Karianne Kalshoven, Utrecht University, Ethics Institute, Heidelberglaan 8, 3584 CS Utrecht, The Netherlands. Email: email@example.com
This study uses a multi-level approach to examine the moderating influence of two aspects of the ethical context on the relationship between ethical leadership and follower helping and courtesy. Using multi-source data from a field sample of leaders and followers and controlling for transformational leadership, we found that shared perceptions of moral awareness and empathic concern of the work group moderated the relationship between ethical leadership and follower helping and courtesy. Relationships between individual and group-level perceptions of ethical leadership and these two follower behaviors were positive when moral awareness was low, whereas these relationships weakened when moral awareness was higher. The relationship between individual and group perceptions of ethical leadership and courtesy was positive when empathic concern was high, whereas this relationship weakened when empathic concern was lower. Thus, although ethical leadership relates positively to follower helping and courtesy, the strength of this relationship differs depending on the level of empathic concern and moral awareness in the work group.
The expense of doing business in a completely unethical environment would preclude any organisation from making a profit (Beu & Buckley, 2004). Organisations increasingly emphasise ethics at work and rely on the ethical behavior of leaders at different levels. Theory suggests that such leaders form role models of appropriate behavior for followers. Followers emulate leaders' fair, open, and honest behaviors by, for example, engaging in pro-social behaviors such as interpersonal citizenship, and through showing such pro-social behaviors they constructively contribute to the organisation (e.g. Brown, Treviño, & Harrison, 2005). Research shows that ethical leadership relates positively to pro-social interpersonal citizenship (e.g. Mayer, Kuenzi, Greenbaum, Bardes, & Salvador, 2009b; Kalshoven, Den Hartog, & De Hoogh, 2011; Piccolo, Greenbaum, Den Hartog, & Folger, 2010). However, the relationship between ethical leadership and follower citizenship varies in strength in the different studies, indicating that moderators of this relationship may exist. Here, we focus on this issue.
We propose that the ethical context may act as a moderator and we propose that such a context can act as a substitute for ethical leadership that reduces the importance of ethical leadership for follower pro-social behavior, operationalised here as helping and courtesy. Substitutes are variables that decrease the ethical leader's influence on follower behavior (e.g. Kerr & Jermier, 1978). In determining what would be appropriate behavior in a given context, employees are likely to look to the leader, but also to other contextual cues (e.g. Lord, Brown, Harvey, & Hall, 2001). Thus, the work group could also form a source of ethical cues that help employees to determine whether to act pro-socially. We expect that the relationship between ethical leadership and follower pro-social behavior is less strong when followers have sufficient ethical cues from the context that help them determine how to act and stronger when such contextual ethical cues are not present. In other words, ethical leaders are likely to be more salient as role models when the context has few ethical cues.
As Brown and Treviño (2006) note, there are multiple ways to think about the ethical work context, yet much of the empirical research in this area has focused on ethical climate. Ethical climate has been defined as “the prevailing perceptions of typical organisational practices and procedures that have ethical content” (Victor & Cullen, 1988, p. 101) and forms a specific psychological climate. Psychological climate perceptions are meaningful representations of the organisational structure, processes, and events (Parker, Baltes, Young, Huff, Altmann, Lacost, & Roberts, 2003). They enable an individual to interpret events, predict possible outcomes, and gauge the appropriateness of subsequent actions (Jones & James, 1979), and have been linked to a variety of important criteria (Parker et al., 2003). Perceptions of specific work climate aspects are important in relation to specific employee outcomes (see e.g. Kuenzi & Schminke, 2009). In this study, we build on leadership substitute theory and climate literature, and investigate two specific aspects of ethical climate that are important to ethical leadership and follower helping and courtesy, namely moral awareness and emphatic concern.
Shared perceptions of empathic concern relate to whether employees are likely to show helping and courtesy towards co-workers. Also, the business ethics literature suggests that when employees are morally aware they are more likely to show pro-social behaviors (Arnaud, 2010; Rest, 1986). We expect that when work group moral awareness and empathic concern are high, the relationship between ethical leadership and follower helping and courtesy is lower. Employees can exhibit helping and courtesy as a reaction to ethical leader behavior, but also in reaction to the strong morality and empathy of the work group. The presence of such ethical climate aspects may be sufficient to stimulate followers to show citizenship and thus the role of leaders in stimulating these behaviors may be less strong in such climates. However, if these ethical climates are weaker, followers are likely to look to the leader for behavioral cues. Thus, the relationship of ethical leadership with helping and courtesy is stronger when work group moral awareness and empathic concern are low.
Overall, we aim to enhance the understanding of the interplay between ethical leadership and context. We empirically investigate a model that incorporates moral awareness and empathic concern as moderators of the relationship between ethical leadership and follower courtesy and helping. We test this in a multi-level, multi-source field study.
