The corporate scandals of Enron, Tyco, Bank of America, WorldCom, and Merck during the past decade have led to an outpouring of popular press articles elaborating on the lack of regulatory oversight, complex and deceptive accounting schemes, and greed of these organizations’ top leaders (McLean and Elkind 2004; Patsuris 2002; Revell and Burke 2003; Toffler and Reingold 2003). Understanding these scandals requires scrutinizing organizational leaders, their ethical sensibilities and the climates they fostered; in fact, the predominant focus of academic literature has been on top leaders’ characters, behaviors, and formative experiences (e.g., Brown and Treviño 2006).

Without a doubt, top leaders are important for setting the moral tone for an organization (cf., Sims and Brinkmann 2002; Viswesvaran et al. 1998; Zahra et al. 2007). As media reports indicate (e.g., Revell and Burke 2003), unethical top leaders can have tremendous impact on the survival of a company, but small, seemingly insignificant acts of dishonesty performed by immediate supervisors may convey an even more insidious message to the organization’s employees. Supervisors who engage in dirty practices confirm local norms encouraging unethical behavior among immediate reports and most likely for other supervisors of the same rank. Yet, little research has assessed the possibility of differing effects of ethical behavior displayed at top levels of an organization versus that displayed by an immediate supervisor (Davis and Rothstein 2006; Mize et al. 2000; Wiley 1998). We surmise that ethical models at the lower ranks could be as important as those of the top leadership. Indeed, the effects of immediate supervisors on employees may very well be more pronounced than top leadership behavior, as immediate supervisors typically control an employee’s performance evaluations, promotions, and pay. If that direct supervisor is seen as ethical, we would anticipate favorable employee attitudes toward that supervisor (Dirks and Ferrin 2002) and a greater likelihood that the supervisor could influence the direct report’s ethical behavior.

Why Differentiate Top Leadership from Immediate Supervision?

For the majority of organizational employees, their immediate supervisor is the primary representation of the organization, providing daily direction, feedback, evaluation, and information about the job, work group and organization (Greller and Herold 1975; Therkelsen and Fiebich 2003). Further, research indicates that differences emerge in satisfaction with, and trust in, immediate supervision and top executives (Ellis and Shockley-Zalabak 2001). We might expect employees would differentially evaluate their top leaders’ and supervisors’ ethical tendencies. These differences may, in turn, be reflected in organizationally relevant individual outcomes such as organizational commitment and organizational citizenship. Finally, these differentiated perceptions are likely to have important implications for how ethical lapses may be prevented. For example, Nyberg (2008) has suggested that at the lower levels of the organization, ethical dilemmas are particular to the situation and abstract, universal ethical codes may, in actuality, impede good ethical decision-making at this level. Quite possibly, the rank and file may hold their top leaders to a universal code of ethics but recognize that their own supervisors have unique, ambiguous situations with which to cope. By extension, the training and education of ethical awareness, knowledge, and decision-making may be different fundamentally for lower level managers relative to top leadership.

Measuring the Perceptions of Ethics of Top Leaders and Immediate Supervisors

There are a couple of notable scales that have been developed to assess leadership ethics and integrity. Craig and Gustafson (1998) developed the Perceived Leader Integrity Scale (PLIS), a 40-itemFootnote 1 survey that asks respondents to rate a leader’s behaviors (e.g., “would lie to me”). In a two-study paper, Craig and Gustafson demonstrated that the PLIS was psychometrically sound, unidimensional, and displayed convergent and discriminant validity. Parry and Proctor-Thomson (2001) found that ratings of transformational leadership behaviors were moderately correlated (rs ranged from 0.34 to 0.46) with the PLIS scores.

More recently, Brown et al. (2005) developed and validated the Ethical Leadership Scale (ELS), using a social learning framework. The authors proposed as a foundation that ethical leaders influence followers to behave ethically, which at its base means that followers must attend to, accept, and role model leadership behavior. Brown et al. constructed a 10-item scale that they validated in seven interlocking studies. These studies indicated that the ELS was unidimensional, correlated with aspects of leadership influence, honesty, fairness, and was predictive of leader effectiveness. In their discussion, Brown et al. asked if “distance from the leader would influence employees’ ratings of ethical leadership” (p. 131). That question is the one we tackle in this article.

With two different samples, we will illustrate that perceptions of immediate supervisor ethics can be differentiated from perceptions of top leadership ethics, and that such differentiation has implications for organizations. Our focal instrument will be the Perceptions of Ethical Leadership Scale (PELS) that was developed and used in a previously published study (Pelletier and Bligh 2006) to assess perceptions of both top leader and immediate supervisor ethics. We will explain the PELS in more detail as we describe the first study.

