Effects of moral reasoning and management level on ratings of charismatic leadership, in-role and extra-role performance of managers: A multi-source examination

a b s t r a c t

This study investigated whether upper and lower-echelon managers’ moral reasoning
(n=377) was associated with the levels of charismatic leadership, in-role and extra-role
performance they displayed as perceived by their subordinates (n=1731), superiors, and self.
Managers completed the Defining Issues Test (Rest, 1990) to assess their moral reasoning
capacity. Self-ratings of managers’ charismatic leadership, in-role and extra-role performance
were collected two weeks later. Subordinate- and superior-ratings of these constructs were
collected one month later. Analysis of covariance indicated that managers possessing the
highest (i.e., postconventional) level of moral reasoning outperformed managers at the lower
preconventional and conventional levels, but displayed lower levels of self-rated charismatic
leadership. Upper-echelon managers displayed higher levels of charismatic leadership and
extra-role performance than lower-echelon managers. Within and between analysis (WABA)
indicated that upper-echelon managers at conventional and postconventional levels of moral
reasoning agreed with their subordinates and superiors that their charismatic leadership
ratings exceed those in lower management, but are lower than those of leaders who possess
preconventional moral reasoning.

The behavior of top managers in corporations has gained widespread attention from both leadership researchers and
practitioners due to numerous recent ethical scandals that have impacted workplaces, industries, and economies. This focus has
prompted a call for managers to strike a balance between ethical reasoning, doing good through organizational citizenship and
social responsibility, and performing well (Palanski & Yammarino, 2009; Savitz & Weber, 2006). Indeed, the requirement for top
managers to be moral beacons for others is not new (Bass, 2008), and is reflected in research on authentic leadership (Avolio &
Gardner, 2005), altruistic leader self-construals, values and behavior (Sosik, Jung, & Dinger, 2009), ethical leadership (Brown,
Trevino, & Harrison, 2005), transformational leadership and moral reasoning (Sivanathan & Fekken, 2002; Turner, Barling,
Epitropaki, Butcher, & Milner, 2002) and leadership developmental level (Strang & Kuhnert, 2009). Yet, there is relatively little
empirical research on links between moral reasoning and leadership, and what exists suffers in part from three problems.
First, the study of charisma has been a key component of leadership literature since it was introduced by Max Weber in the
early decades of the twentieth century (Bass, 2008), and its nebulous nature continues to challenge scholars. Burns (1978) pointed
out “The concept of charisma has fertilized the study of leadership. Its very ambiguity has enabled it to be captured by scholars in
different disciplines and applied to a variety of situations” (p. 243). For example, when one thinks of top corporate executives,
charismatic leaders, such as Steve Jobs, Richard Branson or Carly Fiorina, come to mind. Such corporate leaders provide today’s

modeling, rewarding, and communicating ethics) are equally important at the supervisory level. Moreover, supervisory level
ethical leaders focus on building relationships with employees. In contrast, executive level ethical leaders engage in more
boundary-spanning and relationship building activities with external constituents. Higher levels of moral reasoning may assist
supervisory level ethical leaders to conform to social and corporate expectations and follow ethical rules, and executive level
ethical leaders to role model and reward ethical behavior. The highest levels of moral reasoning may assist executive level moral
leaders to recognize and appreciate alternative values and goals held by external stakeholders because different groups hold
different values, and to communicate them as part of a holistic vision.
Kohlberg’s (1976) three stages of moral development – preconventional, conventional and postconventional (defined below) –
based on the development of moral judgment over an individual’s lifetime as he or she becomes more educated and gains
experience serve here as the definition of leader’s level of cognitive moral reasoning. This perspective of moral reasoning has been
criticized for failing to include other dimensions beyond rights and justice that should be considered in defining morality such as
social obligation and care (Gilligan, 1982), its individual rights, Western-oriented conception of morality (Peterson & Seligman,
2004), moral reasoning not always translating into moral action (Bass, 2008), its failure to recognize aspects of the situation as
moral (Rest, Narvaez, Bebeau, & Thoma, 1999), and differences in the size and consequences of moral breaches are often ignored
(Trevino et al., 2000).
Despite these challenges, evidence exists that supports the use of Kohlberg’s (1976) framework given the relative lack of
alternative frameworks and measures for assessing morality and ethics in leadership contexts. McCauley, Drath, Palus, O’Connor,
and Baker’s (2006, p. 646) review of the cognitive moral reasoning literature highlighted a significant amount of empirical support
for Kohlberg’s framework. Peterson and Seligman (2004) came to a similar conclusion in their review of research on Kohlberg’s
model: “It is clear that the stages exist, that they develop in invariant sequence, that they are related to cognitive and social
cognitive development, and that they are related to other aspects of moral psychology, such as behavioral outcomes…” (p. 396).
Moreover, Kohlberg’s stages of moral reasoning continue to be used to advance the understanding of ethical (Brown & Trevino,
2006), transformational (Kuhnert & Lewis, 1987), authentic transformational (Bass & Steidlmeier, 1999), and spiritual (Fry, 2003)
leadership, and how a leader’s moral development impacts effectiveness and managerial performance (McCauley et al., 2006).
There are three stages of moral development in Kohlberg’s (1976) framework. Preconventional moral reasoning is based on a
self-interested focus on avoiding punishment and seeking rewards, and is typically associated with lower-order transactional
leaders (Kuhnert & Lewis, 1987) and personalized charismatic leaders (House & Howell, 1992). Conventional moral reasoning is
based on laws and rules that represent good interpersonal relationships and maintain social order, and is typically associated with
higher-order transactional leaders (Kuhnert & Lewis, 1987). Postconventional moral reasoning is based on universal principles
and virtues such as justice and care, and establishing and upholding social contracts and individual rights. This type of reasoning,
while relatively rare in adult populations (Rest, 1990), is typically associated with transformational leaders (Kuhnert & Lewis,
1987) and socialized charismatic leaders (House & Howell, 1992).
Charisma is the principle component of transformational leadership because followers are inspired by, identify with, and wish
to emulate leaders with charisma (Bass, 2008). Charisma corresponds to the inspirational motivation and the idealized influence
components of transformational leadership (Antonakis, Avolio, & Sivasubramaniam, 2003). Charismatic leadership is defined as the
ability to inspire enthusiasm and action in followers via personal attributes, behaviors, and exemplary qualities of the leader,
especially in situations ripe for change. Charismatic leaders often engage in role modeling for their followers and frequently talk
about the importance of ethics and values (Bass, 2008). These behaviors represent two key elements of the moral manager aspect
of ethical leadership (Trevino et al., 2000). Through their role modeling and effective communication of ethics and values,
charismatic leaders build strong emotional bonds with their followers based on high levels of trust, identification and
commitment, and produce superior managerial performance results (Sosik, 2005; Sosik et al., 2002).