ETHICAL LEADERSHIP AND HELPING AND COURTESY
Ethical leader behavior is attracting a growing amount of research attention. At first, most work focused on ethical components of leadership styles such as transformational leadership. Bass (1985) argued that transformational leadership takes ethical as well as unethical forms. He later distinguished between authentic transformational leaders, who are ethical, genuine, and use power to attain moral and social end-values, and pseudo-transformational leaders, who are self-interested and lack morality (Bass & Steidlmeier, 1999). Recently, rather than focusing only on ethical aspects of other leadership styles, research has focused on ethical leadership as a set of behaviors or a behavioral style in itself. Indeed, researchers have shown that ethical leadership is empirically related, but distinguishable from transformational and other leadership styles (Brown et al., 2005; Kalshoven et al., 2011; Walumbwa, Avolio, Gardner, Wernsing, & Peterson, 2008).
Brown and colleagues (2005, p. 120) define ethical leadership as a leadership style that entails “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”. Ethical leader behaviors include acting fairly, allowing voice, demonstrating consistency and integrity, taking responsibility for one's actions, promoting ethical conduct, being concerned for others, and rewarding ethical conduct (Brown et al., 2005; De Hoogh & Den Hartog, 2008; Kalshoven et al., 2011; Treviño, Brown, & Hartman, 2003). Ethical leaders use different ways to influence ethics-related behaviors among followers, such as communication, rewards, and punishment (Brown et al., 2005). Ethical leaders send clear messages about what is expected and use reward systems to hold subordinates accountable (Treviño et al., 2003). In addition, ethical leaders act as role models and promote ethical behavior among followers (Brown et al., 2005; Treviño et al., 2003). Role modeling refers to the process of observational learning, imitation, and identification (see Bandura, 1977) and is important for ethical leadership.
So far, most studies on the outcomes of ethical leadership have shown positive relationships with helping (also labeled altruism), a form of interpersonal organisational citizenship behavior (Mayer, Kuenzi, & Greenbaum, 2009a; Piccolo et al., 2010). Smith, Organ, and Near (1983) define altruism as behavior that is directly and intentionally aimed at helping a specific person in face-to-face situations. Helping can be job related, such as cooperation between colleagues. Courtesy, also a form of interpersonal organisational citizenship (Podsakoff, MacKenzie, Paine, & Bachrach, 2000), also relates positively to ethical leadership (Kalshoven et al., 2011). Courtesy reflects avoiding doing harm and being mindful of how one's action affects others (e.g. Organ, 1988). In the current study both courtesy and helping are included and in line with previous studies we expect a positive relationship of both these behaviors with ethical leadership. However, the relationships between ethical leadership and citizenship behaviors reported in the literature to date vary substantially in strength, suggesting that moderators of this relationship may exist.
MORAL AWARENESS AND EMPATHIC CONCERN AS MODERATORS
Climate research studies employees' perceptions of the work environment and how these perceptions drive their behaviors and attitudes (Schneider, 2000). Schneider and Reichers (1983) define work climates as a set of shared perceptions regarding the policies, practices, and procedures that an organisation rewards, supports, and expects. As noted, ethical climate more specifically focuses on perceptions of the ethical content of such organisational practices and procedures (Victor & Cullen, 1988). These definitions identify climate as a perceptual phenomenon rather than an objective element of the organisation (Schneider & Reichers, 1983). Due to the methodological and theoretical problems of focusing on global conceptualisation of work climate (see e.g. Kuenzi & Schminke, 2009), a perceptional and facet specific approach to climate currently dominates in the literature. Here, we investigate two facets of ethical climate that are important to ethical leadership and follower helping and courtesy, namely work group moral awareness and empathic concern.
Empathic concern reflects evaluating the consequences of actions in terms of how they affect others (Arnaud & Schminke, 2007; Davis, 1983). Empathic concern has an emotional focus and concerns stressing sympathy, emotional responsiveness or affective perspective taking (Stephan & Finlay, 1999). A work group is characterised as high on empathic concern if members sympathise with someone who is having difficulties at work. In a high empathic concern climate, group members are oriented towards others rather than the self. High moral awareness reflects sensitivity to ethics and morality at work, and group members are aware of ethical issues and have similar interpretations about the environment in terms of its ethical components (Arnaud, 2010; Rest & Narvaez, 1994). Members recognise moral dilemmas and know how to incorporate ethical considerations into decision-making (Arnaud, 2010).
Based on “substitutes for leadership” theory, we propose a moderating role of empathic concern and moral awareness in the relationship between ethical leadership and follower helping and courtesy. Substitutes for leadership theory suggests that certain (contextual) variables at the organisational, group, or individual level may substitute for, neutralise, or enhance the leader's ability to affect employee attitudes, behaviors, and performance (Podsakoff, MacKenzie, & Bommer, 1996). Different types of moderators may occur. Substitutes have main effects on follower behavior that replace the effects of leader behavior. Neutralisers do not have main effects, but reduce the impact of leader behavior. In contrast, enhancers strengthen relationships between leadership and criteria (Kerr & Jermier, 1978). Podsakoff et al. (1996) found several substitutes to moderate the relationship between transformational leadership and follower citizenship. Among proposed substitutes for transformational leadership are cohesive work groups, formalisation, and indifference to rewards (e.g. Jermier & Kerr, 1997; Howell, Bowen, Dorfman, Kerr, & Podsakoff, 1990).
