FEATURE ARTICLE
Ethical Leadership at Work and with Friends and Family: Within-Person and Between-Raters Variability Matters
Michael Palanski, Rochester Institute of Technology
Michelle Hammond, Oakland University
Jayoung Kim, Binghamton University
Gretchen Vogelgesang Lester, San Jose State University
Rachel Clapp Smith, Purdue University Northwest
ABSTRACT
Despite its theoretical grounding in the personal moral characteristics of leaders, most research on Brown et al.’s (2005) ethical leadership construct has tended to ignore the personal life (friends/family) aspects of leaders. In this study, we consider ethical leadership behavior in both work and non-work (i.e., with friends and family) domains at both the intra-individual (domain) and individual (leader as a whole person) levels of analysis. We examine our research questions with a sample of 104 leaders and their 1,458 raters in executive MBA programs in the United States and Ireland. Our findings demonstrate that ethical leadership operates at the individual level of analysis in both work and non-work contexts, with the implication that researchers should consider both the mean and variation of ethical leadership. Our findings also indicate strong within-domain and limited cross-domain effects of ethical leadership and ethical leadership variation on cognitive trust, affective trust, and abusive supervision.
Keywords: Ethical Leadership, Multi-Domain Leadership, Levels of Analysis, Trust, Affective Trust, Cognitive Trust, Abusive Supervision
Citation: Journal of Character & Leadership Development 2023, 10: 272 - http://dx.doi.org/10.58315/jcld.v10.272
Copyright: © 2023 The author(s). This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
CONTACT Michael Palanski mpalanski@saunders.rit.edu
Published: 25 October 2023
The notion that effective leaders are also ethical persons is nearly axiomatic in leadership studies (i.e., Bedi et al., 2016). Consensus about personal ethical behavior—that is, different people in different situations observing similar levels of ethical behavior of a particular leader—helps to provide credibility, predictability, and trustworthiness, ultimately leading to outcomes such as increased trust in the leader and increased leader effectiveness (Hoch et al., 2018; Ng & Feldman, 2015; Simons et al., 2015). Some even argue that ethical leadership is itself the very essence of effective leadership (Newstead et al., 2021; Sturm et al., 2017). Indeed, according to findings from the GLOBE study of over 17,000 leaders in 62 countries (Javidan et al., 2006), elements relating to a positive sense of ethics are included in most people’s implicit leadership theories (House et al., 2002). For example, universal facilitators of leadership effectiveness included being trustworthy, just, and honest, whereas universal impediments included being self-protective and malevolent.
While many popular leadership theories contain ethical elements (e.g., transformational, authentic, and servant), Brown et al.’s (2005) theory of ethical leadership arguably sets the standard as the theory most focused on the ethical behavior of individual leaders. Ethical leaders are considered to be attractive and credible role models as they demonstrate integrity, set and maintain high ethical standards, engender trust and justice (Brown et al., 2005), and foster an ethical climate (Eisenbeiss et al., 2015). As most of the aspects of ethical leadership are focused on characteristics and consistent behavior of the individual leader, we might assume—as most theorizing seems to do—that the leader more or less acts morally, regardless of context. We note that aspects of the ethical leadership construct theoretically align with virtues, the habitual behaviors that are congruent with living the good life (Newstead et al., 2021). Indeed, ethical leadership has been defined from a characterological perspective as the adherence to the four cardinal virtues across all areas of life (Riggio et al., 2010). As such, virtues are inherently multi-domain in nature; that is, they are concerned with both the leader’s work and non-work lives. As Newstead et al. (2021, p. 3) illustrate, a CEO’s compassion and wisdom in a board meeting should also be present with the waiter at lunch. Specifically, we argue that to truly understand ethical leadership, it is important to consider the context of a leader’s outside-of-work life, particularly relationships with family and friends.
Doing so is important for at least two reasons. First, insofar as ethical behavior reflects overall leader character, then a multi-domain perspective is necessary for the examination of the fullness of the construct. There is much evidence to suggest that individuals are not always consistent in their ethical behavior across life domains. For example, Riggio et al. (2010) ask readers to recall the many religious or political leaders who “railed against certain vices but engaged in those same behaviors in private” (p. 236) believing (and perhaps convincing those in work setting) they were ethical leaders but living personal lives far from it. Inconsistent behavior across domains may represent a dark or “hidden private life” or a systematic situational contingency in which the character traits manifest across domains (Fleeson, 2007). Indeed, research suggests that social roles across domains (parent or professional) represented a systematic situational contingency for the enactment of several virtues—what they referred to as “virtue states” (Bleidorn & Denissen, 2015). Whether it be a hidden personal life or systematic variation, considerations across domains are meaningful and important for consideration.
Second, a multi-domain consideration is especially important for creating a greater understanding of leaders’ actual experiences (Liao et al., 2015). In particular, some studies have suggested that leaders’ ethical leadership might have reach beyond just the work domain. For example, Liao et al. (2015) found a positive relationship between employee perceptions of ethical leadership and spouse’s family satisfaction. Similarly, Zhang and Tu (2018) found that employees of highly ethical supervisors report more work-to-family enrichment, which in turn leads to greater family and life satisfaction, especially when the supervisor also provided family-supportive behaviors. We surmise that these types of cross-domain effects exist because of the (inherently multi-domain) personal aspects of ethical leadership.
Further, because a leader’s work and non-work relationships are different, it is possible that an effective leader might display at least somewhat different ethical behaviors in work and non-work (family and friends) situations. This dynamic raises the possibility that even within the same person, the “ethical leader at work” acts differently than the “ethical leader with friends and family.” Investigation of this possible variation and its implications lends itself to a levels-of-analysis perspective (e.g., Dansereau et al., 1984) in which we might consider a particular leader’s ethical leadership behaviors at work and ethical leadership behaviors with friends and family to be two different—and possibly even independent—levels of analysis.
In short, despite a myriad of reasons why we might assume that a person would be perceived as an ethical leader across all situations, we believe that it is important to test this assumption. We seek to answer the fundamental research question, “What role does variation play with respect to ethical leader behavior?” by focusing on the ubiquitous construct of ethical leadership (Brown et al., 2005). In this article, we address these issues by first exploring the theoretical case for a high level of consensus regarding a leader’s display of ethical leadership. Next, we review research about multi-domain leadership (Hammond et al., 2017) and demonstrate how it is helpful for gaining a better understanding of ethical leader behavior both at work and with friends/family. We then present a way to test our assumptions by explicating a levels-of-analysis framework (Dansereau et al., 1984) based on ethical leader behavior as experienced by relevant others (e.g., peers, direct reports, and family members) and its impact on commonly-studied outcomes such as cognitive and affective trust in the leader (Yang & Mossholder, 2010), as well as leader abusive supervision (Tepper, 2007). In doing so, we present a series of research questions to guide our analysis. Finally, we test our research questions by examining multi-source, multi-domain data from 1,458 raters of 104 leaders in the United States and Ireland.
