Embedding Ethical Leadership Within and Across Organization Levels

© Academy of Manogement Journal2012. Vol. 55. No. 5. 1053-1078.http://dx.doi.org/10.5465/ainj.2011.0064

EMBEDDING ETHICAL LEADERSHIP WITHIN AND ACROSSORGANIZATION LEVELS

JOHN M. SCHAUBROECKMichigan State University

SEAN T. HANNAHWake Forest University

BRUCE J. AVOLIOUniversity of Washington

STEVE W. J. KOZLOWSKIMichigan State University

ROBERT G. LORDAkron University

LINDA K. TREVIÑOPennsylvania State University

NIKOLAOS DIMOTAKISGeorgia State University

ANN C. PENGMichigan State University

We develop and test a model linking ethical leadership with unit ethical culture, bothacross and within organizational levels, examining how both leadership and culturerelate to ethical cognitions and behaviors of lower-level followers. The data werecollected from 2,572 U.S. Army soldiers representing three organizational levels de-ployed in combat. Findings provide limited support for simple trickle-dovm mecha-nisms of ethical leadership but broader support for a multilevel model that takes intoaccount how leaders embed shared understandings through their influence on theethical culture of units at various levels, which in turn influence followers' ethicalcognitions and behavior. The influences of ethical leadership occur not only directly,among immediate followers within a unit, but also indirectly, across hierarchicallevels, through the cascading of ethical culture and senior leaders' influences onsubordinate leader behavior. We discuss scholarly and practical implications forunderstanding how leaders transmit ethical influence both down and across largeorganizations.

Unethical organizational behaviors have been at-tributed to the effects of individual "had apples"(e.g., Hannah, Avolio, & May, 2011), their leaders

The authors acknowledge the United States Army sol-diers who, despite serving their nation in an active com-bat zone, volunteered to participate in this research. Wealso thank the Army chaplains who traveled betweencombat outposts to collect this data, the leadership ofMultinational Force Iraq, and the U.S. Army's Center forthe Army Profession and Ethic (CAPE), which sponsoredthis research. We also want to express our appreciation toMegan Huth for her assistance with the data.

(e.g.. Brown, Harrison, & Treviño, 2005), aspects oforganizational context, such as low levels of "ethi-cal culture" ("bad barrels" [e.g., Treviño & Young-blood, 1990]), and the interaction of individual pre-dispositions and context (e.g., Treviño, 1986). Inthis study, we extend previous research and theoryto examine how ethical cognitions and behaviors oforganization members reflect tbeir responses to aweb of direct and indirect influences of ethical lead-ership and unit-level norms, standards, and sanctionspertaining to ethical behavior transmitted across mul-tiple levels of an organizational hierarchy.

1053Copyright of the Academy of Management, all rights reserved. Contents may not he copied, emailed, posted toa listserv. or otherwise transmitted without the copyright holder's expresswritten permission. Users may print, download, or email articles for individual use only.

1054 Academy of Management Journal October

Scholars have called for taking a multilevel ap-proach to more fully understand how leadership athigher levels of organizations influences lower-levelfollowers (e.g., Waldman & Yammarino, 1999]. Yetprior empirical work on leadership has primarily fo-cused on direct relationships between leader behav-iors or traits and the responses of their immediatefollowers (Yammarino & Dansereau, 2008]. Research-ers have not comprehensively tested the various locithrough which upper-level leaders influence the cog-nitions and behavior of lower-level followers. Thus,currently little is known about how leadership andcontextual factors at higher organization levels influ-ence outcomes at lower levels. From a metatheoreti-cal perspecfive, how a coherent system of social in-fluence is transmitted across multiple levels of anorganization is fundamental to understanding organ-izafional behavior (Kozlowski & Klein, 2000; Si-mon, 1973].

We develop a multilevel model in which ethicalleadership, which is seen as a leader's use of socialinfluence to promote ethical conduct (Brown et al.,2005], exhibited at various levels of an organiza-tional hierarchy, not only directly influences im-mediate followers' ethical conduct, but also indi-rectly influences the ethical beliefs and conduct offollowers at lower levels owing to (a] the replica-tion of ethical leader behaviors among subordinateleaders and (b] the embedding of shared under-standings that represent observable elements ofwork unit ethical culture. Our proposed model istied to a foundational framework of organizationalculture developed by Schein (1985, 2010], whoposited mechanisms through which leaders embedtheir assumptions into the "thinking, behavior, andfeelings" of groups (1985: 223]. We draw fromSchein's "embedding mecbanisms" (1985: 224] and"sbared cultiural elements" (1985: 169] and extendhis theoretical framework by developing and test-ing a multilevel model linking leadership, sharedcultural elements, and their direct and indirect ef-fects on follower ethical cognitions and behaviors.This approach provides a starting framework forunderstanding how different leaders influence eth-ical outcomes within a complex multilevel system.

THEORETICAL FRAMEWORK

Our research model draws from Schein (1985,2010] to explain how senior leaders influence theethical conduct of followers at lower levels by em-bedding their expectations and assumptions intothe observable manifestations of ethical culture.Schein defined culture as a system of shared as-sumptions that can have a strong influence in di-recting followers' behaviors and beliefs. Organiza-

tional culture is represented as varying in layers,with the deepest layer being a broad system ofassumptions and deeply held shared meanings,and the surface layer representing more tangible,observable factors that reflect those assumptions.Schein (1985, 2010] argued that leaders at all levelscan influence the surface layer of an organization'sculture, including visible artifacts such as behav-ioral norms, policies, and standards, particularly asthese relate to specific domains of influence thatSchein described as shared cultural elements. Byinfluencing these surface elements, leaders canhelp others interpret their unit and organization'sculture in terms of how it relates to their roles andexpectations.

Unit Ethical Culture

Shared understandings concerning ethical con-duct can be seen as a shared cultural subelement, ora "microcosm" of an overall organizational culture(Treviño, 1986, 1990]. This aspect is the part of theculture that distinguishes what is "heroic" fromwhat is "sinful" (Schein, 1985: 79] and is composedof patterns of shared understandings related to un-ethical and ethical conduct reflecting the norms,standards, sanctions, and rewards applied to be-haviors deemed desirable and undesirable in anorganization. This focus is consistent with howTreviño (1986, 1990] and Treviño, Butterfield, andMcCabe (1998] defined ethical culture. Specifi-cally, in their view ethical culture is "a subset oforganizational culture, representing a multidimen-sional interplay among various 'formal' and 'infor-mal' systems of behavioral control that are capableof promoting either ethical or unethical behavior"(Treviño et al., 1998: 451-452]. This use of the term"culture" should not be equated with broader def-initions of culture that relate to the shared assump-tions that combine to form deep organizational cul-ture. Rather, in the current study, ethical culturerefers to surface-level cultural artifacts that are ob-servable and were called shared cultural elementsby Schein (1985]. This conceptualization is similarto descriptions of climate in other research (e.g.,Zohar & Luria, 2005]. However, there exists a long-established construct and measure of ethical cli-mate (Victor & Cullen, 1988] in the organizationalethics literature that refers to criteria used in ethi-cal judgment processes and the level of analysisused by parties in making ethical judgments. Thisconstruct has spawned a significant amount of re-search over the last 20 years (Arnaud & Schminke,in press]. Thus, we use the label "ethical culture"rather than "ethical climate" to be consistent withSchein (1985, 2010], noting that we are focusing on

2012 Schaubroeck et al. 1055

the surface level of ethical culture, followingTreviño (1990; Treviño et al. 1998), and to avoidconfusion with the established ethical climate con-struct and its different meaning.

"Formal" ethical culture systems include policies(e.g., codes of ethics that are enforced), authoritystructures, reward systems, and ethics training pro-grams. "Informal" systems include peer behavior, useof ethical language, myths, stories, and ethical norms.When the members of an organization or a unit in anorganization have strong shared understandingsabout such matters, they are expected to be morecognizant of ethical issues, to avoid unethical con-duct personally, and to discourage unethical conductin others (Treviño, 1986). Such units penalize imeth-ical behavior, reward virtuous behavior, and main-tain strong ethical norms. This conceptualizationcontains both global properties pertaining to organi-zational codes and shared social properdes withinthe imit concerning agreed-upon norms and pre-scribed patterns of behavior. Following Schein, werefer to such domains in general as "shared culturalelements" and refer to Treviño and colleagues' (1998)construct as it applies to the domain of ethics atvarious hierarchical levels as "ethical culture."Whereas one may also use the term "ethical culture"to refer to an organization as a whole, we expect thatethical culture varies across different units and levelsof an organization despite common global influences,such as organizational codes of ethics and othernorms, rules, and regulations at the organizafion orsocietal level. In addition, as Schein noted, "'shared'

understanding means that vmit members recognize aparticular feeling, experience or activity as common"(1985: 168; emphasis in original). From a multileveltheory perspective, such constructs are composi-tional in form and exhibit restricted within-imit vari-ation (Kozlowski & Klein, 2000).

Overview of Research Model

In this study, we develop a multilevel model forunderstanding how leadership and culture influ-ence ethics-related cognitions and behaviors, andwe test it using survey data collected from a largesample of deployed soldiers at different hierarchi-cal levels. Specifically, we examined leadership byassessing four separate potential mechanisms forleadership effects across organizational levels, asdisplayed in Figure 1. First, a senior leader's stylemay be mirrored by that leader's followers (Ban-dura, 1971). To the extent that subordinate leaders'behavior mirrors that oftheir senior leaders, a pos-itive downward relationship exists between thestyles of leaders at adjacent levels. Second, indirecteffects of higher-level leaders that involve multipledownward linkages are possible; these effects mayoccur, for example, when subordinate leaders mir-ror senior leaders' styles and those subordinateleaders' styles in turn influence how their followersthink and behave. The passage of direct and indi-rect effects down a hierarchy of authority has beenvariously labeled "the cascade of leadership," "thetrickle-down model," and "the falling dominoes

FIGURE 1Direct and Indirect EÉfects of Ethical Leadership and Ethical Culture across Hierarchical Levels^

Level 3 Ethical Leadership(L3)

Level 2Ethical Leadership

(L2)

Level 1 Ethical Leadership(LI)

EthicalCulture

(G3)

EthicalGulture

(G2)

EthicalGulture

(GI)

Ethical andUnethical

Gognitions andBehaviors

(01)

° Dashed lines denote unexplained direct effects that have been conventionally labeled as "bypass" effects.


effect" (Bass, Waldman, Avolio, & Bebb, 1987;Mayer, Kuenzi, Greenbaum, Bardes, & Salvador,2009; Waldman & Yammarino, 1999). Third, anindirect leadership effect may occur through thehorizontal and lateral linkages shown in Figure 1;for instance, a more senior leader may foster anethical culture that then spreads to influence cul-ture and behavior in lower organizational units.This effect of leaders on unit cultures is a morediffuse and complex process involving multiplecoactors. It may be reflected in multiple effectsshowing mediation at different levels, such asthrough subordinate leader and follower behaviorsand through leaders' effects on unit culture at lowerlevels (Bass & Bass, 2008). Finally, senior leaders'influence can skip levels and directly affect lower-level followers' behavior as a result of their directinteractions (Detert & Treviño, 2010). This creates afourth potential mechanism of leader influence re-ferred to as a bypass processing model (Yamma-rino, 1994).