Performance is operationalized in the present study in terms of the managers’ in-role and extra-role job performance. In-role
job performance assesses managerial behaviors that are part of the formal job description (Williams & Anderson, 1991). Extra-role
job performance reflects managerial helping behaviors that are not specified in the formal job description, but are presumed to be
considered a good organizational citizen (Smith, Organ, & Near, 1983). Reviews of the leadership literature have linked
charismatic leadership to both types of performance (Bass, 2008).
1.1. Moral reasoning, charismatic leadership and performance
Leaders’ level of moral development has been conceptually linked to transformational leadership (Graham, 1995; Kuhnert &
Lewis, 1987). Furthermore, leaders with higher moral reasoning levels have been shown to display greater transformational
leadership behaviors (Turner et al., 2002). In his discussion of moral reasoning, Bass (2008, p. 217) argued that “Commentaries
about ethical charismatic leaders…and transformational leaders…are consistent with each other.” Because inspirational
motivation and idealized influence represent a significant portion of transformational leadership (Bass, 2008), we expected
leaders’ level of moral development to correlate with ratings of charismatic leadership.
As discussed above, postconventional moral reasoning involves the use of ethical principles and virtues by individuals. Leaders
at this stage of moral development are likely to display high levels of ethical and performance standards reflected in idealized
influence behavior and attributions of idealized influence made by followers. Similarly, postconventional moral reasoning is based
on a holistic and broad perspective by focusing on social contracts and a shifting of interest from the self to the collective. Such
reasoning underlies the role modeling of self-sacrifice seen in the idealized influence displayed by leaders to motivate followers to
make personal sacrifices for the good of the group. This type of reasoning also is required for leaders to display inspirational
motivation aimed at building collective efficacy and esprit de corps (Bass & Avolio, 1997). Thus:
Hypothesis 1. Managers’ with higher moral reasoning will display more charismatic leadership.
Performing well on transformative leadership initiatives involves expanding followers’ understanding of leadership contexts
from simple to complex, pointing out interdependencies among followers, building their collective efficacy, and changing their
interpersonal focus from self to others (Bass, 2008; Burns, 1978; Yukl, 2010). According to theories of constructive (Kegan, 1982;
Kuhnert & Lewis, 1987) and moral (Kohlberg, 1976) development, these behaviors are likely to be supported by higher levels of
moral reasoning that integrate personal standards with societal norms to form interpersonal connections, and value openness and
paradox to promote the well-being of others. Strang and Kuhnert (2009) reported positive correlations between leaders’
constructive-developmental stage and superior-, peer-, and subordinate-ratings of leader performance measuring 46 critical
leadership competencies. Rooke and Tolbert (1998) found the ego development stage of a CEO and his senior managers to be
positively related to a successful organizational change initiative. This line of reasoning suggests a positive relationship between
leaders’ moral reasoning level and their in-role job performance:
Hypothesis 2. Managers with higher moral reasoning will outperform managers with lower moral reasoning.
Bass (2008, p. 215) pointed out that “knowing and deciding what is right to do may not result in doing right.” His statement
raises an interesting question regarding whether leaders who reason at a higher level by focusing on the well-being of others
would actually engage in more helping behaviors in the workplace. Graham (1995) suggested that leaders who are postconventional
moral reasoners are active in the responsible governance of organizations, and are likely to display high levels of
organizational citizenship behavior (OCB). Such extra-role job performance may allow them to achieve a satisfying sense of
consistency between their reasoning process, words, and deeds as described by theories of cognitive dissonance (Festinger, 1957)
and authentic leadership (Avolio & Gardner, 2005). Palanski and Yammarino (2009) pointed out that consistency between words
and actions is one definition of integrity, and that integrity is positively related to social aspects of a leader’s job performance, such
as OCB. To the extent that integrity involves aspects of interpersonal cooperation, justice and compassion associated with
principled moral reasoning (Turner et al., 2002), it is likely that managers’ moral reasoning levels will be positively related to
ratings of their extra-role performance or OCB. Thus:
Hypothesis 3. Managers with higher moral reasoning will be associated with higher extra-role job performance (OCB) ratings.
1.2. Management level, charismatic leadership and performance
Upper echelon theory (Hambrick & Mason, 1984) posits that leaders in upper levels of management engage in more visionary
and strategic forms of leadership than managers in lower organizational ranks. Zaccaro (1996) pointed out that executives and
senior managers interact more frequently with external organizational stakeholders and engage in more network development
and consensus building than lower-level leaders. These activities reflect forms of social influence that characterize charismatic
leadership in corporate contexts such as the present study (Sosik et al., 2002). In their review of the literature on charisma and
management level, Waldman, Bass, and Yammarino (1990) presented both theoretical and empirical evidence indicating that
charismatic leadership is more prevalent at higher management levels. They argued that upper-echelon managers are more likely
to be change agents who create and articulate visions of change, and role model organizational values and behavioral expectations.
In contrast, lower-echelon corporate managers typically implement the policies created by their superiors. Moreover, role

modeling and communicating about ethics and values are key behaviors of charismatic leaders (Bass, 2008) and part of how
executives develop a reputation for being an ethical leader (Trevino et al., 2000). Thus:
Hypothesis 4. Upper-echelon managers will display more charismatic leadership than lower-echelon managers.
Many leadership styles, including transformational, charismatic, authentic, ethical, servant, spiritual and self-sacrificial leadership,
emphasize the importance of the ethics of helping others. In each of these styles, leaders serve as rolemodels for followers to promote
theirmoral and professional development and to create a positive culture (Bass, 2008). In their qualitative study of senior executives
and chief ethics officers, Trevino and her colleagues (2000) identified behaviors that show concern for other people as an important
aspect of being a moral person who is capable of demonstrating ethical leadership. Beahr (1992) conducted a job analysis of 1358
managers and found that two of themajor executive functions are developing theworkforce and dealingwith outside contacts and the
community. Contemporary business models call for a balance between maximizing profit, developing associates, and being socially
responsible (Savitz & Weber, 2006), and these functions are likely to involve a great deal of OCB on the part of senior managers.
Managerial rank typically correlates with educational level (Bass, 2008), which Smith et al. (1983) found to be positively correlated
with OCB (altruism).
Managing in the upper echelons also requires a great deal of boundary spanning and task interdependence (Hambrick & Mason,
1984),which “fosters social normsof cooperation, helping and sensitivity to others’ needs andmakes salient a collective sense of social
responsibility” (Smith et al., 1983, p. 655). Consistentwith these ideas, two studies reported a positive relationship betweenmanagers’
helping behavior and job level (Deckop, Cirka, & Andersson, 2003; Sosik et al., 2009), while one study involving leisure service
administrators found opposite results (Anderson & Bedini, 2002). Given that the current study’s sample includes corporatemanagers
and that the majority of theoretical and empirical evidence supports a positive relationship, we propose:
Hypothesis 5. Upper-echelon managers will be associated with higher levels of extra-role job performance (OCB) than lowerechelon
managers.

  1. Method
    2.1. Procedure and participants
    Data were collected from employees of 218 U.S.-based organizations providing graduate-level adult business students to a
    large public university. These corporations represented the technology (38%), manufacturing (12%), and services (50%) sectors.
    We invited 475 managers from these organizations to participate in the study. Ninety-eight managers were not included in the
    final sample because they chose not to participate or did not complete the moral reasoning measure. Thus, the final sample
    included 377 managers, and their 1731 subordinates, and 377 superiors who also participated in this study. Participants and nonparticipants
    did not differ significantly on characteristics such as age, gender and management level. An analysis of early vs. late
    responders on all study variables did not reveal any meaningful differences between these groups, and therefore we did not deem
    non-response bias to be a problem with our data (Armstrong & Overton, 1977).
    Managers were part-time graduate business students employed full-time in positions ranging from first-line managers to
    executives. Of the 377 managers, 216 were lower-echelon (i.e., first-line or middle) managers and 161 were upper-echelon (i.e.,
    director or executive) managers. The average age of managers was approximately 44 years and ranged from 25 to 69. Managers
    worked for an average of four years in their current position and had a range of job tenure of one to 20 years. Company tenure for these
    managers ranged fromone to 35 years,with ten years representing the average company tenure. Sixty percent of the managerswere
    male and 80% were White. There were no significant differences in study variables across organizations, industry or race.
    The average age of subordinates was approximately 36 and ranged from 17 to 76. Subordinates worked for an average of three
    years (SD=3.7) for their manager and 78% were in either non-supervisory or first line management positions. Fifty percent of the
    subordinates were male and 81% were White. The average age of superiors was approximately 49 and ranged from 34 to 65.
    Superiors worked for an average of three years (SD=3.2) for their manager and 72% were in middle or higher levels of
    management. Seventy percent of the superiors were male and 90% were White.
    We introduced the managers to the study in a leadership class. On the first day of class, managers completed the short form of
    the Defining Issues Test (DIT; Rest, 1990) to measure their moral reasoning capacity. We then instructed the managers that
    additional data were to be collected through an Internet-based survey. Managers completed the online survey, measuring
    charismatic leadership, OCB, and job performance, outside of class and then submitted it via the Internet two weeks later. At the
    same time, we asked the managers to identify 10 subordinates and their primary supervisor, who were considered as potential
    raters. An algorithm built into the online survey randomly selected five subordinates as invitees to participate in the study. The
    system sent an email to these subordinates asking them to help the manager complete his/her course requirements by
    participating in a 360-degree feedback evaluation, and completing an online survey assessing their manager’s charismatic
    leadership, OCB and job performance by the sixth week of class. To maximize response rates, the Web site emailed a series of
    reminder notices to subordinates until they submitted their online surveys. Three to five (M=4) subordinates rated each of the
    216 lower-level and 161 upper-level managers. The system sent similar emails and reminders to the managers’ supervisor to
    collect ratings of the manager’s charismatic leadership, OCB and job performance. Data for subordinate and superior ratings were