Theoretical Development
Overview of Ethical Leadership
Ethical leadership is defined “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” (Brown et al., 2005, p. 120). From its original theorizing, ethical leadership has been framed as a process of social learning in which leaders affect the ethical conduct of followers through modeling appropriate behavior and norms (Brown et al., 2005). Ethical leaders gain a reputation as such through two means (Brown & Treviño, 2006). First, an ethical leader is considered a moral person in that he or she demonstrates and embodies ethical principles, care, and fairness. Second, a reputation as an ethical leader is also built on being perceived as a moral manager through specific behaviors that encourage ethical practices in organizations such as discussing ethics and disciplining ethical violations (Brown & Treviño, 2006). Together, ethical leaders are considered to be attractive and credible role models as they demonstrate integrity, set and maintain high ethical standards, and engender trust and justice (Brown et al., 2005) and foster an ethical climate (Eisenbeiss et al., 2015).
An increasing body of literature highlights the positive relationship between ethical leadership and follower outcomes including perceptions of fairness, follower ethical behavior, job satisfaction, self-efficacy, organizational commitment, organizational citizenship behaviors, and well-being (Bedi et al., 2016). Meta-analysis (Bedi et al., 2016) indicated a very high mean corrected correlation (0.77), between perceptions of ethical leadership and leader effectiveness.
Impact of Ethical Leadership Variability
Once viewed as mostly a nuisance resulting from error, more recent leadership research has begun to recognize the possible importance of rater variability (Lester et al., 2021). Despite this newfound interest, relatively few studies have examined rater variability in perceptions of ethical leadership within workgroups or organizations. These early examinations of variation in perceptions of ethical leadership suggest that a focus on the levels of analysis of ethical leadership is important (Bormann et al., 2018). For example, using a within- and between-group analytic procedure (WABA), Den Hartog and De Hoogh (2009) differentiated aspects of ethical leadership into perceptions of fairness and empowering leadership. They found that team members tend to share consistent evaluations of their leaders’ ethical behaviors in terms of fairness, but greater variability for empowering behaviors. In other words, fairness tended to be a group-level phenomenon whereas perceptions of empowerment tended to be an individual-level phenomenon rated in the eye of the beholder. Additionally, Bai and colleagues (2019) found that 35% of the variance of ethical leadership occurred at the group level.
Research in moral behavior also supports the potential importance of variability in ethical leader behavior. For example, in a series of studies, Hannah et al. (2018) demonstrated that moral identity—that is, how individuals construe their moral self (Blasi, 1993)—is a complex construct. They found that moral identity consists of four elements (justice, benevolence, obligation, and integrity), but that the importance of these elements will vary according to sub-identity or role (e.g., son/daughter, friend, coworker, follower). Their findings hint at the notion that variability in ethical leadership might extend beyond a particular role (i.e., the leader at work).
Multi-domain Perspective of Ethical Leadership
Multi-domain leadership (Hammond et al., 2017) takes a whole-person approach to understanding how individuals see themselves as leader and how this leader identity interacts and develops across the leader’s many life domains (Hammond et al., 2017). Hammond and colleagues suggest that the leader identity is unique in that it spans multiple domains: one can be a leader as a manager, but also as a parent, sibling, little league coach, etc. As such, several sub-identities contribute to the overall identity as a leader and thus can involve many different leadership behaviors. By taking a multi-domain leadership perspective (Hammond et al., 2017), we highlight the nuance associated with variability-based constructs to unpack the within- and between-contextual situations that may serve to clarify the patterns of consistency and variability to resolve the tensions between the two. Multi-domain leadership theory is well suited not only to address this tension because it takes a whole-person approach in understanding a leader, on the one hand, but also to address the inherent variability of sub-identities that exist by leading across life domains (work and non-work), on the other hand.
Hammond and colleagues (2017) suggest that in addition to work, there can be many non-work domains in which leaders are embedded that contribute to their sense of self as a leader. We focus primarily on the friends and family domain because, in questions of consensus and variability of ethical leader behavior, friends and family provide a perspective of leaders that are potentially more robust, more reliable, and more deeply held. Work and family roles tend to be the most salient roles held by individuals, and leader behaviors within both domains reciprocally influence each other (Courtright et al., 2016). Friends and family generally see a leader more frequently and over longer time horizons than peers, supervisors, and followers. Even if a leader changes roles or organizations, their friends and family remain the same. Second, relationships in the friends and family domain generally involve greater levels of intimacy and trust—in other words, friends and family tend to know a leader at a deeper level than work colleagues and vice versa. When it comes to leading among friends and family, because the relationships are long-standing and trust an important ingredient, there may be more at stake for a leader among friends and family than at work.
To our knowledge, up until now, there has not been a published investigation that considers the idea that, with respect to ethical leadership, a leader could be viewed similarly in one domain but variable in another domain, or that the pattern of ethical leadership in one domain may impact the variability of outcomes in the other domain. Given all these nuanced complexities that multi-domain leadership theory helps us tackle, we turn now to explaining observer perspectives of leader behavior followed by the various levels of analysis that help us to examine variability in perceptions of ethical leader behavior.
The Importance of Variable Observer Perspectives on Ethical Leadership
One approach to investigating these questions is by examining the variability in ratings of leadership across multiple sources, a common approach in much of the extant leadership literature, especially in the self-other agreement literature (Lee & Carpenter, 2018) and levels of analysis literature (Dansereau et al., 1984). Over the past couple of decades, to address the limitations of single-source assessment, many organizations have adopted multi-source feedback, commonly referred to as 360-degree assessments (Brett & Atwater, 2001). A key assumption of 360-degree feedback is that multiple perspectives from multiple levels in the organization provide a more accurate and holistic assessment of behavior and performance than that provided by the manager alone.
As ethical leadership is grounded in social learning theories, naturally most studies of ethical leadership rely primarily on follower ratings of the leader (Magalhães et al., 2019). However, multiple perspectives of perceived ethical leadership may be informative, as ethical leadership addresses a “multi-faceted network of stakeholders” (Eisenbeiß & Giessner, 2012, p. 17). Further, ethical leadership is defined as “normatively appropriate conduct” (Brown et al., 2005, p. 120), suggesting that appropriate behavior may be context-dependent. What individuals determine as morally appropriate may vary based on their own personal and cultural perspectives.
While Brown and colleagues (2005) describe ethical leadership as a collective phenomenon, in which members of the same team share perceptions, few studies have examined rater agreement in perceptions of ethical leadership within workgroups or organizations. The few studies that have done so indicate that the operative level of analysis for ethical leadership is unclear (see Bai et al., 2019; Bormann et al., 2018; De Hoogh & Den Hartog, 2008; Lin et al., 2019). These early examinations of variation in perceptions of ethical leadership suggest that a focus on the levels of analysis of ethical leadership is worthy of further explanation.