Outcomes of Ethical Leadership andEthical Culture

Prior research on ethical leadership has assesseda limited set of criteria. A recent review of theethical leadership literature by Brown and Mitchell(2010) noted that besides research on organiza-tional citizenship behavior (OCB), surprisingly lit-tle empirical research has examined how ethicalleadership relates to follower ethical or unethicalbehavior. Theory and research suggest that the nor-mative and informational influences of a unit withstrong ethical leadership should support ethicalconduct and cuih organization members' seriousethical transgressions. Collective understandingsabout ethical conduct that become embedded inshared cultural elements should also encouragestronger beliefs among individual members abouttheir moral agency, including (a) moral efficacy and(b) intentions to report others' ethical violations.Hannah et al. defined moral efficacy as "an indi-vidual's belief in his or her capabilities to organizeand mobilize the motivation, cognitive resources,means, and courses of action needed to attain moralperformance within a given moral domain, whilepersisting in the face of moral adversity" (2011: 24).With higher moral efficacy, individuals are morelikely to convert moral judgments and intentionsinto ethical actions. Hannah and Avolio (2010) pro-posed that moral efficacy would be bolstered whenfollowers are immersed in a context that providesthe "means" (e.g., leader support and policies sup-portive of ethical actions) to act ethically.

Individuals' intentions to report the ethical vio-lations of others are also related to moral agency. Ifunethical behavior is to be addressed in organiza-tions, authority figiures must know about it andtherefore must set conditions to promote followerreporting. Followers tend to keep their knowledgeof ethical problems to themselves for a number ofreasons, including fear of retaliation, a sense thatnothing will be done, or habituation to silence inauthority situations (Detert & Edmondson, 2011;Kish-Gephart, Detert, Treviño, & Edmondson,2010). Ethical leadership and a strong ethical cul-ture can be expected to enhance followers' willing-ness to speak up because they are more likely to feelprotected from retaliation and to believe that positiveacüons will be taken to address their concern. In pastresearch, ethical leadership has been posifively re-lated to "voice" and the likelihood that followers willreport problems to authorities (Brown et al., 2005;Walvunbwa & Schaubroeck, 2009).

Beyond being seen as promoting followers' moralefficacy and enabling them to report violations andto refrain from ethical transgressions, ethical lead-ers are also seen as moral role models who encour-age followers to strive toward moral ideals andengage in ethical behavior that exceeds expecta-tions (Brown & Treviño, 2006). We draw fromscholarly work on virtuous ethical behavior (e.g..Walker & Henning, 2004) in defining exemplarybehavior as undertaking imusual personal risk orsacrifice to achieve a desirable collective purpose.

It is important to note that these outcomes—moralefficacy, intentions to report violations, transgres-sions, and exemplary behavior—are conceptualizedas "compilation constructs" (Kozlowski & Klein,2000). Rather than regarding them as constructs thatare emergent and shared at the unit level, as areethical leadership and ethical culture, we conceptu-alize them as configurations of individual beliefstates, intentions, and rates that capture the descrip-tive averages among individuals in a unit.

Transmission of Ethical Leadership throughEthical Culture

Direct effects. The notion that there is a relation-ship between what leaders do and the norms andcodes of their units—such as the direct effects ofethical leadership on ethical culture shown as L3^ C3, L2 -» C2, and Ll -^ Cl in Figure 1, whichsummarizes our model—is quite well acceptedamong organizational scholars (Schein, 2010; Yukl,2002). Leaders establish priorities for their unitsand expectations for what will be rewarded, pun-ished, and tolerated. In addition to employing thesemore transactional means, leaders can also serve as


moral exemplars for followers to help them under-stand what is right and wrong and to be attracted todeveloping a more ethically centered self-image(Walker & Herming, 2004).

However, previous leadership studies have notexplicitly integrated unit shared cultural elements(Schein, 1985, 2010) into either their theoreticalframing or their analyses of how these elementsmay mediate the relationships between leader be-havior and follower outcomes occurring within orbetween organizational levels. Such cultural ele-ments are likely to be important because, in addi-tion to directly impacting followers' beliefs andbehaviors, leaders create shared understandingsthat affect these outcomes (Depret & Fiske, 1993;Dragoni, 2005). With regard to ethics-related out-comes, follower cognitions and bebavior shouldtend to align with these shared normative under-standings, which are represented as ethical culture.Thus, in our study, ethical leadership is expectedto be associated with ethical culture, which in tvirnpromotes more positive beliefs among membersabout their personal capacity to act as moral agents,as well as reduced ethical transgressions and in-creased exemplary ethical behavior.

Ethical leadership transmitted through embed-ding mechanisms. To influence shared cultural el-ements at a particular unit level and then transmitthis influence to lower levels, Schein argued, lead-ers strive to embed their beliefs, values, and as-sumptions into members' shared understandings.Leaders do this by using a range of primary andsecondary "embedding mechanisms" throughwhich they influence the visible artifacts of sharedcultural elements (Schein, 1985: 223). The primarymechanisms are seen as having the most direct andstrongest effects on the beliefs, values, and assump-tions of followers. Among these primary mecha-nisms, Schein described "what leaders pay atten-tion to, measure, and control" as being the mostpowerful (2010: 237; cf. Feldman, 1984). Other pri-mary embedding mechanisms include "deliberaterole modeling, teaching, and coaching" and "allo-cating rewards and status" (Schein, 2010: 236). Fol-lowers' perceptions of a leader's ethical leadershiprelate directly to these primary embedding mecha-nisms. Specifically, Brown and colleagues definedethical leadership as "the demonstration of norma-tively appropriate conduct through personal ac-tions and interpersonal relationships, and thepromotion of such conduct to followers throughtwo-way communication, reinforcement, and deci-sion making" (2005: 120). Ethical leadership in-cludes what has been labeled a "moral manager"aspect and a "moral person" aspect (Brown &Treviño, 2006: 597). As a moral manager, a leader

specifies conduct that is proper and improper, ex-plains the rewards and sanctions associated withbehaving ethically and unethically, appropriatelydisciplines infractions, and rewards virtuous be-havior. As a moral person, the leader models exem-plary behavior and leads in a principled, just, andcaring manner. Ethical leadership as conceptual-ized by Brown et al. (2005) incorporates elementsof Schein's (1985, 2010) primary embedding mech-anisms, including paying attention to, measuring,and controlling followers' behavior, as well as theprimary mechanisms of role modeling and coach-ing followers to exhibit exemplary ethical behaviorand to avoid unethical behavior.

Other embedding mechanisms may also be rep-resented in ethical leadership. Schein describedhow a leader's approach to handling "critical inci-dents" (cf. Feldman, 1984), allocating resources,and "selecting, promoting, and excommunicating"(Schein, 1985: 235) followers are additional pri-mary embedding mechanisms. When a member'sethical infraction comes to a leader's attention, thisevent can be seen as a critical incident. How theleader responds is diagnostic and will influencefollowers' perceptions of the leader as a moral man-ager. Failing to discipline the follower and use thiscritical incident as a lesson to others may under-mine her/his authority in ethical matters and per-suade followers that ethical standards are not en-forced in the unit. Alternatively, a significant lossof status or dismissal ("excommunication") of afollower for an ethical infraction may reinforce per-ceptions of ethical leadership and, in turn, increaseconfidence that sanctions are used to enforce ethi-cal standards. One can also imagine that leaderswho allocate resources to teaching proper ethicalconduct, such as through ethical training, wouldmore likely be seen as moral managers or ethicalleaders.

Ethical culture may also involve secondary em-bedding mechanisms in which leaders engage.Schein (1985, 2010) described these as being lesspowerful and potentially more ambiguous than theprimary embedding mechanisms. These secondarymechanisms refer to such artifacts as stories, work-flow designs, and other organization structures,procedures, and formal statements of leaders. Forexample, ethical leaders may tell stories of how anindividual's exemplary ethical behavior was re-warded, and/or stories of how past organizationmembers' unethical conduct was disciplined. Suchstories may be very salient in members' memories,and they may pass them along to others to whomthe story will serve as an incentive or as a deterrent.These secondary mechanisms are also consistentwith ethical leadership, as stories and formal state-


ments with implications for ethical conduct are apotential tool for managers seeking to lead ethi-cally. Together, the primary and secondary embed-ding mechanisms are expected to shape the forma-tion of shared, unit-level perceptions of ethicalleadership as a compositional construct (Kozlowski& Klein, 2000). Ethical leadership in turn becomesembedded into shared understandings about ethi-cal norms, standards, and sanctions in a luiit—thatis, ethical culture—which is also compositional.

Studies by Browm and colleagues (2005) andother recent work collectively indicate that whenindividuals perceive their leader as a proponentand exemplar of ethical behavior, they report indi-vidual and work unit psychological states that areconducive to more ethical conduct (Walumbwa,Mayer, Wang, Wang, Workman, & Christensen,2011; Walumbwa & Schaubroeck, 2009). Whenthey perceive stronger ethical leadership, followersare more likely to engage in prosocial behaviorsand less likely to engage in deviant or counterpro-ductive bebaviors (Mayer et al., 2009; Neubert,Carlson, Kacmar, Roberts, & Chonko, 2009). How-ever, we are not aware of any research that hasexamined the linkage between ethical leadershipand ethical culture, much less research concerninghow ethical culture may aid in the transmission ofethical leadership in a manner that influences im-portant ethics-related cognitions and behaviors offollowers within and between organizationallevels.

Hypothesis 1. Ethical leadership is positivelyrelated to ethical culture within ahierarchical level.

Hypothesis 2. Through ethical culture, ethicalleadership is indirectly and positively relatedto unit members' moral agency (their belief intheir capacity to act effectively as moral agentsand their intentions to report unethical behav-ior) and exemplary ethical behavior and nega-tively related to their ethical transgressions.

Hypotheses 1 and 2 recognize that differenceslikely exist in leadership and ethical culture amongunits at a particular organizational level. In the nextsection, we discuss how our theoretical frameworkis operationalized at different hierarchical levelsand how higher-level ethical leadership and ethicalculture are associated with outcomes at the low-est level.

Cascading Effects of Ethical Culture andEthical Leadership

Unit-level differences in ethical culture may bedetermined in part by variation in unit ethical lead-

ership as described above. Importantly, however,these unit differences in ethical culture may alsoreflect the impact of higher-level leaders who caninfluence the ethical culture of lower-level units,either directly through their personal contact withmembers, or indirectly through their influence onthe thoughts and behaviors of lower-level leaders.Below, we discuss these direct and indirect pro-cesses further in the context of leader behaviorcascades and bypass processes.

Differentiation and cascading of ethical cul-ture. To date, the limited research on ethical cul-ture has focused chiefly on followers' perceptionsof ethical culture in their overall organization (e.g.,Treviño et al., 1998). Given that culture has bothglobal and local emergent properties, however, eth-ical culture may differ among organizational units(Glisson & James, 2002). For example, differencesmay emerge between units, both within and acrossmanagement levels, in the extent to which normsfor conduct are made clear to members, ethicalstandards are enforced, and norms emerge for tol-erating behaviors that conflict with establishedcodes.

Factors other than leader behavior may also in-fluence ethical culture at lower levels of an organ-ization. These include a superordinate ethical cul-ture of the organization that conveys standards ofconduct, sanctions for their violation, norms forethical conduct, and more for the entire organiza-tional community (Schein, 1985). A mechanismthat Schein does not explicitly examine is howcultural elements in a particular unit in an organi-zation tend to mimic content from cultural ele-ments established in higher-level units. Ethical cul-tures of lower-level units are nested within broaderorganizational units and thus may be affected bythe cultural elements operating at higher levels.These effects may be somewhat independent of theproximal, within-level influences of leadership.For example, shared cultural elements can be ex-plained in part by the past actions of higher-levelleaders, through the creation of designs and struc-tures, systems and procedures, creeds and codes,stories, rituals, and rites (Schein, 1985, 2010).

Ethical cultures at one level, however, are notnecessarily fully replicated at lower levels. In thisregard, Schein (1985, 2010) described how culturalelements of different groups in an organization candiffer, conflict, or align with one another. As Brownand Treviño (2006) explained, organizational lead-ers and groups develop tools and practices for rea-soning and acting upon ethical matters that canbecome embedded over time. Unit ethical culturesthat are operating at higher organizational levelsthat are seen as effective by lower-level leaders


would be expected to result in those lower-levelleaders' embedding them into their own units. Thisis especially likely when lower-level leaders per-ceive that they will be rewarded for reinforcingthese cultural elements in their own units.

