participate in this study as part of their enrollment in a multisection undergraduate management course. The average age of participants was 21.5 years. One hundred two of them (33.4%) were female. All of them reported having full- or part-time job or internship experience, while 194 of them (63.6%) were employed at the time of this experiment. Two hundred forty-five of these participants (80.3%) reported work experience of more than 6 months. The participants were configured into 81 groups with an average group size of 3.77 participants. Upon arriving at the location of the experiment, participants were randomly assigned to groups consisting of 3 or 4 each and were briefed about the procedures of the experiment. Table 1 Means, Standard Deviations, and Intercorrelations of Variables for Study 1a Variable 1 2 3 4 5 6 7 8 9 10 11 1. UM’s age — 2. UM’s gender .10 — 3. UM’s org tenure .33 .06 — 4. LM’s age −.15 .10 .08 — 5. LM’s gender −.16 −.05 −.00 −.13 — 6. LM’s org tenure −.07 −.00 −.07 .54 −.07 — 7. Group size .10 −.02 −.11 −.04 .12 −.00 — 8. Group type −.06 .01 −.05 .07 −.15 .02 −.50 — 9. UM’s ethical leadership −.10 .03 .04 .04 −.05 −.17 −.14 .15 (.93) 10. LM’s ethical leadership .19 .05 −.01 −.01 .19 −.03 −.08 .00 .27 (.90) 11. Group ethical voice −.08 .22 .01 .01 −.12 .03 .00 .08 .41 .26 (.93) M 38.46 1.58 9.25 33.34 1.84 5.00 5.05 1.51 3.89 3.64 3.79 SD 5.24 0.50 4.60 3.96 0.37 3.20 1.39 0.73 0.84 0.48 0.83 Note: N = 689. Group-level units = 134. Store-level units = 40. Gender was coded as 1 = female, 2 = male. Group type was coded as 1 = sales, 2 = customer service, 3 = logistics. UM = upper-level manager; LM = lower-level manager; org = organizational. Coefficient alphas for scales are provided in parentheses. Correlations with absolute values of .08 (.14 for upper level) and above are significant at p < .05. Correlations with absolute values of .11 (.19 for upper level) and above are significant at p < .01. Two-tailed test. Downloaded from jom.sagepub.com by guest on April 3, 2015 14 Journal of Management / Month XXXX All participants first completed a brief survey, which consisted of demographic variables. Upon completing the demographic survey, the students were asked to read a short case. The case was adapted from “A Billing Blind,” a case available on the Giving Voice to Values website. The case describes a group of summer interns at a midsized accounting firm tasked with digitizing client files. The interns are asked to bill clients for more hours than the work actually takes and to do so at the partners’ billable rate even though the work was done by the interns. There were a total of four different versions of the case, allowing us to manipulate two variables: ethical leadership and leader’s level in the hierarchy. To manipulate ethical leadership, the case described either a leader who exemplified (high) each of the items in the ELS (Brown et al., 2005) or one who did not (low). The manipulation for leader’s level of hierarchy consisted of the leader’s title being either CEO (high) or team leader (low). There were 20 groups assigned to each manipulation with the exception of the manipulation with high ethical leadership and low level of hierarchy, which had 21. After reading the case, participants were asked to answer the manipulation check questions on their own. Participants were then asked to openly discuss the case together in groups and discuss how likely they were to speak up as a group to challenge the unethical request. Finally, after the group discussion ended, we asked each participant to complete questions of group ethical voice. Table 2 Results of Hierarchical Linear Modeling for Study 1a Dependent Variables LM’s Ethical Leadership Group Ethical Voice Variable Model 1 Model 2 Model 3 Model 4 Model 5 Intercept 3.65** 3.65** 3.81** 3.81** 3.81** Store-level control variables UM’s age .02 .02* −.02 −.02 −.01 UM’s gender .06 .02 .20 .18 .18 Store-level predictor UM’s ethical leadership .17** .34** Group-level control variables LM’s gender .17 .15 −.10 −.15 −.15 LM’s organizational tenure .00 .01 .01 .02 .02 Group size −.07 −.06 .04 .06 .06 Group type −.04 −.05 .15 .15 .15 Group-level predictors LM’s ethical leadership .32** .31** R2 .05 .16 .05 .19 .25 ∆R2 .11 .14 .06 Note: N = 134 (nested in 40 stores). UM = upper-level manager; LM = lower-level manager. R2 values were calculated using the formula developed by Snijders and Bosker (1999). *p < .05. **p < .01. Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 15 Measures To ensure the effectiveness of the ethical leadership manipulation, we asked participants to complete measures of ethical leadership using the 10-item ELS scale (Brown et al., 2005). Answers were given on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). The Cronbach’s alpha was .95. Given the high levels of agreement (mean rwg[j] = .91) in group members’ reported ethical leadership and the acceptable aggregate reliability (ICC1 = .42, ICC2 = .73), we were able to justify the aggregation of individual ratings of ethical leadership. To ensure the effectiveness of the manipulation of leaders’ level in the hierarchy, we asked participants to answer which role the leader in this case played (CEO vs. team leader). We assessed group ethical voice using the six-item group ethical voice scale described in Study 1. One sample item was “This group would develop and make recommendations concerning the ethical issue discussed in the case study” (1 = strongly disagree, 5 = strongly agree). The Cronbach’s alpha was .94. Given the high levels of agreement (mean rwg[j] = .92) in group members’ reported group ethical voice and the acceptable aggregate reliability (ICC1 = .34, ICC2 = .66), we were able to justify the aggregation of individual ratings of group ethical voice to the group level. Results Manipulation check. We conducted two sets of ANOVA for manipulation checks of leaders’ ethical leadership and their level in the hierarchy. First, participants assigned to conditions in which they read the case of an ethical leader reported significantly higher ratings of ethical leadership than those assigned to conditions in which they read the case of an unethical leader: F(1, 79) = 653.644, p < .001, η2 = .89 (ethical, M = 3.80, SD = 0.39; unethical, M = 1.77, SD = 0.32). Second, the results also showed that participants assigned to conditions in which they read the case of the leader being a CEO reported higher ratings on the role of the leader (coded 1 = team leader, 2 = CEO) than those assigned to conditions in which they read the case of the leader being a team leader: F(1, 303) = 581.231, p < .001, η2 = .66 (ethical, M = 1.87, SD = 0.34; unethical, M = 1.06, SD = 0.25). These results indicated successful manipulations. Main tests. Since we operationalized ethical leadership at upper (CEO) and lower (team manager) levels, we first tested if level of leaders’ hierarchy would moderate the relationship between ethical leadership and ethical voice. Results from a 2 × 2 between-group ANOVA showed that the interaction of ethical leadership and leaders’ level of hierarchy was not significant: F(3, 77) = .091, p = .763, η2 = .001. This result suggested that leaders’ level of hierarchy in this case did not significantly affect the relationship between ethical leadership and group ethical voice. Thus, in the next step, we tested whether ethical leadership in general would predict group ethical voice. The results from regression analyses showed a positive main effect of ethical leadership on ethical voice: F(1, 79) = 5.698, p = .019, η2 = .07. Therefore, we found evidence that ethical leadership positively affected group ethical voice. Study 2 Adding to our initial findings from Study 1a, Study 1b’s experimental design provided evidence of the causal link between ethical leadership and group ethical voice. Following Downloaded from jom.sagepub.com by guest on April 3, 2015 16 Journal of Management / Month XXXX these two studies, Study 2 involves a primary field survey investigation that tests the complete proposed model incorporating two mediating mechanisms that explain how ethical leadership influences group ethical voice. Study 2 also attempts to explore the relationship between ethical voice and ethical performance, as we expect to demonstrate that ethical voice can lead to meaningful positive outcomes for organizations. Sample and Procedures We invited 1,337 employees working at 36 chain stores of a large retail company in China to participate in the present study. Each store of this retail company is relatively small in size so that the management at both upper level (store manager) and lower level (group manager) has direct interactions with employees in their daily work. All invited participants worked in either sales or customer service. Within each group, members share similar job responsibilities and regularly participate in group meetings to discuss work-related issues. We also learned from the management team that employees sometimes discuss unethical work behavior/service (e.g., intentionally withholding product information that might reduce the likelihood of purchases from customers, exaggerating the characteristics/advantages of a product, etc.) they observe from their colleagues or hear from their friends working in the same profession. The top management team has learned through multiple channels (e.g., complaint letters/notes from customers, random service quality sampling by thirdparty independent contractors, mobile-app customer reviews, and exit interviews) that stores differ in the way they serve customers. There were concerns that the management team at local stores had implicitly, or even explicitly, tolerated unethical work behaviors as described above. In sum, we decided that the setting was an ideal environment to explore our hypotheses. We collected survey data in three waves. At Time 1, we distributed surveys to 198 lowerlevel managers, who reported their upper-level managers’ ethical leadership, and 1,337 employees, who reported their lower-level managers’ ethical leadership. Responses were collected from 136 lower-level managers (68.7%) and 835 employee respondents (62.5%). Three weeks later (Time 2), we asked the 835 responding employee participants to complete measures of ethical culture and group ethical voice efficacy. Six hundred forty-seven responses (77.5%) were collected. Three weeks later (Time 3), we asked the assistant manager of each store to complete measure of group ethical voice of those 198 groups. We collected responses from all of the 36 assistant store managers (100%). We were then able to match up responses from 552 employees and those from 128 lower-level managers. Following the same procedure (i.e., 60% group response rate) used in Study 1a to determine the validity of our collected responses in sufficiently representing group properties, we obtained a final sample of 511 employee respondents matched up with 111 lower-level managers at 32 stores. Group size ranged from 3 to 8 members, with an average of 4.6 members. Among the 511 employee respondents, 318 were female (62.2%). The average age and organizational tenure were 27.0 and 4.7 years, respectively. Among the 111 lower-level managers, 42 were female (37.8%). The average age and organizational tenure were 34.0 and 8.6 years, respectively. Among the 32 upper-level managers, 9 were female (28%). The average age and organizational tenure were 36.2 and 8.6 years, respectively. Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 17 Measures Upper-level and lower-level ethical leadership. We assessed upper- and lower-level ethical leadership using ELS (Brown et al., 2005). The Cronbach’s alpha was .94 for upper-level ethical leadership and .92 for lower-level ethical leadership. Ethical culture. We adapted 8 items from the 10-item ethical culture scale developed by Treviño et al. (1998) and used in Schaubroeck et al.’s (2012) study of ethical leadership and outcomes. We were advised by management of the company to omit two items that did not fit into the company setting. The two items we dropped were “People of integrity are rewarded in this organization” and “Ethical behavior is rewarded in this organization.” One sample item was “Ethical behavior is the norm in this store.” The Cronbach’s alpha was .88. Group ethical voice efficacy and group ethical voice. We measured group ethical voice using the same six items described in Study 1a. Following Morrison et al.’s (2011) approach in measuring voice efficacy, we used a referent-shift