Levels of Analysis of Ethical Leadership
Levels of analysis are simply the objects or entities of study (Dansereau et al., 1984; Kozlowski & Klein, 2000; Yammarino et al., 2005). In the current investigation, we focus on the individual leader as our primary level of analysis, with a focus on lower-level “building blocks” (Palanski & Yammarino, 2011) of different domains (work and friends/family) and different persons within those situations. In order to adequately consider the fullness of ethical behavior, we need to consider the variability of such behavior both within and between domains, which we do as a two-step process, as outlined below and summarized in Table 1.
Table 1
Ethical Leadership Levels of Analysis Possibilities
Step 1: Variability of Ethical Leadership Within Domain (Intra-individual) Levels of Analysis
The purpose of Step 1 is to assess the operative level of analysis for each of two domain-level sub-identities (work and friends/family, respectively). Variability at any particular level of analysis may be described as consisting of one of four conditions (Dansereau et al., 1984; also see Klein et al. [1994] for a complementary discussion). First, a “wholes” condition occurs when there is significant variation between entities (e.g., between leaders), but not within entities. Second, a “parts” condition occurs when there is no significant variation between entities, but there is significant variation within entities. Third, an “equivocal” condition occurs when there is significant variation both between and within entities. Fourth, a “null” condition occurs when there is no significant variation either between or within entities. A null condition often indicates the presence of a very strong effect residing at a different (often higher) level of analysis. For example, imagine leaders functioning under a totalitarian regime whose strong norms govern all areas of life, resulting in no meaningful variation of behavior between leaders at work or in personal life.
Given that our two domains (intra-individual levels of analysis) may have any of these four conditions present when comparing the domains we have a total of 16 (4 × 4) possible combinations to consider. To simplify our explanation, in Table 1 we consider six of these possible combinations. We omit discussion of the four combinations with null conditions because a null condition simply indicates that we would need to consider an altogether different level of analysis (something other than an individual leader and the leader’s work and family/friends intra-individual levels). We also simplify the discussion of the remaining 12 combinations down to 6 by noting that the arguments and inferences are the same if we reverse the domain classifications for unlike combinations of levels of analysis (represented as scenarios 4–6 in Table 1). It is important to note that in Step 1, we are considering the appropriate level of analysis in the same way as presented in prior leadership research (e.g., transformational leadership [Yammarino & Bass, 1990]; behavioral integrity [Palanski & Yammarino, 2011]; fairness/empowering leadership [De Hoogh & Den Hartog, 2008]). The only difference is that we are considering it two times (once for each respective domain).
Research Question 1a and 1b: What is the operative level of analysis condition for ethical leadership in the (a) work domain and (b) friends and family domain?
Although not our primary concern, it is also necessary to investigate the operative level of analysis for our outcome variables. While we recognize that many studies naturally focus on the outcome constructs of interest, in this investigation we take the liberty of including three well-established “usual suspect” outcomes because our primary interest is on the multi-domain and multi-level characteristics of ethical leadership as an exogenous (independent) construct. Specifically, we focus on perhaps the most common outcome of ethical leadership (i.e., trust; Mayer et al., 1995). Numerous studies and multiple meta-analyses (Bedi et al., 2016; Hoch et al., 2018; Ng & Feldman, 2015) have firmly established trust in the leader as a robust and enduring result of consistent ethical leadership. Often, studies focus on the role of trust in the leader as a mediator between ethical leadership and other outcomes such as employee well-being (e.g., Chughtai et al., 2015). In this study, we examine trust in the forms of both affective trust and cognitive trust (Newman et al., 2014). Likewise, we also focus on an important “anti-ethical” outcome (i.e., abusive supervision) (Tepper, 2007) because of its ubiquity and importance in the broader ethical leadership literature.
Research Question 1c and 1d: What is the operative level of analysis condition for cognitive trust, affective trust, and abusive supervision, respectively, in the (c) work domain and (d) friends and family domain?
Step 2: Relationship of Ethical Leadership and Outcome Variables
Having assessed the operative level of analysis of ethical leadership for both work and friends/family in Step 1, the purpose of Step 2 is to identify the relevant variables to be used for investigating the impact of ethical leadership on outcome variables (in our investigation, cognitive trust, affective trust, and abusive supervision, respectively) and to decide how to test these relationships.
Identifying the relevant variables is fairly simple. For a wholes condition, one may use the individual-level mean score of ethical leadership; that is, the average score for all raters within a domain. For a parts condition, one needs to choose and calculate some measure of variability of the scores from raters within a particular domain, for example, standard deviation (SD). For an equivocal condition, one may use the mean (representing between-leader variation), a chosen variability measure such as SD (representing within-leader variation), or some combination of these two.
Although identification of the relevant variables is fairly straightforward, there are multiple challenges with deciding how to use these variables in an empirical test with respect to outcome variables. First, there is a challenge in interpreting any combinations of ethical leadership that include a parts condition (e.g., wholes for Domain A with parts for Domain B); second, there is a challenge of interpreting ethical leadership variables that include an equivocal condition; third, even if we consider outcome variables for each domain separately, there is a challenge for interpreting any outcome variables which occur in an equivocal condition. Table 1 provides guidance for interpreting the results in any of these scenarios. With these different scenarios in mind, we ask the following research question.
Research Question 2: What is the relationship between ethical leadership and (a) cognitive trust, (b) affective trust, and (c) abusive supervision with respect to the level (mean) and variability of ethical leadership at work and with friends and family?
Methods
Sample and Procedures
Data were collected from 124 working adults in executive MBA, executive education, and MBA programs. Within their programs, participants completed a multi-domain 360-degree survey administered by the authors at three universities in the United States and one university in Ireland. Participants could opt in to receive feedback in up to three domains (work, community organizations, and friends/family) (NB: data focused on community organizations were not used in this study). Participants who elected to receive feedback in more than one domain were instructed to complete the questions about each respective domain in their own survey on different days in order to focus attention on each domain individually. We received responses from 1,458 raters (926 in work and 532 in friends/family), but only participant leaders who had multiple ratings in both domains (work and friends/family, respectively) were included in our final sample. Thus, our final sample included 1,310 responses (782 in work and 528 in friends/family domain) on 104 participants. Each participant averaged 7.5 raters in the work domain (SD = 4.05, range = 2–23 raters) and 5.1 raters in the friends and family domain (SD = 2.46, range = 2–14 raters).
Measures
Ethical Leadership
Following Mayer et al. (2012), we used the 10-item ethical leadership scale developed by Brown et al. (2005), which includes the two sub-dimensions of moral person (5 items) and moral manager (5 items). However, also like Mayer et al. (2012) and others, the subscales did not hold independently in our study; therefore, we collapsed all 10 items into a single ethical leadership scale. We slightly modified some statements to better suit use across domains. For instance, “discusses business ethics or values with employees” was modified to “discusses ethics and values with others.” The items were assessed on a scale from 1 (strongly disagree) to 5 (strongly agree). Sample items include “this person asks what is the right thing to do when making decisions” and “conducts his/her personal life in an ethical manner” (α = 0.91 in work, α = 0.87 in friends and family).
