An increasing amount of research suggests that morality is a core feature of good leadership (e.g., Mayer et al., 2012; Skubinn & Herzog 2016; Zhu et al., 2011, 2016). In order to be perceived as moral, a leader has to demonstrate moral behaviors that are visible to followers (Treviño et al., 2000); that is, a leader has to show high integrity. However, most previous studies on leader integrity have emphasized consistency between words and actions (Simons, 2002). The risk of focusing only on this behavioral dimension is that it ignores the ethical meaning of integrity: even a tyrant could be described as having high integrity if one looked at word-deed consistency without any regard to the content of the words and actions (Bauman, 2013). Thus, moral leaders should be perceived to both hold moral values and enact those values with a high degree of consistency (Moorman et al., 2013).
One might intuitively assume that a leader who shows high morality and consistent value-aligned behavior could also cultivate positive relationships with their followers. However, so far only few studies have explicitly shown that leader morality can positively relate to the quality of leader-member exchange (Bedi et al., 2016; Giessner et al., 2015; Gu et al., 2015), while no studies have looked how leader integrity might associate with leader-member relationships (Dulebohn et al., 2012; Nahrgang & Seo, 2016). Here it is important to take into account that leaders do not only form one-to-one relationships with individual followers. Rather, in order for the leader to influence follower actions, his or her integrity needs to be recognized by a group of followers (Fields, 2007). Thus, more studies are needed that focus on between-person variability in experiences regarding leaders’ integrity and how these group-related perceptions influence the quality of leader-member exchange (LMX). We aimed to contribute to a better understanding of the followers’ integrity perceptions by simultaneously investigating both (a) the extent of consistency between the followers’ perceptions of their leaders, and (b) individual differences in followers’ perceptions regarding their leaders. We also propose that leaders who are perceived by their followers as showing high integrity are seen to form high quality LMX relationships with them.
The challenge with integrity studies is that they have been held back by an unresolved controversy over construct definition and measurement (Moorman et al., 2013; Palanski & Yammarino, 2009; Tomlinson et al., 2014). Several definitions have emphasized consistency between a leader’s words and deeds, thus focusing on behavioral integrity (Palanski & Yammarino, 2007; Simons, 1999). However, this approach has been criticized for the possibility that integrity can be attributed even to those who act immorally, so long as they act consistently (Bauman, 2013). Thus, the moral dimension should be recognized as an essential part of (leader) integrity.
The existing approaches to moral leadership have also been criticized for their normativity and moral absolutism (for a review, see Lemoine et al., 2019). It is often assumed that there is only one optimal moral behavioral pattern for leaders, although there are actually multiple approaches to what is seen as moral. Lemoine et al. (2019) highlight several unanswered questions related to decision-making where several moral perspectives are present at the same time: Would leaders feel forced to make a choice against one approach and in favor of the other or would they be more likely to find compromises that retain elements of both moral perspectives? Do most leaders make norm- and standard-based decisions, focusing on outcomes and stakeholders, or do they follow their own self-concordant judgments? And finally, would followers perceive their leader as inconsistent, if they are sometimes self-concordant but, in other cases, comply with company policies without personally believing in them?
We approach the question of moral diversity as individual differences in leader integrity. Here, we use leaders’ own descriptions of how they approach and solve moral issues in organizational settings. We used the social-cognitive process model (Berzonsky, 1990) as the theoretical framework of our study to investigate the differences in integrity between leaders. According to this model, there are individual differences in the social-cognitive processes people use to solve personal problems, make decisions, and process relevant information, which can be classified into three orientations (see Berzonsky, 2011 for a review). The first is informational style, which is characterized by openness to actively explore and evaluate different situations based on relevant information. Individuals with informational style negotiate their personal views through active processing, by being skeptical towards their current ideas, interested in expanding their knowledge, and accommodating their self-views when needed. The second is normative style, depicted by conforming to the prescriptions and expectations of significant others. This style is characterized by a relatively automatic tendency to preserve and defend those views that they have adopted and internalized from others. The third is diffuse/avoidant style, where individuals try to avoid confronting problems for as long as possible, for example by procrastinating or postponing making decisions (Berzonsky, 1990). When individuals with a diffuse-avoidant style have to make choices, their actions are primarily influenced by situation-specific determinants, leading to weak commitments, limited self-control, and large variability in their behaviors (Berzonsky & Cieciuch, 2016). Despite their potential to tap into different styles of personal morality and integrity (based on either situational information, prescriptive norms, or avoidance), these approaches have not been used in the moral domain (Lapsley & Hardy, 2017). Our aim was to test the applicability of this processing style framework to capture differences in leaders’ integrity.
We further suggest that integrity styles can act as one possible antecedent of leadership variability in within follower groups (see Bormann et al., 2018). Leaders with the information-oriented style would gather relevant information and then resolve the moral conflict on the basis of their personal values and situational determinants. Previous studies have shown that information orientation is positively associated with several indicators of psychosocial maturity, such as self-reflection, problem-focused coping strategies, cognitive complexity, open-mindedness, and adaptive self-regulation (Berzonsky, 2011). Therefore, we suggest that the informational orientation would depict leaders who are willing to make moral decisions on the basis of relevant information. As moral questions are often multifaceted and complex, this kind of flexibility is needed. However, informational style leaders might not appear as consistent to their followers, as the followers might be unaware of the careful process of gathering and evaluating the available information that the leader has gone through before arriving at a solution. Simons (2002, 2008) and Leroy et al. (2012) have argued that especially if the leader does not communicate their values, it may lead to a perceived incongruence between what they say and do. Thus, these varying, flexible decisions might seem less fair to the followers. Therefore, we propose that leader’s informational integrity style is positively related to follower-rated moral behavior, but negatively related to behavior integrity and consistency (Hypothesis 1).
The diffuse-avoidant style applied to leader integrity would mean avoiding or downplaying morally challenging situations at work, having neither explored nor committed to personal moral values. The diffuse-avoidant style has been shown to associate with weak commitment, and an external locus of control (Berzonsky, 2011). If similar observations can be made in the moral domain, this would represent an integrity style that lacks on both morality and consistency.
The central difference between the informational style and the diffuse-avoidant style should be found from the level of moral commitments. Although both styles are characterized by the tendency to evaluate situations separately and allow the context to affect personal decisions, informational integrity style should include moral commitments that are formed in an informed, self-exploratory fashion that are open to revisions. These leaders would therefore use careful consideration of moral conflicts, through evaluating and prioritizing competing values and expectations. In comparison, diffuse-avoidant style has been linked to low commitments coupled with procrastination, ambivalence, uncertainty, and indecision (Berzonsky & Cieciuch, 2016). Thus, in the moral domain it is more likely to result in being susceptible to external pressures and making decisions that depend strongly on situational influences without having any personal commitments to certain moral values. The diffuse-avoidant style describes leaders who would try to avoid morally problematic situations and find ways to either postpone or refrain from making moral decisions. Thus, it should translate to follower perceptions of low integrity. Therefore, we propose that leader’s diffuse-avoidant integrity style is negatively related to follower-rated moral behavior, behavior integrity, and consistency (Hypothesis 2).