Purpose of This Article

We present data from two studies. Our goal with the first data set was to provide preliminary evidence that PELS could be differentiated into measurable perceptions of top leadership ethics and perceptions of the ethics of the immediate supervisor. The second study provides additional evidence for such a differentiation at the level of the individual follower.

Study (Sample) 1

For this study, we re-analyze the dataFootnote 2 of Pelletier and Bligh (2006), and examine the two aspects of the PELS proposed above, specifically top leadership and immediate supervisor ethics. Because the focus of their study was on testing a model of ethics program effectiveness rather than scale development or differentiating the perceived ethics of immediate supervisor and top leadership, the PELS was developed in situ to assess employee perceptions in an agency that had suffered from several public scandals (see Pelletier and Kottke 2009, for a case analysis of the agency). Thus, no a priori definition was provided for PELS. An examination of the PELS items strongly suggests that the definition utilized to create the items was based on a leader’s ability to foster an ethical culture by articulating organizational values, facilitating the moral development of followers through the implementation of morally appropriate influence strategies and tactics, supporting employees when faced with ethical dilemmas, confronting ethical misconduct in a timely manner, and making ethics a top priority in the organization’s functions. Ethical leadership was considered as a potential precondition for ethics program effectiveness, since ethics programs are likely to be perceived as effective when leaders encourage ethical dialog, create an environment of trust, and provide organizational resources to ensure ethical decisions are made. When an organization is seeking to re-establish social legitimacy as a result of ethical misconduct, the role of leadership is crucial for enhancing an ethical culture. Within this context, Pelletier and Bligh (2006) argued that ethical leadership is a critical complement to ethics codes, decision-making processes, and organizational resources in crafting an ethical organizational culture. As measured in their study, ethical leadership was highly correlated with those variables (i.e., ethical decision-making, informal norms, perceived ethics program effectiveness).

Method

Participants

One thousand employees were randomly selected from a total of 18,000 who worked for a southern California county government agency. Of the 1,000 surveys distributed, 75 were returned as undeliverable as addressed or the employee had left the agency. Of the 925 surveys delivered, 418 employees responded (45.2 % response rate). This response rate is higher than average response rates of 20–30 % that have been reported in other survey studies (cf., Griffis et al. 2003; Hays and Kearney 2001). The majority of the respondents (62.2 %) were female. Participants averaged 45 years of age (SD = 10.1) and had been employed with the agency from less than 1 month to 40 years (M = 9.5 years, SD = 8.5).

Measures

For this article, we are primarily focused on the 10 items that composed the PELS developed for the Pelletier and Bligh (2006) study. The PELS includes a blend of items, some addressing top leadership, others the immediate supervisor. The 10 items constituting the PELS were contained within a 37-item Perceptions of Ethical Climate (PEC) survey constructed to assess employees’ perceptions of the effectiveness of the agency’s ethics program as it related to informal and formal norms, ethical decision-making, and the employees’ awareness of existing ethical codes. We reference the other scales presented in Pelletier and Bligh, as we will conduct a re-analysis of those data, ultimately separating the original PELS items into Top Leadership and Immediate Supervisor subscales, and correlating those subscale scores with the other variables included in the original study. The other scales were the Ethics Code Awareness Scale (6 items, α = 0.82), the Perceptions of Ethical Decision-Making Processes (2 items, α = 0.55), three subscales of the Perceptions of Ethical Resources Scale (time, 3 items, α = 0.65; organizational financial resources, 4 items, α = 0.90, and ethics information, 3 items, α = 0.83), Perceptions of Informal Ethical Norms (5 items, α = 0.82), and Perceptions of Ethics Program Effectiveness (4 items, α = 0.79). All items were presented on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree) with higher values representing higher levels of agreement.

Procedure

Using a random seed algorithm, the payroll department of the agency generated a list of 1,000 names to whom the online survey link or paper surveys were distributed. Because not all employees had intranet access, about a third of the employees received a paper copy of the survey via interoffice mail and were instructed to mail the survey back to the lead researcher via confidential interoffice mail.Footnote 3

Results

Analytic Strategy

Our primary interest was whether the PELS could be segregated meaningfully into two dimensions, so our first analysis was a factor analysis of those items. If the scale factored as expected, the next step was to determine if the subscales would correlate differentially with other study variables that represented concepts corresponding more closely to the top leadership of the agency versus those that were illustrative of immediate supervisor influence.