Trust
The scales for cognitive and affective trust were taken from Yang and Mossholder (2010). Each scale includes five items on a scale of 1 (strongly disagree) to 5 (strongly agree). Yang and Mossholder’s (2010) original scale development research distinguishes parallel scales for affective and cognitive trust across two foci: management and supervisor. For consistency across ratings sources within our 360, we changed the referent to “this person.” Sample items for cognitive trust include “I can depend on this person to meet his/her responsibilities” and “given this person’s track record, I see no reason to doubt his/her competence.” Sample items of affective trust include “I’m confident that this person will always care about my personal needs” and “I’m sure I could openly communicate my feelings to this person” (α = 0.93 for cognitive trust in work, α = 0.87 for cognitive trust in friends and family, α = 0.94 for affective trust in work, and α = 0.93 for affective trust in friends and family).
Abusive Supervision
Following Mitchell and Ambrose (2007), we measured abusive supervision with a shortened (5-item) scale from Tepper’s (2000) original abusive supervision measure. Raters indicated the frequency to which the individual demonstrated abusive behaviors such as “ridicules me” and “puts me down in front of others.” The items were rated on a 1–5 scale from 1= I cannot remember him/her ever using this behavior with me to 5 = he/she uses this behavior very often with me (α =.83 in work and α = 0.82 in friends and family).
Variability
Similar to Bormann et al. (2018), we used SD across raters as an operationalization of variability. SD is a robust indicator of dispersion especially when also examining strength or interaction effects (Roberson et al., 2007).
Analytic Strategy and Results
Step 1: Variability Within Domain (Intra-individual) Levels
To address Research Question 1, which focuses on understanding the levels of analysis of ethical leadership, cognitive trust, affective trust, and abusive supervision in two domains (i.e., work, friends and family), we used Within and Between Analysis (WABA) (Dansereau et al., 1984) via the DETECT software package.1 Similar to intraclass correlation (ICC), WABA examines within- and between-group variance and can be utilized to make decisions on aggregation. However, unlike ICC, WABA can provide additional insights into group-level effects as it also considers heterogeneity within groups (group parts levels of analysis) (Dixon & Cunningham, 2006). As we were not solely focused on testing for aggregation, but rather we were exploring and open to different possibilities of levels of analysis, WABA was more suitable for our analysis purpose.
In WABA, variations within an entity (e.g., dyad, group) and variations between entities (e.g., dyad, group) are compared and the ratio of the two variations is tested for practical (E-tests) and statistical significance (F-tests). As noted in Dansereau et al. (1984), practical significance tests (E-tests), which are free from the effect of sample size, are conducted first to check which eta-correlation was larger and whether the ratio (i.e., E ratio) is significant by examining its geometric properties via 15-degree test or 30-degree test for a more conservative test (see Dansereau et al., 1984, p. 170 for more information). Based on the E-test results, statistical significance (F-tests), which accounts for the degrees of freedom/sample size, is calculated. Specifically, if between-eta correlation is larger than within-eta correlation, traditional F is calculated, and if within-eta correlation is larger than between-eta correlation, corrected F (inverse F) is computed. Results of both significance tests are used to make inference about whether the level of the variable of interest is wholes, parts, or equivocal (Dansereau et al., 1984; Yammarino, 1998). Wholes are inferred if between-entity variations are significantly greater than within-entity variations (E ≥ 1.30); parts are inferred if within-entity variations are significantly greater than between-entity variations (0.77 ≥ E ≥ 0); equivocal is inferred if either of between-entity variations or within-entity variations is not significantly greater than the other (1.30 > E > 0.77), meaning that both variations are meaningful.
The analysis was conducted separately for each domain. The input data included scale scores of all variables (i.e., ethical leadership, cognitive trust, affective trust, and abusive supervision) and were organized by each leader. The unit of each cell was the individual leader and the unit within each cell was raters’ responses on each leader. In other words, the cell size of each leader was the number of raters for each leader. For each variable in both domains, between-eta correlations (i.e., variations between leaders) and within-eta correlations (i.e., variations within each leader) were calculated and the ratio of the two eta correlations was tested for practical (E-tests) and statistical significance (F-tests). Inferences on the levels of the variables were made based on both practical and statistical significance test results. Descriptive statistics of variables in work domain and friends/family domain are presented in Table 2 and Table 3 3, respectively.
Research Question 1 explored the level of ethical leadership (RQ 1a) and outcome variables (RQ 1c) in the work domain, as well as ethical leadership (RQ 1b) and outcome variables (RQ 1d) in the friends/family domain. For the work domain, as presented in Table 4, difference between the within-eta correlations and the between-eta correlations for ethical leadership (E = 0.60, + θ > 15°; F = 0.43, p = 1.00), cognitive trust (E = 0.51, ++ θ > 30°; F = 0.58, p = 1.00), affective trust (E = 0.52, ++ θ > 30°; F = 0.56, p = 1.00), and abusive supervision (E = 0.68, + θ > 15°; F = 0.33, p = 1.00) were all practically significant but not statistically significant. Although E-test results alone suggested significant leader parts/within-leader effects (i.e., significant variation within each leader but no variation between leaders) as the E ratios for all four variables were between the 0.77 and 1.30 range, the within-leader effects should be considered weak statistically, given that corrected F-tests were not statistically significant. In other words, within variations were not significantly greater than between variations. Considering the E- and F-test results together, it would be safe to say that the operative level of analysis for all four variables was equivocal, meaning significant variability in both within and between leaders.
| Eta Correlations | E | F | Inference | ||
| Between | Within | ||||
| EL | 0.51 | 0.86 | 0.60+ | 0.43 | Equivocal |
| CT | 0.46 | 0.89 | 0.51++ | 0.58 | Equivocal |
| AT | 0.46 | 0.89 | 0.52++ | 0.56 | Equivocal |
| AS | 0.56 | 0.83 | 0.68+ | 0.33 | Equivocal |
| Note: df for (between) F = 103, 678; df for (within) F = 678, 103. +θ > 15°. ++θ > 30°. EL, ethical leadership; CT, cognitive trust; AT, affective trust; AS, abusive supervision. | |||||
Likewise, regarding the friends/family domain as shown in Table 5, the difference between the within-eta correlations and the between-eta correlations for ethical leadership (E = 0.63, + θ > 15°; F = 0.61, p = 1.00), cognitive trust (E = 0.56, ++ θ > 30°; F = 0.76, p = 0.97), affective trust (E = 0.63, + θ > 15°; F = 0.61, p = 1.00), and abusive supervision (E = 0.69, + θ > 15°; F = 0.52, p = 1.00) were also practically significant. Specifically, E ratios were in the 0.77 and 1.30 range, which suggested leader parts/within-leader effects. However, the results were not statistically significant given the corrected F-tests results. Again, the E- and F-test results together suggested that the operative level of analysis for all four variables was equivocal.
| Eta Correlations | E | F | Inference | ||
| Between | Within | ||||
| EL | 0.53 | 0.84 | 0.63+ | 0.61 | Equivocal |
| CT | 0.49 | 0.87 | 0.56++ | 0.76 | Equivocal |
| AT | 0.53 | 0.85 | 0.63+ | 0.61 | Equivocal |
| AS | 0.57 | 0.82 | 0.69+ | 0.52 | Equivocal |
| Note: df for (between) F = 103, 424; df for (within) F = 424, 103. +θ > 15°. ++θ > 30°. EL, ethical leadership; CT, cognitive trust; AT, affective trust; AS, abusive supervision. | |||||
In summary, both variability within leader and between leaders for ethical leadership and all three outcome variables (cognitive trust, affective trust, and abusive supervision, respectively) should be considered in further analysis. Thus, in Step 2, we proceed with linear regression as indicated in Scenario 3 (equivocal/equivocal).