method (Chan, 1998) to capture grouplevel shared beliefs. Specifically, we asked the respondents to indicate the extent to which they felt that their group as a whole was capable of effectively doing the things as specified in the six-item group ethical voice scale. One sample item was “. . . develops and makes recommendations concerning ethical issues that affect their work.” The Cronbach’s alpha was .87 for group ethical voice efficacy and .85 for group ethical voice. Ethical performance. To assess ethical performance, we obtained company records regarding customer ratings of employees’ service behaviors. Specifically, in its chain stores, the company used 10 items each to measure the service performance of its sales and customer service groups. Customers are requested to offer their evaluations at the checkout counter or the service center counter via an electronic device for a promotional coupon valued at 5% off regular-priced products in the store (with conditions and limitations). At the end of each month, customer ratings of individual employees’ service performance are aggregated to indicate customer evaluations of the service performance of the whole group. For both of these two groups, there is one item that specifically captures the ethical performance of the group. Customers are asked to indicate the extent to which they feel the item is characteristic of the behaviors of the employee who offered his or her service (from 1 = completely uncharacteristic to 10 = completely characteristic). The item for measuring ethical performance of the sales employee was “The sales associate truthfully explained the features of the product without overly attempting to convince me to purchase.” The item for measuring ethical performance of the customer service employee was “The customer service associate resolved my complaint or request in a sincere and ethical manner.” While the use of a one-item measure might raise concerns with respect to psychometric validity, prior studies have suggested that the first and the foremost concern while using a single-item objective measure from a company setting is whether the instrument can represent and measure the construct (e.g., single item used in Kirkman, Rosen, Tesluk, & Gibson, 2004, to measure customers’ satisfaction with the service provided by a team). That said, we argue that the one-item measure used by the company in this study is an adequate indicator of the construct of ethical performance of the sales and the customer service groups. Downloaded from jom.sagepub.com by guest on April 3, 2015 18 Journal of Management / Month XXXX Control variables. Similar to Study 1a, we controlled for age, gender (female = 1, male = 2), and organizational tenure of both upper-level managers and lower-level managers; the size of the group; and the type of group (i.e., sales and customer service). Measurement and Analytic Procedures Data aggregation. Similar to the procedures described in Study 1a, we evaluated the appropriateness of data aggregation for the study variables. The median rwg(j) values were .83 (.71-.92) for upper-level ethical leadership, .80 (.72-.84) for store-level ethical culture, .88 (.81-.93) for lower-level ethical leadership, .81 (.72-.87) for group ethical voice efficacy, and .84 (.75-.90) for group moral identity. All of these values exceeded the suggested cutoff value of .70 (Chen et al., 2004), indicating that respondents in our sample had sufficient agreement on their reported measures. We further calculated the ICC values to determine if there was sufficient between-level variance for these above measures. The ICC values for these measures, including upper-level ethical leadership (ICC1 = .35, ICC2 = .65), ethical culture (ICC1 = .15, ICC2 = .74), lower-level ethical leadership (ICC1 = .28, ICC2 = .64), group ethical voice efficacy (ICC1 = .37, ICC2 = .73), and group moral identity (ICC1 = .26, ICC2 = .62), all fell into an acceptable range for aggregation (James, 1982; Klein & Kozlowski, 2000) and therefore rendered further support for data aggregation. Finally, following suggestions from Mathieu and Taylor (2007) in testing cross-level mediation, we also conducted ANOVA tests and found that there was sufficient between-store variance in the lower-level variables (i.e., lower-level manager ethical leadership, group ethical voice efficacy, group moral identity, and group ethical voice). Analytic strategy. Prior to testing study hypotheses, we first performed a set of CFAs to examine the fit of all multi-item scales and our proposed measurement model. Second, we tested our hypotheses using multilevel modeling techniques (Hofmann, 1997) through HLM 7 (Raudenbush et al., 2011). Following prior literature in testing multilevel mediation models, we assessed the “cross-level mediation-lower-level mediator relationship” (Mathieu & Taylor, 2007: 158). Third, to test the significance of the indirect effect of upper-level ethical leadership on group ethical voice via upper-level (i.e., store-level ethical culture) and lower-level (i.e., group ethical voice efficacy) mediators, we constructed 95% confidence intervals using the Monte Carlo estimation approach (Selig & Preacher, 2008), which allowed us to estimate the proposed indirect effect described above with greater statistical power and more accurate Type I error rates in the current multilevel setting (MacKinnon, Lockwood, & Williams, 2004). Results Measurement model tests. Using Mplus 6.1 (Muthén & Muthén, 2010), we first examined the construct validity of group ethical voice. Consistent with results from Study 1a, the CFA results in this study showed that the construct of group ethical voice produced acceptable fit with the data: χ2 = 12.75, df = 9, CFI = .98, TLI = .96, RMSEA = .06, SRMR = .04. We then performed a series of CFAs to assess the fit of our proposed measurement model. The CFA results suggested that the five-factor baseline model produced acceptable fit: χ2 = 1486.28, df = 730, CFI = .95, TLI = .94, RMSEA = .05, SRMR = .04. We compared this baseline model with several alternative models. The best competing model was the one in which we loaded Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 19 all the items of upper-level ethical leadership and ethical culture onto