Substitutes need to be closely linked to the leader behavior they are supposed to substitute for in order to be able to replace that specific behavior. Substitutes for ethical leadership thus are variables that would be able to replace the impact that the ethical leader has on followers. Specially, we focus on the ethical climate dimensions of moral awareness and empathic concern as potential “substitutes” for ethical leadership in relation to pro-social behaviors such as helping and courtesy (e.g. Brown et al., 2005; Kerr & Jermier, 1978; Treviño et al., 2003). Perceived shared ethical climates within the work group prescribe how individuals within the groups should behave and, as such, ethical climates may substitute for ethical leadership as they also provide ethical guidance for followers. Thus, we propose that shared perceptions of moral awareness and empathic concern may act as substitutes, such that perceived ethical leadership relates more strongly to follower helping and courtesy when moral awareness and empathic concern are low rather than high.
Moral Awareness as a Moderator
We propose that when moral awareness in a work group is high, leaders have less impact on followers' helping and courtesy and, vice versa, that leaders' ethical guidance has more impact on such behavior when moral awareness is low. If the work context is high on moral awareness, work group members know what is right and wrong and understand what the ethical choices are. Contextual cues exist on when and how one should take responsibility and behave pro-socially. The leader's ethical modeling and rewarding is less needed to direct follower behavior as followers are aware of ethical issues and helping and courtesy are likely already elicited through the high moral awareness.
Social learning theory suggests that when employees observe the consequences of others' behavior (such as that of peers), they use this information to guide their own behavior. Bandura (1977) stresses that attention is important in observational learning and is necessary before an individual can successfully model the behavior of others. Attention to the desirability of pro-social behaviors such as helping and courtesy behavior can come from guidance offered by the ethical leader, but it could also come from the shared perception of moral awareness. Thus, if the work group context emphasises moral awareness, individuals are motivated to show helping and courtesy, and additional encouragement from ethical leaders is not needed to show such behaviors.
In contrast, if the context does not offer sufficient cues, leaders are likely to have a stronger impact on followers' behavioral choices (e.g. Shamir & Howell, 1999). A climate low on moral awareness is characterised by little shared attention for ethical issues and this climate provides less guidance on how to interpret or decide on ethical issues or deal with ethical dilemmas. If ethical guidance does not come from the context, employees are likely to look to their leader for this. An ethical leader is likely to stimulate and reward pro-social behaviors such as helping and courtesy. Social learning theory suggests that when employees observe their leader acting as an ethical role model, they are likely to copy the leader's fair, ethical, and trustworthy behavior and translate this into their own behavior by showing helping and courtesy. Thus, we expect that in a low moral awareness climate, ethical leadership relates more strongly to followers' helping and courtesy than in a high moral awareness climate in which other proximal cues for how to behave are present and followers more easily recognise the need to act pro-socially without the leader's role modeling.
Hypothesis 1: The positive relationship between ethical leadership and followers' helping and courtesy will be stronger in work groups low on moral awareness than in work groups high on moral awareness.
Empathic Concern as a Moderator
Empathic concern reflects the degree to which behaviors such as sympathy and affective perspective taking are valued in the work group. Work groups high on empathic concern endorse individuals making sacrifices for others and take care of each other when treated unfairly. Group members are likely to be motivated to help each other and show helping and courtesy in these contexts, regardless of the behavior of leaders. The social bond employees experience with their colleagues and the shared emphasis on pro-social behavior in a high empathic concern climate sufficiently motivates employees to show helping and courtesy. In contrast, when empathic concern is low, helping and courtesy are less obvious or salient behavioral options for followers. In that context, role modeling of leaders becomes more relevant for guiding follower behavior. Ethical leaders model concern for others, fairness and respect and through this are likely to enhance followers' helping and courtesy. Thus, we expect that when employees perceive low empathic concern in the context, the relationship between ethical leadership and helping and courtesy is stronger, whereas when they perceive high empathic concern, this relationship is weaker as helping and courtesy are already strongly emphasised in the work group in such a context.
Hypothesis 2: The positive relationship between ethical leadership and followers' helping and courtesy will be stronger in work groups low on empathic concern than in work groups high on empathic concern.