Step 2: Relationship of Ethical Leadership and Outcome Variables
Given the WABA I results, which suggested that all variables in both domains had significant variance within the individual leader and between leaders, aggregated mean and SD (LeBreton & Senter, 2008) of all variables for each leader were calculated and used in our analysis. Following our suggestion presented earlier regarding Scenario 3 (equivocal/equivocal), we used linear regression to explore our research questions on the relationship between variability of ethical leadership and various outcomes (RQ 2). As the aggregated means and SDs of all independent and dependent variables were at the same level (i.e., leader) and our models did not involve any cross-level effects, we applied multiple linear regression. Specifically, we regressed each outcome variable (the mean and SD, respectively, for cognitive trust, affective trust, and abusive supervision, respectively, for both the work and family/friends domains, respectively—a total of 12 outcome variables) on the mean of ethical leadership at work, the SD of ethical leadership at work, the mean of ethical leadership with friends/family, and the SD of ethical leadership of friends/family (a total of four predictor variables).
Means, SDs, and correlations among variables are presented in Table 6. Table 7 shows the results for all relationships tested. Rather than repeat the same information from Table 7 in written form, we instead highlight some key findings according to the framework we suggested for answering Research Questions 2a, 2b, and 2c.
| Variables | Cognitive Trust | Affective Trust | Abusive Supervision | |||||||||
| Work | Friends/Family | Work | Friends/Family | Work | Friends/Family | |||||||
| W_CT_m | W_CT_sd | F_CT_m | F_CT_sd | W_AT_m | W_AT_sd | F_AT_m | F_AT_sd | W_AS_m | W_AS_sd | F_AS_m | F_AS_sd | |
| Constant | 2.12*** | 1.10** | 2.55*** | 1.68*** | 0.46 | 0.93+ | 2.14*** | 0.56 | 3.54*** | 2.19*** | 4.66*** | 2.01** |
| W_EL_m | 0.63*** | −0.21*** | −0.04 | 0.09 | 0.95*** | −0.05 | −0.10 | 0.28** | −0.54*** | −0.38*** | 0.01 | 0.04 |
| W_EL_sd | −0.03 | 0.46*** | 0.13 | −0.14 | 0.04 | 0.76*** | 0.05 | 0.00 | 0.11 | 0.33+ | 0.34* | 0.30+ |
| F_EL_m | −0.05 | 0.01 | 0.54*** | −0.41*** | −0.08 | −0.08 | 0.63*** | −0.36*** | −0.01 | −0.09 | −0.80*** | −0.47*** |
| F_EL_sd | −0.17 | 0.09 | −0.16+ | 0.44*** | 0.02 | −0.22+ | −0.59*** | 0.84*** | −0.02 | −0.07 | −0.04 | 0.31* |
| R2 | 0.44 | 0.26 | 0.48 | 0.42 | 0.51 | 0.18 | 0.51 | 0.45 | 0.28 | 0.20 | 0.41 | 0.27 |
| F | 21.29 | 9.95 | 25 | 19.37 | 27.73 | 6.84 | 27.64 | 21.97 | 10.98 | 7.33 | 18.98 | 10.42 |
| Note: N = 104 leaders. Unstandardized regression coefficients are reported. +p < 0.1. *p < 0.05. **p < 0.01. ***p < 0.001. w, work domain; f, friends and family domain; EL, ethical leadership; CT, cognitive trust; AT, affective trust; AS, abusive supervision; m, mean; SD, standard deviation. | ||||||||||||
Mean to Mean
Results indicated a positive and significant relationship between the mean of ethical leadership at work and the means of cognitive trust (β = 0.63, p < 0.001) and affective trust (β = 0.95, p < 0.001) at work and a negative and significant relationship between the mean of ethical leadership at work and abusive supervision (β = −0.54, p < 0.001) at work. The same pattern of results held for the friends and family domain as well. Specifically, the mean of ethical leadership in friends/family domain had a positive and significant relationship with the means of cognitive trust (β = 0.54, p < 0.001) and affective trust (β = 0.63, p < 0.001) in the same domain, and a negative and significant relationship with abusive supervision (β = −0.80, p < 0.001) in friends/family domain. There were no significant cross-domain relationships between ethical leadership and outcome variables.
Variation to Mean
Results indicated no significant relationship between the SD of ethical leadership at work and the means of cognitive trust (β = −0.03, p = n.s.), affective trust (β = 0.04, p = n.s.), and abusive supervision (β = 0.11, p = n.s.) at work. However, in the friends/family domain, there was a negative and significant relationship between the SD of ethical leadership and the mean of both cognitive (β = −0.16, p < 0.10 [marginally significant]) and affective trust (β = −0.59, p < 0.001), but not abusive supervision (β = −0.04, p = n.s.). There were no significant cross-domain relationships between ethical leadership and outcome variables. Interestingly, there is a significant and positive cross-domain relationship between the SD of ethical leadership at work and the mean of abusive supervision (β = 0.34, p < 0.05) in the friends/family domain.
Mean to Variation
Results indicated a negative and significant within-domain relationship between the mean of ethical leadership and SD of cognitive trust in both domains (work domain: β = −0.21, p < 0.001; friends/family domain: β = −0.41, p < 0.001), affective trust (β = −0.36, p < 0.001) in friends/family domain, and abusive supervision in both domains (work domain: β = −0.38, p < 0.001; friends/family domain: β = −0.47, p < 0.001). However, the relationship between the mean ethical leadership and the SD of affective trust (β = −0.05, p = n.s.) at work was not significant. In terms of cross-domain relationship, results indicated a significant and positive cross-domain relationship of the mean ethical leadership at work and the SD of affective trust (β = 0.28, p < 0.01) with friends/family.