Another potential lever helping ethical culture tohave cascading influences at lower levels derivesfrom leader contact and peer-to-peer contact acrossunits. Such interactions can reinforce key aspectsof the culture of the higher-level unit to which theyall belong and are nested through normative andinformational influences (Cialdini & Trost, 1998].For example, leaders and members at a particularlevel collaborate and socialize with peers and lead-ers in the context of meetings held at a higherorganizational level. In this way, workers develop acommon point of reference for embedding ethicalculture into their respective units. Thus, we expectthat higher-level-unit ethical culture will be asso-ciated with ethical culture at adjacent levels, asillustrated by the vertical lines linking C3, C2, andCl in Figure 1.

Hypothesis 3. Ethical culture at higher hierar-chical levels is positively related to ethical cul-ture at lower hierarchical levels.

Influences of ethical leadership on ethical cul-ture at lovi^er levels. The literature further suggeststhat ethical leadership at higher levels may be as-sociated with ethical culture at lower levels. Insome cases, these effects may be direct influencesof bigher-level leaders on lower-level ethical cul-ture through direct interactions with lower-levelfollowers. Such direct effects have been called"bypass effects" because they bypass the effectsof lower-level leaders (Yammarino, 1994] or low-er-level cultural elements (L3 -^ C2 or L2 -^ Cl].Detert and Treviño (2010] reported qualitativeevidence of such bypass effects of senior leaderbehavior on voice-related practices of lower-levelfollowers. There is also some evidence of bypasseffects from quantitative field studies (Dvir,Eden, Avolio, & Shamir, 2002; Yang, Zhang, &Tsui, 2010]. According to Yammarino, this mayoccur because leaders often interact directly withmembers at lower levels. Schein (1985, 2010]observed that senior leaders' expectations arealso conveyed by other means to the broader or-ganization, through secondary embedding mecb-anisms such as formal communications, policies,and stories. Such mechanisms may reinforce theleaders' reputation and expectations amongmembers, thereby influencing their cognitionsand actions in a way that is distinct from theinfluence of their immediate leader. Any exami-nation of such processes nevertheless presup-

poses a relationship between higher-level ethicalleadership and lower-level ethical culture. Thus,we propose the following:

Hypothesis 4. Ethical leadership is positivelyrelated to ethical culture at lower hierarchicallevels.

However, what appears to be a bypass effect mayactually be a product of embedding processes thatare omitted from an analysis. Below, we considertwo different indirect pathways through which se-nior leaders may embed their ethical influence atlower levels. Studies presenting evidence of bypasseffects have not examined the extent to which high-er-level leaders have embedded their assumptionsand expectations in shared cultural elements attheir own (higher] level. Given that ethical leader-ship is expected to be positively related to ethicalculture within each organization level (Hypothesis1] and these shared culture elements are in tiirnexpected to cascade to lower levels (Hypothesis 3],one indirect way that higher-level leaders can in-fluence ethical conduct at a lower level is by influ-encing the ethical culture at their own level, whichthen cascades to lower levels (i.e., L3 ̂ C3 -> C2 orL2 -» C2 -* Cl paths].

Hypothesis 5. Higher-level ethical leadershipindirectly influences ethical culture at lowerhierarchical levels through ethical culture atthe same (higher) hierarchical levels.

Another indirect path linking ethical leadershipto ethical culture at lower levels is the conventionalleader behavior cascade. Leaders at higher levelsindirectly affect the beliefs and behaviors of lower-level followers through their subordinate leaders'emulation of their bebavior (as shown by the verti-cal lines connecting L3, L2, and Ll in Figure 1].Drawing on social learning theory (Bandura, 1971],Mayer and colleagues (2009] examined the indirecteffects of ethical leadership on members' behaviortransmitted through common ethical leadershippractices across two organization levels. Theyasked followers in multiple organizations to eachrate both their immediate supervisor's ethical lead-ership and the ethical leadership of "top manage-ment." These ratings were positively correlated,and the ethical leadership of both top managementand immediate supervisor were positively relatedto OCB and negatively related to levels of deviantbehavior followers observed in their units. More-over, the ethical leadership of the immediate super-visor mediated the effects of top management eth-ical leadership on group deviance and group OCB.

Additional research has examined how otherleadership constructs, such as transformational


leadership, correlate across levels (Bass et al., 1987;Chun, Yammarino, Dionne, Sosik, & Moon, 2009;Yang, Zhang, & Tsui, 2010). For example, to ex-plain how the effects of leader behavior cascadeacross managerial levels, Bass and colleagues (Basset al., 1987) concluded that junior supervisorsmimic the behavior of their senior supervisors, asdescribed by social learning theory (cf. Weiss,1977), and then become models for their own fol-lowers. Chun et al. (2009) found that subordinateleaders' personal identiflcation with their leadersand internalization of their values explained indi-rect effects of the latter's charismatic and transfor-mational leadership on follower outcomes of satis-faction, helping behavior, and job performance.

Normative and informational influences provideanother potential avenue for explaining the cascad-ing replication of leaders' behaviors. People tend toconform their behavior to the expectations of oth-ers, either to be liked or respected (normative in-fluence) or to be accurate or correct (informationalinfluence) (Cialdini & Trost, 1989). Ethical leadersmake it clear that they will not condone unethicalbehavior while also presenting persuasive argu-ments as to the beneflts of ethical practices to sub-ordinate leaders (Brown & Treviño, 2006). Whetherthrough behavior modeling, informational influ-ence, or compliance with authority figures, ethicalleadership is expected to be positively correlatedacross hierarchical levels. Civen the positive rela-tionship we have already proposed between ethicalleadership and ethical culture at the same level(Hypothesis 1), this cascading effect of ethical lead-ership provides an additional mechanism throughwhich higher-level leaders may indirectly influ-ence the ethical culture in lower-level units (see,e.g., paths L3 ^ L2 ^ C2 and L2 ^ Ll ^ Cl inFigure 1).

Hypothesis 6. Ethical leadership at higher or-ganizational levels is indirectly related to eth-ical culture at lower hierarchical levelsthrough ethical leadership at lower levels.

Higher-Level Facilitation of Ethical Leadership

To the extent that leader behavior can influenceshared understandings about ethical conduct atlower levels, as we have argued above, influencefrom a higher-level leader may potentially reinforceor countervail the influence that the lower-levelleader has on his or her unit. When the higher-levelleader is not a proponent of ethical conduct, mem-bers of lower-level units may be less likely to de-velop strong norms and standards for their ownethical behavior (i.e., strong ethical culture), even

though their immediate supervisor exhibits a highlevel of ethical leadership (Detert & Edmondson,2011). When the leader of a lower-level unit seeksto develop a more ethical culture in her unit, herown leader may facilitate that by reinforcing theimportance of such behaviors to her subordinates.A higher-level leader's influence in facilitating thebehavior of the leader at the next lower level is aleadership enhancer (Howell, Dorfman, & Kerr,1986), which suggests a cross-level interaction pre-dicting ethical culture at the lower level (e.g., L2 XLl -^ Cl in Figure 1).

Hypothesis 7. Ethical leadership at higher hi-erarchical levels moderates the influence ofsubordinate ethical leadership: A high level ofethical leadership at a higher organizationlevel is associated with a stronger influence ofsubordinate leaders' ethical leadership on eth-ical culture in their units.

METHODS

Sample and Procedures

This study was commissioned by the U.S. Army,which asked us to evaluate the ethical conduct,cognitions, attitudes, and well-being of soldiersduring their combat deployment in Iraq in May2009. The surveys were designed by the authorsand were administered through the Inspector Gen-eral (IG; the Army's official investigative office) andchaplains. The aim was to maintain a "chain ofcustody" with a trusted agent to reassure soldiersthat their leaders would not have access to theirsurvey responses. To achieve broad representation,the IG randomly selected two brigades from each ofthe four Army divisions conducting combat opera-tions in Iraq. Two battalions were randomly se-lected from each of those two brigades. Three com-panies were randomly selected from each of thosebattalions, as were three platoons from those com-panies and three squads from each platoon. Westudied leaders and followers from the lowest threemanagerial levels in the U.S. Army: the typicallynine-person squad (lowest level), the three- to four-squad platoon (middle level), and the three- to four-platoon company (highest level). Each level had itsown leader who reported to the leader at the nexthigher level.

The questionnaires were administered to mem-bers of the randomly selected units in person insmall assemblies by Army chaplains who traveledbetween combat outposts in regions with activecombat operations. The chaplains received a stan-dardized training course that covered the purposeof the research, protocol, and the survey items. The


chaplains returned the completed surveys to the IG,who delivered them to the research team. All re-spondents provided informed consent, and soldiersunderstood they were free to decline to participatein this study.

The sampling frame included approximately 4percent of the soldiers in the theater of operationsat that time, or about 6,000 servicemen and service-women. A total of 2,572 surveys were returned, foran estimated 43 percent response rate. For squad-and platoon-level aggregate variables, to ensure ad-equate representation at each nested organizationallevel, we designated that platoons must have atleast 2 squads reporting, and each squad in turnmust have a minimum of 4 members reporting. Ofthe 322 squads sampled, 280 (90%) met this crite-rion. Of 130 platoons sampled, 96 (78%) met thiscriterion. These platoons represented 42 compa-nies. Those units meeting our criteria for the anal-yses required for this study totaled 2,048 soldiers.Missing data on certain variables (and combina-tions of variables across levels) limited our finalanalysis sample sizes to 172 squads, 78 platoons,and 39 companies for cross-level analyses, and 243squads, 85 platoons, and 40 companies for within-level analyses. Mean unit sizes for the latter were6.09 (s.d. = 2.24) for squads, 3.61 (s.d. = 1.50) forplatoons, and 4.57 (s.d. = 1.45) for companies.

Most respondents were male (88%), and theiraverage age was 27.5 years. About half (54%) weremarried, and 48 percent reported having children.All had high school degrees or the equivalent, and14 percent had college degrees. These statistics ap-proximately mirror the demographic profile of theoverall U.S. Army contingent in Iraq (MHAT-IV,2006). Nearly 68 percent of the respondents were

active duty U.S. Army soldiers. The remainderwere activated members of the U.S. Army NationalGuard or the U.S. Army Reserve. Eight percentwere commissioned officers.

Measures

All outcome variables were reported by squadmembers and aggregated to the squad level. Predic-tors and mediators were reported by multiple andseparate rating sources at each of the squad, pla-toon, and company levels, and we aggregated rat-ings at each level. Squad-, platoon-, and company-level variables were reported by squad members,squad leaders, and platoon leaders and sergeants,respectively. As shown on the diagonal in Table 1,coefficient alpha reliabilities for the measures wereall above .70. We examined the outcomes only atthe squad level. These frontline units had the great-est opportunities and risks for serious ethical mis-conduct arising from contact with combatants, non-combatants, and other operational events. Inaddition, our research questions concerned onlythe effects of higher-level leader behavior and eth-ical culture on outcomes at the lowest organiza-tional level.