one latent variable. The chi-square difference test suggested that our baseline model produced significantly better fit than this best competing model: ∆χ2(4) = 697.33, p < .01. We thus decided to retain our baseline model. Hypotheses tests. As mentioned earlier, with preliminary results from Study 1a and 1b in support of positive relationships between ethical leadership and group ethical voice, in Study 2, we tested all the hypothesized relationships via HLM. We report means, standard deviations, and correlations of study variables in Table 3. Among the control variables, five of them, including upper-level managers’ gender and organizational tenure, lower-level managers’ age and organizational tenure, and the size of the group, correlated significantly with at least one major study variable. Therefore, we included these five control variables in further multilevel analyses. Hypotheses 1 proposed that ethical leadership would be positively associated with group ethical voice. As shown in Table 4, the HLM results showed that upper-level (γ = .28, p < .01) and lower-level ethical leadership (γ = .45, p < .01) were both positively associated with group ethical voice. We thus found support for Hypotheses 1. Table 3 Means, Standard Deviations, and Intercorrelations of Variables for Study 2 Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. UM’s age — 2. UM’s gender .20 — 3. UM’s org tenure .65 .23 — 4. LM’s age .08 .06 .01 — 5. LM’s gender −.00 −.11 −.04 .16 — 6. LM’s org tenure .07 −.01 .06 .62 .14 — 7. Group size −.13 −.08 −.18 −.08 −.04 −.13 — 8. Group type −.01 −.08 −.00 .07 −.28 .04 −.41 — 9. UM’s ethical leadership −.08 .01 .03 −.01 .00 .06 .21 .01 (.94) 10. LM’s ethical leadership −.07 .07 −.08 −.03 .02 −.11 .08 −.08 .28 (.92) 11. Ethical culture −.06 .07 .12 −.05 −.05 −.05 .15 −.04 .46 .26 (.88) 12. Ethical voice efficacy .01 .07 −.14 −.03 −.06 −.10 .09 −.07 .23 .39 .11 (.81) 13. Group ethical voice −.03 −.05 −.08 .06 .08 .09 −.01 .03 .38 .45 .33 .44 (.85) 14. Ethical service quality .06 .18 −.05 .08 .01 .07 .08 .07 .15 .13 .05 .02 .21 — M 39.78 1.65 12.36 33.96 1.62 8.59 4.60 1.43 3.46 3.53 3.54 3.54 3.53 7.63 SD 4.71 0.48 4.47 3.79 0.49 3.39 1.16 0.50 0.76 0.64 0.60 0.66 0.73 1.73 Note: N = 511. Group-level units = 111. Store-level units = 32. Gender was coded as 1 = female, 2 = male. Group type was coded as 1 = sales, 2 = customer service. UM = upper-level manager; LM = lower-level manager; org = organizational. Coefficient alphas for scales are provided in parentheses. Correlations with absolute values of .09 (.14 for upper level) and above are significant at p < .05. Correlations with absolute values of .12 (.19 for upper level) and above are significant at p < .01. Two-tailed test. Downloaded from jom.sagepub.com by guest on April 3, 2015 20 Table 4 Results of Hierarchical Linear Modeling for Study 2 Dependent Variables Store-Level Ethical Culture Group Ethical Voice Efficacy Group Ethical Voice Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Intercept 3.54** 3.53** 3.54** 3.54** 3.54** 3.53** 3.53** 3.53** 3.53** Store-level control variables UM’s gender .05 .06 .07 .03 .03 −.07 −.11 −.12 −.14 UM’s organizational tenure .01 .01 .02 .02 .02 −.02 −.01 −.02 −.02 Store-level predictors UM’s ethical leadership .40** .10 .28** .20* .16 Ethical culture .26** .28** Group-level control variables LM’s age .00 .00 .00 .01 .01 .00 .01 LM’s organizational tenure −.03 −.02 −.02 .01 .02 .02 .03 Group size .07 .06 .04 −.02 −.06 −.07 −.09* Group-level predictors LM’s ethical leadership .41** .38** .45** .42** .30** Group ethical voice efficacy .33** R2 .02 .19 .04 .20 .20 .03 .27 .35 .39 ∆R2 .17 .16 .00 .24 .08 .04 Note: N = 111 groups (nested in 32 stores). UM = upper-level manager; LM = lower-level manager. R2 values were calculated using the formula developed by Snijders and Bosker (1999). *p < .05. **p < .01. Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 21 In Hypotheses 2 and 3, we proposed two mediating mechanisms linking multilevel ethical leadership with group ethical voice. We tested our model in HLM by investigating crosslevel mediation/lower-level relationships as described in Mathieu and Taylor (2007). Hypothesis 2 proposed that ethical culture would mediate the positive relationship between upper-level ethical leadership and group ethical voice. As shown in Table 4, we found that ethical culture was positively associated with upper-level ethical leadership (γ = .40, p < .01) and group ethical voice (γ = .26, p < .01). Following suggestions by Preacher and his colleagues (Zhang et al., 2009) in computing multilevel mediation models and assessing the significance of indirect effects in those mediation models, we then used the Monte Carlo estimation approach to construct 95% confidence intervals (CIs) for our hypothesized indirect effect (Selig & Preacher, 2008). The results showed that the indirect effect of upper-level ethical leadership on group ethical voice via ethical culture (ρ = .104) was significant with 95% CI [.048, .176], which excludes zero. We thus found support for Hypothesis 2. Hypothesis 3 proposed that both upper- and lower-level ethical leadership would have significant indirect effects on group ethical voice via group ethical voice efficacy. As shown in Table 4, upper-level ethical leadership did not significantly predict group ethical voice efficacy (γ = .10, ns), while lower-level ethical leadership was positively associated with group ethical voice efficacy (γ = .41, p < .01). We further found that group ethical voice efficacy significantly predicted group ethical voice (γ = .33, p < .01). Using Monte Carlo estimation, we were able to detect a significant indirect effect of lower-level ethical leadership on group ethical voice via group ethical voice efficacy (ρ = .135) with 95% CI [.045, .265]. However, the indirect effect of upper-level ethical leadership on group ethical voice via group ethical voice efficacy (ρ = .033) was insignificant because the 95% CI [–.014, .108] includes zero. Thus, we only found partial support for Hypothesis 3. Hypothesis 4 proposed that group ethical voice would be positively associated with ethical performance. Given that the ratings of ethical performance were skewed, prior to model testing, we used logarithmic