Finally, other leaders can make decisions in moral conflicts on the basis of normative expectations, which can come from significant others (e.g., managers or colleagues) or from official sources, such as company ethical codes. These individuals approach conflicts with a fixed value set and are very unlikely to show flexibility in response to situational information. Thus, they follow a normative style (Berzonsky, 1990). Leaders with a normative style should show a strong sense of commitment and purpose in their moral decision-making, but their primary goal is to defend and preserve their existing self-views (Berzonsky, 2011), thus making them less open to alternative viewpoints. As the normative style has been found to associate with high commitment levels, self-control, a need for structure and cognitive closure, inflexibility, and low tolerance of ambiguity (Berzonsky, 2011), we assume that leaders with these characteristics will be seen as highly consistent across contexts by their followers. However, whether or not the leader will be perceived to show moral behavior and behavioral integrity by their followers might depend on the expectations, rules and norms available in the organization. Thus, although normative leaders’ decision-making is based on firm, even rigid commitments, the motives behind them depend on the affecting “forces” that the leader is exposed to. Therefore, these decisions and actions might not always be based on moral causes or reflect integrity (if the leader has adopted company norms that strive towards economic success without any ethical considerations, for example) to the followers. This led us to hypothesize that leaders’ normative integrity style is positively related to follower-rated consistency (Hypothesis 3).
When investigating the individual differences between leaders who have been assigned with different integrity styles and how their followers perceive them, one central question is the level of agreement versus disagreement in followers’ views of their leader (Fields, 2007; Henderson et al., 2008, 2009). For example, even if a leader might consider him- or herself to be highly consistent between his/her moral beliefs and actions, this internal experience of integrity may not always be recognized and agreed upon by the followers as a collective. Thus far, very limited attention has been given to variability between and within leaders regarding their followers’ perceptions in leadership research (Henderson, 2008, 2009; Jennings et al., 2015; Zagenczyk et al., 2015). Hence, the present study focused on investigating this variability in followers’ leader ratings and relating this variability to leaders’ integrity styles.
In order to assess followers’ perceptions of their leader’s integrity we use the three-dimensional definition of perceived leader integrity by Moorman et al. (2013), which is based on a broad review and synthesis of previous theoretical frameworks and empirical studies. The first dimension assesses moral behavior (e.g., the leader treats people fairly), the second taps into behavioral integrity (e.g., the leader delivers what is promised), and the third dimension assesses consistency across contexts (e.g., the leader does right even when others disagree). This multidimensional model enables to distinguish whether some employee outcomes are more likely to associate with the moral component while others might be more related to the consistency component (Moorman et al., 2013). A more detailed understanding of how the different aspects of perceived integrity relate to follower outcomes will help to develop and refine theories about leader integrity. In this study, it allowed us to examine the separate contributions that moral behavior, integrity, and consistency make on follower attitudes, that is, their perceptions of leader-member exchange relationship quality in the current study.
Our final aim was to investigate how followers’ perceptions of leader integrity associate with their experienced leader-member exchange quality (LMX), which is defined as a social exchange process between a leader and the leader’s subordinates (Dulebohn et al., 2012; Graen & Uhl-Bien, 1995). These relationships can vary in their quality from low (e.g., the relationship is based only on the employment contract) to high (e.g., the relationship is characterized by mutual respect and trust). According to the most used definitions of LMX, the degree of reciprocity varies, so that leaders are more closely linked with some subordinates than with others (Graen & Uhl-Bien, 1995). LMX theory has previously been focused at the dyad level of analysis, where leaders are seen to treat subordinates differently and independently of other dyadic relationships within the same group. During the past decade LMX has been conceptualized also at the team level (LMX differentiation; Henderson et al., 2009), which has been typically operationalized as the within-team variability of individual LMX (e.g., Li & Liao, 2014).
The different ways that leaders approach moral questions (i.e., their integrity style) are likely to influence how followers perceive the leader and what kind of relationship they form. Leaders who show morally committed leadership to the followers, should be positively related to a reciprocal, high quality LMX (Mahsud et al., 2010; Walumbwa et al. 2011). The two integrity components, moral behavior and behavioral integrity, have previously been demonstrated to associate with followers’ higher trust in and satisfaction with the leader (Moorman et al., 2013; Simons et al., 2015), whereas consistency across contexts has not been significantly related to the follower trust. Thus, we assume that among those follower groups who perceive their leader to act with integrity in a moral way would also have a higher team average rating on their LMX relationships with the leader. In addition, we investigate similar associations among individual followers. We expect that follower-rated (leader) moral behavior and behavior integrity, but not consistency, are positively related to their LXM relationship quality at both levels (Hypothesis 4).
We began with a qualitative investigation of leaders’ self-reported short descriptions of how they approach and solve moral conflicts in their work. The aim was to search for similarities (and potential differences) between the leaders’ self-reports and the social-cognitive processing styles (Berzonsky, 1990). After this theory-driven thematic analysis, we examined whether leaders who were categorized into different integrity styles got different ratings from their followers, measure with the three-dimensional perceived leader integrity measure. Finally, we investigated whether leader integrity related to the followers’ experiences of the LMX relationship, both at a dyadic and group level (see Figure 1). We used a multilevel modeling approach in Mplus, because it enables a simultaneous investigation of both individual and group level associations.
The sample was collected from various sources in order to produce data that would broadly represent different leaders in Finland. As a majority of employees in Finland are members of labor unions organized according to industry (64.5% in 2013; Ahtiainen, 2015), five trade unions were chosen as the primary collaborative partners in the data collection: the Finnish Union of University Professors, Finnish Union of University Researchers and Teachers, Finnish Business School Graduates, Academic Architects and Engineers in Finland, and the Confederation of Unions for Professional and Managerial Staff in Finland (Akava; a confederation of trade unions for those with a university degree or other higher education). An electronic questionnaire was sent to all members aged 18–65 years of the first two trade unions mentioned here, and an electronic questionnaire was sent to a random sample of 3000 members of the latter two unions, with the response rates of 45%, 26%, 17%, and 13%, respectively. In Akava, the questionnaire was sent out as an open invitation with a link to an electronic questionnaire via Akava’s leader network, and altogether 141 responses were collected.
Participants were also recruited from an executive MBA (EMBA) program. Contact persons from the EMBA program delivered the questionnaire to potential participants (n = 644), of whom 161 responded (response rate 25%). In the final phase, psychology students volunteered to recruit highly educated leaders (n = 23) from among their acquaintances, as a part of their studies. Participants from the EMBA program and the leaders recruited by students represented various sectors (e.g., media and marketing, finance and insurance, industry, and the service sector). Each of the study invitations were sent by email and they included a description of the study aims and procedure, as well as information about confidentiality. All the individuals participated on a voluntary basis.