collected two and four weeks later than manager self-ratings, respectively. All participants completed their surveys confidentially
and returned them directly via the research Web site for processing.
2.2. Measures
2.2.1. Moral reasoning
The moral reasoning of managers was measured with the Defining Issues Test (DIT; Rest, 1990), which represents the most
widely used measure of moral reasoning (McCauley et al., 2006). The shortened form of the reliable and valid DIT test measuring
moral reasoning was used (Bass, 2008; Turner et al., 2002). According to Rest (1990), the DIT test-retest reliabilities are generally
in the high .70 to .80 range and Cronbach’s alpha are in the high .70 range. While some researchers have criticized the DIT on the
grounds that it lacks brevity, clarity and validity across cultures (cf. Rest et al., 1999), Peterson and Seligman (2004) and McCauley
et al. (2006) concluded that the psychometric properties of the DIT are well established and appropriate for assessing what
constitutes moral reasoning.
Managers were presented with three moral dilemmas and asked to make morally exigent decisions on each of the three
scenarios. Following a procedure outlined in Turner et al. (2002, p. 307), we used the principled score (P score, reflecting the
degree to which a respondent reasons in a morally developed manner in terms of rights and justice) from the DIT to identify
superiors’ preconventional (P score≤27), conventional (28≤P score≤41), and postconventional (P score≥42) levels of
Kohlberg’s (1976) stages of moral development. We followed Rest (1990) and Turner et al. (2002)’s approach on scoring the DIT
and determining the respondents’ level of moral reasoning.
2.2.2. Charismatic leadership
Self-, subordinate-, and superior-ratings of managers’ charismatic leadership behaviors were measured with the inspirational
motivation, idealized influence-behavior, and idealized influence-attributes subscales of the Multifactor Leadership Questionnaire
(MLQ-5X; Bass & Avolio, 1997). The idealized influence and inspirational motivation subscales contain items that reflect the role
modeling and communicating about ethics and values behaviors identified by Trevino et al. (2000) as aspects of ethical leadership.
The 12 items in these subscales were assessed on a 5-point frequency scale ranging from 0=not at all to 4=frequently, if not
always. These items “measure several leader behaviors and attributes described in the theory of Shamir, House, and Arthur (1993)
of charismatic leadership” (Sosik & Dinger, 2007, p. 141) and were averaged to form a composite measure (Self α=.85,
Subordinate α=.90, and Superior α=.88) of charismatic leadership. Despite its criticisms for failing to empirically demonstrate
consistently the factor structure of transformational leadership (see Yukl, 2010 for an overview of these issues), prior research
(e.g., Antonakis et al., 2003; Avolio, Bass, & Jung, 1999) has shown the MLQ-5X to be a psychometrically sound instrument in terms
of measuring the construct of charismatic leadership.
2.2.3. In-role and extra-role managerial performance
Williams’ (1988) 3-item measure was used in the present study to assess in-role performance of managers based on self,
subordinate, and superior-ratings. Williams’ scale has demonstrated acceptable psychometric properties in prior research (e.g.,
Hui, Law, & Chen, 1999; Sosik, 2005; Williams & Anderson, 1991). A sample item reads “Completes the duties specified in his/her
job description in a timely, accurate, and complete manner” (Self α=.76, Subordinate α=.83, and Superior α=.87).
Respondents rated each item on a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree).
Similarly, using self-, subordinate-, and superior-ratings, we measured extra-role (OCB) performance of managers using four
subscales developed and validated by Hui et al. (1999): altruism toward associates (3 items, e.g., “Is willing to assist new associates
in adjusting to the work environment”), conscientiousness (3 items, e.g., “Complies with company rules and procedures even
when nobody is watching and no evidence can be traced”), identification with company (3 items, e.g., “Actively attends company
meetings”), and interpersonal harmony (3 items, e.g., “Uses smooth tactics to seek personal influence and gain with harmful
effects on interpersonal harmony in the organization” — reverse coded). We averaged the items of the four subscales to form a
composite measure of extra-role (OCB) performance for the manager (Self α=.70, Subordinate α=.82, and Superior α=.80).
Our approach is consistent with results of a meta-analysis conducted by Hoffman, Blair, Meriac, and Woehr (2007) who concluded
that “current operationalizations of OCB are best viewed as indicators of a general OCB factor…, there is likely little to be gained
through the use of separate dimensional measures as opposed to an overall composite measure” (p. 562).
2.2.4. Confirmatory factor, measurement invariance and aggregation analyses
Given the prior theoretical and empirical evidence for the measures used in the present study, we conducted a series of
confirmatory factor analyses (CFAs) to examine the validity of all measures before testing the hypotheses. We tested a higher
order model of three factors: charismatic leadership (i.e., idealized influence-behavior, idealized influence-attributes, and
inspirational motivation), in-role performance, and extra-role performance (i.e., altruism, conscientiousness, identification, and
interpersonal harmony). This higher order, three factor model fit the data well across self (CFI=.99, TLI=.98, RMSEA=.06),
subordinate (CFI=.98, TLI=.97, RMSEA=.08), and superior (CFI=.98, TLI=.97, RMSEA=.08) ratings. The CFA with the onefactor
model, where all items were loaded on a single factor, revealed a poor fit with the data across self (CFI=.59, TLI=.62,
RMSEA=.10), subordinate (CFI=.69, TLI=.67, RMSEA=.13), and superior (CFI=.69, TLI=.66, RMSEA=.11) ratings.
We also examined whether the factorial structure of this measurement model was invariant across the three rating sources,
using a multi-group measurement invariance analysis (Vandenberg & Lance, 2000). A three-group baseline model was first

estimated, in which factor patterns were equal and all parameters were set free across self, subordinate, and superior groups. This
baseline model was then compared to another three-group model where factor patterns were equal but factor loadings were
constrained to be equal across the three groups. To demonstrate the invariance of factorial measurement structure, the baseline
model should not be different from the constrained model. Fit indices for both models were identical and indicated good model fit
(CFI=.98, TLI=.98, RMSEA=.05), and the chi-square difference between the models was not significant (Δχ2(df)=48.53 (38),
pN.05). Taken together, the results of CFAs and multi-group measurement tests indicate that the measures for each construct can
be differentiated and that the factorial structure with this distinction is invariant across three rating sources.
Aggregation of scales for subordinate-ratings of charismatic leadership, in-role job performance, and extra-role job
performance were based on results of ANOVA and rwg analyses. Aggregation is appropriate when within-group variances are
minimal or insignificant while variances across groups are significantly different (Dansereau, Alutto, & Yammarino, 1984), and
when agreement or consensus exists among raters portrayed by rwg index scores ≥.7 (James, Demaree, &Wolfe, 1984). Results of
ANOVA were for charismatic leadership, F (376, 1729)=2.40, pb.001, η2=.40; for in-role job performance, F (376, 1729)=2.10,
pb.001, η2=.35; and for extra-role job performance, F (376, 1729)=1.90, pb.001, η2=.29. The following percentages of groups
had rwg scores of greater than .7: for charismatic leadership, 93%; for in-role job performance, 91%; and for extra-role job
performance, 90%. Thus, the data analyses described below are based on a sample size of 377 focal leaders.
2.2.5. Control variables
Prior research suggests that age, gender, and socially desirable responding can influence moral reasoning (Rest, 1994) and
leadership processes and outcomes (Bass, 2008). To control for these potential effects, we entered leader’s age, gender (1=male
and 2=female), and socially desirable responding in our data analysis as covariates. Socially desirable responding was measured
using Reynolds (1982) 13-item short form of the Crowne and Marlowe (1960) Social Desirability Scale (α=.80), which has
demonstrated sound psychometric properties in prior research (e.g., Paulhus, 1988; Sosik, 2005; Sosik et al., 2002). A sample item
reads “I am always courteous, even to people who are disagreeable.” Managers rated each item on a 7-point scale ranging from 1
(strongly disagree) to 7 (strongly agree).