Outcome variables. Ethical transgressions ofdifferent degrees of severity were measured with aset of items used in a previous study of the ethicalbehavior of U.S. soldiers during previous years ofthe Iraq war (MHAT IV, 2006). These items seekpersonal observations of unethical behavior exhib-ited by members of a respondent's squad. Theywere adapted for use in this study. The target be-haviors presented after the stem "rate the numberof times you are aware that members of your

TABLE 1Variance Gomponents"

Variables

Variance Components''

Level 1: Squad Level 2: Platoon Level 3: Company

Platoon levelEthical cultureEthical leadership

Squad levelEthical cultureEthical leadershipMoral efficacyPeer exemplary behaviorTransgressions vs. the ArmyTransgressions vs. noncombatantsReport intentions

56.74%99.8278.7860.6987.1287.0172.54

67.85%87.85

25.520.08

21.180.03

12.802.650.01

32.15%12.15

17.740.100.04

39.280.08

10.3527.44

" n = 243 for the squad level, n = 85 for the platoon level, and n = 40 for company level.^ Percentage of variance between groups at different hierarchical levels.


squad . . . " represented two categories of transgres-sions, which in this context refer to a total of sevensignificant and formally punishable organizationalactions of two types: transgressions against non-combatants (i.e., "mistreated non-combatants,""unnecessarily damaged non-combatant's prop-erty," "killed or injured a non-combatant when itwas not necessary") and transgressions against theArmy (i.e., "defied the ROE [Rules of Engagement,i.e., the policies given by commanders as to tbeproper use of force in combat] to accomplish themission," "lied or falsified reports," "stole," "de-fied legal orders").

Peer exemplary behavior was measured usingitems developed specifically for this study. Theyinvolve observations of peers undertaking personalrisks and engaging in self-sacrifice in ways that areconventionally regarded in the military as virtuousand necessary to accomplish collective goals. Re-spondents were asked to "rate the number of timesthey had seen fellow squad members "put them-selves in physical danger to protect a fellow Sol-dier/Marine," "put themselves in physical dangerto protect a non-combatant," and "challenge lead-ers' orders in order to protect a non-combatant."Transgressions and exemplary behavior frequen-cies were reported on a 0-5 scale ("never" to "fivetimes or more").

Moral efficacy is the belief that one is capable ofacting effectively as a moral agent. It was measuredwith five items developed by Hannah and Avolio(2010) and rated on a scale ranging from 0, "not atall confident," to 10, "totally confident." The itemsspecified an individual perception. Sample itemsinclude "I am confident that I can confront otherswho behave unethically to resolve the issue" and "Iam confident that I can determine what needs to bedone when I face a moral/ethical decision." Weused five items to assess report intentions, whichmeasure an individual's intention to report uneth-ical conduct of other soldiers when it is observed.The items were tailored to the unique context ofwar in an environment with close proximity to anonconjbatant population. As with moral efficacy,the items specified an individual perception. Asample item is "I would report a unit member if Isaw him or her violating ROEs [rules of engage-ment]." The other items referred to "injuring orkilling an innocent non-combatant," "unnecessar-ily destroying private property," "mistreating anon-combatant," and "stealing from a noncomba-tant." Report intentions, together with ethical lead-ership and ethical culture (see below), were ratedfrom 1 ("strongly disagree") to 5 ("strongly agree").

Ethical leadership and ethical culture. We mea-sured ethical culture as a reference shift composition

construct (i.e., items targeted the tmit level [Chan,1998]) using the ten-item ethical culture scale devel-oped by Treviño et al. (1998). Sample items include"Penalties for unethical behavior are stricfly enforcedin this unit" and "The expressed ethical standardsserve as "window dressing" only in this unit" (re-verse-coded). Ethical leadership was also measuredas a reference shift composition consfruct, via theten-item scale developed by Bro-wn et al. (2005). Sam-ple items include "My leader discusses ethics or val-ues with unit members" and "My leader sets an ex-ample of how to do things the right way in terms ofethics." One item ("disciplines unit members whoviolate ethical standards") was omitted from the scalein the analyses because it too closely overlapped withcontent in the ethical culture scale. Ethical cultureand ethical leadership were rated independently ateach level. Squad members rated squad ethical lead-ership and ethical culture; squad and section leadersrated platoon ethical leadership and ethical cultiu'e;and platoon leaders and platoon sergeants rated com-pany ethical leadership and ethical culture.

Combat exposure. All hypothesis tests con-trolled for combat exposure, which we measuredusing items adapted from an instrument developedfor combat research by Hoge, Castro, Messer,McCurk, Cotting, and Kofftnan (2004). Respon-dents were asked to rate the number of times theyhad the following experiences (0, "never," to 5,"five times or more"): "an improvised explosivedevice/booby trap has exploded near you," "re-ceived direct fire," "received indirect fire," "havebeen in a threatening situation where you wereunable to respond due to ROE," "other members ofyour unit were killed or seriously wounded in the-ater," and "you have personally seen individualskilled or seriously wounded in theater." For controlpurposes we used the mean of these items for in-dividuals' reports within each unit, at the lowest(squad) level of analysis. We controlled for combatexposure to reduce the likelihood that effects ofethical leadership on our dependent variablescould be explained by psychological distress re-sulting from combat exposure.

Analysis Procedure

We used hierarchical linear modeling (HLM;Bryk & Raudenbush, 1992) to test the hypothesizedrelationships among our study variables. Severalanalyses required simultaneous estimation of re-gression models at three distinct levels of analysis(company, platoon, and squad) to predict the studyoutcomes at the first level of analysis. Because ourhypotheses concern comparisons among groupsboth within and between hierarchical levels, we


grand-mean-centered the independent variables ateach level. For indexes of variance explained, wepresent a pseudo-ñ^ [~R^) statistic (Snijders &Bosker, 1999). This statistic indicates the propor-tional reduction in the total variance of each vari-able across levels of analysis that is derived fromintroducing independent variables at higher levels.

RESULTS

Measurement Models

A number of confirmatory factor analyses (CFAs)were conducted to test the distinctiveness of theconstructs in the study. Given our focus on squad-level ethical cognitions and behaviors, we used themean of squad members' ratings to conduct the firstset of CFAs. Because of the relatively small samplesizes, we reduced the total number of indicators ineach CFA model by using item parcels as indicatorsof the latent variables (Landis, Beal, & Tesluk,2000). We formed item parcels by randomly com-bining two or three items into one parcel. Thisyielded a total of 24 indicators for the eight latentvariables, which is suitable for a sample of 243squads. The congeneric measurement model,which specifies that indicators load only onto theircorresponding latent variables, fit the data well(A^[224] = 393.86, CFI = .97, NNFI = .97, SRMR =.05). The average variance extracted (AVE) for eachlatent factor ranges from .56 (transgressions againstthe Army) to .91 (ethical leadership) with a mean of.74. The AVE for each construct exceeds the con-ventional criterion of .50 and is larger than thesquare of any correlation between the constructs,thus supporting the convergent and discriminantvalidity of the construct measures (Fornell &Larcker, 1981). In addition, we tested alternativemodels by specifying the indicators of two or moreof the constructs as loading onto the same latentvariable. None of the possible alternative models fitthe data as well as the hypothesized, conge-neric model.

We then tested a CFA at the platoon level [n =96) that specified a congeneric model of ethicalleadership and ethical culture only, as these are theonly factors at the platoon level. The results indi-cate a good fit for this model (/[8] = 10.43, CFI =.99, NNFI = .99, SRMR = .04). The AVEs for theethical leadership and ethical culture latent vari-ables are .74 and .91, which is larger than thesquared correlation between them (.31). The smallsample of companies (n = 42) provides less confi-dence in the robustness of CFA findings. We nev-ertheless specified the CFA model at the companylevel using the same item parcels for ethical lead-

ership and ethical culture. The fit with the data wasquite strong (x^[8] = 11.45, p = .18, SRMR = .05,CFI = .99, NNFI = .97). The AVEs for the two latentfactors of ethical leadership and ethical culture are.85 and .64, which is substantially higher than thesquared correlation between them (.42).

Data Aggregation

We conceptualized ethical leadership and ethicalculture as composition constructs, assessing theextent to which members of a unit share percep-tions of these constructs. Thus it was necessary toexamine evidence for restricted within-unit vari-ance as a justification for aggregating measures ofthese constructs to the relevant unit level (James,1982; Kozlowski & Hattrup, 1992). Two indexesassess restricted within-unit variance: (a) r^^giß, anindex of within-group consensus or agreement(James, Demaree, & Wolfe, 1984) and (b) the intra-class correlation coefficient, ICCl, an index ofwithin-group consistency, or interrater reliability(Bliese, 2000). The r^g^ß index is known to be bi-ased downwards when unit sizes are small (e.g., n= 5) (Kozlowski & Hattrup, 1992). Because ourunits were relatively small, we followed other lead-ership research (Den Hartog, House, Hanges, &Ruiz-Quintanilla, 1999) and used the maximum ob-served variance as the "expected value" (Lindell &Brandt, 1997) in the computation of r^g^j^ for eachconstruct within each unit of interest. Kozlowskiand Klein characterized this method of justifyingaggregation as a "construct-by-group approach"(2000: 35). Median agreement for squad-level com-position constructs was .89 for ethical leadershipand .84 for ethical culture; platoon-level medianagreement was .82 for ethical leadership and .81 forethical culture; and company-level median agree-ment was .90 for ethical leadership and .78 forethical culture. Thus, unit consensus exceededconventional standards for data aggregation(Bliese, 2000).

ICCl indexes the effect size of unit membership,and ICC2 indexes the reliability of an aggregatedunit-level variable (Bliese, 2000). Kozlowski andKlein (2000) characterized this theoretical ap-proach to aggregation as a "construct-by-sample ap-proach" (2000: 35), as the indexes are based on theratio of between-unit and total overall variance,rather than being unit specific. In the case of oursampling plan, however, we have assessments ofthree different target levels by different raters; thus,the "sample" is different at each level. The ICCland ICC2 values for ethical leadership and ethicalcultural elements were .12 and .36 (ethical leader-ship), and .14 and .41 (ethical culture) for the squad


level; 09 and .19 (ethical leadership] and .32 and.52 (ethical culture] for the platoon level; and .17and .29 (ethical leadership] and .22 and .36 (ethicalculture] for the company level. The range of theseICC values is comparable to those reported in theliterature and indicates that these data are suitablefor aggregation (Bliese, 2000].

Whereas ethical leadership and ethical culturalelements were conceptualized as composition con-structs that necessitated empirical justification foraggregation, we conceptualized the outcome con-structs as compilation variables that are descriptiveof individual-level beliefs, intentions, and rates ofobservations in a given unit. These constructs arenot necessarily shared phenomena and thus an ex-amination of restricted within-unit variance is notneeded or proper to meaningfully aggregate them tothe squad level (Bliese, 2000; Kozlowski &Klein, 2000].

Variance Components

In a final preliminary step, we conducted a vari-ance components analysis to examine whethervariance at the lower hierarchical levels was ade-quate to justify multilevel analyses predicting out-comes at these levels. We examined two nestedmodels: platoons nested in companies, and squadsnested in platoons and companies (see Table 1]. Forthe platoons nested in companies model, both vari-ables included at this level (ethical leadership andethical cultinre] showed adequate variance at theplatoon level (67.9 and 87.9 percent, respectively].Similarly, all squad-level ethical outcomes dis-played adequate variance at the higher and lower

levels of analysis, as did squad-level ethical cul-ture. Almost all squad-level variance in ethicalleadership, however, was at the level of the squad.We nevertheless proceeded in testing Hypothesis6, which pertains to platoon ethical leadership ef-fects on squad ethical leadership, because there issubstantial variance in the outcomes and predictorsof squad ethical leadership, and thus multilevelanalyses are required to correctly test the appropri-ate models (see LaHuis & Ferguson, 2007].

Hypothesis Testing

Table 2 shows the means, reliabilities, standarddeviations, and correlations of the variables. Hy-pothesis 1 posits direct effects of leaders that aresuch that ethical leadership will be positively re-lated to ethical culture within each hierarchicallevel. As Table 3 shows, ethical leadership wasassociated with ethical culture at the company, pla-toon, and squad levels [y = .69, p < .01, y = .56,p < .01, and y = .45, p < .01, respectively). Hy-pothesis 1 is therefore supported.