All leaders who had given the permission to be contacted for the leader-follower study during the first phase of the data collection received a second email. It included a description of the aims, procedure, and confidentiality of the leader-follower study and a link for the electronic follower survey. The leaders were asked to forward the link and study information to their followers (defined as employees who worked as their direct reports). The follower responses were delivered directly to the research group through the electronic survey system without being brought to the attention of their leaders. The link included an identification code that was derived based on each leaders’ personalized code. These ID codes were used to match the data between leaders and their followers. This matching procedure was described to all participants in the information letters.
This study focused on all those leaders who had answered the open-ended question about their moral conflict solving (n = 234). Of these participants, 41% were leaders either from the EMBA program or recruited by students, 26% social and health sector leaders, 22% professors, 6% university teachers and other academics, 3% business sector leaders, and 2% engineers. Of the participants, 55% were women. The average age of the participants was 50.1 years (SD = 7.8), the mean of past leadership experience was 13.59 years (SD = 8.45), 96% had a permanent job and 99% were working full-time.
The hierarchical sample included altogether 987 followers from 233 leaders (one participant did not recruit any followers, others forwarded an electronic link received from the researchers to their direct subordinates). The followers’ ratings were combined with the data of the closest supervisor who had recruited them to participate in the study. The number of follower participants per leader ranged between 1 and 14 (M = 4.2). Of the followers studied here, 67.4% were women, the majority (58%) were aged 31–50 years, and the average duration of the relationship with the supervisor who had given the invitation to take part in the survey was 3.5 years (SD = 3.4).
Because the social-cognitive processing styles (Berzonsky et al., 2013) have not previously been applied in the integrity domain, we designed an open-ended question to capture leaders’ self-generated descriptions of their moral decision-making, which allowed us to test the applicability of the processing styles into our data. The questionnaire included the following definition (see Hiekkataipale & Lämsä, 2017): “We can often face ethical problems in our work. Ethical problems refer to situations where the decision-maker does not know what is the right way to act (choosing between alternatives that are equally good or equally bad) or where they feel that for some reason they cannot act as they would see right. These decisions have consequences for someone: for the person who is the target of the decision, for other individuals who are involved, or for the decision-maker. Time pressure and conflicting expectations, interests or values also often depict these situations.” After this definition came the following: “On what grounds do you usually make decisions in these kinds of situations?” followed by an empty space (without a word limit) for the leaders to type their answers. The question was deliberately left at a general level so that it would prompt the participants to reflect on and summarize their thoughts about their moral decisions as they personally thought fit, focusing on their most common style of approaching moral conflicts.
Followers evaluated their leader using the perceived leader integrity (PLI) scale (Moorman et al., 2013), which consists of three dimensions: moral behavior (6 items, e.g., “My supervisor treats people fairly”), behavioral integrity (6 items, e.g., “My supervisor practices what he/she preaches”), and consistency (4 items, e.g., “My supervisor stands by principles no matter the price”). All the items were measured using a 5-point scale ranging from strongly disagree to strongly agree, higher values representing higher integrity.
As the PLI scale was being used for the first time in a Finnish context, we conducted a confirmatory factor analysis to ensure that the structure of the measure would fit the sample of our study. However, the hypothesized 16-item, three-factor model yielded a poor fit to the data: χ2(101) = 973.89, p < 0.001, RMSEA = 0.215, SRMR = 0.137, CFI = 0.671, TLI = 0.562. Therefore, we modified the model in order to reach a factor structure in which goodness of fit indexes meet acceptable standards. In the selection of items for the modified factor model, we used the guidelines by Marsh et al. (2005). We chose items that best measured the intended construct as inferred on the highest standardized factor loadings in CFA, items that had minimal cross-loadings (based on the modification indexes in Mplus, which indicate the extent to which the fit would be improved if an item were allowed to load on a factor other than the one that it was intended to measure), and a sufficient number of items in each subscale to maintain a coefficient alpha estimate of reliability of at least 0.80. Our goal was to ensure that the final model sufficiently distinguished between each factor (i.e., dimension) but using as few items as possible.
Based on the aforementioned criteria for item selection, we proceeded to test the fit of a model that included three items for each factor (see Figure 2). This 9-item model (henceforth referred to as PLI-9) provided a good fit with the follower data: χ2(24) = 129.81, p < 0.001, RMSEA = 0.067, SRMR = 0.064, CFI = 0.972, TLI = 0.958, with standardized factor loadings between 0.70 and 0.94, and with Cronbach’s alphas of 0.87 (moral behavior), 0.91 (behavioral integrity), and 0.85 (consistency). The Satorra-Bentler scaled χ2-difference test supported choosing this model over the original model with 3 dimensions and 16 items [χ2 (77) = 844.21, p < 0.001].
As a final step to ensure the fit of the PLI-9 with our data, we conducted a multilevel CFA (see Figure 3). A freely estimated model between the two levels provided a good fit with the data [χ2 (51) = 234.65, p < 0.001, RMSEA = 0.060, SRMR(W) = 0.070, SRMR(B) = 0.311, CFI = 0.964, TLI = 0.949], but the restricted model where factor loadings were set equal between levels provided an even better fit [χ2 (57) = 216.54, p < 0.001, RMSEA = 0.053, SRMR(W) = 0.067, SRMR(B) = 0.125, CFI = 0.968, TLI = 0.960]. We compared the fit of these nested models (the restricted model to the previous, less restricted one) by using the Satorra-Bentler scaled difference chi-square test (Satorra & Bentler, 2001), which produces a nonsignificant loss of fit between the models if the restriction assumption is supported. The test supported choosing the model with equal factor loadings rather than the freely estimated model [χ2(5) = 8.14, p = 0.149]. Thus, the PLI-9 captures follower perceptions of integrity at both the individual and the shared level in a similar manner. The PLI-9 was therefore used in further analyses.
The followers’ perceptions of their leader-member exchange relationship were measured by using the LMX-7 scale (Graen & Uhl-Bien, 1995). This shortened scale has been supported for its validity over other LMX measures according to both the results of a meta-analysis (Gerstner & Day, 1997) and a broad review (Graen & Uhl-Bien, 1995). The scale includes items such as “My supervisor recognizes my potential” and “I would characterize my working relationship with my supervisor as extremely effective”, which followers rate on a 5-point Likert scale, ranging from “strongly disagree” (1) to “strongly agree” (5), higher values indicating a better LMX relationship. The Cronbach’s alpha for the scale was 0.89.
We investigated if the leaders’ demographic variables associated with the qualitatively assigned integrity styles. These variables were the leader’s gender (0 = female, 1 = male), leader’s age (in years), tenure as a leader (in years), whether the leader had taken part in any kind of additional training as a leader (e.g., participating in short or long-term courses, 1 = yes, 2 = no). We included these variables, because there might be differences in integrity and morality between men and women (see, e.g., Aquino & Reed, 2002) and because personal (leadership) experience could relate to more mature forms of integrity (Dane & Sonenshein, 2015). Finally, to address the heterogeneity of our leader sample, we tested if the trade union membership had any associations with the integrity coding.