  1. Results
    Table 1 presents the descriptive statistics and product-moment correlations for this study’s measures. Table 2 presents means
    and standard deviations for the study’s dependent variables for each rating source (self, aggregated subordinate, and superior) for
    managers in each moral reasoning group for both lower and upper echelon manager groups.
    The hypotheses were tested using a series of 3×2 analysis of covariance (ANCOVA) with charismatic leadership, in-role job
    performance, and extra-role job performance as dependent variables, moral reasoning group (preconventional vs. conventional
    vs. postconventional) and management level (lower echelon vs. upper echelon) as independent variables, and leader’s age,
    gender, and socially desirable responding as covariates. Details of the main effects summarized in Table 2 are provided below.
    There were no significant interaction effects.

3.1.1. Leaders’ self-ratings
After controlling for the covariates (significant effects for socially desirable responding, F (1,368)=66.02, pb.0001, η2=.15;
age and gender nonsignificant, pN.05), ANCOVA revealed a significant main effect for moral reasoning group (F (2,368)=8.56,
pb.001, η2=.04) and management level (F (1,368)=16.59, pb.001, η2=.04) on leader’s self-ratings of charismatic leadership.
Results of planned comparisons indicated that charismatic leadership scores for the preconventional group (M=3.01, SD=.51)
were significantly greater (t (374)=2.09, pb.04) than those for the conventional group (M=2.87, SD=.49), which were greater
(t (374)=2.09, pb.04) than those for the postconventional group (M=2.74, SD=.52). Charismatic leadership self-ratings for
upper echelon management (M=2.99, SD=.50) were significantly greater than those for lower echelon management (M=2.78,
SD=.51).
3.1.2. Subordinates’ ratings
After controlling for the covariates (all of which had nonsignificant effects, pN.05), ANCOVA revealed a nonsignificant effect for
moral reasoning group (F (2,368)=2.37, p=.09) and a significant main effect for management level (F (1,368)=9.87, pb.01,
η2=.03) on subordinates’ ratings of charismatic leadership. Subordinates’ ratings of charismatic leadership for upper echelon
management (M=2.93, SD=.43) were significantly greater than those for lower echelon management (M=2.81, SD=.43).
3.1.3. Superiors’ ratings
After controlling for the covariates (all of which had nonsignificant effects), ANCOVA revealed a significant main effect for
management level (F (1,368)=15.21, pb.0001, η2=.04) and a nonsignificant effect for moral reasoning group (F (2,368)=1.99,
p=.14) on superiors’ ratings of charismatic leadership. Superiors’ ratings of charismatic leadership for upper level management
(M=2.95, SD=.55) were significantly greater than those for lower level management (M=2.73, SD=.65).
Taken together, results from each of the rating sources do not support Hypothesis 1 stating that managers with higher moral
reasoning would display greater charismatic leadership. However, results from each rating source fully support Hypothesis 4
stating that upper echelon managers would display more charismatic leadership than lower echelon managers.
3.2. In-role job performance
3.2.1. Leaders’ self-ratings
After controlling for the covariates (significant effects for age, F (1,368)=4.66, pb.03, η2=.03, socially desirable responding,
F (1,368)=20.69, pb.0001, η2=.05, and gender, F (1,368)=7.19, pb.01, η2=.02), ANCOVA revealed nonsignificant effects for
moral reasoning group (F (2,368)=.03, p=.97) and management level (F (1,368)=.23, p=.63) on leader’s self-ratings of
performance.
3.2.2. Subordinates’ ratings
After controlling for the covariates (significant effects for age, F (1,368)=14.94, pb.001, η2=.04; others nonsignificant, pN.05),
ANCOVA revealed nonsignificant effects for moral reasoning group (F (2,368)=.74, p=.48) and management level (F (1,368)=.01,
p=.97) on subordinates’ ratings of performance.
3.2.3. Superiors’ ratings
After controlling for the covariates (significant effects for age, F (1,368)=4.11, pb.05, η2=.01, others nonsignificant, pN.05),
ANCOVA revealed a significant main effect for moral reasoning group (F (2,368)=3.35, pb.04, η2=.02) and a nonsignificant effect
for management level (F (1,368)=1.15, p=.29) on superiors’ ratings of performance. Results of planned comparisons indicated
that superiors’ ratings of performance for the postconventional group (M=6.33, SD=.74) were significantly greater (t (374)=1.
96, pb.05) than those for the conventional group (M=6.08, SD=1.00) and the preconventional group (M=6.14, SD=.81;
t (374)=2. 28, pb.05).
Taken together, these results partially support Hypothesis 2 stating that managers with higher moral reasoning would
outperform managers with low moral reasoning, based on the superior rating source, but not the leader and subordinate sources.
3.3. Extra-role (OCB) job performance
3.3.1. Leaders’ self-ratings
After controlling for the covariates (significant effects for socially desirable responding, F (1,368)=53.91, pb.0001, η2=.13,
and gender, F (1,368)=7.63, pb.01, η2=.02; age nonsignificant, pN.05), ANCOVA revealed significant main effects for moral
reasoning group (F (2,368)=3.80, pb.02, η2=.02) and management level (F (1,368)=4.72, pb.03, η2=.01) on leaders’ selfratings
of OCB. Results of planned comparisons indicated that OCB scores for the preconventional (M=6.20, SD=.47) and
conventional (M=6.21, SD=.58) groups were significantly greater (t (374)=2. 33, pb.01 and t (374)=2. 49, pb.01,
respectively) than those for the postconventional group (M=6.05, SD=.53). Leaders’ self-ratings of OCB for upper echelon
management (M=6.22, SD=.45) were significantly greater than those for lower echelon management (M=6.10, SD=.58).

3.3.2. Subordinates’ ratings
After controlling for the covariates (significant effects for gender, F (1,368)=3.80, pb.05, η2=.01, and socially desirable
responding, F (1,368)=6.59, pb.01, η2=.02; age nonsignificant, pN.05), ANCOVA revealed nonsignificant effects for moral
reasoning group (F (2,368)=.34, p=.71) and management level (F (1,368)=.85, p=.36) on subordinates’ ratings of OCB.
3.3.3. Superiors’ ratings
After controlling for the covariates (significant effects for gender, F (1,368)=3.90, pb.05. η2=.01, socially desirable
responding, F (1,368)=5.51, pb.02, η2=.02; age nonsignificant, pN.05), ANCOVA revealed a nonsignificant effect for moral
reasoning group (F (2,368)=1.47, p=.23) and a significant main effect for management level (F (1,368)=9.17, pb.01, η2=.03)
on superiors’ ratings of OCB. Superiors’ ratings of OCB for upper echelon management (M=6.23, SD=.59) were significantly
greater than those for lower echelon management (M=6.08, SD=.63).
Taken together, results from each of the rating sources do not support Hypothesis 3 stating that managers with higher moral
reasoning would display more OCB than managers with lower moral reasoning. However, results from leaders’ self-ratings and
superiors’ ratings, but not subordinates’ ratings, provide partial support for Hypothesis 5 stating that upper-echelon managers
would display more OCB than lower-echelon managers.
3.4. Post-hoc analyses
Given these mixed results in testing the hypotheses and ratings of the leadership and performance variables provided by three
independent sources, we summarized each hypothesis and its results in Table 3. The results shown in Table 3 indicate that leaders,
subordinates and superiors viewed leadership and performance somewhat differently, supporting the idea that focal leaders at
various levels of moral reasoning and management may differ in their ability to provide accurate self-ratings compared to ratings
provided by other sources.
It is well-established that self-ratings of leadership and performance are generally less reliable than ratings provided by others
(Atwater & Yammarino, 1997; Harris & Schaubroeck, 1988), but may offer developmental insights if they are closely aligned with
ratings provided by others (Hogan et al., 1994; van Knippenberg et al., 2005). What remains unclear is whether some of the
unexpected findings in the present study are a function of the leader’smoral reasoning andmanagement levels. According to Kohlberg
(1976), moral reasoning levels progress from a self-focus (preconventional), to a comparison with the views and expectations of
society (conventional), and ultimately to an integrated view of self and others within a social community (postconventional);
however, not all individuals progress to higher levels. Thus, focal leaders who are conventional or postconventional moral reasoners
are likely to provide ratings that are more consistent with ratings provided by their subordinates and superiors. Similarly, upperechelon
leaders aremore cognitively complex and able (Ones & Dilchert, 2009), have access to more information (Hambrick &Mason,
1984), and receive more feedback on their behavior and performance (Atwater & Yammarino, 1997) than lower-echelon leaders.
These resources are likely to allowformorally developed upper-echelon leaders’ self-ratings that aremore closely alignedwith ratings
provided by their subordinates and superiors.
To explore these ideas and attempt to explain the unexpected findings summarized in Table 3, we used Within and between
analysis (WABA; Dansereau et al., 1984) to assess the self-other rating agreement between leaders’, subordinates’ and superiors’
ratings of charismatic leadership, in-role job performance, and extra-role job performance. Following guidelines outlined in
Atwater and Yammarino (1997, pp. 154–155), leader–subordinate aggregated dyads and leader–superior dyads were assessed at
the dyad level using WABA I. This procedure indicates whether the variable of interest varies primarily between, within, or both
between and within dyads. The source of variation is determined based on within- and between-dyad etas, which are tested for
statistical significance with F-tests and practical significance with E-tests. WABA I results where variation is primarily between selfother
dyads indicates agreement within each dyad (whole view). Variation primarily within self-other dyads indicates “patterned
disagreement” within each dyad (parts view). Variation both within and between self-other dyads indicates that self and other
ratings are unconnected (equivocal view) (Atwater & Yammarino, 1997).