transformation to recalculate this measure. Also, since ethical performance was measured using different items for the sales and the customer service groups in our sample, we estimated the relationship between group ethical voice and ethical performance of these two groups separately. As shown in Table 5, we found a positive relationship between group ethical voice and ethical performance of sales groups (γ = .22, p < .05). Yet, the effect of group ethical voice on ethical performance of customer service groups was only marginally significant (γ = .13, p = .053). Thus, we found partial support for Hypothesis 4. Discussion In the present study, we introduced the concept of group ethical voice, which focuses the group voice construct on ethical issues. Relying on social learning theory (Bandura, 1977) and social cognitive theory (Bandura, 1986), we conceptualize group ethical voice as an important outcome of ethical leadership and a precursor to ethical performance. We also proposed two mediating mechanisms through which ethical leadership indirectly affects group ethical voice. Using field survey data in Study 1a, we validated the construct of group ethical voice and found initial support for a positive relationship between ethical leadership and ethical voice. In Study 1b, results from an experiment provided evidence for the causal link between ethical leadership and group ethical voice. Using field survey data in Study 2, Downloaded from jom.sagepub.com by guest on April 3, 2015 22 Table 5 Results of Hierarchical Linear Modeling for Study 2 With Ethical Performance as Dependent Variable Dependent Variables Ethical Performance (Sales) Ethical Performance (Customer Service) Variable Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16 Model 17 Intercept 1.97** 1.97** 1.97** 1.97** 2.02** 2.02** 2.02** 2.02** Store-level control variables UM’s gender .19† .16† .17† .18* .09 .10 .09 .08 UM’s organizational tenure −.01 −.01 −.01 −.01 −.00 −.00 −.00 .00 Store-level predictors UM’s ethical leadership .13* .11 .10 .07 .07 .05 Ethical culture .03 .02 −.00 −.04 Group-level control variables LM’s age .00 .01 .00 .00 .00 .00 .00 .00 LM’s organizational tenure −.01 −.01 −.01 −.01 .00 .00 .00 .00 Group size .03 .01 .01 .01 .04 .03 .03 .03 Group-level predictors LM’s ethical leadership .05 .03 .01 −.01 −.01 −.03 Group ethical voice efficacy .02 −.00 .02 .02 Group ethical voice .22* .13† R2 .04 .06 .06 .12 .01 .01 .01 .04 ∆R2 .02 .00 .06 .00 .00 .03 Note: N = 111 groups (63 sales groups and 48 customer service groups nested in 32 stores). UM = upper-level manager; LM = lower-level manager. R2 values were calculated using the formula developed by Snijders and Bosker (1999). †p < .10. *p < .05. **p < .01. Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 23 we tested the complete proposed model. We found that at the store level, ethical culture mediated the positive relationship between upper-level ethical leadership and group ethical voice. At the group level, group ethical voice efficacy mediated the relationship between lowerlevel, but not upper-level, ethical leadership and group ethical voice. We further found support that group ethical voice positively influenced group ethical performance. Theoretical Implications The present research has produced a number of theoretical implications regarding voice and ethical leadership research. First, we answered the call from Morrison (2011) to specify the content of voice by considering a particular dimension (ethical) under the construct of group voice. We suggest that group ethical voice changes the status quo by presenting a group’s collective view of an ethical dilemma or perceived misconduct observed by group members in the work environment. Group ethical voice enables a group to take advantage of its within-group discussions and informal processes to respond to ethical issues that concern one or more group members. Therefore, group ethical voice is manifested when group members are aware of ethical issues in the workplace and raise the awareness of others with regard to those issues. Instead of merely criticizing ethical misconduct, group ethical voice produces meaningful suggestions about how to change current practices and/or how to prevent ethical misconduct from happening in the first place. The concept of group ethical voice builds on existing work on group-level voice (Frazier & Bowler, in press; Walumbwa et al., 2012) and ethical voice (Gentile, 2010, 2013) to emphasize the inherent advantages of group-level voice behaviors especially in the context of ethical dilemmas. Second, our study has extended the current understanding of the outcomes of ethical leadership by demonstrating how ethical leadership facilitates group voice concerning ethical issues. By finding positive direct (including initial evidence of causality in Study 1b) and indirect effects of ethical leadership on group ethical voice via two mediating mechanisms, this study provides additional evidence of the value of ethical leadership in organizational settings. This is an especially needed contribution to the ethical leadership literature because little research attention has been devoted to discussions about how ethical leadership can lead to specific ethical outcomes (see Mayer et al., 2009, 2012; Schaubroeck et al., 2012, for exceptions). That said, answering the call from Mayer et al. (2012: 167), this research contributes to the ethical leadership research by expanding the “nomological network of potential dependent variables” of ethical leadership. Moreover, drawing from social cognitive theory, our study extends the current understandings of voice behavior that consider voice as proactive and discretionary (Van Dyne & LePine, 1998). Specifically, our study indicates that ethical leadership can activate people’s agency in interacting with others in the group to regulate the collective behaviors of the group in a way aligned with the conduct of ethical leaders as well as the norm of the work environment affected by such leaders. Third, our study also adds to the voice literature by specifying group voice efficacy as an important mediator through which leadership can activate group voice behaviors. Our findings from Study 2 show that upper-level ethical leadership