Data Cleaning

Prior to conducting the factor analysis, item data for the PELS were screened for univariate and multivariate outliers. No univariate outliers were found. Using a χ2 critical value of 29.6 (p < 0.001), six multivariate outliers were detected; these cases were excluded from the subsequent factor and reliability analyses, resulting in n = 412. In anticipation of the follow-up correlation analysis, the other scales used in the Pelletier and Bligh study were also examined for univariate outliers. One outlier that exceeded the recommended 3.29 z (Tabachnick and Fidell 2001) was found for the Perceptions of Decision-Making Processes Scale and was excluded from subsequent analyses.

Factor Analysis

A principal axis factor analysis with an oblique (SPSS oblimin) rotation was conducted. The items representing top leadership and immediate supervisor factored neatly into two factors with two distinct eigenvalues (EV) greater than 1 (EV1 = 5.5 and EV2 = 1.8). Factor loadings ranged from |0.67 to 0.88|. See Table 1 for the text of the items, factor loadings, item means, and standard deviations. In addition, reliability analyses were conducted on the two factors. αs for the top leadership and immediate supervisor ethics subscales were 0.93 and 0.88, respectively. Table 1 also includes the item–total correlations for the two subscales. These values range from 0.67 to 0.83.

Table 1 Means, standard deviations, factor loadings, and item–total correlations for data from Pelletier and Bligh (2006)
Full size table

Correlational Analysis of Subscales with Original Perceptions of Ethical Climate Scales

Subscale scores were computed and correlated with the seven scales originally used in Pelletier and Bligh. To determine if the strength of the correlations was different by level of leadership ethics, the correlations of top leader and immediate supervisor ethics with the seven PEC subscales were compared (Steiger 1980). These correlations with statistically significant differences are shown in Table 2. The information and financial subscales of the ethical resources scale, informal ethical norms scale, and ethics program effectiveness scale scores were all more highly correlated with the perceptions of top leadership ethics than with immediate supervisor ethics. The perceptions of ethical decision-making processes scale was more highly correlated with the perceptions of immediate supervisor ethics than with top leadership ethics.

Table 2 Correlations of Pelletier and Bligh perceptions of ethical climate scales with top leader and immediate supervisor PELS subscales
Full size table

Discussion

The items that constitute the PELS can be categorized statistically and meaningfully into perceptions of the ethics of top leadership and immediate supervisor. The follow-up reliability analysis indicated that the two sets of items exceeded a commonly cited minimum for α (0.70, Nunnally and Bernstein 1994), suggesting that the PELS subscales can be used to assess two levels of organizational ethics.

The pattern of correlations with measures of the PEC (Pelletier and Bligh 2006) was consistent with the measurements of top leadership and immediate supervisor ethics. Since it is the top leaders who typically embed culture (Schein 1990), it is not surprising that employees would associate the formal and informal norms and values of an organization with the upper echelon. Although immediate supervisors are likely to have an influence on the subcultures of the organization, employees recognize that it is the top leaders who create the “climate” and determine ethics policies. With regard to informational resources, immediate supervisors typically do not have the decision-making latitude to develop unit policies; they are responsible for the enforcement of these policies. The newly instituted ethics resource office was the entity that provided information such as the ethics code, procedures for filing an ethics complaint, ethics training calendar, and other resources designed to aid an employee when faced with an ethical dilemma. Since this office was instituted at the direction of the governing board and top leaders of the agency, rather than the middle management team, employees who perceived the ethics program to be effective also had more favorable ratings of the top leaders.

The employees in this organization were also likely to understand with whom financial decisions resided. In a government agency, top leaders and government officials determine how the money will be spent. Immediate supervisors work within their respective budgets, but only top leaders had access to the general fund; consequently, the top leaders had control of discretionary funds. At the time of the corruption scandal, unethical decisions based on greed were made by the top leaders, and these decisions were well publicized (Pelletier and Kottke 2009). Employees were aware of the leaders who were implicated in the scandal, and why they were being implicated. An example of an item that assessed perceptions of financial decision-making was “This organization is willing to do the right thing even if it costs more money.” It makes sense then that if employees felt the top leaders were willing to sacrifice money to ensure ethical conduct, they would perceive those leaders to be modeling the behavior they espoused, and adhering to the values they embedded. Conversely, when employees realized that decisions were being made based on monetary and personal gain by the top leadership, they were less likely to perceive the ethics program as being effective.

Perceptions of sound decision-making processes were related more to immediate supervision than to top leaders. For example, the decision-making scale included an item that addressed consultation with an immediate supervisor (i.e., “When faced with making a decision that has an ethical implication, I feel I can discuss the matter with my immediate supervisor”), which clearly could be expected to relate to an employee’s perceptions of his or her immediate supervisor’s ethical standards or the supervisor’s positive influence on the work unit’s subculture. In contrast, the four other scales within the PEC—information, financial resources, informal norms, and ethics program effectiveness—flowed from the top leadership of the organization.