Tables 4 and 5 present leader–subordinate and leader–superior dyad-level WABA I results by moral reasoning group for lowerechelon
and upper-echelon management level, respectively. With respect to charismatic leadership ratings displayed in Table 4,
lower-echelon leader–subordinate and leader–superior dyads in each moral reasoning group produced self and other scores that
are independent of each other. In contrast, results displayed in Table 5 indicate that upper-echelon leader–superior dyads in the
conventional moral reasoning group produced congruent self and other charismatic leadership scores. Similarly, upper-echelon
leader–subordinate and leader–superior dyads in the post-conventional moral reasoning group produced congruent self and other
charismatic leadership scores. These results suggest that conventional and postconventional moral reasoning focal leaders in
upper management positions generally agree with their raters that their charismatic leadership ratings exceed those in lower
management, but their ratings are lower than those leaders who are preconventional moral reasoners.
With respect to in-role job performance ratings displayed in Table 4, lower-echelon leader–subordinate and leader–superior
dyads in only the post-conventional moral reasoning group produced congruent self and other scores. In contrast, results
displayed in Table 5 indicate that upper-echelon leader–superior dyads in the preconventional and conventional moral reasoning
groups produced congruent self and other performance ratings. These results suggest that postconventional moral reasoning focal
leaders in lower management positions generally agree with their raters that their performance exceeds the performance of those
focal leaders who are preconventional and conventional moral reasoners. This agreement disappears in upper management
positions where leaders who are precoventional or conventional reasoners agree with their superiors that their performance level
is lower than that of leaders who are postconventional moral reasoners.
With respect to extra-role OCB ratings displayed in Table 4, lower-echelon leader–superior dyads in only the conventional
moral reasoning group produced congruent self and other scores. Results displayed in Table 5 indicate that upper-echelon leader–
superior dyads in preconventional and conventional moral reasoning groups produced congruent self and other OCB scores. These
results suggest that focal leaders who are conventional moral reasoners agree with their superiors that those in upper
management positions display more OCB than those in lower management positions.

  1. Discussion
    This study examined how corporate managers’ cognitive moral reasoning and management level are associated with self-,
    subordinate- and superior-ratings of charismatic leadership and in-role and extra-role job performance at both supervisory and
    executive management levels. It yielded several theoretically interesting and practically applicable findings implying that senior
    leaders of U.S. corporations may be more charismatic but potentially less ethical.
    First, study results confirmed that a manager’s level of moral reasoning is associated with his or her ratings of charismatic
    leadership. However, results diverge from theoretical considerations (Kuhnert & Lewis, 1987) and empirical results demonstrated
    by Turner et al. (2002), in which higher levels of moral reasoning are linked to transformational leadership ratings. Specifically,
    managers who were preconventional moral reasoners reported displaying charismatic leadership more frequently than those who
    reasoned at the conventional and postconventional levels.
    One possible explanation for this finding is that managers who are preconventional moral reasoners are more narcissistic than
    managers who are conventional or postconventional moral reasoners. Preconventional moral reasoning is based on a selfinterested
    focus and seeking rewards (Kohlberg, 1976), which bears a resemblance to narcissism. Narcissistic leaders may be selfdeceptive
    (Maccoby, 2004), and are characterized by self-absorption, entitlement, egocentricity, and needing rewards such as
    power, privileges, status, and admiration (Rosenthal & Pittinsky, 2006). These characteristics of narcissism seem to parallel the
    rigid rule-based and self-centered orientation of the preconventional stage of moral reasoning. This interpretation appears to be
    consistent with the results of our post-hoc analyses suggesting that managers who are preconventional moral reasoners may have
    less self-awareness and more distorted self-conceptions regarding their charismatic leadership than managers at higher stages of
    moral development.