In addition, both leader and follower characteristics are shown to influence the LMX relationship between them (for a review, see Dulebohn et al., 2012). Therefore we included the available demographic variables from leaders (gender and age) and followers (gender, 0 = female, 1 = male; age in years, and past experience working with the current leader in years) in our multilevel models in order to investigate if these variables relate to the follower-rated LMX quality.
The leaders’ open-ended answers were analysed using a theory-driven content analysis that included comparing different coded themes from the answers with the social-cognitive processing styles (Berzonsky, 1990) (see Table 1). Three independent coders were used: a post-doctoral researcher and a PhD student from the field of work and organizational psychology who both had expertise in leadership studies, and a Master’s level psychology student who was trained for the coding as a part of her analysis for her Master’s thesis. The coders were introduced the following content criteria, based on which they independently categorized each participant’s answer. Based on Berzonsky’s (1990) original concept, the diffuse-avoidant style was coded for leaders who described a tendency to avoid facing and dealing with moral problems or who lacked the capability to be aware of moral issues in one’s work. Example keywords for this coding included “(based on) facts”, “moving the responsibility”, and “ignore the situation”. Normative orientation was coded for leaders who described following and conforming to the formal standards and norms, such as the company’s official guidelines. Coders were instructed to look for key terms that included external rules, norms, and instructions (instead of personal moral values) and inflexible decision-making (e.g., “I always follow company rules”). Informational orientation was coded for leaders who reported actively and flexibly seeking out, processing, and evaluating relevant information before making a moral decision. Here, the coders were instructed to look for key terms such as “contextual”, “situational”, and “flexible”, but at the same time the answers had to include some personal reflection of values that were applied to make decisions (to differentiate informational orientation from diffuse-avoidant).
|IDENTIFIED THEMES||INTERPRETATION||INTEGRITY STYLE|
|I always aim to be objective.||Decisions are controlled by situational demands or factors. A moral issue can be reframed as a factual one. Postponing the decision (reluctance to confront and deal with the moral conflict).||Diffuse-avoidant|
|I look at the facts.|
|I make practical decisions.|
|I haven’t faced such situations.|
|I use common sense.|
|I postpone making a decision.|
|The decision happens quickly in the moment.|
|I look at the numbers.|
|I rationalize the decisions.|
|I compare the situation to a similar one in the past.|
|I try to find the way that would cause least trouble.|
|I trust my intuition and my interpretation of the situation.|
|I think what is best for the organization.||Making decisions by conforming to other people’s prescriptions and complying with rules and official guidelines.
Instead of acknowledging the consequences from a broader perspective, decisions are made solely based on what benefits the organization.
|I follow the law and official protocols.|
|According to the values of my employer.|
|Based on the values and culture of the organization.|
|I make my decisions based on equality.|
|I follow commonly agreed rules.|
|I strive to be as consistent as possible in my decisions.|
|I follow norms: a similar problem produces a similar solution without any personal consideration.|
|I always try to find what is best for the target of the decision (e.g., client).|
|I look at the whole picture.||Looking at the moral situation broadly from different perspectives and processing the solution through assessing the fairness of the potential solution. Using an open-minded approach to explore a variety of options, evaluating the potential consequences of each solution.||Informational|
|I spend a lot of time finding out different viewpoints before making a decision.|
|I discuss with different parties in order to get a broader picture of the problem.|
|I try to evaluate what is fair, I have discussions with the people who will be affected by my decision.|
|I find out the background of the situation as well as I can and evaluate the solution as it concerns different parties.|
|I analyze different options and their consequences.|
|After hearing different sides of the situation I try to find a fair and just decision.|
|After several rounds of discussions I try to find the best or the least bad option.|
The coders reached full agreement on 60% of the leaders’ answers (the same integrity style code was given by all three coders). For 36% of the leaders, the code that got two similar codes out of three was selected as the final code. Finally, 4% of the leaders had written such ambiguous answers that all three coders disagreed on their integrity style. These leaders and their subordinates were omitted from the final sample. Thus, the final data set that was used in the multilevel analysis included 223 leaders with altogether 963 followers.
Inter-rater reliability and consistency were assessed using a two-way mixed, consistency-type, average-measures ICC (McGraw & Wong, 1996). The resulting ICC was 0.75, which indicates that the coders had a high degree of agreement (see Cicchetti, 1994). As the ICC suggested that integrity style was rated similarly across coders, the amount of measurement error based on the coding procedure was low, and therefore statistical power for subsequent analyses was not substantially reduced.
The statistical analyses on leader-follower hierarchical data were performed using a multilevel modeling technique (Heck & Thomas, 2015), where individuals are nested within higher levels of classification, each to their own leader in the present study. The data was modeled at two levels (see Figure 1) using multilevel mediation modeling. At the “between” level we modeled the variation that is due to similarity among followers’ perceptions from the same leader (i.e., variation between leader means). At the “within” level, we modeled variation that is due to individual differences in followers’ perceptions within leaders.
We performed all the analyses using the Mplus program (version 8.0; Muthén & Muthén, 1998–2017), as it offers the possibility of performing the aforementioned hierarchical analysis by separating out the variance in the observed variables into two components. First, we investigated the extent of individual (within-level) and leader-related (between-level) variance of each study variable. The interclass correlations (ICC; see Heck & Thomas, 2015) were calculated by dividing the between-level variance by the total variance (total variance = between-leader variance + within-leader variance). The ICCs provide an estimate of what percentage of the total variance is accounted for by the leader level (i.e., the extent of shared views between followers of the same leader). If the observed variables show statistically significant between-level variation (i.e., within-leader homogeneity), it is acceptable to proceed to further multilevel analyses.
After establishing that there was variance on both levels, we tested the fit of our full intended two-level measurement model to our data. We estimated two individual multilevel models, in which we used the leader-assigned integrity style as a dummy variable (whether or not the leader was coded to the informational, normative, or diffuse-avoidant group). Two models were used to get results from the opposite reference categories, where two out of three dummy variables were entered in the same model simultaneously. We compared informational and normative leaders to diffuse-avoidant style leaders, and then normative and diffuse-avoidant leaders to achieved leaders. Because the integrity style was assigned to the leaders, it had no follower-related variation, and thus the associations with it were estimated only at the between level. The follower-rated PLI dimensions were used as three separate but correlated variables: each dimension represents a distinctive concept, which was only moderately correlated with the other two PLI dimensions. Follower-rated LMX was used as one variable. In addition, we tested if the control variables related to the PLI or LMX evaluations. Only the gender of the followers was found to have such significant associations, and it was therefore controlled for in the final models.
Of the different integrity styles, the informational style was assigned to 94 leaders (42%); they described actively seeking and evaluating information before making a moral decision. Leaders who were coded as normative (87; 39%) approached moral conflicts with predetermined values and beliefs that seemed change-resistant. Finally, 42 leaders (19%) were identified as using the diffuse-avoidant style, which was characterized by avoidance and/or reluctance to confront and face the moral aspects of the situation.