Study (Sample) 2

The re-analysis of the Pelletier and Bligh (2006) data revealed that the PELS could be used to gauge employee perceptions of their top leadership as well as their immediate supervisor ethics. We employed a different sample of employees in our second study, and sought additional evidence that the perceptions of ethical behavior would be differentiated for top leadership and immediate supervisors. To that end, we identified several variables that we hypothesized would be related differentially to PELS at the top level of the organization, as well as, perceptions of an immediate supervisor’s ethics.

Expected Corollaries of Ethical Leadership

Organizational Climate

Among the most extensively studied variables, employee perceptions of work climates has been found to be related to individual level work attitudes, motivation, and performance (Parker et al. 2003). Considerable journal space has been devoted to the proper definition of climate, with no one agreed upon designation or label (e.g., psychological climate, work climate, organizational climate), although there has been coalescence around the climate concept representing an individual’s subjective description of an organization’s structure, procedures, and polices taken as an aggregate impression (James et al. 1978). We propose that this aggregate description reflects an individual’s psychological appraisal of the organization as interpreted through that individual’s needs and values (cf., Schein 1990). Among the organizational domains that are likely to be interpreted through the lens of the individual are aspects of its leadership (e.g., support, strategic direction; Jones and James 1979). A meta-analysis (Parker et al. 2003) indicated that leader behaviors were significantly related to perceptions of organizational climate. Similarly, we would expect that employees’ perceptions of top leaders’ integrity could have a critical impact on the climate of the organization. Therefore, we anticipated a positive relationship between organizational climate and perceptions of leader ethics. Further, because climate is a variable defining the organization, of which top leaders may be more strongly associated, we expected the correlations between the perceived ethics of top leaders to be more highly correlated with climate than the perceived ethics of immediate supervisors.

Top Leadership Direction

The top leadership of an organization sets the model for employees. Follower confidence in top leadership has significant implications for organizational success. Employees who perceive that their top echelon is competently guiding the organization are more likely to offer innovative ideas (Clegg et al. 2002), be more committed to the organization (Mathieu and Zajac 1990), and less likely to quit (Niehoff et al. 1990). For our purpose, we would anticipate that employee confidence in their leadership would be extremely relevant to their perceptions of their top leadership’s ethics. More directly, we hypothesized that perceptions of top leadership ethics would correlate strongly with confidence in top leadership. The relationship of immediate supervisory ethics to confidence in top leadership, we predict to be modest.

Organizational Commitment

Organizational commitment is related to a host of psychological work variables. Relative to non-committed employees, committed employees are more satisfied with and involved in their jobs, are more motivated, and are less likely to think of quitting or being absent from work (Mathieu and Zajac 1990). When the ethical values of top managers conflict with employees’ values, employee commitment suffers and turnover intentions increase (Schwepker 1999; Sims and Kroeck 1994). The presence of ethical leaders would be expected to encourage an atmosphere in which employees feel loyal to the organization. Because organizational commitment is measured as an attachment to the organization as a whole, we would expect top leader ethics would have a stronger relationship to organizational commitment than the perceived ethical values of the immediate supervisor. We hypothesized that employees who see their top leaders and immediate supervisors as ethical would be more likely to identify with and feel attached to their organizations. One could argue that top leaders through their creation of organizational culture, lay the foundation for employee commitment; yet, from a social exchange perspective, the immediate supervisor is likely to have a similar impact on an employee’s attachment to the organization, as the supervisor typically has considerable leverage on the employee’s immediate work environment. Therefore, we do not predict a significant difference in the relationships between commitment and perceptions of top leader ethics versus perceptions of immediate supervisory ethics.

Commitment can further be categorized into affective, continuance, and normative (Meyer and Allen 1991). Affective commitment encompasses the emotional attachment and identification with the organization. Continuance commitment is based on the recognition that leaving the organization has costs. Normative continuance refers to a sense of obligation or moral expectation to remain with the organization. Affective and normative commitment have similar antecedents, such as job challenge, role clarity and management receptiveness. In contrast, continuance commitment is most strongly related to organizational tenure (Allen and Meyer 1990). Consequently, we expected affective and normative commitment would relate positively to leader ethics, but continuance commitment would not be affected by leader ethics. As before, we do not expect different relationships by level (top leader versus immediate supervisor).