did not significantly influence group ethical voice efficacy, suggesting that more proximal leaders (e.g., lower-level managers) have a greater influence on work groups’ judgment that they can speak up effectively on ethical issues. This finding also calls for more discussions regarding the nature of group ethical voice efficacy. We suggest that the concept of group ethical voice efficacy might Downloaded from jom.sagepub.com by guest on April 3, 2015 24 Journal of Management / Month XXXX incorporate two components. First, as described by Morrison et al. (2011), group-level voice efficacy relies on a shared feeling among group members such that they understand the issues they discuss and they can communicate with their leaders in a way that leaders can understand. Consistent with this notion, group ethical voice efficacy indicates group members’ shared understanding about the ethical problems they discuss (e.g., whether the problem they see refers to an ethical matter and the potential consequence of it) and their confidence in expressing it to their leaders in an understandable way. Second, consistent with writings in the social cognitive theory that consider goal attainment as an important motivational and regulatory factor driving human agency (Bandura, 1989, 2001), group ethical voice efficacy reflects shared beliefs of group members such that ethical voice can lead to specific goal attainment, such as resolution of an ethical issue. Finally, we offer initial evidence that group ethical voice should be of concern to organizational researchers and practitioners because of its relationship with an objective rating of ethical performance (particularly for the sales group in our Study 2 sample). We have relied on a deontological approach (Treviño & Nelson, 2014) to conceptualize and measure ethical performance. We posit that ethical performance reflects the extent to which group members attend to their perceived duties of serving the organization and protecting the rights of those that they serve. This conceptualization aligns well with the context of our two field studies (retail sales). For example, customers expect that ethical sales and customer service representatives will honor certain duties (provide them with accurate information about products, treat them fairly, etc.) and respect their rights as customers (to be told the truth, to be treated with respect, to have their problems resolved in a timely manner, etc.). Since ethical performance at the group level is a relatively new concept (discussed by Gatewood and Carroll, 1991, but little work has been done since), it is unknown what impact ethical performance has on more traditional outcomes, such as financial performance. While some speculate that ethical performance might negatively impact financial performance, we contend that in the long run, ethical conduct functions to promote the company’s reputation, which will have a positive impact on firm financials. This is also consistent with prior findings at the company level showing a positive relationship between companies’ compliance with the ethical code and its financial performance (Verschoor, 1998). In sum, our findings add to the voice literature by showing that group ethical voice behaviors assist work groups in regulating their own ethical performance as rated by customers. Practical Implications The major ethical lapses highlighted in the introduction have resulted in a series of regulations that put increased responsibility on top management to be aware of any ethical breakdowns occurring within the organization. However, no manager, no matter how capable, is able to single-handedly stay abreast of the potential ethical problems occurring in a large organization. The attractiveness of group ethical voice behaviors is that they offer these managers a mechanism through which concerns of an ethical nature are expressed openly and eventually garner their attention. Our findings demonstrate that these behaviors do occur in organizations and that ethical leadership is instrumental in providing work groups with the culture and confidence necessary to voice up. Moreover, our findings indicate that groups that speak up about ethical issues also perform in a more ethical manner overall. While ethical performance may not be as often studied or cared about as financial performance, recent Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 25 high-profile ethical lapses suggest that perhaps it should be. A win-at-all-costs mentality has proven the downfall of numerous seemingly profitable organizations that compromised their ethics for short-term financial gains. Limitations and Future Directions The present study, with strengths in design and theoretical contribution, also is subject to several limitations. First, our cross-sectional research design prevents us from making causal claims about all of the proposed effects. However, we attempted to alleviate concerns about generating causal effects in mediation models by collecting data from multiple time periods, from multiple sources, and using an experimental design in Study 1b. As advocated by Mayer et al. (2012), future research might benefit from employing a longitudinal design investigating the causal effects of ethical leadership on outcomes, such as group ethical voice, as well as the mediating mechanisms linking these effects. Second, although in the present study we provide some initial theoretical and empirical foundations for group ethical voice by specifying its antecedents and outcomes, there is still room for further development of the construct. Future research might explore if group ethical voice is a unidimensional construct, as we operationalize it in this paper, or if it has multiple dimensions. Moreover, future research on group ethical voice could utilize social network analysis to identify the social mechanisms at work with regard to ethical issues and decisions to speak up as a group. Such analyses would allow for a more nuanced understanding of the likelihood of work groups reaching consensus on ethical issues and how such consensus is achieved. In addition, future research may explore group processes that help (or prevent) group members to reach a consensus view regarding what is ethical or unethical