The full measurement model included three correlated factors for perceived leader integrity dimensions and one factor for LMX. The measurement model was estimated at both within- and between-levels. The results of the CFA showed that the intended structure fit the data well: χ2(199) = 755.62, p < 0.001, CFI = 0.945, TLI = 0.933, RMSEA = 0.054, SRMR(W) = 0.060, SRMR(B) = 0.236. Correlations between all study variables at both levels are presented in Table 2.
|4. Moral behavior||4.04||.557||–||–||–||1.00||.56***||.04||.81***||.23|
|5. Behavioral integrity||4.14||.554||–||–||–||.70***||1.00||.13||.57***||.33*|
In the hierarchical data the ICCs were 0.17 for moral behavior, 0.13 for behavioral integrity, and 0.09 for consistency, and 0.11 for LMX. These results indicate that 9–17% of variance in followers’ leader integrity ratings can be attributed to between-group differences. The remaining 83–91% of variance in perceived leader integrity occurs at the within-group level. All ICCs were significant at the level of p < 0.001 (except for consistency, which had the p-value below 0.01), thus justifying the examination of both individual level (within followers) and leader level (between followers, average cluster size = 4.32) effects.
The between (leader) level results of the two-level model (corresponding to the theoretical model presented in Figure 1) showed that that followers of leaders with the normative processing style gave higher ratings regarding leader’s consistency than followers of leaders with informative (standardized path estimate β = 0.126, p < 0.05) and diffuse-avoidant (β = 0.151, p < 0.05) styles. Thus, Hypothesis 3 was confirmed. However, no significant effects were found from informational or diffuse-avoidant integrity styles to follower-rated integrity (path estimates varied between -0.049 and 0.023, p > 0.05). Thus, Hypotheses 1 and 2 were not supported.
The results at between (leader) level showed further that shared perceptions within follower groups of their leader’s moral behavior were associated with positive evaluations of their LMX relationship (standardized path estimate β = 0.644, p < 0.001). Leader moral behavior explained 79% of the between-level LMX variance. Behavioral integrity (β = 0.070, p > 0.05) and consistency (β = 0.230, p > 0.05) did not associate with between-level LMX. These findings partly supported Hypothesis 4.
At the within (individual) level, the more the followers perceived their leader to behave morally and to show behavioral integrity, the better they evaluated the quality of their LMX relationship. These level 1 path estimates were β = 0.501 and β = 0.337 at the level of p < 0.001, respectively. However, leader consistency did not have a significant effect with follower LMX (β = 0.021, p > 0.05). These results supported Hypothesis 4.
Our study offers an extension to the leadership integrity literature by establishing that the social-cognitive processing styles (Berzonsky, 1990) are applicable to the moral domain at work. Furthermore, we showed that especially the moral dimension of integrity is important in cultivating high quality relationships with followers. We predicted that leaders who process moral conflicts with a normative style would be perceived to show consistency by their followers. Our findings supported this first hypothesis. However, we did not find support to our assumption that leaders with informative or diffuse-avoidant styles would be perceived to have high integrity by their followers, although the associations were in the hypothesized directions. Thus, Hypotheses 2 and 3 were not supported. Finally, we found that leader moral behavior and behavioral integrity related with better ratings on LMX relationship quality among individual followers, and leader moral behavior related with better group-level LMX. These findings partially supported Hypothesis 4.
Our findings extend the current understanding of leader integrity by broadening the concept towards capturing individual differences in social-cognitive processing styles. As we identified these styles based on how leaders approached and solved moral questions at work, we can revisit the question presented by Lemoine et al. (2019): “Do most leaders make norm- and standard-based decisions, focusing on outcomes and stakeholders, or do they follow their own self-concordant judgments?” Based on our findings, both of these styles were almost equally represented among the leaders in our study. Of the leaders, 42% were identified to follow an informative integrity style, which was characterized by flexible commitments that were used rationally and reflectively when solving moral conflicts. One participant summarized this style as trying to find a fair and just decision after hearing different sides of the situation.
In comparison, 39% of the leaders described making primarily norm- and standard-based decisions. This was identified as the normative integrity style, which represented more of an automatic style that was based on adopting moral commitments from others (such as the company’s values) and internalizing and maintaining them. An illustrative example came from one participant, who gave the following description to his or her moral decision-making: “I follow norms: a similar problem produces a similar solution without any personal consideration”.
The risk that comes with the normative style is that it can also be inflexible, rigid, and resistant to change, and in some situations, even lead to unethical decisions. For example, if a leader emphasizes what is best for the client or the company (aiming to maximize profit and performance with no reference to critical and moral reasoning), this could result in acting unethically or in the “moral gray zone” (see also Shaw & Liao, 2020). Also, adopting values from others without personal reflection and situational consideration could mean that the leader might end up complying with principles that are not morally sustainable. In contrast, leaders with the informational style will try to find the best solution by viewing the available information broadly, but this can lead to the followers’ perceiving their leader as inconsistent.
Finally, a minority of the leaders (19%) were identified to follow a diffuse-avoidant style. When a person uses procrastination and defensive avoidant decisional strategies when faced with moral conflicts, he or she is not likely to attribute much importance to moral values or to show consistency in their decision-making.
Our second theoretical contribution stems from explicitly testing two relevant questions related to leader integrity and follower outcomes. First, we examined how leaders’ integrity was recognized by follower groups (consistency between followers’ perceptions) and by individual followers (differences in followers’ perceptions). Second, we looked how the different integrity dimensions (moral behavior, integrity, and consistency) related to the quality of leader-member exchange.
We found that the amount of shared perceptions among followers of the same leader varied from 9–17% percent for the perceived leader integrity dimensions. This indicates that there were more individual differences between followers in their ways of evaluating their leader than a common understanding that was shared among all the followers. A similar finding appeared concerning the quality of LMX relationships: only 11% of variance in followers’ LMX ratings was attributed to between-group differences. Thus, both leader integrity and LMX quality are more likely to depend on how the leader behaves with each individual follower. These findings support the theory of LMX differentiation (for a review, see Henderson et al., 2009) – a process in which a leader engages in different exchange patterns with followers, which results in different quality exchange relationships with them. This process creates a group-level context in which there is variability within each group in the nature and quality of leader–follower relationships (Henderson et al., 2009).
By using the three-dimensional model of leader integrity, we were able to contribute to a more detailed understanding of how the different leader integrity components relate to follower outcomes (Moorman et al., 2013). As expected, the normative leaders were evaluated by their followers as showing more consistency than were informational or diffuse-avoidant leaders. However, the shared perceptions of leader consistency did not associate with LMX quality either among follower-leader dyads or follower groups. Thus, our study confirms the importance of the ethical meaning (i.e., the moral dimension) of integrity. In order to foster high-quality LMX relationships, the consistency component alone is not be enough. Rather, the values that the leader consistently practices should be of a moral nature.