Organizational Citizenship Behavior

Organ (1988) originally defined Organizational Citizenship Behavior (OCB) as “individual behavior that is discretionary, not directly or explicitly recognized by the formal reward system and that in the aggregate promotes the effective functioning of the organization” (p. 4). Since Organ first introduced the concept, considerable debate has ensued about the definitional nature of OCBs. In his later work, Organ redefined the construct as activities that contribute “to the maintenance and enhancement of the social and psychological context that supports task performance” (Organ 1997, p. 91). Though the dimensionality of OCB has been debated over the past 20 years (cf., LePine et al. 2002), the five categories that Organ originally proposed have endured. These five types of OCB are altruism, conscientiousness, sportsmanship, courtesy, and civic virtue. The first, altruism, represents behaviors that have the effect of helping a specific other person with an organizationally relevant task or problem. Conscientiousness involves employee behaviors that go beyond the minimum role requirements of the organization, in the areas of attendance, conserving resources, and obeying rules and regulations. The third, sportsmanship, is the willingness of the employee to tolerate less than ideal circumstances without complaining. Courtesy includes behaviors aimed at preventing work-related disputes. Finally, civic virtue is characterized by behaviors by an individual that indicates he or she participates in, is involved in, or is concerned about the life of the company (Podsakoff et al. 1990).

Aligned with the tenets of social exchange theory (Blau 1964), followers of ethical leaders would want to support their leaders because of the expectation that they—the followers—would be supported in turn. If employees perceive top leaders, as well as immediate supervisors, as being ethical, we would expect them to be willing to “step up” to fill the “white space” of the organizational chart. Because this aspect of social exchange would be more salient at the local level (Ilies et al. 2007; Settoon et al. 1996), we would expect the immediate supervisors’ ethical stance to have more impact than top leaders on those behaviors. Therefore, we expected OCB and its subdimensions would be moderately (and positively) correlated with leader ethics, with the relationship stronger for immediate supervisory ethics.

Method

Sample

Participants were recruited from a variety of organizations including an advertising agency, a collections bureau, a medical services corporation, mortgage companies, and one mid-sized research institution. Some participants were recruited through a “snowball” method that utilized students from upper division psychology classes at a regional university in southern California; students were asked to distribute one or two paper surveys to working adults they knew, who, in turn, would distribute a survey to a coworker.Footnote 4 Two forms of the survey were administered: an online version and a paper and pencil version.Footnote 5

A total of 371 surveys were returned to the researchers or submitted online (online n = 170; paper n = 201). The majority of participants were women (69 %). There was considerable variation regarding participants’ tenure, department and company size, and type of industry. The modal values for length of employment were 1–3 years with the organization (35 %), working for a company smaller than 100 employees (22 %) and in a department of 10–24 employees (37 %). The distribution of participants’ tenure, department and company size, and type of industry are shown in Table 3.

Table 3 Participants’ tenure, department and company size, and type of industry
Full size table

Because organizational climate was a key variable, and we believed employees require a history with a given organization to become fully cognizant of that climate, only data from those who had been working in their organizations for at least 1 year were used in the statistical analyses. The sample contained 301 participants after this criterion was applied.

Measures

Perceptions of Ethical Leadership

The focal measure for this study, the PELS, contains 10 items, six that address top leadership ethics and four that compose the immediate supervisor ethics subscale.

Organizational Climate

The Organizational Climate Scale (Lawler et al. 1974) is a 15-item measure designed to capture organizational climate as perceived by those who work within the organization. Respondents were instructed to use their organization (i.e., not work group or department) as the frame of reference. Each item consists of two bipolar adjectives anchoring a 7-point Likert scale. We selected this measure as its format (modified semantic differential) was expected to elicit a potent representation of the participant’s affectation toward his or her organization’s overall climate. α was 0.80 for this study.

Top Leadership Direction

Four items from the Top Leadership Direction Scale (TLDS; Kottke et al. in press) were included. These items assess follower confidence in the direction provided by top leadership and perceived instrumentality of that direction. An example item is “The leadership inspires confidence in the future success of the organization.” α for the scale was 0.86.

Organizational Commitment

The Organization Commitment Scale (Meyer and Allen 1997) consists of 24 items with three subscales measuring affective, continuance, and normative commitment. αs for this study were 0.83 for affective, 0.79 for continuance, and 0.84 for normative.

Organizational Citizenship Behavior Scale

The Podsakoff et al. (1990) OCB Scale is composed of 19 items.Footnote 6 The scale measures the degree to which employees perform behaviors that contribute to the welfare of the organization, but are not included in the duties and responsibilities of their position. The five types of OCBs articulated by Organ (1988) are measured. αs for this study: altruism, 0.79, conscientiousness, 0.74, sportsmanship, 0.64, courtesy, 0.44, and civic virtue, 0.71.

Results

Analytic Strategy

First, the proposed two-factor structure of the PELS was examined with a confirmatory factor analysis (CFA); this structure was compared to a single factor structure, the rationale being that the scale originally was used to create a single scale score. Subsequently, subscale scores were calculated and correlated with other measures to establish convergent validity for the proposed differentiation of perceptions of ethics by level of supervision.