and why the group as a whole should speak up to challenge ethical issues. Third, although the present study has tested two mediating mechanisms linking ethical leadership with ethical voice and ethical performance as an outcome of group ethical voice, we did not explore the boundary conditions (i.e., moderators) that might affect these relationships. Future research could specify boundary conditions that might strengthen or hinder these relationships. For example, our results showed that group ethical voice efficacy was positively predicted by lower-level, not upper-level, ethical leadership. This might be a meaningful direction for future research regarding the necessity of specifying the target of voice behavior. Moreover, it will be interesting to explore if leaders’ personality traits influence the relationship between their ethical leadership and employees’ ethical voice. Finally, in this study, we have adopted a deontological approach to conceptualize ethical performance, which is also aligned with the way it was assessed by the company we studied in Study 2. Yet we recognize that there are different normative frameworks researchers can build on to explore how ethical voice might affect ethical performance in the organization. We thus call for future research to approach the concept of ethical performance from other perspectives (e.g., justice, utilitarianism) and consider the implications of these different conceptualizations and measures of ethical performance in the ethical voice context. Conclusion This study produces a number of important implications to both voice and ethical leadership research. We first contribute to the development of the voice construct by specifying an Downloaded from jom.sagepub.com by guest on April 3, 2015 26 Journal of Management / Month XXXX ethical dimension. We suggest that ethical voice at the group level is a meaningful tool managers can leverage to identify problems related to ethical issues in the workplace. We have further identified two mediating mechanisms that help to transmit the effect of ethical leadership on ethical voice. This adds important knowledge to ethical leadership research regarding the influence of ethical leadership on ethical outcomes. Finally, we have offered initial evidence showing that ethical voice can have a positive impact on ethical performance. We suggest that group ethical voice acts as “eyes and ears” throughout the organization, which can help not only work groups but also the whole organization to improve overall ethical performance and warn management of ethical failings before they reach headline-producing scale. References Ajzen, I. 1991. The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50: 179-211. Andrews, K. 1989. Ethics in practice. Harvard Business Review, 67: 99-104. Ardichvili, A., Mitchell, J. A., & Jondle, D. 2009. Characteristics of ethical business cultures. Journal of Business Ethics, 85: 445-451. Bandura, A. 1977. Social learning theory. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. 1982. Self-efficacy mechanism in human agency. American Psychologist, 37: 122-147. Bandura, A. 1986. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. 1989. Human agency in social cognitive theory. American Psychologist, 44: 1175-1185. Bandura, A. 2001. Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52: 1-26. Becker, T. E. 1998. Integrity in organizations: Beyond honesty and conscientiousness. Academy of Management Review, 23: 154-161. Bliese, P. D. 2000. Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions: 349-381. San Francisco: Jossey-Bass. Bliese, P. D., & Hanges, P. J. 2004. Being both too liberal and too conservative: The perils of treating grouped data as though they were independent. Organizational Research Methods, 7: 400-417. Brislin, R. W., Lonner, W. J., & Thorndike, R. M. 1973. Cross-cultural research methods. New York: John Wiley. Brooks, L. J. 1989. Corporate ethical performance: Trends, forecasts and outlooks. Journal of Business Ethics, 8: 31-38. Brown, M. E., & Treviño, L. K. 2006. Ethical leadership: A review and future directions. Leadership Quarterly, 17: 595-616. Brown, M. E., Treviño, L. K., & Harrison, D. A. 2005. Ethical leadership: A social learning perspective for construct development and testing. Organizational Behavior and Human Decision Processes, 97: 117-134. Burris, E. R. 2012. The risks and rewards of speaking up: Managerial responses to employee voice. Academy of Management Journal, 55: 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. Burris, E. R., Detert, J. R., & Romney, A. C. 2013. Speaking up vs. being heard: The disagreement around and outcomes of employee voice. Organization Science, 24: 22-38. Chan, D. 1998. Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83: 234-246. Chen, G., Mathieu, J. E., & Bliese, P. D. 2004. A framework for conducting multi-level construct validation. In F. J. Dansereau & F. Yammarino (Eds.), Research in multi-level issues: The many faces of multi-level issues, vol. 3: 273-303. Oxford, UK: Elsevier Science. DeGeorge, R. T. 1990. Business ethics, 3rd ed. New York: MacMillan Press. Detert, J. R., & Burris, E. R. 2007. Leadership behavior and employee voice: Is the door really open? Academy of Management Journal, 50: 869-884. Downloaded from jom.sagepub.com by guest on April 3, 2015 Huang, Paterson / Group Ethical Voice 27 Detert, J. R., & Treviño, L. K. 2010. Speaking up to higher-ups: How supervisors and skip-level leaders influence employee voice. Organization Science, 21: 249-270. Falkenberg, L., & Herremans, I. 1995. Ethical behaviors in organizations: Directed by the formal or informal systems? Journal of Business Ethics, 14: 133-143. Frazier, M. L., & Bowler, W. M. in press. Voice climate, supervisor undermining, and work outcomes: A grouplevel examination. Journal of Management. Gatewood, R. D., & Carroll, A. B. 1991. Assessment of ethical performance of organization members: A conceptual framework. Academy of Management Review, 16: 667-690. Gentile, M. C. 2010. Giving voice to values. New Haven, CT: Yale University Press. Gentile, M. C. 2013. Educating for values-drive leadership: Giving voice to values across the curriculum. New York: Business Expert Press. Hannah, S. T., Avolio, B. J., Luthans, F., & Harms, P. D. 2008. Leadership efficacy: Review and future directions. Leadership Quarterly, 19: 669-692. Hofmann, D. A. 1997. An overview of the logic and rationale of hierarchical linear models. Journal of Management, 23: 723-744. Hu, L., & Bentler, P. M. 