Participants and Procedures
Individual-level data for the present study were requested from 256 employees working in different organisations (e.g. health care, government, insurance) located in the Netherlands. All organisations had formally defined work groups of employees at the same hierarchical level. At the work group level, 88 supervisors leading such a group and several members were requested to participate in the study. Management team members were our organisational contacts. They randomly selected the employees from each work group that the supervisor would be asked to rate and provided us with contact information on both supervisors and these direct reports. Employees completed surveys rating their immediate supervisors' ethical leadership, and the work group level of moral awareness and empathic concern. Supervisors were asked to rate the work behaviors of three followers. Supervisors were given the names of the employees, and employees that of the supervisor, that they needed to evaluate. The cover letter (or email) explained the study purpose, the voluntary nature of participation, assured confidential treatment of data, promised a general report at the close of the study, and included approval and support from each organisation's management team. Surveys could be completed during work time and after two weeks a reminder was sent.
For 20 work groups, supervisor surveys were not received. As a result, the employees from these work groups (n= 60) were excluded from the analyses. In 27 cases only one employee survey was returned and both the supervisor and employee data were excluded from the analyses. For testing our hypotheses we needed at least two employees per supervisor to indicate the work group level of moral awareness and empathic concern (i.e. 36 employee questionnaires were only used to measure ethical leadership and climate as there were no supervisor ratings for these employees). The follower sample size was 170 and we obtained supervisor ratings of follower behavior for 133 followers. These were used to test our hypotheses. This resulted in a final sample of 53 work groups and supervisors. On average supervisors rated 2.5 employees each. Response rates are 60 per cent for the supervisor and 50 per cent for the employees. The average age of supervisors was 50 years (SD= 9); 30 per cent were women. The average age of employees was 46 (SD= 9); 42 per cent were women. The work groups were mostly smaller than 10 people (87%). For 84 per cent, supervisor-employee tenure was over six months and 81 per cent of the supervisors were employed for at least six months. All measures used a 5-point response scale, ranging from 1 (strongly disagree) to 5 (strongly agree).
Ethical Leadership. Ethical leadership was assessed with the 38-item Ethical Leadership at Work questionnaire (ELW) developed and validated in the Netherlands by Kalshoven et al. (2011). The ELW overall scale includes several ethical leader behaviors: fairness, integrity, ethical guidance, people orientation, power sharing, role clarification, and concern for sustainability (based on Brown et al., 2005; De Hoogh & Den Hartog, 2008; Treviño et al., 2003). Kalshoven et al. (2011) show that these behaviors can be combined in an overall score for ethical leadership as we do in this study. Measuring leadership styles as a combined second-order construct including several behavioral components is common in leadership research (for transformational leadership, see e.g. Podsakoff, MacKenzie, Moorman, & Fetter, 1990; for authentic leadership, see e.g. Walumbwa et al., 2008).
The behaviors measured with the ELW (see Kalshoven et al., 2011) include fairness (six items), which measures leaders' honesty, acting responsibly, treating followers equally, and being dependable. Integrity (four items) reflects being consistent in word and deed and keeping promises. Ethical guidance (seven items) assesses acting according to ethical standards, role modeling, and setting expectations about work-related ethical issues. People orientation (seven items) refers to caring about people, respecting others, and taking an interest in their welfare. Power sharing (six items) reflects providing voice and opportunities for input. Role clarification (five items) refers to clarification of expectations and responsibilities and engaging in open communication. Concern for sustainability (three items) reflects being sensitive to environmental issues and sustainability. Sample items are: “Can be trusted to do the things (s)he says (s)he will do”; “Clearly explains integrity-related codes of conduct”; “Initiates recycling of items and materials in our department”; and “Allows subordinates to influence critical decisions”. The overall ethical leadership scale had a Cronbach's α of .95. To test the appropriateness of using the overall scale, we performed a second-order confirmatory factor analysis (CFA), in which individual items were modeled as indicators of their underlying dimensions (fairness, integrity, people orientation, power sharing, role clarification, ethical guidance, concern for sustainability), and these were modeled as indicators of an overall latent ethical leadership construct. The CFA showed a good fit, χ2 second-order factor model (658, N= 139) = 961.27, p < .01, CFI = .98; NNFI = .97; RMSEA = .05; SRMR = .08 (see Hu & Bentler, 1999).
Helping and Courtesy. Employee helping referred to helping co-workers and was assessed with four items from MacKenzie, Podsakoff, and Fetter (1991). Items were reformulated as supervisor ratings, for example: “Is always ready to help or lend a helping hand to those around him/her”. Cronbach's alpha was .87. Courtesy was assessed with four items from MacKenzie et al. (1991). Courtesy reflects polite and respectful behavior intended to prevent work-related conflicts. A sample item is: “‘Touches base’ with others (informs them in advance) before initiating actions that might affect them”. Cronbach's alpha was .81.
Empathic Concern and Moral Awareness. Empathic concern and moral awareness were each assessed with three items of the Ethical Climate Index developed and validated by Arnaud (2010). This index is based on the ethical decision model of Rest (1986). So far, most studies use Victor and Cullen's (1987) measure of ethical climate. Mayer et al. (2009a) argue that although this measure has been useful, other measures should also be used as the Victor and Cullen framework has received extensive criticism relating to conceptual, operational, methodological, and analytical issues (see Mayer et al., 2009a; Arnaud & Schminke, 2007). Thus, in line with their call we used the two scales from Arnaud (2010).