Hypothesis 2 proposes that ethical leadershipwill be related to outcomes indirectly through eth-ical culture at the lowest level (e.g., Ll -» Cl ^ 01].Table 4 shows the results of our tests of Hypothesis2. We foimd that squad ethical culture provided anindirect path from squad ethical leadership tosquad ethical outcomes in terms of report inten-tions (7 = .06, p < .05], moral efficacy (7 = .09,p < .05], transgressions against the Army (7= -.06, p < .05], and peer exemplary behavior (7= .11, p < .01]. There was no significant indirecteffect on transgressions against noncombatants (7

TABLE 2Means, Standard Deviations, and Correlations among Study Variables"

Variable Mean s.d. 10 11 12

1. Squad ethical leadership2. Squad ethical culture3. Moral efficacy4. Peer exemplary behavior5. Transgressions vs. the Army6. Transgressions vs. noncombatants7. Report intentions8. Combat exposure9. Platoon ethical leadership

10. Platoon ethical culture11. Company ethical leadership12. Company ethical culture

3.222.966.530.460.410.263.600.903.403.033.613.32

0.580.361.260.410.380.340.610.620.670.470.620.49

.96

.50*

.19**-.01-.28**-.17**

.27**-.18**

.17*

.11

.13

.15**

.74

.23**

.10-.30**-.13

.15*-.08

.36**

.20*

.26**

.35**

.95

.01-.11-.20**

.37**-.07

.10

.06

.05-.04

.72

.32**

.39**-.08

.64**

.22**

.12

.00

.07

.82

.69**

.11

.35**

.01-.17**-.11-.12

.80-.37**

.39**

.03-.13-.06

.01

.96

.42**

.08-.05-.09-.06

.88

.16*-.01-.06

.05

.95

.57**

.14

.53**

.87

.09

.21*.94.69 .88

" Rows 1-8 represent within-squad correlations. Rows 9-10 represent correlations between platoon leadership and culture variables andsquad variables aggregated to the platoon level. Rows 11-12 represent correlations between company leadership and culture variables andsquad and platoon variables aggregated to the company level.

Values on the diagonal in italic are alpha reliability esümates. For the company-level correlations, n = 39; for the platoon-levelcorrelations, n = 78; and for the squad-level correlations, n = 172.

* p < .05**p < .01


TABLE 3Relationships between Ethical Leadership and Ethical

Gulture within Hierarchical Levels (Hjqpothesis 1)"

Dependent Variable: Ethical Culture''

Variables

Company-level modelEthical leadership~R^Platoon-level modelEthical leadership-R^Squad-level modelEthical leadership

ß

.69**

.47

.56**

.24

.45**

.31

5.74

4.18

6.29

" All effects are standardized. In all the analyses, combatexposure was entered as a control variahle (not included in thistable). For the squad level, n = 243; for the platoon level, n = 85;and for the company level, Í7 = 40.

^ Ethical culture refers to the culture at the same level as theindependent variable; e.g., for the company-level model, thedependent variable is company ethical culture.

**p < .01

= - . 0 1 , p < .20). Hypothesis 2 was therefore sup-ported for all outcomes except one. In addition tothese indirect effects, ethical leadership also hadsignificant direct effects on report intentions (7= .14, p < .05) and transgressions against the Army(7 = -.26, p < .05).

Hypothesis 3 predicts a positive relationship be-tween higher- and lower-level ethical culture (C3-^ C2 and C2 —> Cl). To more rigorously test the

cascading relationships of ethical culture as repre-sented in Hypothesis 3 and also to test subsequenthypotheses involving multilevel effects, we simul-taneously estimated the effects of the higher-levelvariables (platoon and company ethical culture) oncorresponding variables measured at the lower lev-els. This relationship was significant for ethicalculture in all three hierarchical paths: company toplatoon (7 = .26, p < .05), company to squad (7= .37, p < .01), and platoon to squad (7 = .23, p <.01). Hypothesis 3, pertaining to the cascading ef-fects of ethical culture, is therefore supported. Ta-ble 5 presents a simimary of these analyses.

Hypothesis 4 predicts a positive relationship be-tween higher-level ethical leadership and lower-level ethical culture (L3 -^ C2 or L2 -^ Cl). Table 6displays the results of a series of models testing thishypothesis. Higher-level ethical leadership was re-lated to ethical culture at lower levels (7 = .21, p <.10, for the company to platoon model; 7 = .22, p <.05, for the company to squad model; and 7 = .41,p < .01, for the platoon to squad model). Thesefindings together support Hypothesis 4.

Hypotheses 5 and 6 elaborate Hypothesis 4 byspecifying indirect relationships between ethicalleadership and lower-level ethical culture transmit-ted through ethical culture at the same (higher)level (H5; L3 ^ C3 -» C2 or L2 ^ C2 -^ Cl) andthrough lower-level ethical leadership (H6; L3 -^L2 -^ C2 or L2 ^ Ll ^ Cl). In tests of Hypothesis5, only one out of the three possible indirect rela-

TABLE 4Ethical Leadership Effects on Squad Member Outcomes through Squad Ethical Culture (Hypothesis 2)°

Variables

Squad-level main effectsEthical leadership~R^Ethical culture~R^Mediated modelEthical leadershipEthical culture~R^Decomposition of ethical

leadership effectsTotalDirectIndirect

ReportIntentions

ß

.20**

.14

.22**

.14

.14*

.14*

.15

.20

.14

.06*

t

3.23

3.06

2.241.90

1.84

MoralEfiBcacy

ß t

.06 0.75

.00

.17** 2.29

.01

-.05 -0.50.20* 2.02.01

.04-.05

.09* 1.95

Transgressionsvs.

Noncombatants

ß *

-.14* -2.02.17

-.10 -1.52.16

-.12 -1.38-.03 -0.43

.17

-.13-.12-.01 0.43

Transgressionsvs.

the Army

ß t

-.32** -4.33.22

-.28** -4.25.19

-.26* -2.56-.13* -1.78

.24

-.32-.26-.06* 1.73

Peer ExemplaryBehavior

ß t

.07 0.90

.22

.22** 3.08

.26

.17 -0.38

.24** 3.21

.26

.28

.17

.11** 2.88

All coefficients are standardized. Combat exposure was included as a control variable (not shown in this table), n = 243.

* p < .05**p < .01


TABLE 5Standardized Relationships between Higher- and

Lower-Level Ethical Culture (H5rpothesis 3)"

Dependent Variable:Lower-Level Ethical Culture"*

TABLE 6Standardized Effects of Higher-Level EthicalLeadership on Lower-Level Ethical Culture

(Hypotheses 4 and 5)"

Variables

Company to platoon modelCompany ethical culture

Company to squad modelCompany ethical culture

Platoon to squad modelPlatoon ethical culture

.26*

.04

.37*«

.18

.23*

.10

2.31

5.30

3.12

" All coefficients are standardized. Combat exposure is in-cluded as a control variable (not shown in this table), n = 7Ö forthe platoon level and JI = 39 for the company level.

^ Lower-level ethical culture refers to the lower level of themodel tested; e.g., for the company to squad model, the depen-dent variable is squad-level ethical culture.

* p < .05** p < .01

tionships was significant. Specifically, companyethical leadership was indirectly related to squad-level ethical culture through company ethical cul-ture. There was no significant indirect effect ofhigher-level ethical leadership on lower-level ethi-cal culture for the company to platoon or the pla-toon to squad models. Thus, Hypothesis 5 is onlypartially supported. Regarding Hypothesis 6, theindirect path from company-level ethical leader-ship to platoon-level ethical culture through pla-toon-level ethical leadership was significant (7= .11, p < .05). The indirect path from platoon-level ethical leadership to squad-level ethical cul-ture through squad ethical leadership was also sig-nificant [y = .08, p < .01). Table 7 presents thesefindings, which overall support Hypothesis 6.

In supplementary analyses, we examinedwhether ethical leadership has any direct influenceon lower-level ethical culture after the indirect ef-fects specified in Hypotheses 5 and 6 have beenaccounted for. Accounting for the cascading influ-ence of company and platoon ethical culture (Hy-pothesis 5) fully explained the direct effect of com-pany ethical leadership on squad ethical culture.However, the direct influence of platoon ethicalleadership on squad ethical culture remained, evenafter we further accounted for the indirect effectthrough squad ethical leadership (Hypothesis 6).Insofar as neither lower-level ethical leadership norsame-level ethical culture explain this platoon tosquad relationship, it can legitimately be labeled a"bypass effect."

Variables

Company to platoon model—Main effectsCompany ethical leadership~R^Company to squad model—Main effectsCompany ethical leadership~R^Platoon to squad model—Main effectsPlatoon ethical leadership~R^Company to platoon model—Indirect

effectsCompany ethical leadershipCompany ethical culture~R^Decomposition of company ethical

leadership effectsTotalDirectIndirectCompany to squad model—Indirect

effectsCompany ethical leadershipCompany ethical culture

Decomposition of company ethicalleadership effects

TotalDirectIndirectPlatoon to squad effects—Indirect

effectsPlatoon ethical leadershipPlatoon ethical culture~R^Decomposition of platoon ethical

leadership effectsTotalDirectIndirect

Dependent'Variable:Lower-Level

Ethical Culture""

ß

.21^

.01

.22*

.07

.41**

.18

-.03.25.01

.21-.03

.15

-.03.41**.18

.22-.03

.28**

.35**

.10

.19

.41

.35

.05

f

1.78

2.55

5.26

-0.091.18

1.17

-0.463.88

3.24

3.721.13

1.12

" Combat exposure was entered as a control variable (notshown in this table), n = 172 for the squad level, n = 78 for theplatoon level, and n = 39 for the company level.

^ Lower-level ethical culture refers to the lower level of themodel tested; e.g., for the company to squad model, the depen-dent variable is squad-level ethical culture,

^p < .10* p < .05

** p < .01

Hypothesis 7 proposes that ethical leadership athigher hierarchical levels will moderate the rela-tionship between lower-level ethical leadershipand lower-level ethical culture. As shown in Table


TABLE 7Standardized Higher-Level Ethical Leadership Effects on Lower-Level Ethical Culture through

Lower-Level Ethical Leadership (Hypothesis 6)"

Dependent Variable: Lower-Level Ethical Culture""

Variables ß

Company to platoon model—Main effectsCompany ethical leadership-R^Company to squad model—Main effectsCompany ethical leadership~R^Platoon to squad model—Main effectsPlatoon etbical leadership

Company to platoon model—Indirect effectsCompany ethical leadershipPlatoon ethical leadership

Decomposition of ethical leadership effectsTotalDirectIndirectCompany to squad model—Indirect effectsCompany ethical leadershipSquad ethical leadership~R^Decomposition of ethical leadership effectsTotalDirectIndirectPlatoon to squad model—Indirect effectsPlatoon ethical leadershipSquad ethical leadership~R^Decomposition of ethical leadership effectsTotalDirectIndirect

.02

.15

.03

.22*

.05

1.69

1.39

2.99

.011.78

22"07

41**18

0854**33

200811*

19**45**18

454503

32**42**20

413208**

2.55

5.26

0.784.10

1.58

2.806.31

1.37

4.605.99

2.71

" Combat exposure was entered as a control variable (not shown in this table), n = 172 for the squad level, n = 76 for the platoon level,and /I = 39 for the company level.

^ Lower-level ethical culture refers to the lower level of the model tested; e.g., for the company to squad model, the dependent variableis squad-level ethical culture,

^p < .10* p < .05

**p < .01

8, a significant moderating effect was found forboth the company to platoon and the platoon tosquad model (interaction y = .22, p < .05, and y— .08, p < .05, respectively). Specifically, althoughthe effects of lower-level ethical leadership are pos-itive and significant in each case, the slopes aresignificantly stronger when higher-level ethicalleadership is high rather than low, thus supportingHypothesis 7. Figures 2 and 3 illustrate theserelationships.