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6: 1-55. James, L. R. 1982. Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67: 219-229. James, L. R., Demaree, R. G., & Wolf, G. 1984. Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69: 85-98. Kirkman, B. L., Rosen, B., Tesluk, P., & Gibson, C. B. 2004. The impact of team empowerment on virtual team performance: The moderating role of face-to-face interaction. Academy of Management Journal, 47: 175-192. Klaas, B. S., Olson-Buchanan, J. B., & Ward, A.K. 2012. The determinants of alternative forms of workplace voice: An integrative perspective. Journal of Management, 38: 314-345. Klein, K. J., & Kozlowski, S. W. 2000. From micro to meso: Critical steps in conceptualizing and conducting multilevel research. Organizational Research Methods, 3: 211-236. Kozlowski, S. W., & Klein, K. J. 2000. A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions: 3-90. San Francisco: Jossey-Bass. Liang, J., Farh, C. I., & Farh, J.-L. 2012. Psychological antecedents of promotive and prohibitive voice: A two-wave examination. Academy of Management Journal, 55: 71-92. MacKinnon, D. P., Lockwood, C. M., & Williams, J. 2004. Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39: 99-128. Manz, C. C., & Sims, H. P., Jr. 1981. Vicarious learning: The influence of modeling on organizational behavior. Academy of Management Review, 6: 105-113. Mathieu, J. E., & Taylor, S. R. 2007. A framework for testing meso-mediational relationships in organizational behavior. Journal of Organizational Behavior, 28: 141-172. Mayer, D. M., Aquino, K., Greenbaum, R. L., & Kuenzi, M. 2012. Who displays ethical leadership, and why does it matter? An examination of antecedents and consequences of ethical leadership. Academy of Management Journal, 55: 151-171. Mayer, D. M., Kuenzi, M., Greenbaum, R., Bardes, M., & Salvador, R. B. 2009. How low does ethical leadership flow? Test of a trickle-down model. Organizational Behavior and Human Decision Processes, 108: 1-13. 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. Morrison, E. W. 2011. Employee voice behavior: Integration and directions for future research. Academy of Management Annals, 5: 373-412. Morrison, E. W. 2014. Employee voice and silence. Annual Review of Organizational Psychology and Organizational Behavior, 1: 173-197. 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., & Milliken, F. J. 2003. Speaking up, remaining silent: The dynamics of voice and silence in organizations. Journal of Management Studies, 40: 1353-1358. Morrison, E. W., Wheeler-Smith, S. L., & Kamdar, D. 2011. Speaking up in groups: A cross-level study of group voice climate and voice. Journal of Applied Psychology, 96: 183-191. Downloaded from jom.sagepub.com by guest on April 3, 2015 28 Journal of Management / Month XXXX Muthén, L. K., & Muthén, B. O. 2010. Mplus: Statistical analysis with latent variables. User’s guide, 6th ed. Los Angeles: Author. Parker, S. K. 1998. Enhancing role breadth self-efficacy: The roles of job enrichment and other organizational interventions. Journal of Applied Psychology, 83: 835-852. Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Congdon, R. T., & du Toit, M. 2011. HLM 7: Hierarchical linear and nonlinear modeling. Chicago: Scientific Software International. Roth, P. L., & BeVier, C. A. 1998. Response rates in HRM/OB survey research: Norms and correlates, 1990-1994. Journal of Management, 24: 97-117. Schaubroeck, J., Hannah, S., Avolio, B., Kozlowski, S., Lord, R., Treviño, L., Dimotakis, N., & Peng, A. 2012. Embedding ethical leadership within and across organization levels. Academy of Management Journal, 55: 1053-1078. Schein, E. H. 1985. Organizational culture and leadership. San Francisco: Jossey-Bass. Selig, J., & Preacher, K. 2008. Monte Carlo method for assessing mediation: An interactive tool for creating confidence intervals for indirect effects. Kansas City: Selig & Preacher. Snijders, T., & Bosker, R. 1999. Multilevel modeling: An introduction to basic and advanced multilevel modeling. London, UK: Sage. Tierney, P., & Farmer, S. M. 2002. Creative self-efficacy: Its potential antecedents and relationship to creative performance. Academy of Management Journal, 45: 1137-1148. Timmerman, T. A. 2005. Missing persons in the study of groups. Journal of Organizational Behavior, 26: 21-36. Treviño, L. K. 1990. A cultural perspective on changing and developing organizational ethics. In R. Woodman & B. Passmore (Eds.), Research in organizational change and development: 195-230. Greenwich, CT: JAI Press. Treviño, L. K., & Brown, M. E. 2004. Managing to be ethical: Debunking five business ethics myths. Academy of Management Executive, 18(2): 69-81. Treviño, L. K., Butterfield, K. D., & McCabe, D. L. 1998. The ethical context in organizations: Influences on employee attitudes and behaviors. Business Ethics Quarterly, 8: 447-476. Treviño, L. K., & Nelson, K. A. 2014. Managing business ethics: Straight talk about how to do it right, 6th ed. New York: Wiley. 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., 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. Verschoor, C. C. 1998. A study of the link between a corporation’s financial performance and its commitment to ethics. Journal of Business Ethics, 17: 1509-1516. Walumbwa, F. O., Morrison, E. W., & Christensen, A. L. 2012. Ethical leadership and group in-role performance: The mediating roles of group conscientiousness and group voice. Leadership Quarterly, 23: 953-964. 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. Wood, R., & Bandura, A. 1989. Social cognitive theory of organizational management. Academy of Management Review, 14: 361-384. Yandle, B. 2010. Lost trust: The real cause of the financial meltdown. Independent Review, 14: 341-361. Zhang, Z., Zyphur, M. J., & Preacher, K. J. 2009. Testing multilevel mediation using hierarchical linear models: Problems and solutions. Organizational Research Methods, 12: 695-719. Downloaded from jom.sagepub.com by guest on April 3, 2015