Relatedly, our findings showed that of the three integrity dimensions, leader moral behavior had the strongest association with both individual and group-level perceptions on LMX quality. The more morally the leader was perceived to behave, the better the followers evaluated their relationship quality with their leader at both the dyadic and group levels. In comparison, leader behavioral integrity had a positive association only with dyad-level LMX quality. It is possible that moral behaviors (such as treating people fairly and protecting their rights) are focused on acting for the best of all followers more than are behavioral integrity (such as delivering on promises – promises made can vary from one follower to another) or consistency (e.g., standing up for personal beliefs may not always lead to the advantage of all followers). Therefore, leaders who engage in moral behaviors may be more likely to support high quality LMX relationships in their work groups (see also Henderson et al., 2009). These findings relate to the discussion on the dual effects of leadership: leading followers as individuals and leading followers as a collective at the same time (Kark & Shamir, 2002; see also Bormann et al., 2018). Although a leader who follows strong personal integrity might have strong moral motives for his or her behavior, high integrity can be more likely to lead to perceived inconsistency by the followers. This could lead to negative outcomes on group-level satisfaction with the leader.
These findings are in line with a previous study by Moorman et al. (2013), which showed that moral behavior and behavioral integrity might be more relevant to follower outcomes (such as having trust in the leader), whereas consistency across contexts does not necessarily make any additional contribution to these associations. Surprisingly, none of the leaders’ integrity styles had any direct associations with followers’ moral behavior evaluations. One explanation can be found from what we know about the judgement-action gap (Walker, 2004): even though the leaders describe making their moral decisions in a certain way (such as in an information-oriented way), this might not be realized in the actual moral behaviors that are visible to their followers.
Finally, we contributed to the measurement and conceptualization of perceived leader integrity. Based on our multilevel factor modeling, a shortened, nine-item scale might capture the phenomenon with a more compact and accurate manner compared to the original measure (Moorman et al., 2013). Furthermore, our study showed that this PLI-9 scale captures follower perceptions of integrity at both the individual and the shared level in a similar manner. This means that the scale can be recommended for future studies that wish to simultaneously investigate follower perceptions as differences between individuals and similarities within teams.
Drawing together our findings with previous evidence on the differences between social-cognitive processing styles (for a review, see Berzonsky, 2011), the informational style appears to be the most mature one – also when applied to the integrity domain. Thus, it becomes an important goal to support leaders in acquiring this flexible yet morally committed approach to processing and solving moral conflicts. A moral leader should have the ability to recognize and reflect on personal values and commitments, which we found to be missing from the diffuse-avoidant leaders, who do not commit to moral decisions. These moral commitments should be combined with a flexibility to look at each moral conflict individually, collect and evaluate relevant information, be open to diverse views, and be willing to evaluate and modify one’s own views if necessary. This, based on our qualitative findings, was missing from the leaders with a normative style, who instead based their decisions on more rigid adherence to espoused norms and rules. However, it is worth noting that in our study the informative integrity style did not associate with followers’ ratings of leader integrity or LMX. Therefore leaders should recognize that if their decision-making is based on processing situational information flexibly, it might not always appear as consistent or fair to their subordinates. One solution could be enhancing transparency within work groups. The leader could consider making his or her decisional process and justifications more visible to followers, which might help them to understand the moral background behind the decisions.
A diffuse-avoidant integrity style, in contrast, appears to be the least desirable one for a moral leader. Trying not to confront morally challenging situations can lead to poor outcomes for followers. If the leader acts like a “weathervane” and bases his or her decisions on short-term acts of compliance in order to satisfy immediate external demands, or refuses to make any choices (Berzonsky, 1989), this leadership style evidences a lack of coherent moral commitments, characterized by an absence of continuity and consistency between choices and actions. It is important to provide these leaders with leadership interventions, mentoring, training, or other measures that would help them to become more aware of the moral aspects that arise in work situations and to find ways of accepting personal responsibility even when facing difficult decisions. Here it would be important to have an organization that serves as a learning environment, providing opportunities for leaders to engage in reflection and discussion about issues of personal and moral values (Dane & Sonenshein, 2015). Similar interventions could also be helpful for normative-oriented leaders, who could benefit from learning a more critical perspective towards their adopted values and being more open to differing ideas. One way to achieve this could be to include normative style leaders in heterogeneous teams, where they might observe more various perspectives as a basis for moral decisions from their colleagues.
Finally, our qualitative analysis can be of practical use for future studies that seek to gain a broader picture of integrity. The coding principles described in Table 1 can be used as a qualitative rating scheme that can provide a starting point in analyzing different qualitative material such as open-ended answers or interviews and recognizing differences in integrity styles from them. Our rating scheme could be further developed into an integrity style measure, which would make it possible to collect more evidence on the validity of the concept.
As the leaders gave only a short summary about their moral decision-making, some of the answers were not straightforwardly identifiable in terms of integrity styles. In these cases the coders had to make some interpretations in order to identify the best fitting category, which may have lowered inter-rater reliability. However, including three coders in the process, and by accepting in the subsequent analyses only answers that got at least a two-thirds majority increased the reliability.
These short summaries were also not likely to have captured all the nuances in integrity. For example, even though Berzonsky’s original concept refers to stable inter-individual differences in short-term processes, and these differences in social-cognitive strategies have shown relative stability (Berzonsky, 1990), it could be that some work-related moral conflicts could be better solved by following the official norms even when the individual mostly used an information-oriented style.
It is also possible that if the leaders follow norms that are more team- or unit-specific, their decision-making might appear less consistent between the different contexts (because the norms that they adopt in each context are different). In a similar manner, informative style leaders who use relevant situational information in their decision-making can appear more consistent if they choose prioritize information that concerns the organization as a whole (such as company rules). Therefore, future studies could test in more detail to what extent integrity is a stable trait or a more fluctuating state, or if individuals could represent several integrity styles simultaneously, some being more dominant than others. One possibility would be to modify the measure that has been developed to capture the original processing styles (see Berzonsky et al., 2013) and to apply it to the integrity domain more directly. This would make it possible to get personal scores on each of the three integrity styles and to compare their prevalence within and between individuals. This would give a more detailed understanding of the individual differences that people have when they approach and solve moral issues at work.
One of the strongest points of our study is the hierarchical design, which allowed us to test in real-life work groups how followers perceive their leaders. Acquiring information from multiple sources (leader and followers) reduced the potential for common method bias (Podsakoff et al., 2003). However, we found only partial support for our hypotheses, which could mean that although leaders can be identified as representing different integrity styles, this does not directly translate to how their followers perceive them as moral leaders. Thus, future research might consider investigating how different integrity styles are espoused in actual leader behaviors. Such investigations could examine, for example, if leaders with informative, normative, or diffuse-avoidant style differ in their moral attentiveness, moral reasoning, propensity to morally disengage, or expedient actions (seeing that moral principles can be flexible in order to take advantage of profitable opportunities).
It should also be noted that because the leaders invited their employees to this study, there is potential for selection bias. This kind of bias could occur if the leaders wanted to present themselves in a positive way. Thus, especially leaders with low integrity could have selected employees who they expected to give more favorable ratings. In contrast, leaders with higher integrity may have invited a broader range of employees, who could give a more realistic evaluation of them. Therefore, it is possible that the selection of followers artificially created a higher inconsistency for leaders higher in integrity.