Empathic concern items focus on role and perspective taking. They are: “People in my department sympathise with someone who is having difficulties in their job”; “For the most part, when people around here see that someone is treated unfairly, they feel pity for that person”; and “People around here feel bad for someone who is being taken advantage of”. Cronbach's α was .74. The items we used to measure moral awareness are: “People around here are aware of ethical issues”; “People in my department recognise a moral dilemma right away”; and “People in my department are very sensitive to ethical problems”. Cronbach's α was .73. Participants indicated the degree to which items described their work group.
Control Variables. Previous research shows that ethical leadership correlates with transformational leadership (Brown et al., 2005; Kalshoven et al., 2011). Thus, we controlled for perceived transformational leadership measured with the Dutch CLIO scale (De Hoogh, Den Hartog, & Koopman, 2004). Sample items are: “Has a vision and imagination of the future” and “Mobilises a collective sense of a mission” (five items). Cronbach's α was .80.
Measurement Model. We tested a confirmatory factor analysis involving the ethical leadership, empathic concern, moral awareness, helping, and courtesy measures. To increase indicator stability (e.g. West, Finch, & Curran, 1995) and meet sample size guidelines for parameter estimation (see Landis, Beal, & Tesluk, 2000), we used the seven behavioral dimensions or sub-scales of ethical leadership as indicators to form a reduced set of indicators to seven (instead of 38 at the item level) for the latent variable ethical leadership. The other latent variables had three or four indicators (item level) each.
Results for the proposed five-factor measurement model provided a good fit to the data (χ2= 225.72, df= 160, p < .01; CFI = .95; NNFI = .94; RMSEA = .055; SRMR = .069; see Hu & Bentler, 1999). In order to test the discriminant validity of helping and courtesy, we compared the fit of our measurement model with competing models. For example, helping and courtesy items were set to load on a single factor. The fit of this competing four-factor model (ethical leadership, moral awareness, empathic concern, and citizenship) was inferior to our measurement model (χ2= 288.26, df= 164, p < .01; CFI = .91; NNFI = .89; RMSEA = .08; SRMR = .08) as was a three-factor model that included ethical leadership, combined moral awareness and empathic concern into one factor, and combined helping and courtesy into one factor (χ2= 317.26, df= 167, p < .01; CFI = .89; NNFI = .88; RMSEA = .09; SRMR = .09). The chi-square difference tests, Δχ2= 63.46, df= 4, and Δχ2= 92.46, df= 7, indicate that the five-factor model fits the data better than the four- and three-factor models. Finally, a one-factor model showed poor fit to the data (χ2= 624.20, df= 170, p < .01; CFI = .68; NNFI = .64; RMSEA = .17; SRMR = .15). The difference in chi-square, Δχ2= 398.48, df= 10, suggests that the a priori five-factor model fits significantly better than the one-factor model. These analyses supported the empirical distinctiveness of all study variables.
Data Aggregation. To assess the appropriateness of aggregating individual scores to the work group level, we calculated within-team agreement (rwg; LeBreton & Senter, 2008) and intraclass correlations (ICC); Bliese, 2000). These tests yielded sufficient support to aggregate our data to the work group level of analysis (moral awareness: ICC= .19, rwg ranged from .75 to .89; empathic concern: ICC= .20, rwg ranged from .79 to .90; ethical leadership: ICC= .20, rwg ranged from .83 to .92).
Our data were multi-level in nature because employees were nested within work groups and our hypotheses focus on both the individual and group levels (Bliese, Halverson, & Schriesheim, 2002). Therefore, we performed a two-level Hierarchical Linear Modeling (HLM) analysis to decompose the variance into within- and between-group components. In addition, HLM is valuable in modeling cross-level interaction effects between group-level predictors and individual-level independent variables on outcome variables (Hofmann, Griffin, & Gavin, 2000; Kozlowski & Klein, 2000). The model was tested for helping and courtesy. We investigated whether the moderators have a cross-level or between-group effect (Hofmann & Gavin, 1998; Hofmann, Morgeson, & Gerras, 2003). To test the cross-level interaction, we used group-mean centering with the between-group variance in ethical leadership included in the Level 2 intercepts model. This means that ethical leadership will be included in the models as a Level 1 and Level 2 variable. The other variables were grand-mean-centered (Hofmann & Gavin, 1998). We also analysed the data using more traditional OLS cross-level analysis. Results were consistent with the multi-level results reported below.
Means, standard deviations, and correlations are reported in Table 1. Ethical leadership at the individual level correlates with helping (r= .25, p < .01) and courtesy (r= .20, p < .05). At the group level, ethical leadership correlates significantly with helping (r= .19, p < .05), but not courtesy (r= .12, p < .10). Transformational and ethical leadership were highly related (r= .75, p < .01) as in previous studies (Kalshoven et al., 2011). Empathic concern was not related to ethical leadership at either level (both r= .07, ns). Moral awareness was positively associated with ethical leadership at both levels (r= .35, p < .01; r= .45, p < .01).