Variation to Variation
Results indicated a positive and significant within-domain relationship between the SDs of ethical leadership and the SDs of cognitive trust (work domain: β = 0.46, p < 0.001; friends/family domain: β = 0.44, p < 0.001), affective trust (work domain: β = 0.76, p < 0.001; friends/family domain: β = 0.84, p < 0.001), and abusive supervision (work domain: β = 0.33, p < 0.10 [marginally significant]; friends/family domain: β = 0.31, p < 0.05) both at work and in friends/family (note: the work domain ethical leadership–abusive supervision relationship is marginally significant). There are also two marginally significant cross-domain relationships: first between work ethical leadership and friends/family abusive supervision (positive) (β = 0.30, p < 0.10) and second between friends/family ethical leadership and work affective trust (negative) (β = −0.22, p < 0.10).
Discussion
The purpose of the present study was to examine questions of variability in ethical leadership and leadership outcomes. Using data from a 360-degree survey that includes ratings from observers both at work (i.e., bosses, peers, and direct reports) and non-work (i.e., friends and family members), we first examined the operative level of analysis condition for ethical leadership. Our WABA results suggest that an “equivocal” inference is appropriate for both ethical leadership and our outcome variables of cognitive trust, affective trust, and abusive supervision. These results highlight the importance of both the level and the variation in ratings. In general, approximately 25%–30% of the variance resided between leaders, whereas 70%–75% resided within leaders, as rated by others. Our findings roughly correspond to previous WABA results of leadership behaviors and outcomes (e.g., Den Hartog & De Hoogh, 2009; Palanski & Yammarino, 2011) and Bormann and colleague’s (2018) study of within-unit variability. While others have examined constructs related to ethical leadership such as behavioral integrity (Palanski & Yammarino, 2011) and empowering leadership and fairness (Den Hartog & De Hoogh, 2009), we know of no WABA analyses on ethical leadership itself, particularly using the most frequently used measure—that of Brown et al. (2005). Likewise, we know of no studies that extend beyond a leader’s direct subordinates to consider variability in the larger work and non-work contexts. Dionne and colleagues (2014) highlighted a lack of levels reflected in data analysis for ethical leadership, which is a contribution of this study.
First, we consider mean ratings of ethical leadership and our outcome variables. These analyses are the most common approach. Our findings are very consistent with previous findings highlighting the positive relationships of ethical leadership and trust (Bedi et al., 2016; Hoch et al., 2018; Ng & Feldman, 2015) and we show evidence that these relationships extend within the friends/family domain. The relationships were slightly stronger for affective than cognitive-based trust, which is the opposite of Newman and colleagues’ (2014) findings at an individual level. This pattern also held in the friends and family domain, although the relationships were generally weaker. Likewise, we found a negative relationship between mean ethical leadership and abusive supervision, consistent with previous research (Lin et al., 2016). Interestingly, this relationship was even stronger in friends and family domains than in work. Taken together, these findings highlight the robustness of these relationships even beyond work.
Second, we consider the variability of ethical leadership on mean level outcomes. These analyses most directly describe the impact of ethical leadership variability on outcomes. Interestingly, we found no relationships between the deviations of ethical leadership on outcomes within the work domain. This contrasts with the only other study we know of examining ethical leadership variability (Bormann et al., 2018), who found small negative relationships between variability and trust in the leader in general. Perhaps these relationships fade when considering other relationships (peers and supervisors) and when distinguishing dimensions of trust. However, these relationships did hold in the friends and family domain, such that greater variability in perceptions of ethical leadership in the family/friends domain was related to reduced affective trust and approached significance for cognitive-based trust. Inconsistent attention to ethics and morality in personal relationships may be more harmful to emotion-based trust, which may be a more personally vulnerable form of trust—“trust from the heart” (Chua et al., 2008).
When individuals are assessing their leaders’ ethical leadership, they may be holding observed behaviors against their own implicit moral and ethical ideals, leading to increased variance in ethical leadership across raters (Keck et al., 2018). Indeed, ratings of ethical leadership reflect the extent to which a leaders’ relational style matches followers’ ideal interactional norms which form the basis of moral imperatives in relationships (Keck et al., 2018). These findings suggest that assessments of ethical leadership are determined in part by individually determined normally appropriate conduct in leader-follower relationships. Beyond different determinants of morally appropriate behavior, variance in observer ratings of ethical leadership may also reflect unique relationships within work teams or departments or individual differences in perception of similar behavior. For example, some individuals may have higher-quality relationships with their leaders and assess them more globally positive.
Third, we examined the relationships between the mean levels of ethical leadership on the variability of the outcomes. With the exception of affective trust in work, all pairs of relationships were significant and negative such that higher ethical ratings were related to greater consensus in ratings of outcomes. That is, leaders perceived to be more ethical overall had more similar ratings of trust and abusive supervision. We could argue that these findings are in line with an integrity-based view of ethical leadership, such that higher levels of ethical leadership produce a similar response across raters. We know of no studies examining this type of relationship, which is worthy of future study as the relationships seem to be robust.
Fourth, we examined the relationships among the variability within ethical leadership and in outcomes. These relationships also appeared to be quite robust, yet we know of no research primarily focused on variation. In general, these relationships were strong and positive across all outcomes. The relationships were strongest with affective trust in both domains, followed by cognitive-based trust in both domains. The relationships were weakest with variability in abusive supervision and only approached significance in the work domain.
Finally, our data collection in work and family/friends networks allowed us to examine cross-domain effects. In their original description of ethical leaders, Brown and Treviño (2006, p. 597) describe ethical leaders as individuals “who behave ethically in their personal and professional lives”; therefore, we were curious to see if there is support for cross-domain relationships, particularly from the family/friends domain to work outcomes. However, only two of the possible 24 cross-domain relationships were significant (and one additional approached significance), suggesting these relationships reside largely within domain. Interestingly, the two significant relationships were from work ethical leadership to family/friend outcomes. The mean of work ethical leadership was positively related to variability within affective trust in friends and family domain, which was the opposite direction from the within-person effects. Perhaps this reflects something akin to the “kick the dog effect” in which work stressors spillover in home life. Additionally, the deviation in work ethical leadership was positively related to mean levels of abusive supervision in the family/friends domain, such that greater deviation in perceptions of work ethical leadership was associated with more abusive supervision with family and friends. While some literature has begun to examine the effects of leadership on follower work-family outcomes, very little has examined leaders’ own experiences outside of work. We echo calls to examine leaders’ own cross-domain experiences and relationships (Hammond et al., 2017).
Future Research Directions, Limitations, and Conclusion
As noted earlier, there are interesting opportunities for future research particularly examining rater variability and multiple domains, but there are also some variations in this approach that might be worth considering. First, it might be interesting to include a leader’s self-rating of his/her ethical leadership in both domains. It is possible that the relative level of agreement between self- and other ratings might impact the nature of the relationships under investigation. For example, a leader who is rated similarly by others across domains and whose self-ratings agree with observer ratings might indicate more intentionality in ethical leadership than a leader who has significant disagreement between self- and other ratings.
Additionally, although prior studies have generally not found an empirical distinction between the moral manager and moral person aspects of ethical leadership, it is possible that there is a meaningful distinction in the family/friends domain. For example, the moral manager aspect of “disciplining others who violate ethical standards” may not be nearly as relevant in personal relationships. If such a distinction holds, then it may be desirable to consider a third level of analysis (based on moral manager and moral person) for the family/friends domain.