Data Screening

Prior to conducting the factor analysis, item data for the PELS were screened for univariate and multivariate outliers. No univariate outliers were found. Using a χ2 critical value of 29.6 (p < 0.001), eight multivariate outliers were detected; these cases were excluded from the subsequent factor and reliability analyses (resulting n = 293). In addition, the scaled variables were also examined for univariate outliers. Eight outliers were detected and excluded from the follow-up correlational analyses.

Structure of the PELS

CFA was conducted using EQS (Bentler, 1995) to evaluate the fit of the 10 items to a two-factor model. We examined multiple indices of fit (Bollen 1989), the χ2, the root mean square error of approximation (RMSEA), the comparative fit index (CFI) and the non-normed fit index (NNFI).

Ultimately, two confirmatory factor analyses were performed, first with the proposed two-factor structure, and then with a single factor solution for comparison. In the first solution, factors were permitted to covary based on the rationale that there may be an underlying relationship between the perceptions of the ethics of top leadership and immediate supervisor. The fit of the items to their respective two factors was adequate when evaluated by the Satorra-Bentler χ2 ([34, N = 244) = 137.55, p < 0.001], the CFI (0.93), and NNFI (0.90). The RMSEA statistic (0.10, 90 % CI 0.09–0.12) suggested less of a fit (<0.06 is recommended. Hu and Bentler 1999; Chen et al. 2008, however, suggest that the 0.06 criterion may be too rigid). The two-factor solution was superior to the one factor (Satorra-Bentler χ2 (35, N = 243) = 203.07, p < 0.001, RMSEA = 0.13, 90 % CI 0.11–0.15, CFI = 0.88, and NNFI = 0.85.). The χ2 difference between the two models was statistically significant: χ2(1) = 28.08, p < 0.001.

Standardized factor loadings (See Table 4) for the two-factor structure ranged from |0.61 to 0.85|. Because composite subscale scores were to be used, coefficient αs were also computed. Cronbach’s α for the two subscales, top leadership and immediate supervisor, were 0.93 and 0.88, respectively. These αs are well above the typically recommended minimum of 0.70 (Nunnally and Bernstein 1994).

Table 4 Means, standard deviations, factor loadings, and item–total correlations for sample 2
Full size table

Convergent Validity

Table 5 contains means, standard deviations, αs, and correlations of the hypothesized correlates of leader ethics. As predicted, organizational climate was highly correlated with top leader ethics, 0.64, and immediate supervisor ethics, 0.47, with the relationship significantly greater for the top leader ethics (difference in rs: z = 3.88, p < 0.001). Ratings of top leadership direction were also differentially correlated (z = 8.65, p < 0.001) with top leadership (0.73) and immediate supervisor ethics (0.43). Correlations of affective, continuance, and normative organizational commitment with top leader ethics were 0.55 (p < 0.001), 0.02 (ns), and 0.50 (p < 0.001), respectively. For immediate supervisor ethics, the correlations were 0.46 (p < 0.001) with affective, 0.04 (ns) with continuance, and 0.42 (p < 0.001) with normative commitment. The direction of both sets of correlations was supportive of our hypothesis that ethical leadership would be related to organizational commitment; the lack of differences in the correlation pairs demonstrate that organizational commitment was not differently related to top and immediate leader ethics. Subdimensions of OCB, altruism, courtesy, sportsmanship, civic virtue, and conscientiousness, correlated with top leader ethics, 0.17 (p < 0.01), 0.10 (ns), −0.14 (p < 0.05), 0.27 (p < 0.001), and 0.07 (ns), respectively; with immediate supervisory ethics, the correlations were 0.18 (p < 0.01), 0.04 (ns), −0.12 (ns), 28 (p < 0.001), and 0.11 (ns). With the exception of civic virtue, the OCB results were not supportive of the hypothesis that OCB would be moderately related to leader ethics, nor were there differences in the correlations by level of leadership.

Table 5 Scale item means, standard deviations, αs and correlations
Full size table

Comparison of PELS Subscales by Sample

Finally, because the context for the data collected for the first study of this article included an organization that had faced several public scandals, we compared the means of the top leadership and immediate supervisor subscales for the two samples. For both top leadership and immediate supervisor measures, there were significant differences. Specifically, the top leadership subscale mean was significantly lower in the data set from sample 1 (M = 3.83, SD = 1.43) than from the data in sample 2 (M = 4.56, SD = 1.48; t[699] = 6.57, p < 0.001). In contrast, the mean for supervisor ethics from sample 1 (5.07, SD = 1.52) was higher than the mean in sample 2 (M = 4.58, SD = 0.96; t[699] = 4.86, p < 0.001). These differences also suggest construct validity for the use of the two subscales, as well as convergent validity in that the first sample consisted of employees whose agency had been shaken by ethical scandals involving their top leaders. The second sample was collected from a variety of organizations whose employees most likely would have had a mix of ethical and unethical top leaders and immediate supervisors.