Table 1. Means, Standard Deviations, and Correlations
Note: N= 133. Moral awareness and emphatic concern are work group level variables.
Before testing the hypotheses using HLM, we examined null models (no individual- or group-level predictors) to examine whether there was significant systematic within- and between-work group variance in supervisor-rated helping and courtesy. This null model forces all within-group variance in the outcome variable into Level 1 and all of the between-group variance in the outcome variable into the Level 2 residual term (Gavin & Hofmann, 2002). The null models were used to calculate the ICC1, which were .26 for helping and .21 for courtesy, indicating that 26 per cent of the variance in helping and 21 per cent of the variance in courtesy is between work groups. The chi-square tests also indicated between-group variance for courtesy (τ00= .08, χ2(54) = 240.60, p= .05) and helping (τ00= .16, p < .01). In other words, the intercept terms significantly varied across groups and support cross-level analyses.
Table 2 provides an overview of the tested models and the results of hypotheses tests.1 First, individual-level ethical leadership was entered in the model, after controlling for transformational leadership. Results demonstrated that individual perceptions of ethical leadership were significantly related to helping, γ20= .40, p < .05, and courtesy, γ20= .49, p < .05. Also, shared perceptions of ethical leadership were significantly related to helping, γ01= .48, p < .05, but not to courtesy, γ01= .17, ns. Second, for testing interactions, Hofmann and Gavin (1998) note that it is important to distinguish between cross-level and between-group interaction as the variable being moderated (ethical leadership) contains both individual and group-level variance. To account for this, we added the interaction term of group-mean ethical leadership, moral awareness, and empathic concern as predictors of the intercept and we added moral awareness and empathic concern as predictors of the variance in the slopes relating ethical leadership to helping and courtesy.
Table 2. Results of Hierarchical Liniear Models of Interactions on Follower Behaviors
Note: a EL = Ethical Leadership; ns= not significant
The results of the model for helping revealed that the between-group interaction and the cross-level interaction of empathic concern were both not significant (respectively, γ05= 1.31, and γ23= .60, ns), which is not in line with Hypothesis 2. For moral awareness, both the between-group and the cross-level interaction were significant in relation to helping (respectively, γ04=−1.04 and γ22=−1.18, p < .05). The significant cross-level interaction is depicted in Figure 1 and the between-group-level interaction in Figure 2. Both were plotted at one standard deviation above and below the mean. The slope for the relationship between ethical leadership and helping was positive and significant for employees who experience a low (t(126) = 2.75, p < .01) but not for those experiencing a high level of moral awareness (t(126) =−.19, ns). A similar pattern arises when calculating the simple slopes for ethical leadership at the group level. That is, they are significant for low levels of moral awareness, t(126) = 2.96, p < .01, but not for high levels of moral awareness, t(126) =−.21, ns.
Figure 1 illustrates that the relationship between ethical leadership and helping was positive when moral awareness was low, but this relationship weakened and became non-significant when moral awareness of the work group became higher. The cross-level interaction represents how the within-group relationship between ethical leadership and helping changes as a function of moral awareness. Thus, the relationship between ethical leadership and helping within work groups changes as a function of between work group differences in moral awareness. The results presented above support Hypothesis 1.
The results of the model for courtesy revealed that the between-group interactions of moral awareness and empathic concern were both significant (respectively, γ04=−1.63, p < .01; γ05= 1.73, p < .05). Furthermore, both the cross-level interactions of empathic concern and of moral awareness were significant (γ23= 1.32, p < .05 and γ24=−1.24, p < .05). The slope for the relationship between ethical leadership and courtesy was positive and significant for low moral awareness, t(126) = 2.98, p < .01, but not for high moral awareness, t(126) = .02, ns (see Figure 3). Simple slope analyses for ethical leadership at the group level also indicated that the slope for the relationship between ethical leadership and courtesy was significant for low levels, t(126) = 2.60, p= .01, but not for high levels of moral awareness, t(126) =−1.27, ns (see Figure 4). Thus, in line with Hypothesis 1, the relationship between both individual- and group-level perceptions of ethical leadership and courtesy were positive when moral awareness was low, but weakened and became non-significant when work group moral awareness became higher.
Further, the slope for ethical leadership and courtesy was positive and significant for high levels of empathic concern, t(126) = 2.81, p < .01, but not for low levels of empathic concern, t(126) = .47, ns (see Figure 5). The slopes for the relationship between ethical leadership at the group level and courtesy were significant for high levels of empathic concern, t(126) = 1.81, p < .05, but not for low levels of empathic concern, t(126) =−.67, ns (see Figure 6). The direction of the slopes is not in line with Hypothesis 2 as the relationship between individual- and group-level perceptions of ethical leadership and courtesy were positive when empathic concern was high.