While unexpected, this result is consistent with Atwater, Dionne, Avolio, Camobreco, and Lau (1999), who found that higher
levels of moral reasoning of U.S. military cadets were associated with the use of contingent punishment, which in turn was
positively related to leader effectiveness. In military settings, being moral may mean being fair by using contingent rewards and
punishment, which prior review results show to be highly correlated with transformational/charismatic leadership (Antonakis et
al., 2003), and is consistent with preconventional moral reasoning. Some traditional corporate settings, such as those in the
present study, are similar to military settings (e.g., hierarchical, rule-based, and compliance-oriented). From this perspective and
discussions of ethical leadership (Trevino, 1986), situational consideration appears critical to what is considered to be “moral” and
“charismatic.” Results of the present study suggest that senior corporate leaders are more charismatic but potentially less ethical.
This implication is consistent with Trevino et al.’s (2000) finding that “many executives failed to recognize the importance of
others’ perceptions and of developing a reputation for ethical leadership” (p. 129). The corporate setting and manager sample of
both Trevino et al.’s (2000) study and the present study seem quite different in terms of ethical standards and values from Turner
et al.’s (2002) study which included a university and hospital ward sample. Thus, leadership context appears to be quite important
in interpreting research on forms of ethical leadership.
Another possible explanation for the divergent results involves the theoretical composition of the leadership constructs and its
operationalization in the two studies. Turner et al. (2002) included the MLQ subscale factors of charisma-inspiration,
individualized consideration, intellectual stimulation, and contingent reward in their measure of transformational leadership.
Contingent reward is a transactional leadership measure that is not typically included in conceptualizations of transformational
leadership (Antonakis et al., 2003; Bass, 2008). Including contingent reward in the measure of transformational leadership
introduces a key element of transactional leadership into the construct, and thus may have confounded the results reported in
Turner et al. (2002). The measure of charismatic leadership used in the present study does not include a contingent reward
subfactor, nor does it include the individualized consideration, intellectual stimulation subfactors. Moreover, prior research (e.g.,
Brown & Trevino, 2006; Trevino et al., 2000) posits that ethical leadership includes a transactional component (i.e., rewards and
discipline). While the present study did not measure ethical leadership per se, future research can build upon our work by
measuring both the transformational and transactional components of ethical leadership.
According to Bass (2008), transformational leadership includes two additional components not associated with charismatic
leadership — intellectual stimulation and individually considerate behavior. Intellectual stimulation involves questioning
assumptions about the status quo and examining problems from different perspectives. Individualized consideration recognizes,
appreciates and leverages differences in followers’ knowledge, skills and abilities (Bass & Avolio, 1997). These behaviors are
consistent with postconventional moral reasoning which takes a holistic and other-oriented approach to making decisions
(Kohlberg, 1976). Because these behaviors are not part of charismatic leadership and were not measured in the present study, it is
possible that managers’ self-centered images and messages that elicit awe and respect were more important in forming leadership
(self) perceptions (House & Howell, 1992; Sosik et al., 2002). The positive relationship discovered in the present study between
the preconventional moral reasoning and charismatic leadership ratings supports this explanation.
A second major finding of this study was that the level of moral reasoning possessed by managers was positively associated
with their performance. Specifically, superiors’ ratings of managers’ in-role job performance were significantly greater for
managers reasoning at the postconventional level than managers reasoning at the preconventional or conventional levels. This
finding provides empirical support for theories of authentic (Avolio & Gardner, 2005) and ethical (Brown & Trevino, 2006)
leadership proposing that ethical leader behavior results in positive outcomes. A practical implication of this finding is that
superiors appear to reward managers with high in-role job performance ratings for reasoning ethically. If high performance
translates into career advancement opportunity, managers who reason ethically may help themselves while they help their
organization. Organizations that highlight such findings and reward managers for demonstrating ethical decision-making may
create a culture of ethical leadership and high performance at all managerial ranks (Trevino et al., 2000).
Third, we confirmed that charismatic leadership tends to be more prevalent in the upper echelons of corporations than in the
lower ranks (Waldman et al., 1990). Specifically, charismatic leadership ratings from each of the three sources (self, subordinate,
and superior) were significantly higher for upper-level managers than lower-level managers. Despite our multi-source data
collection and its tendency to produce divergent ratings (Atwater & Yammarino, 1997), results of this study could be cited as
strong evidence that charismatic leadership is especially prevalent in the upper echelons of firms where promoting organizational
change initiatives and creating policy within a long time horizon are essential.
Yet, the prevalence of charismatic leadership in the upper echelons of corporations may not be comforting since managers in
our sample who viewed themselves as charismatic also possessed the lowest level of cognitive moral reasoning, which can lead to
unethical leadership (Trevino, 1986). Corporate boards run the risk of dealing with narcissistic, self-centered executives, who may
handicap their organization in today’s competitive environment and decrease returns to the very shareholders the board wishes to
protect. This implies that organizations should consider implementing accountability systems to manage executives’ ethics and
delivering ethical training programs to groom managers at lower levels to move up to the next rung of management.
Accountability systems can include Audit Committees that monitor executive ethics, policies of stewardship to shareholders, and
fiscal responsibility control systems. Chief Ethics Officers responsible for creating an ethical organizational climate can also be
appointed to help managers at all organizational ranks value a reputation for ethical leadership (Trevino et al., 2000).
Indeed, possessing a charismatic self-perception may be a prerequisite for individuals who aspire to the ranks of senior
management. Alternatively, such self-perceptionsmay develop as a result of leading in the upper-echelons of corporations. However,
in some organizations and industries such as banking and finance (e.g., Countrywide), energy (e.g., Enron), healthcare (e.g.,
HealthSouth) and telecommunications (e.g., WorldCom), high-level positions may elicit charismatic leadership that is unethical or personalized (House & Howell, 1992). Therefore, corporate training on charismatic leadership should includemodules on ethics and
positive role modeling in addition to framing messages, inspirational communication, and public speaking in order to temper any
negative situational influences that may promote the dark side of charismatic leadership. Such training also may help to create an
environment to filter and weed out those unable to meet the challenging demands of top corporate positions.
To better understand the apparent trade-off between charismatic leadership self-ratings and ethical reasoning in our sample of
corporate managers, we also directly compared self- and other-ratings of charismatic leadership and in-role and extra-role
performance to assess the degree of agreement acrossmanagement levels and moral reasoning groups in our post-hoc analyses. On
the basis of our assumption that rating agreement is in part influenced by information access, cognitive complexity, and reasoning that
integrates multiple viewpoints (Atwater & Yammarino, 1997), we had expected that morally developed upper-echelon leaders’ selfratings
would be more closely aligned with ratings provided by their subordinates and superiors. This was true for ratings of
charismatic leadership at the uppermanagement level formanagers in the postconventional moral reasoning group, but at the lower
management level in all moral reasoning groups, the ratings were independent or not in agreement.
To us, it still remains an open question why self-other rating agreement emerges on charismatic leadership in postcoventional
moral reasoners in upper management levels. Finding a conclusive answer to this question would be important because it would
help place conditional boundaries on prior research indicating that as management level increases, so does the gap between selfand
others’ ratings (e.g., Gentry et al., 2007; Sala & Dwight, 2002). It would suggest that cognitive moral reasoning may be an
individual characteristic that influences the self-other rating agreement process, and therefore should be included in theoretical
models of self-other rating agreement and corporate training of self-awareness (Atwater & Yammarino, 1997). One explanation
for this result is that upper-level managers typically possess cognitive complexity required for developing a mental map of the
organization and its environment. Postconventional moral reasoning requires integrating multiple perspectives into a holistic
solution by acknowledging that individuals hold different values, and this ability is a hallmark of cognitive complexity (Rest, 1994)
associated with upper echelon managers (Ones & Dilchert, 2009). Subordinates pay special attention to managers’ credibility and
trust that can stem from high levels of moral reasoning coupled with charismatic leadership (Hogan et al., 1994). This line of
reasoning suggests that being aware of others’ perspectives and taking them into account may be prerequisites for being selfaware
and may correspondingly yield self-other rating agreement (Atwater & Yammarino, 1997).
Another explanation is that levels of analysis can change over time (Dansereau, Yammarino, & Kohles, 1999). Unconnected
self-other ratings at the individual level can evolve into self-other rating agreement at the dyadic level as social interaction events
between members are accumulated over time. Leader–superior and leader–subordinate dyads at upper levels of management can
be characterized by a longer tenure of the relationships than those at lower levels. Upper echelon managers have gained more
interpersonal experience with their tenure, and therefore their self-ratings may become more attuned to others’ perspectives on
them. It is also possible that subordinates and superiors may be more lenient in their ratings of upper echelon managers for
political reasons, thereby providing more agreement between self- and other-ratings. Future longitudinal research designs will be
necessary to explore these issues.
4.1. Limitations
The abovementioned findings should be interpreted within the context of the following study limitations. Sample
characteristics present a first set of limitations. The focal leaders’ initial selection of 10 subordinates as potential respondents
could have introduced bias because they may have chosen subordinates who view them favorably. In addition, this study’s use of a
convenience as opposed to a random sample of managers presents a limitation. Convenience samples make it difficult to
determine the extent to which a sample is representative. This concern is somewhat allayed given that study participants were
part-time graduate adult business students employed full-time in positions covering a wide range of ages, backgrounds, industries,
organizations and managerial experience. These sample characteristics add to the generalizability of the research, in that it is not
the typical “university student” sample. Given the age and experience of the participants, many had worked for several entities
across the U.S. and other countries as well during their careers. Thus, the traditional geographic limitations may not apply to the
sample and the study results may be more indicative of the general population.
Such a sample was judged preferable to using employees within the same organization due to the potential for the data
reflecting shared participant pool, organizationally specific values, or leadership practices that may not be representative of the
general population. Nevertheless, we acknowledge limitations of generalizations from convenience sample data. For example, we
advise caution in making inappropriate generalizations of our findings to non-corporate settings or industries not represented in
our sample, such as educational institutions, faith communities, or governmental agencies.
Another limitation involves the association found between managers’ moral reasoning level and self-ratings of charismatic
leadership, which may be partially a function of common source bias. However, concern for this issue may be attenuated by the
study’s data collection that was separated in time, controlling for social desirability bias, and nature of the DIT as a scored test as
opposed to a survey (Rest, 1990). Given the short time-frame of the study’s longitudinal data collection and the small effect sizes
associated with our results, the causal relationships reported here must be viewed as somewhat arbitrary. Nevertheless, the
organizational level of the managers in this study and their moral reasoning level do have temporal precedence over the ratings of
leadership behavior and performance collected in this study based on theoretical models of ethical leadership (cf. Brown &
Trevino, 2006; Trevino, 1986).
Followers’ level of moral reasoning was not examined in this study, which may have an impact on ascribing charisma to leaders
based on needs and self-conceptions (e.g., Shamir et al., 1993; Sosik et al., 2002). Also, the corporate setting from which we drew