Finally, we want to address the issues of endogeneity related to our study design (Ketokivi & McIntosh, 2017). We acknowledge that there is a self-selection issue regarding our sampling procedure. The leaders chose whether or not to participate, and sent the questionnaire link to their followers, which could have led them to choose who rated them. However, we did not find strong shared variance among the followers nor strong correlations between all estimated associations, which would have raised doubt that there might be an issue of cognitive alignment stemming from a biased sample (e.g., followers who give a shared, favorable impression of their leader). It is also possible that the hypothesized relationships may be partly caused by environmental or personal effects that were not included in our model. Although we controlled for the effects of the relevant covariates available in our data, including exogeneous sources of variance (such as leader and follower age and gender), the design was nevertheless cross-sectional and hence we cannot make causal claims about the direction of the associations. Future research should therefore study the phenomenon in a more causally defensible way, for example by using panel data (Antonakis et al., 2010).
Despite its limitations, our study provides an important and novel step towards understanding the complexity of integrity within the work context and the relationship between leader integrity and follower perceptions. In order to understand in more detail how leaders’ morality and integrity operate in complex organizational contexts and how they affect followers, it is important to focus on both value commitment and flexibility. Strong commitment alone could direct the leaders towards making rigid decisions, whereas leaders who lack value commitments are vulnerable to situational pressures and external expectations, or they try to avoid solving conflicts by one means or another. The information-oriented style represents the “golden mean”, where the person has moral commitments, but is at the same time open to relevant new knowledge and can review his or her perspectives critically, when necessary. With regard to cultivating high quality relationships with followers, leaders should show moral behavior and behavioral integrity both in dyadic interactions and within their work groups.
The data has not been made available, as the open-ended answers include ethically sensitive material (e.g., detailed information about moral conflicts at work and organization-specific information).
We complied with the American Psychological Association Ethical Principles in the treatment of the study participants. The research followed the guidelines of the Finnish National Board on Research Integrity and there was no approval by an ethical or institutional review board required to conduct the research, which was based on adult participants informed consent.
This study was supported by the Academy of Finland (grant number 294428) and the Finnish Work Environment Fund (project number 116163).
The authors have no competing interests to declare.
Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A review and recommendations. The Leadership Quarterly, 21, 1086–1120. DOI: https://doi.org/10.1016/j.leaqua.2010.10.010
Aquino, K., & Reed, A., II. (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83(6), 1423–1440. DOI: https://doi.org/10.1037/0022-35188.8.131.523
Bauman, D. C. (2013). Leadership and the three faces of integrity. The Leadership Quarterly, 24(3), 414–426. DOI: https://doi.org/10.1016/j.leaqua.2013.01.005
Bedi, A., Alpaslan, C. M., & Green, S. (2016). A meta-analytic review of ethical leadership outcomes and moderators. Journal of Business Ethics, 139(3), 517–536. DOI: https://doi.org/10.1007/s10551-015-2625-1
Berzonsky, M. D. (1990). Self-construction over the life-span: A process perspective on identity formation. In G. J. Neimeyer, & R. A. Neimeyer (Eds.). Advances in personal construct theory (pp. 155–186). JAI Press.
Berzonsky, M. D. (2011). A social-cognitive perspective on identity construction. In S. J. Schwartz, K. Luyckx, & V. L. Vignoles, (Eds.). Handbook of identity theory and research (pp. 55–76). 0: Springer. DOI: https://doi.org/10.1007/978-1-4419-7988-9_3
Berzonsky, M. D., & Cieciuch, J. (2016). Mediational role of identity commitment in relationships between identity processing style and psychological well-being. Journal of Happiness Studies, 17(1), 145–162. DOI: https://doi.org/10.1007/s10902-014-9588-2
Berzonsky, M. D., Soenens, B., Luyckx, K., Smits, I., Papini, D. R., & Goossens, L. (2013). Development and validation of the revised Identity Style Inventory (ISI-5): Factor structure, reliability, and validity. Psychological Assessment, 25(3), 893–904. DOI: https://doi.org/10.1037/a0032642
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. DOI: https://doi.org/10.1080/1359432X.2018.1491550
Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6(4), 284–290. DOI: https://doi.org/10.1037/1040-35184.108.40.2064
Dane, E., & Sonenshein, S. (2015). On the role of experience in ethical decision making at work: An ethical expertise perspective. Organizational Psychology Review, 5(1), 74–96. DOI: https://doi.org/10.1177/2041386614543733
Dulebohn, J. H., Bommer, W. H., Liden, R. C., Brouer, R. L., & Ferris, G. R. (2012). A meta-analysis of antecedents and consequences of leader-member exchange: Integrating the past with an eye toward the future. Journal of Management, 38(6), 1715–1759. DOI: https://doi.org/10.1177/0149206311415280
Fields, D. L. (2007). Determinants of follower perceptions of a leader’s authenticity and integrity. European Management Journal, 25(3), 195–206. DOI: https://doi.org/10.1016/j.emj.2007.04.005
Gerstner, C. R., & Day, D. V. (1997). Meta-Analytic review of leader-member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82(6), 827–844. DOI: https://doi.org/10.1037/0021-9010.82.6.827
Giessner, S. R., van Quaquebeke, N., van Gils, S., van Knippenberg, D., & Kollée, J. A. (2015). In the moral eye of the beholder: The interactive effects of leader and follower moral identity on perceptions of ethical leadership and LMX quality. Frontiers in Psychology, 6, 1126. DOI: https://doi.org/10.3389/fpsyg.2015.01126
Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader-member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. The Leadership Quarterly, 6(2), 219–247. DOI: https://doi.org/10.1016/1048-9843(95)90036-5
Gu, Q., Tang, T. L. P., & Jiang, W. (2015). Does moral leadership enhance employee creativity? Employee identification with leader and leader-member exchange (LMX) in the Chinese context. Journal of Business Ethics, 126(3), 513–529. DOI: https://doi.org/10.1007/s10551-013-1967-9
Heck, R. H., & Thomas, S. L. (2015). An introduction to multilevel modeling techniques: MLM and SEM approaches using Mplus. 0: Routledge. DOI: https://doi.org/10.4324/9781315746494
Henderson, D. J., Liden, R. C., Glibkowski, B. C., & Chaudhry, A. (2009). LMX differentiation: A multilevel review and examination of its antecedents and outcomes. The Leadership Quarterly, 20, 517–534. DOI: https://doi.org/10.1016/j.leaqua.2009.04.003