To follow up on our analyses of Hypothesis 7, weconducted two supplementary analyses. First, weexamined whether higher-level ethical culturemoderates the relationship between lower-level

ethical leadership and ethical culture at the lowerlevel (similar to the hypothesized moderating roleof higher-level ethical leadership). Second, we as-sessed whether the moderating effects of higher-level ethical leadership discussed in Hypothesis 7are mediated through higher-level ethical culture.Higher-level ethical culture was found to moderatethe effects of lower-level ethical leadership in away that was similar to the moderating effect ofhigher-level ethical leadership. Platoon-level ethi-cal leadership and squad-level ethical leadershipexhibit a stronger relationship to platoon and squadethical culture when company and platoon ethicalculture levels are both high (interaction y = .34, p


TABLE 8Interactions of Higher- and Lower-Level Ethical

Leadership Predicting Lower-Level Ethical Culture(Hypothesis 7)"

Variables

Company to platoon modelCompany ethical leadershipPlatoon ethical leadershipCross-level interaction term~R^Company to squad modelCompany ethical leadershipSquad ethical leadershipCross-level interaction term~R^Platoon to squad modelPlatoon ethical leadershipSquad ethical leadershipCross-level interaction term~R^

Dependent Variable:Lower-Level Ethical Culture""

Y

.05

.39**

.22*

.56

.11**

.27**

.04

.34

.17**

.26**

.08*

.37

f

0.545.292.01

2.745.570.68

4.776.692.06

" Combat exposure was entered as a control variable (notshown in this table), n = 172 for the squad level, n = 78 for theplatoon level, and n = 39 for the company level.

^ Lower-level ethical culture refers to the lower level of themodel tested; e.g., for the company to squad model, tbe depen-dent variable is squad-level ethical culture.

* p < .05** p < .01

< .05, platoon, and y = .16, p < .05, squad, respec-tively). The simple slopes, although significant ineach case, were significantly stronger for high eth-

ical culture than for low ethical culture (p < .05);Figures 4 and 5 illustrate. Findings from this set ofmediated moderation analyses indicate higher-level ethical culture mediates the moderating ef-fects of higher-level ethical leadership. This sug-gests that higher-level ethical culture, as a moreproximal moderator of lower-level ethical leader-ship effects, is a mechanism that may more fullyexplain the effects of ethical leadership that werespecified in Hypothesis 7.

DISCUSSION

Ethical leadership represents a domain of leaderbehavior that accords with what Schein (1985,2010) described as primary mechanisms throughwhich leaders embed their expectations and as-sumptions into the fabric of an organization.Schein's model, however, does not precisely spec-ify how these embedded understandings and ex-pectations of leaders become suffused tbrough mul-tiple organizational levels and influence followeroutcomes. Our study builds on Schein's theoreticalframework in a way that begins to lay out the spe-cific modes of transmission through which ethicalleadership at different hierarchical levels may per-meate an organization. In our model, the role ofethical culture is essential to understanding theserelationships.

We described and tested three basic categories ofethical influence transmission down and acrossan organizational hierarchy. First, we found that

FIGURE 2Interaction of Company-Level Ethical Leadership and Platoon-Level Ethical Leadership Predicting

Platoon-Level Ethical Culture"

3.5

PlatoonEthicalCulture

3.3

3.1

2.9

2.7

2.5Low Platoon Ethical Leadership High Platoon Ethical Leadership

•Low company ethical leadership High company ethical leadership

" For low company ethical leadership, simple slope = .26, f = 2.54, p < .05. For high company ethical leadership, simple slope = .53,Í = 4.29, p < .01.


FIGURE 3Interaction of Platoon-Level Ethical Leadership and Squad-Level Ethical Leadership Predicting

Squad-Level Ethical Culture*

SquadEthicalCulture

3.5

3.3

3.1

2.9

2.7

2.5Low Squad Ethical Leadership High Squad Ethical Leadership

• Low platoon ethical leadership • • •• • High platoon ethical leadership

" For low platoon ethical leadership, simple slope = .20, t = 3.81, p < .01. For high platoon ethical leadership, simple slope = .31.t = 8.07, p < .01.

FIGURE 4Interaction of Company-Level Ethical Culture and Platoon-Level Ethical Leadership Predicting

Platoon-Level Ethical Culture"

3.5

3.3

Platoon 3.1EthicalCulture

2.9

2.7

2.5

Low Platoon Ethical Leadership High Platoon Ethical Leadership

»Low company ethical culture • • #t • High company ethical culture

° For low company ethical culture, simple slope = .21, í = 2.12, p < .05. For high company ethical culture, simple slope = .56, í = 4.12,p < .01.


FIGURE 5Interaction of Platoon-Level Ethical Culture and Squad-Level Ethical Culture Predicting

Squad-Level Ethical Culture"

3.5

3.3

3.1

SquadEthicalCulture 2.9

2.7

2.5Low Squad Ethical Leadership High Squad Ethical Leadership

»Low platoon ethical culture • • •• • High platoon ethical cultiu-e

" For low platoon ethical cultiire, simple slope = .06, f = 1.02, p > .20. For high platoon ethical culture, simple slope = .23, f = 3.90,p < .01.

ethical leaders embed their expectations into theethical culture that is shared and understood bymembers of their own units (Hypothesis 1).Through this mechanism, ethical leaders indirectlyinfiuence their immediate followers' cognifions (e.g.,moral efficacy) and behaviors (e.g., transgressions)through unit ethical culture (Hypothesis 2). Thus,ethical leadership appears to infiuence ethical cul-ture within each hierarchical level, and much ofthe infiuence of leaders at each level is indirect,transmitted through ethical cultvu'e at the same hi-erarchical level. Second, ethical culture cascadesacross hierarchical levels to a substantial extent(Hypothesis 3), thus enabling a new and previouslyunstudied pathway for leadership influence totransmit across levels. Further, we predicted (Hy-pothesis 4) and found that ethical leadership ispositively related to ethical culture at lower levels.Direct influences of leader behavior on the behav-ior of followers at lower levels have been examinedpreviously (e.g., Dvir et al., 2002). Of greater inter-est in this study was to determine whether ethicalleadership's influence is embedded through effectsof ethical culture at the same level (Hypothesis 5)and/or through effects of lower-level ethical lead-ership (Hypothesis 6) in ways that explain theseoverall cross-level effects of ethical leadership on

the lower-level ethical culture. Both hypotheseswere partially supported, and of three overall cross-level direct effects stated in Hypothesis 4, onlyone—the connection between platoon ethical lead-ership and squad ethical culture—was not ex-plained by indirect processes transmitted throughlower-level leadership or higher-level ethicalcu.lture.

Third, we found that leaders who exhibit a highlevel of ethical leadership may facilitate the influ-ence of subordinate leaders' ethical leadership ontheir followers (Hypothesis 7). Post hoc analysesshowed that ethical culture at the higher level me-diated the facilitating effect of higher-level ethicalleadership. This highlights the importance of con-gruence in ethical leadership across multiple levelsand how ethical leaders depend on support fromtheir principals. Such cross-level interactions arethought to be critical for understanding complexsystems (Page, 2007), but they have rarely beenexamined generally, nor specifically with respectto ethical leadership and culture. These interac-tions and the direct and indirect effects of ethicalleadership transmitted through ethical culture in-dicate that a more holistic view of both proximaland distal leadership processes that impact fol-lower behavior is required to more fully appreciate


the locus and transmission of leadership influencein complex organizations.

Table 9 summarizes all the findings concerningthe hypotheses. The most interesting of our find-ings was that local leadership, higher-level ethicalculture, and higher-level ethical leadership influ-ence local luiderstandings of ethical culture. Thesefindings suggest that much more attention shouldbe paid to the role of shared understandings aboutethical conduct (ethical culture) at each hierarchi-cal level in an organization.

A Multilevel Systems Perspective on EthicalLeadership and Ethical Culture

From a meta-theoretical perspective, the notionof organizations as social entities of interacting andcoacting elements operating at multiple levels of asystem is one with a long history in the organiza-tional sciences, one going back at least to Roethlis-berger and Dickson (1939) and apparent in the writ-ings of the many systems theorists of the 1950s and1960s (e.g., Likert, 1961; Katz & Kahn, 1966). Apersistent theme in the writings of these and otherscholars is the central role of leaders in the enact-ment of a system of human relations across levels of

organizations. Likert's concept of leaders as linkingpins who transmit values and shape cultural inte-gration goes to the core of "classic" organizationalsystems theory. Although leadership is widelyviewed as central for linking organizational sys-tems together, for the most part it has not beenresearched from a multilevel theory and integratedsystems perspective (Yammarino & Dansereau,2008). Such research is critical to evaluating systemcohesion as a fundamental assumption of organiza-tional behavior (Kozlowski & Klein, 2000).

Owe model included three levels of leadership ina complex organization which, combined with alarge sample size, allowed us to establish how eth-ical leadership becomes embedded in a multilevel,multiperson system with both direct effects of lead-ers on their immediate followers and indirect,cross-level effects on lower-level organizationmembers. Research and scholarship about leader-ship has thus far examined the processes throughwhich leaders suffuse their influence in an organi-zation primarily by referencing the trickle-downrelationships between leaders' behavior at differentlevels (Bass et al., 1987; Mayer et al., 2009) and/ordirect effects of behavior across levels (e.g., Dvir etal., 2002). By integrating horizontal (within-level)

TABLE 9Summary of Hypothesis Test Results"

Hypothesis Figure 1 Relationships Results

HI. Ethical leadership is positively related to ethical culture within eachhierarchical level.

H2. Through ethical culture, ethical leadership is indirectly and positivelyrelated to unit members' moral agency (moral efficacy and reportintentions) and exemplary ethical behavior and negatively related totheir ethical transgressions.

H3. Ethical culture at higher hierarchical levels is positively related toethical culture at lower hierarchical levels.

H4. Ethical leadership is positively related to ethical culture at lowerhierarchical levels.

H5. Higher-level ethical leadership indirectly influences ethical culture atlower hierarchical levels through ethical culture at the same (higher)hierarchical levels.

H6. Ethical leadership at higher organizational levels is indirectly relatedto ethical culture at lower hierarchical levels through ethicalleadership at lower levels.

H7. Ethical leadership at higher hierarchical levels moderates theinfluence of subordinate ethical leadership in such a way that highethical leadership at a higher organization level is associated with astronger influence of subordinate leaders' ethical leadership on ethicalculture in their units.

L l ^ C lLl -^ Cl ^ 01 (moral efficacy)

Ll —> Cl ^ 01 (report intentions)Ll —> Cl —>• 0 1 (transgressions vs. Army)

Ll -» Cl -» 0 1 (transgressions v. NC)Ll —> Cl - • 0 1 (exemplary behavior)

C3 ^ C 2

C3 ^ C l

L2->C1L3 ^ C3 -> C2L3 ^ C3 -^ ClL2 ^ C2 -^ ClL3 ^ L2 -* C2L2 -• Ll ^ Cl

L3 X L2L2 X Ll

•C2•Cl

SS

SSNSSSSSsssNSNSS

SS

" See Figure 1 for path diagram. Abbreviations: C = "ethical culture"; L = "ethical leadership"; N = "not supported"; NC ="noncombatants"; O = "outcomes"; S = "supported." Numbers after abbreviations denote hierarchical level; e.g., L3 denotes ethicalleadership at company (3rd) level.


influence processes with vertical and diagonalcross-level linkages, we are proposing ways notonly to study ethical leadership in organizations,but also to highlight the different transmissionmechanisms for leadership constructs moregenerally.