Discussion

The purpose of this second study was to assess the relationship of several individual attitudes and behaviors that could be expected to emanate from ethical leaders, both distal and immediate. We also wished to assess the relationship of organizational climate to top leader and immediate supervisor ethics. Finally, we hoped to evaluate the differential importance of top and immediate leader ethics.

Climate is a shared perception of the nature of the organization, its policies, procedures and its “way of doing things” (Vardi 2001, p. 327). Organizational climate and ethics of leaders, at both levels, were clearly linked, providing some of the strongest correlations of the study. How employees perceived their general organizational climate had considerable overlap with their assessment of the ethical perspectives of their top leaders and supervisors. Our expectation that top leader ethics would be more highly correlated with climate was confirmed, supporting that top leader ethics are an important part of the overall organizational fabric. The strength of the link between supervisor ethics and climate was somewhat surprising as climate is predominantly an organizational focus, which logically would be more associated with leaders at the top than an immediate supervisor. Perhaps the strength of the correlation even at the supervisory level demonstrates the potency or saliency of ethical issues regardless of the leader being evaluated for ethical behavior.

The other variable that displayed a strong, differentiating correlation with top leader and immediate supervisor ethics was the top leadership direction variable. Scores on the TLDS represent confidence in top leadership. Those who rated their top leaders as more ethical also expressed more confidence in their ratings of the top leadership direction. These results are consistent with other research that finds that employees who rate their leaders as ethical also rate those leaders as having more effective leadership styles (Brown et al. 2005). These results support the mounting evidence that ethical leadership is inexorably bound to follower perceptions of good leadership (e.g., Bass and Steidlmeier 1999; Brown and Treviño 2006; Simons 1999).

As expected, affective and normative commitment were strongly related to perceptions of ethics for both top leaders and immediate supervisors. Continuous commitment, faithful to its construct meaning, was not related to PELS at either level. Employees who evaluate their commitment in terms of the behavior necessary to remain employed are not likely to see the ethical attributes of either their top leaders or immediate supervisor as necessarily relevant. In contrast, the moderate to strong correlations of affective and normative commitment with both subscales of leader ethics indicate that the tone sounded by top management and the immediate supervisor is relevant for employees’ sense of attachment and obligation to their organization. Though their meta-analysis did not address ethical leader behaviors, Mathieu and Zajac (1990) found that leader behavior was a moderate predictor of organizational commitment. Our results are also consistent with previous research in which ethical leadership led to desirable individual employee outcomes, especially affective attitudes (Koh and Boo 2001; Vitell and Davis 1990).

The relationship of OCBs to leader ethics was small with the exception of civic virtue, which correlated 0.27 and 0.28, respectively, with top leader and immediate supervisor ethics. Employees who perceived their top leaders and their immediate supervisors to be ethical were also the ones most willing to support the organization beyond their formal job requirements. Research by Vey and Campbell (2004) has indicated that most attitudinal items tapping OCBs are considered by independent observers to represent in-role, as opposed to extra role, behavior. Vey and Campbell found, however, that altruism and civic virtue were seen as reliably extra role, with the results for civic virtue strongly suggesting it to be the most “contextual” of the OCBs that have been studied intensively. Its relationship here with ethics perceptions lends credence to this notion.

General Discussion

The results of these two studies provide ample evidence that the confidence that employees have in the integrity of their top leaders and immediate supervisors leads to differential, and important, individual outcomes. The integrity of top leaders and of immediate supervisors can be construed as part of the fabric of an ethical organizational culture that engenders benefits for the organization and its employees. At a minimum, one hopes the ethical organization is able to avoid the public backlash and subsequent collapse that accompanied Enron and its ilk. At best, one expects the organization will prosper because its employees and stakeholders can place trust in their leaders to “do the right thing” with the corollary that employees who trust their leaders’ probity can be expected to put forth more effort, support management’s ideas, and be good organizational citizens.

Taken together, these two studies demonstrated that perceptions of ethics at the immediate supervisor level can be differentiated meaningfully from perceptions of top leadership ethics by employees. In both studies, perceptions of top leadership ethics correlated more strongly with aspects of the organization associated with climate, leadership direction, financial resources, informal norms, and ethical culture—those organizational aspects are more often within the control of top leadership. Ethical decision-making processes were more related to immediate supervisory ethics than to top leader ethics, suggesting that role modeling by immediate supervisors is also quite important. Brown and Treviño (2006), for example, have suggested that ethical behavior develops more from close working relationships than from observing distal models.