To estimate the level of variance in helping and courtesy accounted for by the interactions, hierarchical regression analyses were conducted to estimate the R2 change when the interaction terms were added in the model (Hofmann et al., 2003). Interactions of both moral awareness and empathic concern with ethical leadership explained 8 per cent of the variance in helping. These interactions explained 10 per cent of the variance in courtesy. According to Cohen, Cohen, West, and Aiken (2003), interactions typically explain 1–3 per cent of the variance in outcome variables. Thus, the R2 changes in our models can be considered relatively high.
In this study, we focused on the interaction between ethical leadership and climate and found that work group moral awareness and empathic concern acted as moderators of the relationship of ethical leadership with follower helping and courtesy. These interaction effects were found both within and between work groups and were partially in line with expectations. First, as expected, ethical leadership is less strongly related to followers' helping and courtesy behavior when moral awareness climate is high than when it is low. Second, not in line with expectations, when both ethical leadership and empathic concern were high, followers show more courtesy. When empathic concern is low, no relationship was found between ethical leadership and courtesy. For helping, no interaction with empathic concern was found.
The most important contribution of this study to the ethical leadership literature is that it investigates whether ethical climate acts as a “substitute” for ethical leadership in the relationship with followers' helping and courtesy. So far, potential contextual moderators of the relationship between ethical leadership and follower behavior such as ethical climate have hardly been considered in research. Moderator effects in leadership research are hard to detect (Villa, Howell, Dorfman, & Daniel, 2003). Yet, there is a need to understand how employees process various contextual cues in determining appropriate behavior and while both leadership and climate are clearly relevant to this issue, ethical leadership and ethical climate are currently separate bodies of research. In this study, we start filling this gap by starting to integrate the ethical leadership and facet-specific ethical climate literatures.
Based on substitutes for leadership and social learning, we proposed a substituting role for both moral awareness and empathic concern climates in the relationship between ethical leadership and follower helping and courtesy. However, we found different moderating effects for these two facets of ethical climate. Moral awareness acted more (but not entirely) like a substitute and empathic concern acted as an enhancer. Previous research has also shown that similar context elements can have different effects on the relationship between leader and follower behaviors (cf. Erdogan, Liden, & Kraimer, 2006; Cole, Bruch, & Shamir, 2009). For example, Cole et al. showed that one contextual variable (social distance) had differential effects, enhancing some relationships with leadership while neutralising others, already suggesting that the effects of specific moderators for specific outcomes may differ.
The pattern we found for moral awareness is partly in line with a substitute role in which moral awareness is not related to follower helping and courtesy but does reduce the impact of ethical leadership on helping and courtesy. Indeed, we did not find a direct relationship between moral awareness and helping and courtesy and thus moral awareness alone may not be enough for individuals to show citizenship. This is in line with recent literature suggesting that ethical violation may be recognised, but this does not automatically mean that it will be acted upon (Tenbrunsel & Smith-Crowe, 2008). Also, in line with the proposed substitute role, we found that ethical leaders had a strong impact on follower helping and courtesy when moral awareness was low and that when moral awareness was high, the relationship between ethical leadership and helping became non-significant.
However, inspecting the plots also shows that the shape of our interactions is not entirely in line with a substitute explanation. If moral awareness acted perfectly as a substitute, a ceiling effect for moral awareness would be present, that is the high moral awareness bar would be flat, but as high as the level of low moral awareness with high ethical leadership. And although we do indeed find the expected flat line for high moral awareness (showing that as expected there is no relationship between leadership and citizenship when awareness is high), we also see that in low moral awareness contexts, the relationship of ethical leadership with outcomes is even stronger than expected (as in the plots that line goes beyond the ceiling rather than up to it). Perhaps the salience of ethical behavior by the leader is even higher than expected in a context low on moral awareness and such leaders may contrast or stand out strongly positively in the eyes of followers when moral awareness is low. Also, our study has a dyadic focus, involving both leaders and individual followers and this too may have stressed the relationship between leader and follower and increased leader salience in this context.
Empathic concern acts as an enhancer rather than a substitute for ethical leadership in relation to courtesy. No significant effects were found for helping. The findings show that followers react with more courtesy when both empathic concern and ethical leadership are high. This suggests that ethical leaders may only be able to affect follower courtesy when employees are willing or able to act pro-socially and be ethical, for example, when there is an ethical climate for empathic concern. To use a metaphor: only when soil is nutritious will sowing seeds lead to the growth of plants, and without either the seed or nutritious soil no plants will grow. In other words, ethical leadership may be more effective in influencing follower courtesy when there is a “breeding ground” or an ethical climate characterised by empathic concern.2 Employees are not always equally open to leadership (Lord et al., 2001) and when it comes to acting with courtesy, followers may be more open and responsive to ethical leader behavior when shared perceptions of empathic concern exist.