Horizontal (within-level) effects. We found sfronghorizontal, within-level effects of ethical leadershipon ethical culture at each of the three levels. Suchresults are consistent with the widely held assimip-tion that the ethical leadership of an individual'simmediate supervisor affects the proximal ethicalculture of the individual's unit. Further, at the lowestlevel, squad-level ethical culture was significanflyrelated to fovir out of five of the outcome measures,including report intentions, moral efficacy, transgres-sions against the Army, and exemplary ethical behav-ior. These simple horizontal effects indicate the valueof selecting and developing ethical leaders at lowerlevels who proactively maintain favorable ethicalnorms and uphold high standards for ethical conductin their own units.

Vertical direct and indirect effects of embed-ded ethical culture and leadership. Our modelspecifying the indirect transmission of leadershipinfluence has two core elements: First, leader be-haviors at higher levels tend to be similar to leaderbehaviors at adjacent lower levels. Second, we ob-served that more senior leaders influence lower-level leaders' and followers' behaviors and cogni-tions via indirect pathways. One of these indirectpathways observed in the cvu"rent study wasthrough the ethical culture at adjacent lower levels.Ethical culture at the highest hierarchical level(company, in this study) was positively related toethical culture at both lower levels, as per Hypoth-esis 3. This served as a foundation for subsequenthypotheses concerning indirect effects of ethicalleadership on lower-level outcomes.

Cross-level indirect effects. In addition to thevertical and horizontal effects proposed in ovirmodel, the diagonal dashed lines in Figure 1 rep-resent direct effects of higher-level ethical leader-ship on lower-level ethical culture. Company-levelethical leadership exhibited significant diagonal ef-fects on ethical culture at both the platoon and thesquad levels, and platoon-level ethical leadershipalso directly influenced squad ethical culture (Hy-pothesis 4). This relationship would convention-ally be interpreted as a bypass of the lower-levelleaders, but we suggest that is a rather liberal inter-pretation unless one also accounts for mechanismsthat may explain such direct cross-level effects,such as mediation through lower-level leaders' be-havior or through the cascading influence of sharedculture. We found that the indirect relationship of

company ethical leadership through platoon ethi-cal leadership fully accounted for the relationshipbetween the former and platoon ethical culture. Inaddition, accounting for the cascading influencesof ethical culture explained the aforementioneddirect effect of company ethical leadership onsquad ethical culture. The authors of one previ-ous study observed bypass effects of leader be-havior after accounting first for cascading leader-ship to the next level (Yang et al., 2010), but noprevious study has simultaneously accounted forindirect pathways through shared cultural ele-ments in which leaders embed their influence.This suggests the possibility that bypass effects ofleader influence across levels may be at best aweak phenomenon once all embedding processes(indirect relationships through lower-level lead-ers and same-level culture elements) are takenfully into consideration.

One direct cross-level relationship, that betweenplatoon ethical leadership and squad ethical cul-ture, persisted after we accounted for the otherhypothesized influences on squad ethical culture(i.e., the simultaneous effects of L2 ^ Ll —> Cl andL2 —> C2 —» Cl). Even using our more conservativedefinition and test, this qualifies as a leadershipbypass. This finding is similar to the results pre-sented by Dvir et al. (2002) in that the platoonleaders' transformational leadership in their studyhad a significant effect on their indirect followers'behavior, but not on the behavior of their immedi-ate followers (squad leaders). Explaining direct ef-fects of this nature remains an interesting questionfor future research. One interpretation may be thatsenior leaders serve as more powerful role modelswhen they are similar to and represent a potentialpossible self for subordinate leaders (Bandura,1971). In the current study, the leaders at level 1(squad leaders) were all noncommissioned officers(NCOs), whereas the formal leaders at level 2 (pla-toon leaders) were commissioned officers. Themost salient role models to the squad leaders mayhave instead been the informal leaders who werethe more senior NCOs in the organization (i.e., pla-toon sergeants and first [company] sergeants). Thisalternative senior leadership referent may have at-tenuated squad leaders' emulation of their platoonleaders, which in turn enabled platoon leaders toinfluence squad ethical culture in distinct ways. Aswe noted above under the heading "Variance com-ponents," although platoon ethical leadership ex-plained a significant portion of the between-squadvariance at tbe platoon level, the total variance ofsquad ethical leadership at the platoon level wasvery small.

Outcomes of ethical leadership and ethical cul-


ture. Our study examined the extent to which eth-ical leadership at different levels, both directly andthrough ethical culture, is associated with ethicaltransgressions, exemplary behaviors, and the moralagency cognitions of moral efficacy and reportingintentions. In measuring transgressions and peerexemplary behavior, we aggregated individuals' re-ports about frequencies of certain bebaviors occvir-ring among members of their squads. We expectedthis approach to enhance reliability of reportingand encourage more honest reporting than might beexpected if soldiers were reporting about them-selves or specific peers. Transgressions includedmistreatment of noncombatants as well as the moregeneral rule breaking that typifies behavior in manyother organizations, such as thievery, misreportingperformance, and defying directives. Peer exem-plary behavior and moral efficacy showed thestrongest positive relationships with squad ethicalculture, and transgressions against the organization(e.g., stealing) were negatively related to squad-level ethical culture.

Moral efficacy and report intentions are manifes-tations of moral agency, reflecting orientations torespond in purposeful ways to opportunities toconfront unethical behavior (Hannah et al., 2011).Ethical context may influence these dispositions,as we observed significant, positive effects of bothsquad- and platoon-level ethical culture on squadmean levels of moral efficacy and report intentions.To ovu" knowledge, this is the first empirical studylinking ethical leadership to such constructs ofmoral agency. To better test causality, future re-search might profitably examine baseline levels ofmoral efficacy and report intentions before follow-ers are exposed to particular levels of ethical lead-ership and ethical culture that may encoiuage themto be more (or less) effective moral agents.

Importance of Context: Ethical Cultureand Multilevel Leadership

In keeping with the embedding of leadership asdescribed by Schein (1985, 2010), much of the in-fluence of ethical leadership on ethical outcomesthat was observed in this study was mediated byunit-level ethical culture. Consequently, models ofleadership and ethical behavior that omit the ef-fects of ethical culture at different hierarchical lev-els may be underspecified. For example, one migbtconclude that senior leaders have a direct influenceon outcomes at a lower level that results from directinteractions between these leaders and lower-levelfollowers (Yammarino, 1994), whereas the influ-ence of these leaders may in fact be indirect, trans-

mitted through their influence on culture at theirown levels, which then cascades to lower levels.

This study also begins to address calls for exam-ining leadership in context (e.g., Avolio, 2007); inthis instance, shared cultural elements (Schein,1985) relating to ethical conduct represents con-text. Importantly, the effects of lower-level ethicalleadership on ethical culture at the same level werecontingent on ethical leadership at the higher level.Ethical leadership at a lower level had a strongerpositive influence on ethical culture at that levelwhen the leader at the next higher level was re-ported to exhibit a high level of ethical leadership.Additional analyses indicated that this facilitatingeffect of higher-level ethical leadership may occurthrough the mediating role of ethical culture. Spe-cifically, ethical leadership at higher levels pro-motes ethical culture at those levels, which in turnfacilitates a more favorable impact of etbical lead-ership on ethical culture at the next lower level.Thus, when studying ethical leadership in the con-ventional way, focusing chiefly on the influences ofthe immediate leader, one can potentially be mis-led into drawing inferences about the magnitude ofthese effects, when in fact they may be to someextent contingent on qualities of ethical leadershipat higher levels. This moderating effect of higher-level ethical leadership suggests that senior leadersplay a significant role in cultivating shared ethicalunderstandings at lower levels, working not onlythrough direct effects and indirect influencesthrough ethical culture, but also by constraining orfacilitating the influence of ethical leadership atlower levels.

Practical Implications

Leaders should be encouraged to embed tbeirassumptions and expectations concerning ethicalconduct among organizational members by engag-ing in ethical leadership and by seeking to createstrong ethical cultures in their units. Through themechanisms examined in the current study, thisleadership strategy may influence ethical culture atlower levels. As conceptualized by Brown and col-leagues (2005), ethical leadership involves many ofthe primary and secondary mechanisms thatSchein (1985, 2010) noted are critical for a leader toembed influence. Specifically, ethical leaders payattention to providing contingent consequences fordesired eind undesired bebavior. Tbey serve asmodels and coaches to followers in ways that en-courage them to exhibit exemplary ethical behaviorand to avoid unethical behavior. Critical incidentsconcerning ethical conduct, such as severe em-ployee infractions, can be bighlighted as a key


means to teach followers about ethical standardsand sanctions and to reinforce the message thatthey can be effective moral managers and agents.Ethical leaders can also allocate resources to im-prove followers' ethical understandings, while alsousing levers such as selection, promotion, training,and performance appraisals in ways that reinforcethese actions. An ethical leader can further rein-force such understandings by utilizing Schein'ssecondary mechanisms, such as telling engagingstories about ethics that are likely to be repeated byothers, or using formal means, such as speechesand policies, to communicate expectations.

Using the concept of ethical culture was helpfulto our understanding and ability to explain theinfiuences of ethical leaders within and across lev-els. Organizational leaders might pay more atten-tion to the role that ethical culture plays in estab-lishing the standards and boundaries for ethicalconduct in their organizations. These shared ele-ments can be tracked by surveying organizationmembers to determine whether leaders are creatingunit ethical cultures that support ethical behavioror whether they are creating obstacles to it thatmight reduce followers' moral agency and increasethe frequency and severity of unethical conduct.

Another important practical implication of thisstudy is that to understand the infiuence of a directleader, it may be necessary to look up one or morelevels in an organization to obtain a more completepicture of the locus ofthat leader's style and effects.Our results suggest that building the full leadershipcapacity of an organization requires viewing lead-ership more as an integrated system of relation-ships that operate across hierarchical levels, drivensubstantially by both leader behaviors and ethicalculture that are present within and across levels.

It is also important for leaders to understand howtheir inñuence on their direct reports who are lead-ers themselves can be used to transmit and rein-force their indirect leadership to followers at lowerlevels of their organization. Such leadership mayfocus on aligning performance and ethical expecta-tions or emphasize the importance of a particular"guiding" message that the more senior leaderwants all lower-level leaders and organizationmembers to embrace. To lead, therefore, requiresthat an individual think about both direct and in-direct impacts as represented through his or hersubordinate leaders and the assumptions and ex-pectations the individual is embedding in the eth-ical culture. Furthermore, knowing how seniorleaders can facilitate the effects of subordinateleaders' ethical leadership by exhibiting strong eth-ical leadership themselves reinforces how impor-tant it is for organizations to ensure senior leaders

are selected and developed to exhibit high levels ofethical leadership.

Study Strengths and Limitations

A number of distinct empirical strengths are as-sociated with this study. First, unlike most priorstudies of leadership transmission across levels(e.g., Chun et al., 2009; Mayer et al., 2009], we usedindependent ratings of leader behavior at each or-ganizational level. This may at least partially ex-plain why the leader bebavior cascade effects inour study are somewhat weaker than those found inprevious studies. We also encourage researchers toobtain independent measures of ethical culture atdifferent levels, as we did. Relying upon the samelower-level participants to evaluate shared under-standings about cultural elements (e.g., Zohar &Luria, 2005] or the behavior of leaders (e.g., Mayeret al., 2009] at multiple higher levels potentiallyinflates levels of agreement. This inflation of ob-served relationsbips may occur because of same-source bias but also because participants at lowerlevels may lack adequate information with whichto gauge the behavior of leaders and the cultures ofunits at higher levels.