Though no mean difference was hypothesized, means for the leader ethics subscales were compared between the samples and were significantly different for both subscales. For the employees of the agency that had endured public scandal, the employees rated their top leaders as significantly less ethical than the participants in the second sample, which included employees from many different organizations. Interestingly, employees of the scandal-ridden agency perceived their own supervisors as more ethical than did the employees from the second sample. Whether this difference is a result of a contrast effect with the agency’s top leadership viewed as unethical, or simply an indication that employees are more likely to trust their immediate supervisor with whom they are likely to have a close relationship, is not discernable from these data. The implication of this mean difference, taken together with Brown and Treviño’s (2006) review, is that impression management on the part of top leaders is necessary but not sufficient to convince organizational members that they are ethical. Working with an immediate supervisor gives a definitive assessment of his or her ethical behavior.

Although Pelletier and Bligh (2006) did not intend for the PELS to be used solely as a measure of top and immediate leader ethics, the subscales demonstrated desirable psychometric properties: good reliabilities and convergent validity with relevant organizational attitudes. Of particular interest, these two subscales of the PELS show promise for investigating employee perceptions of top leader and immediate supervisor integrity, adding an additional scale to the available measures (e.g., the PLIS, Craig and Gustafson 1998, and ELS, Brown et al. 2005).

Directions for Future Research

Longitudinal studies of multiple organizations would be very desirable to gauge recovery from scandal, or retrospectively, antecedents of ethical slips to determine if immediate supervisors are the first line of defense against ethical breaches. Though their sample of supervisors was small (1.6 % of the total sample) and authoritative statistical analyses could not be conducted, Velthouse and Kandogan (2007) found that first line supervisors accorded ethical issues less importance than did middle managers and executives. Their results, taken together with ours indicating that employees of a scandal-ridden agency rated their immediate supervisors as more ethical, suggest a potentially troubling disconnect in the ethical links within organizations.

Other scales and constructs need to be included in future research to investigate the nature of proximal and distal leader ethics. Meta-analytic moderator analysis to extract the effect of immediate supervisory ethics would be valuable, once sufficient data are available to make such comparisons. In their meta-analysis, Davis and Rothstein (2006) located only seven studies that provided enough information to be categorized as being about the ethics of top management or immediate bosses. The number of studies was small—four studies of immediate bosses, three about top management. The population correlations of manager behavioral integrity with employee attitudes were stronger for immediate bosses (0.50) than for top management (0.29). Despite the small number of samples, their data support that level of leadership ethics moderates the relationship between employee perceptions of leader ethics and attitudes. Clearly, more research is needed in this area.

Limitations

The data collected for these studies cannot imply causation. All data are cross-sectional and were analyzed using correlational statistics. Hence, any interpretations must be salted with caution. Further, other variables need to be assessed to verify the meaning of these correlational results. Other variables for future study include stage of development of the firm, the tenure of the top management team, the CEO in particular, the historical turnover of leaders and the centralization of power within the top management team. These variables cannot be assessed with a survey; they require different methods.

Because all data were collected via survey, common method variance could have inflated the correlations (Podsakoff, MacKenzie, Lee, and Podsakoff, 2003). The pattern of correlations in two separate samples, however, mitigates against our results being solely the consequence of method bias. In addition, Spector (1987, 2006) has strongly argued that common, systematic method variance is not as problematic as originally imagined in work attitude research.

In Study 2, we used an affective, evaluative measure of organization climate, which permitted us to assess general organizational climate. We acknowledge that several researchers (cf., Patterson et al. 2005) have suggested multiple climates operate simultaneously within an organization, a possibility we could not detect with our measure. Because our intent was to examine immediate supervisor ethics and top leader ethics, the existence of multiple climates, especially ethical climates (Victor and Cullen 1988) is particularly relevant and should be examined in future research.

Conclusion

Using two samples, we examined and found meaningful relationships between employee perceptions of the ethics of top leaders and their immediate supervisors, with important individual outcomes—perceptions of climate, top leadership direction, organizational commitment, and the OCB of civic virtue. Employees draw distinctions between top leader and immediate supervisor ethics; in the first sample, these distinctions were related to their perceptions of ethics program effectiveness, organizational norms, decision-making processes, and financial resources. Having found differential effects of leader level, the impact of overall leader ethics on employees is compelling. The serious and public collapses revealed in the cases of Enron, Adelphia, and Tyco may represent the final consequence of leader ethics that had in fact been eroding employee perceptions, attitudes, and behaviors for some time. Our research suggests that long before such large-scale organizational consequences, leader ethics at both top and immediate supervisor levels have an important impact on employees.