Finally, future research is needed to address limitations of this study. While the multi-domain 360-degree survey provided an excellent opportunity to collect responses from many raters both in work and family/friends domains, the data were collected cross sectionally. Longitudinal designs are necessary for examining any causal claiming such as the effect of variation on outcomes. Additionally, as part of the 360-degree survey, leaders self-nominated their raters. While leaders were encouraged to invite raters across a variety of relationships, it is possible that some leaders selected those with whom they held close relationships or might view them more favorably.
In conclusion, in this article, we have explored both the work and non-work (friends/family) aspects of ethical leadership, as well as the levels of analysis of ethical leadership. Our findings emphasized the importance of examining not only levels of leader behaviors and outcomes, but also variation in ratings as a meaningful variable of interest both in terms of leader behaviors (Bormann et al., 2018) and follower outcomes. We also provided several recommendations on issues and analytic strategies associated with these examinations in Table 1. Likewise, the study contributed to the multi-domain leadership theory (Hammond et al., 2017) by replicating and extending previous research findings of leader behaviors on outcomes to domains outside of work as well as highlighting cross-domain relationships.
References
| Bai, Y., Lin, L., & Liu, J. T. (2019). Leveraging the employee voice: A multi-level social learning perspective of ethical leadership. The International Journal of Human Resource Management, 30(12), 1869–1901. https://doi.org/10.1080/09585192.2017.1308414 |
| Bedi, A. C., Alpaslan, M., & Green, S. (2016). A meta-analytic review of ethical leadership outcomes and moderators. Journal of Business Ethics, 139(3), 517–536. https://doi.org/10.1007/s10551-015-2625-1 |
| Blasi, A. (1993). The development of identity: Some implications for moral functioning. In: G. G. Noam & T. E. Wren (Eds.), The moral self (pp. 99–122). MIT Press. |
| Bleidorn, W., & Denissen, J. (2015). Virtues in action—The new look of character traits. British Journal of Psychology, 106, 700–723. https://doi.org/10.1111/bjop.12117 |
| Bormann, K. C., Poethke, U., Cohrs, C., & Rowold, J. (2018). Doing bad through being selective in doing good: The role of within-unit variability in ethical leadership. European Journal of Work and Organizational Psychology, 27(6), 683–699. https://doi.org/10.1080/1359432X.2018.1491550 |
| Brett, J. F., & Atwater, L. E. (2001). 360° feedback: Accuracy, reactions, and perceptions of usefulness. Journal of Applied Psychology, 86(5), 930–942. https://doi.org/10.1037/0021-9010.86.5.930 |
| Brown, M. E., & Treviño, L. K. (2006). Ethical leadership: A review and future directions. The Leadership Quarterly, 17, 595–616. https://doi.org/10.1016/j.leaqua.2006.10.004 |
| Brown, M. E., Treviño, L. K., & Harrison, D. (2005). Ethical leadership: A social learning perspective for construct development and testing. Organizational Behavior and Human Decision Processes, 97, 117−134. https://doi.org/10.1016/j.obhdp.2005.03.002 |
| Chua, R. Y. J., Ingram, P., & Morris, M. W. (2008). From the head and the heart: Locating cognition- and affect-based trust in managers’ professional networks. Academy of Management Journal, 51, 436–452. https://doi.org/10.5465/amj.2008.32625956 |
| Chughtai, A., Byrne, M., & Flood, B. (2015). Linking ethical leadership to employee well-being: The role of trust in supervisor. Journal of Business Ethics, 128(3), 653–663. https://doi.org/10.1007/s10551-014-2126-7 |
| Courtright, S. H., Gardner, R. G., Smith, T. A., McCormick, B. W., & Colbert, A. E. (2016). My family made me do it: A cross-domain, self-regulatory perspective on antecedents to abusive supervision. Academy of Management Journal, 59(5), 1630–1652. https://doi.org/10.5465/amj.2013.1009 |
| Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Prentice Hall. |
| De Hoogh, A. H. B., & Den Hartog, D. N. (2008). Ethical and despotic leadership relationships with leader’s social responsibility, top management effectiveness and subordinates’ optimism: A multimethod study. The Leadership Quarterly, 19, 297–311. https://doi.org/10.1016/j.leaqua.2008.03.002 |
| Den Hartog, D. N., & De Hoogh, A. H. B. (2009). Empowering behaviour and leader fairness and integrity: Studying perceptions of ethical leader behaviour from a levels-of-analysis perspective. European Journal of Work and Organizational Psychology, 18(2), 199–230. https://doi.org/10.1080/13594320802362688 |
| Dionne, S. D., Gupta, A., Sotak, K. L., Shirreffs, K. A., Serban, A., Hao, C., & Yammarino, F. J. (2014). A 25-year perspective on levels of analysis in leadership research. The Leadership Quarterly, 25(1), 6–35. https://doi.org/10.1016/j.leaqua.2013.11.002 |
| Dixon, M. A., & Cunningham, G. B. (2006). Data aggregation in multilevel analysis: A review of conceptual and statistical issues. Measurement in Physical Education and Exercise Science, 10(2), 85–107. https://doi.org/10.1207/s15327841mpee1002_2 |
| Eisenbeiß, S. A., & Giessner, S. R. (2012). The emergence and maintenance of ethical leadership in organizations. Journal of Personnel Psychology, 11(1), 7–19. https://doi.org/10.1027/1866-5888/a000055 |
| Eisenbeiss, S. A., van Knippenberg, D., & Fahrbach, C. M. (2015). Doing well by doing good? Analyzing the relationship between CEO ethical leadership and firm performance. Journal of Business Ethics, 128(3), 635–651. https://doi.org/10.1007/s10551-014-2124-9 |
| Fleeson, W. (2007). Situation-based contingencies underlying trait content manifestation in behavior. Journal of Personality, 75, 825–861. https://doi.org/10.1111/j.1467-6494.2007.00458.x |
| Hammond, M., Clapp-Smith, R., & Palanski, M. (2017). Beyond (just) the workplace: A theory of leader development across multiple domains. Academy of Management Review, 42(3), 481–498. https://doi.org/10.5465/amr.2014.0431 |
| Hannah, S. T., Thompson, R. L., & Herbst, K. C. (2018). Moral identity complexity: Situated morality within and across work and social roles. Journal of Management, 46(5), 726–757. https://doi.org/10.1177/0149206318814166 |
| Hoch, J. E., Bommer, W. H., Dulebohn, J. H., & Wu, D. (2018). Do ethical, authentic, and servant leadership explain variance above and beyond transformational leadership? A meta-analysis. Journal of Management, 44(2), 501–529. https://doi.org/10.1177/0149206316665461 |