our sample may have produced somewhat of a restriction in the range of moral reasoning levels. Bass’ (2008) review of the moral
reasoning literature indicates that most managers, and adults in general, are likely to be conventional moral reasoners. Bass notes
that the average P score for adults is 40, and for graduate business students, 42.8. The average P score for managers in this study is
36.3, suggesting that our sample may limit the generalizability of study results to other settings.
In addition to the range restriction of DIT P score in our study sample, potential moderators were not considered to better
understand how managers’ moral reasoning levels translate into their leadership behaviors and performance. Because conventional
moral reasoning managers tend to be susceptible to situational influence, such as organizational culture (Trevino, 1986) and the
distance between leaders and followers (Brown & Trevino, 2006), results of the present study must be interpretedwith caution, given
the omission of situational characteristics.
4.2. Future research and conclusion
Despite these limitations, this study offers a foundation for future research in several areas. First, the unexpected finding that selfratings
of charismatic leadershipwere highest for preconventionalmoral reasoningmanagers calls into question Turner et al.’s (2002)
speculation that “manifesting postconventional moral reasoning might predispose individuals to choose to use transformational
leadership” (p. 309). Our study suggests that using the transformational leadership model to understandmoral development’s role in
gaining attributions of charismamay be insufficient, andwe should nowlook to othermodels, such as ethical (Brown & Trevino, 2006)
or spiritual (Fry, 2003) leadership, to guide for future research. For example, ethical leadership models and measures (Brown et al.,
2005) which consider elements of being a moral person and moralmanager may provide opportunities for greater understanding of
ethical decision-making processes. Future research should also assess personality factors such as narcissism, and situational factors
such as ethical climate, thatmight predispose both upper- and lower-echelonmanagers to reason unethically and display personalized
charismatic leadership behaviors.
If charismatic executives do lean toward preconventional levels of moral reasoning, companies should consider emphasizing
ethical conduct and raise the moral thought processes of their managers through training, education and research-based
leadership development programs (Day, Harrison, & Halpin, 2009). Can leadership development programs be designed to elevate
managers’ moral reasoning capacity over time? Trevino et al. (2000) suggest that executives can cultivate a reputation for ethical
leadership through discussions of shared values and ethics, role modeling, rewards and education at all organizational levels.
Future research examining the effectiveness of such initiatives in organizations seems appropriate.
Second, to further elucidate how our results diverged from those reported by Turner et al. (2002), a more finely grained
examination of all transformational and transactional leadership behaviors in Bass and Avolio’s (1997) Full Range Leadershipmodel is
required. For example, relationships between a manager’s moral reasoning capacity and ratings of idealized influence, inspirational
motivation, individualized consideration and intellectual stimulation should be examined for differences and similarities. Also,
transactional contingent reward leadership should be considered separately fromtransformational and charismatic leadership in this
evaluation.
Third, althoughmanagers’moral reasoning capacity did not interactwithmanagement level to influence the dependent variables
in this study, it is possible for this to occur in different contexts with diverse samples. For example, Poneman and Gabhart’s (1994)
review indicated that accounting partners in CPA firms possess lower levels of moral development than staff accountants in those
firms, and that their behavior differs across management levels.While moral reasoning levels of managers in the present study was
unrelated to their management level (r=−.04, ns), it is possible that in other industries, the amount of OCB displayed by managers
may depend on their management level, as demonstrated in therapeutic recreation industry by Anderson and Bedini (2002). This
relationship may be moderated by the manager’s moral reasoning capacity given that individuals in helping professions such as
theologians, seminarians, doctors, and nurses typically score higher on the DIT than adults in general (Rest, 1994). Future research
should explore these ideas in organizations from industries other than those examined in the present study.
Fourth, this study’s mixed results suggest a need for alternative measures of leadership ethics. While measures exist for ethical
(Ethical Leadership Scale; Brown et al., 2005) and transformational/charismatic (MLQ-5X; Bass & Avolio, 1997) leadership, they fail to
distinguish between transformational and pseudo-transformational leadership (Bass & Steidlmeier, 1999) and socialized and
personalized charismatic leadership (House& Howell, 1992) behaviors. There is a need to developmeasures that tap into the dark side
of such leadership. In addition, with all due respect to the work of Kohlberg and Rest, perhaps the time has come for a new, more
contemporary measure of moral reasoning given the criticisms of DIT. As Appiah (2006) points out, our world has become much
smaller, somuch so that today “every human community has gradually been drawn into a singleweb of trade and a global network of
information,we have cometo a pointwhere each of us can realistically imagine contacting any other of our six billion conspecifics and
sending that person something worth having” (p. xii). Appiah argues that our technologically connected and complexmultinational,
multiculturalworld has placed significant demands on our ethical and moral fabric,with increasingly complex choices of actions and
dilemmas. The traditional measures of ethics and morality may be falling short, as evidenced by the limited amount of empirical
leadership research in this area, and the complexities of examining these topics as demonstrated in the present study. Thus,we believe
that the development of amoremodern, contemporary measure of ethics and morality would be a useful avenue for future research.
Fifth, in terms of self-other rating agreement, future research might examine how the distance between leaders and followers in
terms of physical distance, social distance, or frequency of interaction impacts howleaders are perceived and evaluated (Antonakis &
Atwater, 2002). Upper echelon leaders typically have little face-to-face interaction with followers in the lower ranks, and therefore
may engage inmore impressionmanagement and inauthentic behavior. These practicesmay exacerbate gaps between self- and otherratings
of leadership and performancewhich typically emerge in the higher organizational ranks (Gentry et al., 2007). Followers who

our sample may have produced somewhat of a restriction in the range of moral reasoning levels. Bass’ (2008) review of the moral
reasoning literature indicates that most managers, and adults in general, are likely to be conventional moral reasoners. Bass notes
that the average P score for adults is 40, and for graduate business students, 42.8. The average P score for managers in this study is
36.3, suggesting that our sample may limit the generalizability of study results to other settings.
In addition to the range restriction of DIT P score in our study sample, potential moderators were not considered to better
understand how managers’ moral reasoning levels translate into their leadership behaviors and performance. Because conventional
moral reasoning managers tend to be susceptible to situational influence, such as organizational culture (Trevino, 1986) and the
distance between leaders and followers (Brown & Trevino, 2006), results of the present study must be interpretedwith caution, given
the omission of situational characteristics.
4.2. Future research and conclusion
Despite these limitations, this study offers a foundation for future research in several areas. First, the unexpected finding that selfratings
of charismatic leadershipwere highest for preconventionalmoral reasoningmanagers calls into question Turner et al.’s (2002)
speculation that “manifesting postconventional moral reasoning might predispose individuals to choose to use transformational
leadership” (p. 309). Our study suggests that using the transformational leadership model to understandmoral development’s role in
gaining attributions of charismamay be insufficient, andwe should nowlook to othermodels, such as ethical (Brown & Trevino, 2006)
or spiritual (Fry, 2003) leadership, to guide for future research. For example, ethical leadership models and measures (Brown et al.,
2005) which consider elements of being a moral person and moralmanager may provide opportunities for greater understanding of
ethical decision-making processes. Future research should also assess personality factors such as narcissism, and situational factors
such as ethical climate, thatmight predispose both upper- and lower-echelonmanagers to reason unethically and display personalized
charismatic leadership behaviors.
If charismatic executives do lean toward preconventional levels of moral reasoning, companies should consider emphasizing
ethical conduct and raise the moral thought processes of their managers through training, education and research-based
leadership development programs (Day, Harrison, & Halpin, 2009). Can leadership development programs be designed to elevate
managers’ moral reasoning capacity over time? Trevino et al. (2000) suggest that executives can cultivate a reputation for ethical
leadership through discussions of shared values and ethics, role modeling, rewards and education at all organizational levels.
Future research examining the effectiveness of such initiatives in organizations seems appropriate.
Second, to further elucidate how our results diverged from those reported by Turner et al. (2002), a more finely grained
examination of all transformational and transactional leadership behaviors in Bass and Avolio’s (1997) Full Range Leadershipmodel is
required. For example, relationships between a manager’s moral reasoning capacity and ratings of idealized influence, inspirational
motivation, individualized consideration and intellectual stimulation should be examined for differences and similarities. Also,
transactional contingent reward leadership should be considered separately fromtransformational and charismatic leadership in this
evaluation.
Third, althoughmanagers’moral reasoning capacity did not interactwithmanagement level to influence the dependent variables
in this study, it is possible for this to occur in different contexts with diverse samples. For example, Poneman and Gabhart’s (1994)
review indicated that accounting partners in CPA firms possess lower levels of moral development than staff accountants in those
firms, and that their behavior differs across management levels.While moral reasoning levels of managers in the present study was
unrelated to their management level (r=−.04, ns), it is possible that in other industries, the amount of OCB displayed by managers
may depend on their management level, as demonstrated in therapeutic recreation industry by Anderson and Bedini (2002). This
relationship may be moderated by the manager’s moral reasoning capacity given that individuals in helping professions such as
theologians, seminarians, doctors, and nurses typically score higher on the DIT than adults in general (Rest, 1994). Future research
should explore these ideas in organizations from industries other than those examined in the present study.
Fourth, this study’s mixed results suggest a need for alternative measures of leadership ethics. While measures exist for ethical
(Ethical Leadership Scale; Brown et al., 2005) and transformational/charismatic (MLQ-5X; Bass & Avolio, 1997) leadership, they fail to
distinguish between transformational and pseudo-transformational leadership (Bass & Steidlmeier, 1999) and socialized and
personalized charismatic leadership (House& Howell, 1992) behaviors. There is a need to developmeasures that tap into the dark side
of such leadership. In addition, with all due respect to the work of Kohlberg and Rest, perhaps the time has come for a new, more
contemporary measure of moral reasoning given the criticisms of DIT. As Appiah (2006) points out, our world has become much
smaller, somuch so that today “every human community has gradually been drawn into a singleweb of trade and a global network of
information,we have cometo a pointwhere each of us can realistically imagine contacting any other of our six billion conspecifics and
sending that person something worth having” (p. xii). Appiah argues that our technologically connected and complexmultinational,
multiculturalworld has placed significant demands on our ethical and moral fabric,with increasingly complex choices of actions and
dilemmas. The traditional measures of ethics and morality may be falling short, as evidenced by the limited amount of empirical
leadership research in this area, and the complexities of examining these topics as demonstrated in the present study. Thus,we believe
that the development of amoremodern, contemporary measure of ethics and morality would be a useful avenue for future research.
Fifth, in terms of self-other rating agreement, future research might examine how the distance between leaders and followers in
terms of physical distance, social distance, or frequency of interaction impacts howleaders are perceived and evaluated (Antonakis &
Atwater, 2002). Upper echelon leaders typically have little face-to-face interaction with followers in the lower ranks, and therefore
may engage inmore impressionmanagement and inauthentic behavior. These practicesmay exacerbate gaps between self- and otherratings
of leadership and performancewhich typically emerge in the higher organizational ranks (Gentry et al., 2007). Followerswho