Henderson, D. J., Wayne, S. J., Shore, L. M., Bommer, W. H., & Tetrick, L. (2008). Leader–member exchange, differentiation, and psychological contract fulfillment: A multilevel examination. Journal of Applied Psychology, 93, 1208–1229. DOI: https://doi.org/10.1037/a0012678
Hiekkataipale, M. M., & Lämsä, A. M. (2017). What should a manager like me do in a situation like this? Strategies for handling ethical problems from the viewpoint of the logic of appropriateness. Journal of Business Ethics, 145(3), 457–479. DOI: https://doi.org/10.1007/s10551-015-2911-y
Jennings, P. L., Mitchell, M. S., & Hannah, S. T. (2015). The moral self: A review and integration of the literature. Journal of Organizational Behavior, 36, S104–S168. DOI: https://doi.org/10.1002/job.1919
Kark, R., & Shamir, B. (2002). The influence of transformational leadership on followers’ relational versus collective self-concept. 0: Academy of Management Proceedings. DOI: https://doi.org/10.5465/apbpp.2002.7517557
Ketokivi, M., & McIntosh, C. N. (2017). Addressing the endogeneity dilemma in operations management research: Theoretical, empirical, and pragmatic considerations. Journal of Operations Management, 52, 1–14. DOI: https://doi.org/10.1016/j.jom.2017.05.001
Lapsley, D., & Hardy, S. A. (2017). Identity formation and moral development in emerging adulthood. In L. M. Padilla-Walker, & L. J. Nelson (Eds.), Flourishing in emerging adulthood: Positive development during the third decade of life (pp. 14–39). 0: Oxford University Press. DOI: https://doi.org/10.1093/acprof:oso/9780190260637.003.0002
Lemoine, G. J., Hartnell, C. A., & Leroy, H. (2019). Taking stock of moral approaches to leadership: An integrative review of ethical, authentic, and servant leadership. Academy of Management Annals, 13(1), 148–187. DOI: https://doi.org/10.5465/annals.2016.0121
Leroy, H., Palanski, M. E. & Simons, T. (2012). Authentic leadership and behavioral integrity as drivers of follower commitment and performance. Journal of Business Ethics, 107, 255–264. DOI: https://doi.org/10.1007/s10551-011-1036-1
Li, A. N., & Liao, H. (2014). How do leader-member exchange quality and differentiation affect performance in teams? An integrated multilevel dual process model. Journal of Applied Psychology, 99, 847–866. DOI: https://doi.org/10.1037/a0037233
Mahsud, R., Yukl, G., & Prussia, G. (2010). Leader empathy, ethical leadership, and relations-oriented behaviors as antecedents of leader-member exchange quality. Journal of Managerial Psychology, 25, 561–577. DOI: https://doi.org/10.1108/02683941011056932
Marsh, H. W., Ellis, L. A., Parada, R. H., Richards, G., & Heubeck, B. G. (2005). A short version of the Self-Description Questionnaire II: Operationalizing criteria for short-form evaluation with new applications of confirmatory factor analyses. Psychological Assessment, 17, 81–102. DOI: https://doi.org/10.1037/1040-35220.127.116.11
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. DOI: https://doi.org/10.5465/amj.2008.0276
McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass correlation coefficients. Psychological Methods, 1(1), 30–46. DOI: https://doi.org/10.1037/1082-989X.1.1.30
Moorman, R. H., Darnold, T. C., & Priesemuth, M. (2013). Perceived leader integrity: Supporting the construct validity and utility of a multi-dimensional measure in two samples. The Leadership Quarterly, 24(3), 427–444. DOI: https://doi.org/10.1016/j.leaqua.2013.02.003
Nahrgang, J. D., & Seo, J. J. (2016). How and why high leader-member exchange (LMX) relationships develop: Examining the antecedents of LMX. In T. N. Bauer, & B. Erdogan (Eds.), The Oxford handbook of leader-member exchange (pp. 87–118). 0: Oxford University Press.
Palanski, M. E., & Yammarino, F. J. (2007). Integrity and Leadership: Clearing the Conceptual Confusion. European Management Journal, 25(3), 171–184. DOI: https://doi.org/10.1016/j.emj.2007.04.006
Palanski, M. E., & Yammarino, F. J. (2009). Integrity and leadership: A multi-level conceptual framework. The Leadership Quarterly, 20(3), 405–420. DOI: https://doi.org/10.1016/j.leaqua.2009.03.008
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879–903. DOI: https://doi.org/10.1037/0021-9010.88.5.879
Satorra, A., & Bentler, P. M. (2001). A scaled difference chi-square test statistic for moment structure analysis. Psychometrika, 66(4), 507–514. DOI: https://doi.org/10.1007/BF02296192
Shaw, K. H., & Liao, H. Y. (2020). Does benevolent leadership promote follower unethical pro-organizational behavior? A social identity perspective. Journal of Leadership & Organizational Studies. 0: Advance online publication. DOI: https://doi.org/10.1177/1548051820957996
Simons, T. (2002). Behavioral integrity: The perceived alignment between managers’ words and deeds as a research focus. Organization Science, 13(1), 18–35. DOI: https://doi.org/10.1287/orsc.18.104.22.1683
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. DOI: https://doi.org/10.1007/s10551-014-2332-3
Simons, T. L. (1999). Behavioral integrity as a critical ingredient for transformational leadership. Journal of Organizational Change Management, 12(2), 89–104. DOI: https://doi.org/10.1108/09534819910263640
Skubinn, R., & Herzog, L. (2016). Internalized moral identity in ethical leadership. Journal of Business Ethics, 133(2), 249–260. DOI: https://doi.org/10.1007/s10551-014-2369-3
Tomlinson, E. C., Lewicki, R. J., & Ash, S. R. (2014). Disentangling the moral integrity construct: Values congruence as a moderator of the behavioral integrity-citizenship relationship. Group & Organization Management, 39(6), 720–743. DOI: https://doi.org/10.1177/1059601114551023
Treviño, L. K., Hartman, L. P., & Brown, M. (2000). Moral person and moral manager: How executives develop a reputation for ethical leadership. California Management Review, 42(4), 128–142. DOI: https://doi.org/10.2307/41166057
Walker, L. (2004). Gus in the gap: Bridging the judgment-action gap in moral functioning. In D. Lapsley, & D. Narvaez (Eds.), Moral development, self and identity (pp. 1–20). 0: Lawrence Erlbaum Associates.
Walumbwa, F. O., Mayer, D. M., Wang, P., Wang, H., Workman, K., & Christensen, A. L. (2011). Linking ethical leadership to employee performance: The roles of leader-member exchange, self-efficacy, and organizational identification. Organizational Behavior and Human Decision Processes, 115, 204–213. DOI: https://doi.org/10.1016/j.obhdp.2010.11.002
Zagenczyk, T. J., Purvis, R. L., Shoss, M. K., Scott, K. L., & Cruz, K. S. (2015). Social influence and leader perceptions: Multiplex social network ties and similarity in leader–member exchange. Journal of Business and Psychology, 30(1), 105–117. DOI: https://doi.org/10.1007/s10869-013-9332-7
Zhu, W., Riggio, R. E., Avolio, B. J., & Sosik, J. J. (2011). The effect of leadership on follower moral identity: Does transformational/transactional style make a difference? Journal of Leadership & Organizational Studies, 18(2), 150–163. DOI: https://doi.org/10.1177/1548051810396714
Zhu, W., Treviño, L. K., & Zheng, X. (2016). Ethical leaders and their followers: The transmission of moral identity and moral attentiveness. Business Ethics Quarterly, 26(1), 95–115. DOI: https://doi.org/10.1017/beq.2016.11