| House, R., Javidan, M., Hanges, P., & Dorfman, P. (2002). Understanding cultures and implicit leadership theories across the globe: An introduction to project GLOBE. Journal of World Business, 37(1), 3–10. https://doi.org/10.1016/S1090-9516(01)00069-4 |
| Javidan, M., Dorfman, P. W., De Luque, M. S., & House, R. J. (2006). In the eye of the beholder: Cross cultural lessons in leadership from Project GLOBE. The Academy of Management Perspectives, 20(1), 67–90. https://doi.org/10.4324/9781315668703 |
| Keck, N., Giessner, S. R., Van Quaquebeke, N., & Kruijff, E. (2018). When do followers perceive their leaders as ethical? A relational models perspective of normatively appropriate conduct. Journal of Business Ethics, 164(3), 1–17. https://doi.org/10.1007/s10551-018-4055-3 |
| Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19(2), 195–229. https://doi.org/10.5465/amr.1994.9410210745 |
| Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal, and emergent processes. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 3–90). Jossey-Bass. https://psycnet.apa.org/record/2000-16936-001 |
| LeBreton, J. M., & Senter, J. L. (2008). Answers to 20 questions about interrater reliability and interrater agreement. Organizational Research Methods, 11(4), 815–852. https://doi.org/10.1177/1094428106296642 |
| Lee, A., & Carpenter, N. C. (2018). Seeing eye to eye: A meta-analysis of self-other agreement of leadership. The Leadership Quarterly, 29(2), 253–275. https://doi.org/10.1016/j.leaqua.2017.06.002 |
| Lester, H. F., Cullen-Lester, K. L., & Walters, R. W. (2021). From nuisance to novel research questions: Using multilevel models to predict heterogeneous variances. Organizational Research Methods, 24(2), 342–388. https://doi.org/10.1177/1094428119887434 |
| Liao, Y., Liu, X. Y., Kwan, H. K., & Li, J. (2015). Work–family effects of ethical leadership. Journal of Business Ethics, 128(3), 535–545. https://doi.org/10.1007/s10551-014-2119-6 |
| Lin, S. H., Scott, B. A., & Matta, F. K. (2019). The dark side of transformational leader behaviors for leaders themselves: A conservation of resources perspective. Academy of Management Journal, 62(5), 1556–1582. https://doi.org/10.5465/amj.2016.1255 |
| Lin, S. H. J., Ma, J., & Johnson, R. E. (2016). When ethical leader behavior breaks bad: How ethical leader behavior can turn abusive via ego depletion and moral licensing. Journal of Applied Psychology, 101(6), 815–830. https://psycnet.apa.org/doi/10.1037/apl0000098 |
| Magalhães, A., dos Santos, N. R., & Pais, L. (2019). Multi-source research designs on ethical leadership: A literature review. Business and Society Review, 124(3), 345–364. https://doi.org/10.1111/basr.12179 |
| 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(1), 151–171. https://doi.org/10.5465/amj.2008.0276 |
| Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709–734. https://doi.org/10.5465/amr.1995.9508080335 |
| Mitchell, M. S., & Ambrose, M. L. (2007). Abusive supervision and workplace deviance and the moderating effects of negative reciprocity beliefs. Journal of Applied Psychology, 92(4), 1159–1168. https://doi.org/10.1037/0021-9010.92.4.1159 |
| Newman, A., Kiazad, K., Miao, Q., & Cooper, B. (2014). Examining the cognitive and affective trust-based mechanisms underlying the relationship between ethical leadership and organisational citizenship: A case of the head leading the heart? Journal of Business Ethics, 123(1), 113–123. https://doi.org/10.1007/s10551-013-1803-2 |
| Newstead, T., Dawkins, S., Macklin, R., & Martin, A. (2021). We don’t need more leaders—We need more good leaders Advancing a virtues-based approach to leader (ship) development. J The Leadership Quarterly, 32, 101312. https://doi.org/10.1016/j.leaqua.2019.101312 |
| Ng, T. W., & Feldman, D. C. (2015). Ethical leadership: Meta-analytic evidence of criterion-related and incremental validity. Journal of Applied Psychology, 100(3), 948–965. https://doi.org/10.1037/a0038246 |
| Palanski, M. E., & Yammarino, F. J. (2011). Impact of behavioral integrity on follower job performance: A three-study examination. The Leadership Quarterly, 22(4), 765–786. https://doi.org/10.1016/j.leaqua.2011.05.014 |
| Riggio, R. E., Zhu, W., Reina, C., & Maroosis, J. A. (2010). Virtue-based measurement of ethical leadership: The Leadership Virtues Questionnaire. Consulting Psychology Journal: Practice and Research, 62(4), 235–250. https://doi.org/10.1037/a0022286 |
| Roberson, Q. M., Sturman, M. C., & Simons, T. L. (2007). Does the measure of dispersion matter in multilevel research? A comparison of the relative performance of dispersion indexes. Organizational Research Methods, 10(4), 564–588. https://doi.org/10.1177/1094428106294746 |
| Simons, T., Leroy, H., Collewaert, V., & Masschelein, S. (2015). How leader alignment of words and deeds affects followers: A meta-analysis of behavioral integrity research. Journal of Business Ethics, 132(4), 831–844. https://doi.org/10.1007/s10551-014-2332-3 |
| Sturm, R. E., Vera, D., & Crossan, M. (2017). The entanglement of leader character and leader competence and its impact on performance. The Leadership Quarterly, 28(3), 349–366. https://doi.org/10.1016/j.leaqua.2016.11.007 |
| Tepper, B. J. (2000). Consequences of abusive supervision. Academy of Management Journal, 43(2), 178–190. https://doi.org/10.5465/1556375 |
| Tepper, B. J. (2007). Abusive supervision in work organizations: Review, synthesis, and research agenda. Journal of Management, 33(3), 261–289. https://doi.org/10.1177/0149206307300812 |
| Yammarino, F. J. (1998). Multivariate aspects of the varient/WABA approach: A discussion and leadership illustration. The Leadership Quarterly, 9(2), 203–227. https://doi.org/10.1016/S1048-9843(98)90005-4 |
| Yammarino, F. J., & Bass, B. M. (1990). Transformational leadership and multiple levels of analysis. Human Relations, 43(10), 975–995. https://doi.org/10.1177/001872679004301003 |
| Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. The Leadership Quarterly, 16(6), 879–919. https://doi.org/10.1016/j.leaqua.2005.09.002 |
| Yang, J., & Mossholder, K. W. (2010). Examining the effects of trust in leaders: A bases-and-foci approach. The Leadership Quarterly, 21(1), 50–63. https://doi.org/10.1016/j.leaqua.2009.10.004 |
| Zhang, S., & Tu, Y. (2018). Cross-domain effects of ethical leadership on employee family and life satisfaction: The moderating role of family-supportive supervisor behaviors. Journal of Business Ethics, 152(4), 1085–1097. https://doi.org/10.1007/s10551-016-3306-4 |
Footnote
1 Available at: https://www.binghamton.edu/som/research/cls/resources.html