In different contexts, people have different cognitive representations about what leader behaviors are appropriate (Lord et al., 2001) and although our results confirm the importance of considering both, they also show that the interrelationships between follower perceptions of leadership and climate are not straightforward. We find differences for the two moderators and the two outcomes. Future research is needed to explore whether empathic concern and moral awareness may act as substitutes for or enhancers of other behavioral outcomes of ethical leadership as well. Different moderating roles may exist for different outcomes. For example, could empathic concern climate enhance relationships between ethical leadership and pro-social behavior and substitute or neutralise relationships between ethical leadership and anti-social behaviors? Another important area in this respect is to understand more broadly how individuals process different contextual cues in determining appropriate behavior and what the impact is of cues not being in line with one another.
The present study also extends previous research on ethical leadership through the finding that individual perceptions of ethical leadership are more strongly related to helping than shared perceptions of ethical leadership. Den Hartog and De Hoogh (2009) assessed individual and group-level effects of ethical leadership on follower attitudes and also found more individual than group-level effects. As noted, the stronger relationship found between individual perceptions of leadership and follower behavior may be encouraged by the design of the study in which leaders rate follower behaviors as this design stresses the relationship between leaders and individual followers. Additionally though, ethical leadership is likely to occur at the group level, because individuals within a work group are more homogeneous in their ethical leadership perceptions. As ethical leadership can be a within- and between-group variable, theory should determine the level of analysis (Kozlowski & Klein, 2000). Future research may incorporate different levels of analysis depending on theoretical purposes.
In organisational settings ethical leader behavior may enhance follower helping and courtesy. For organisations it is thus worthwhile to increase the level of ethical leadership, for example through training. Creating a climate of empathic concern is also helpful in stimulating courtesy. Training empathic concern seems a challenge, yet interactive role-playing focusing on learning to be receptive and interactive in order to understand another's need might be helpful (Schminke, Arnaud, & Kuenzi, 2007). According to Schminke et al. (2007), shared perceptions of moral awareness can be improved by clear and open communication about what constitutes (un)ethical behavior and desired behavior. Since the moderators empathic concern and moral awareness are differentially related to ethical leadership and have a different moderating role (i.e. enhance and substitute), it will pay off if a leader can recognise both these aspects of the ethical climate present in the work group in order to develop a successful strategy for influencing follower helping and courtesy behavior.
Strengths, Potential Limitations, and Future Research
Strengths of our study include its multi-level field research design and multi-source nature. Most research on ethical leadership has been conducted at a single level without taking the influence of contextual variables into account. By using different raters of leader and follower behaviors common source bias was reduced (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Although two antecedents of this study (ethical leadership and ethical climate dimensions) were provided by the same employees, common method bias is unlikely to result in statistical interactions which are the main purpose of this study (Epitropaki & Martin, 2005). Thus, common method bias is not likely to be a concern. Also, we controlled for transformational leadership in our analyses.
Our study has potential limitations as well. First, recognition of the potential limits of the generalisability of the study is important. We investigated a sample of leaders and followers in various organisations in the Netherlands. The organisations and work groups had different tasks and goals and these characteristics may have affected the moral awareness and empathic concern variables. Weber (1995) showed that employees from different departments within an organisation showed different preferences for ethical climate related to their core tasks. In our study, followers within a work group sometimes had different core tasks and thus may have had different climate preferences. Future research might use a more structured setting and is needed to replicate these findings within other cultures and contexts.
Next, we found an interaction effect between ethical leadership and empathic concern on courtesy; but not, however, on helping. Our study was limited to two dimensions of the facet-specific climates regarding ethics. Although these two dimensions are important, the ethical context is a complex and broad construct and additional ethical climate facets could be studied in relation to leadership to further understanding of the circumstances under which ethical leaders are effective. Future research could also investigate procedural justice climate (e.g. Naumann & Bennett, 2000) as a potential moderator for the ethical leadership and follower behavior relationship. Similarly, a limitation is that the current study focuses narrowly on helping and courtesy, and future research is needed to extend the relationship between ethical leadership and other (deviant) behaviors. Further, as noted, we controlled for transformational leadership and did not find a direct relationship between such leadership and follower helping. This has been found before (e.g. Podsakoff et al., 1990), but is not in line with a meta-analysis performed by Podsakoff et al. (2000) that does report a positive relationship between these two variables. Finally, the cross-sectional design of this study does not allow testing for the direction of causality, and supplementing field studies with experimental studies that also address this issue could help tackle this in future work.
In conclusion, stimulating ethical behavior is important for organisations. Leaders form role models of appropriate behavior and can motivate followers to help others and show courtesy at work. This study shows that the roles of climate facets of empathic concern and moral awareness are important in this regard. Specifically, this study contributes to the field by starting to unravel the complex relationship between ethical leadership and ethical climate facets, as both are important in providing ethical guidance and behavioral cues for employees.
We also performed the same analysis for the moderating effects of climate on ethical leadership by entering one climate dimension at a time in the equation. Conducting the analyses in this way yielded the same pattern of results.
We thank an anonymous reviewer for raising this point.