our sample may have produced somewhat of a restriction in the range of moral reasoning levels. Bass’ (2008) review of the moral
reasoning literature indicates that most managers, and adults in general, are likely to be conventional moral reasoners. Bass notes
that the average P score for adults is 40, and for graduate business students, 42.8. The average P score for managers in this study is
36.3, suggesting that our sample may limit the generalizability of study results to other settings.
In addition to the range restriction of DIT P score in our study sample, potential moderators were not considered to better
understand how managers’ moral reasoning levels translate into their leadership behaviors and performance. Because conventional
moral reasoning managers tend to be susceptible to situational influence, such as organizational culture (Trevino, 1986) and the
distance between leaders and followers (Brown & Trevino, 2006), results of the present study must be interpretedwith caution, given
the omission of situational characteristics.
4.2. Future research and conclusion
Despite these limitations, this study offers a foundation for future research in several areas. First, the unexpected finding that selfratings
of charismatic leadershipwere highest for preconventionalmoral reasoningmanagers calls into question Turner et al.’s (2002)
speculation that “manifesting postconventional moral reasoning might predispose individuals to choose to use transformational
leadership” (p. 309). Our study suggests that using the transformational leadership model to understandmoral development’s role in
gaining attributions of charismamay be insufficient, andwe should nowlook to othermodels, such as ethical (Brown & Trevino, 2006)
or spiritual (Fry, 2003) leadership, to guide for future research. For example, ethical leadership models and measures (Brown et al.,
2005) which consider elements of being a moral person and moralmanager may provide opportunities for greater understanding of
ethical decision-making processes. Future research should also assess personality factors such as narcissism, and situational factors
such as ethical climate, thatmight predispose both upper- and lower-echelonmanagers to reason unethically and display personalized
charismatic leadership behaviors.
If charismatic executives do lean toward preconventional levels of moral reasoning, companies should consider emphasizing
ethical conduct and raise the moral thought processes of their managers through training, education and research-based
leadership development programs (Day, Harrison, & Halpin, 2009). Can leadership development programs be designed to elevate
managers’ moral reasoning capacity over time? Trevino et al. (2000) suggest that executives can cultivate a reputation for ethical
leadership through discussions of shared values and ethics, role modeling, rewards and education at all organizational levels.
Future research examining the effectiveness of such initiatives in organizations seems appropriate.
Second, to further elucidate how our results diverged from those reported by Turner et al. (2002), a more finely grained
examination of all transformational and transactional leadership behaviors in Bass and Avolio’s (1997) Full Range Leadershipmodel is
required. For example, relationships between a manager’s moral reasoning capacity and ratings of idealized influence, inspirational
motivation, individualized consideration and intellectual stimulation should be examined for differences and similarities. Also,
transactional contingent reward leadership should be considered separately fromtransformational and charismatic leadership in this
evaluation.
Third, althoughmanagers’moral reasoning capacity did not interactwithmanagement level to influence the dependent variables
in this study, it is possible for this to occur in different contexts with diverse samples. For example, Poneman and Gabhart’s (1994)
review indicated that accounting partners in CPA firms possess lower levels of moral development than staff accountants in those
firms, and that their behavior differs across management levels.While moral reasoning levels of managers in the present study was
unrelated to their management level (r=−.04, ns), it is possible that in other industries, the amount of OCB displayed by managers
may depend on their management level, as demonstrated in therapeutic recreation industry by Anderson and Bedini (2002). This
relationship may be moderated by the manager’s moral reasoning capacity given that individuals in helping professions such as
theologians, seminarians, doctors, and nurses typically score higher on the DIT than adults in general (Rest, 1994). Future research
should explore these ideas in organizations from industries other than those examined in the present study.
Fourth, this study’s mixed results suggest a need for alternative measures of leadership ethics. While measures exist for ethical
(Ethical Leadership Scale; Brown et al., 2005) and transformational/charismatic (MLQ-5X; Bass & Avolio, 1997) leadership, they fail to
distinguish between transformational and pseudo-transformational leadership (Bass & Steidlmeier, 1999) and socialized and
personalized charismatic leadership (House& Howell, 1992) behaviors. There is a need to developmeasures that tap into the dark side
of such leadership. In addition, with all due respect to the work of Kohlberg and Rest, perhaps the time has come for a new, more
contemporary measure of moral reasoning given the criticisms of DIT. As Appiah (2006) points out, our world has become much
smaller, somuch so that today “every human community has gradually been drawn into a singleweb of trade and a global network of
information,we have cometo a pointwhere each of us can realistically imagine contacting any other of our six billion conspecifics and
sending that person something worth having” (p. xii). Appiah argues that our technologically connected and complexmultinational,
multiculturalworld has placed significant demands on our ethical and moral fabric,with increasingly complex choices of actions and
dilemmas. The traditional measures of ethics and morality may be falling short, as evidenced by the limited amount of empirical
leadership research in this area, and the complexities of examining these topics as demonstrated in the present study. Thus,we believe
that the development of amoremodern, contemporary measure of ethics and morality would be a useful avenue for future research.
Fifth, in terms of self-other rating agreement, future research might examine how the distance between leaders and followers in
terms of physical distance, social distance, or frequency of interaction impacts howleaders are perceived and evaluated (Antonakis &
Atwater, 2002). Upper echelon leaders typically have little face-to-face interaction with followers in the lower ranks, and therefore
may engage inmore impressionmanagement and inauthentic behavior. These practicesmay exacerbate gaps between self- and otherratings
of leadership and performancewhich typically emerge in the higher organizational ranks (Gentry et al., 2007). Followerswho