By measuring ethical culture using distinctsources, we were also able to identify very satis-factory levels of discriminant validity of etbicalculture at different levels. Whereas a general or-ganizational culture may explain a portion of thevariance in ethical culture shared across levels,these cross-level relationships were low to mod-erate in magnitude, indicating that global organ-izational culture could not explain the homoge-neity within units. We also examined whethervariance in company (platoon] ethical leadershipand in ethical culture ratings was related to vari-ance in platoon (squad] ethical leadership or pla-toon (squad] ethical culture ratings. The highestcorrelation, which concerned the relationship be-tween tbe variances of squad and platoon ethicalculture ratings, was quite low (r = .15]. Second,unlike earlier multilevel leadership studies, ouranalyses related to ethical leadership also ac-counted for the effects of subunit ethical culture.Not only did the influences of ethical culturefrom higher levels explain additional variance inoutcomes, but also, after controlling for theseelements, we no longer observed some direct (by-pass] effects of higher-level leadership on out-comes that have been the subject of previousresearch and theory (Yammarino, 1994; Yang etal., 2010]. Third, although the effects of ethicalleadership transmitting through lower-level lead-ers' behavior explained only a small portion of


the variance in outcomes among lower-levelmembers, when also considering indirect ethicalleadership effects that were transmitted throughethical culture at different levels, we found thathigher-level ethical leadership was broadly re-lated to squad members' ethical cognitions andbehavior.

A fourth advantage of our study involved ourexamination of the various links in the chain ofinfluence from company-level leader behavior tosquad-level beliefs and behavior. We took a conser-vative approach toward examining these paths, bytaking only two-segment paths (e.g., variable A tovariable B and variable B to variable C) that werefound to represent statistically significant indirecteffects when we calculated total indirect effects.We pursued this strategy because sampling varia-tion around the indirect effects is too substantial ina multilevel study of this scope to cumulate indi-rect effects that were expanded across multiple bi-variate relationships.

As with any field research project, however, anumber of limitations could impact the interpre-tation of our findings. One issue is the potentialto generalize the findings to other organizationalcontexts. Over 89 million people serve in militar-ies worldwide (International Institute for Strate-gic Studies, 2010). Police officers and firefight-ers, who constitute a large segment of the globalworkforce, in some ways share the often extremecontext of military work (Hannah, Uhl-Bien, Avo-lio, & Cavarretta, 2009). Nevertheless, the U.S.military, like other militaries around the world,is a unique organization. Additional researchwill need to be conducted in other types of or-ganizations before one can be confident as to thegeneralizability of our findings. For example,militaries have hierarchical structures, and thusit would be useful to study more organicallystructured organizations to determine if themeans by which higher-level ethical leadership isconveyed to lower-level followers is different insuch organizations. Another example of how or-ganizational context may influence the general-izability of our findings concerns leader selec-tion. Yammarino (1994) noted that to the extentthat leaders play a role in selecting their subor-dinate leaders, they may be inclined to chooseindividuals who are more likely to behave in amanner similar to how they themselves wouldbehave. Subordinate leaders may also tend tohave a closer relationship with leaders who haveselected them, and this close interaction maypromote behavioral similarity. In the U.S. Army,however, at lower levels of leadership the assign-ments of leaders are normally made at higher

levels. Future studies examining samples inwhich such selection processes are more preva-lent may be able to determine if leaders' selectionof their immediate junior leaders enhances thecascading of leader behavior.

A number of unmeasured variables may help toexplain the pattern of results we reported. Forexample, it would be useful in future research toincorporate other potential influences on unitethical culture, such as the presence or absenceof formal ethical codes at each level. In addition,other potential mechanisms could be assessed infuture research, such as leaders' influencing theidentities of followers at lower levels (Chun et al.,2009). Future studies may also examine whetherdisagreement among members about cultural el-ements, as reflected in within-unit variance onsuch measures (Zohar & Luria, 2005), is itself amoderator variable that influences the strength ofcascades of such elements or relationships be-tween leader behavior and the mean levels ofcultural elements of a given unit or lower-level units.

Given that our data collection was based on across-sectional design, we cannot make definitiveinferences about causal processes. To address alter-native models, we conducted supplementary anal-yses in which the order of ethical leadership andethical culture was varied at each level. These anal-yses determined that the direct effects of ethicalculture on squad-level ethical outcomes werestronger than the effects of ethical leadership. Nev-ertheless, we cannot rule out alternative causal or-ders and mechanisms. In addition, the dependentvariables examined in this study were obtainedfrom self-reported data provided by squad mem-bers. Although there was reasonable agreementamong members of a squad as to the frequency withwhich their members engaged in transgressions orexemplary behaviors, we cannot be certain as to thevalidity of their reports, and it is likely that sometransgressions were not reported accurately or wereunderreported. Aggregating to the squad level,however, made it less likely that individuals' idio-syncratic attributions about the same observed be-haviors would bias the study findings.

Conclusions

We developed a multilevel model to guide astudy of the effects of ethical leadership and sharedunderstandings about ethical conduct (ethical cul-ture) on lower-level followers' ethical cognitionsand behavior. Otir findings indicate that ethicalcognitions and conduct reflect multiple normativeand informational influences, including not only


ethical leadership and ethical culture inherent inan immediate unit, but also reflecting the influenceof leaders and ethical cultiu-e at higher levels. Theunderstandings concerning norms, standards, andsanctions for ethical behavior (that is, ethical cul-ture), which reflected the expectations and desiresof leaders at various levels, were found to be themost potent and proximal influence on followercognitions and behavior.

REFERENCES

Arnaud, A., & Schminke, M. In press. The ethical climateand context of organizations; A comprehensivemodel. Organization Science, orsc.1110.0698.

Avolio, B. J. 2007. Promoting more integrative strategiesfor leadership theory building. American Psycholo-gist, 62: 25-33.

Bandura, A. 1971. Social learning theory. Morristown,N): General Learning Press.

Bass, B., & Bass R. 2008. The Bass handbook of leader-ship: Theory, research, and managerial applica-tions. New York: Free Press.

Bass, B. M., Waldman, D. A., Avolio, B. I., & Bebb, M.1987. Transformational leadership and the fallingdominoes effect. Group & Organization Studies, 12:73-87.

Bliese, P. D. 2000. Within-group agreement, non-inde-pendence, and reliability: Implications for data ag-gregation and analysis. In K. ). Klein & S. W. Koz-lowski (Eds.), Multilevel theory, research, andmethods in organizations: 349-381. San Francisco:Jossey-Bass.

Brown, M. E., & Mitchell, M. S. 2010. Ethical and uneth-ical leadership: Exploring new avenues for futureresearch. Business Ethics Quarterly, 20: 583-616.

Brown, M. E., & Treviño, L. K. 2006. Ethical leadership: Areview and future directions. Leadership Quarterly,17: 595-616.

Brown, M., Harrison, D., & Treviño, L. 2005. Ethical lead-ership: A social learning perspective for construct de-velopment and testing. Organizational Behaviorand Human Decision Processes, 97: 117-134.

Bryk, A. S., & Raudenbush, S. W. 1992. Hierarchicallinear models. Newbury Park, CA: Sage.

Chan, D. 1998. Fiuictional relations among constructs inthe same content domain at different levels of anal-ysis: A typology of composition models. Journal ofApplied Psychology, 83: 234-246.

Chun, ). U., Yammarino, F. J., Dionne, S. D., Sosik, ). J., &Moon, H. K. 2009. Leadership across hierarchicallevels: Multiple levels of management and multiplelevels of analysis. Leadership Quarterly, 20: 689-707.

Cialdini, R., & Trost, M. 1998. Social influence: Socialnorms, conformity, and compliance. In D. T. Gilbert,S. T. Fiske, & G. Lindzey (Eds.), Handbook of socialpsychology, vol. 2: 151-192. Boston: McGraw-Hill.

Den Hartog, D. N., House, R. )., Hanges, P. J., & Ruiz-Quintanilla, S. A. 1999. Culture specific and cross-culturally generalizable implicit leadership theories:Are attributes of charismatic/ transformational lead-ership universally endorsed? Leadership Quarterly,10: 219-256.

Depret, E. F., & Fiske, S. T. 1993. Social cognition andpower: Some cognitive consequences of social struc-tures as a source of control deprivation. In G. Weaiy,F. Gleicher & R. Marsh (Eds.), Control motivationand social cognition: 176-202. New York: SpringerVerlag.

Detert, ). R., & Edmondson, A. C. 2011. Implicit voicetheories: Taken-for-granted rules of self-censorshipat work. Academy of Management Journal, 54:461-488.

Detert, I. R., & Treviño, L. K. 2010. Speaking up to higherups: How supervisors and skip-level leaders influ-ence employee voice. Organization Science, 21:249-270.

Dragoni, L. 2005. Understanding the emergence of stategoal orientation in organizational work groups: Therole of leadership and multilevel climate percep-tions. Journal of Applied Psychology, 90: 1084-1095.

Dvir, T., Eden, D., Avolio, B. )., & Shamir, B. 2002. Impactof transformational leadership on follower develop-ment and performance: A field experiment. Acad-emy of Management Journal, 45: 735-744.

Feldman, D. C. 1984. The development and enforcementof group norms. Academy of Management Review,9: 47-53.

Fornell, C, & Larcker, D. 1981. Evaluating structuralequation models with unobservable variables andmeasurement error. Journal of Marketing Research,18: 39-50.

Glisson, G., & James, L. R. 2002. The cross-level effects ofculture and climate in human service teams. Journalof Organizational Behavior, 23: 767-794.

Hannah, S. T., & Avolio, B. ). 2010. Moral potency: Build-ing the capacity for character-based leadership. Con-sulting Psychology Journal, 62: 291-310.

Hannah, S. T., AvoHo, B. )., & May, D. R. 2011. Moralmaturation and moral conation: A capacity approachto explaining moral thought and action. Academy ofManagement Review, 36: 663-685.

Hannah, S. T., Uhl-Bien, M., Avolio, B. J., & Cavarretta, F.2009. A framework for examining leadership in ex-treme contexts. Leadership Quarterly, 20: 897-919.

Hoge, C. W., Castro, C. A., Messer, S. G., McGurk, D.,


Cotting, D. I., & Koffman, R. L. 2004. Combat duty inIraq and Afghanistan, mental health problems, andbarriers to care. New England Journal of Medicine,351: 13-22.

Howell, J. P., Dorfman, P. W., & Kerr, S. 1986. Moderatorvariables in leadership research. Academy of Man-agement Review, 11: 88-102.

International Institute for Strategic Studies. 2010. Themilitary balance. London: Routledge.

James, L. R. 1982. Aggregation bias in estimates of per-ceptual agreement. Journal of Applied Psychology,67: 219-229.

James, L. R., Demaree, R. G., & Wolfe, G. 1984. Estimatingwithin-group inter-rater reliahility with and withoutresponse bias. Journal of Applied Psychology, 69:85-98.

Katz, D., & Kahn, R. L. 1966. The social psychology oforganizations. New York: Wiley.

Kish-Gephart, J., Detert, J., Treviño, L. K., & Edmondson,A. 2010. Silenced by fear: The nature, sources, andconsequences of fear at work. In A. Brief & B. Staw(Eds.), Research in organizational behavior, vol.29: 163-193. Bingley, U.K.: Emerald.

Kozlowski, S. W. J., & Hattrup, K. 1992. A disagreementabout within-group agreement: Disentangling issuesof consistency versus consensus. Journal of AppliedPsychology, 77: 161-167.

Kozlowski, S. W. J., & Klein, K. J. 2000. A multilevelapproach to theory and research in orgeinizations:Contextual, temporal, and emergent processes. InK. J. Klein & S. W. J. Kozlowski (Eds.), Multileveltheory, research and methods in organizations:Foundations, extensions, and new directions: 3-90.San Francisco: Jossey-Bass.

LaHuis, D. M., & Ferguson, M. W. 2007. The accuracy ofsignificance tests for slope variance components inmultilevel random coefficient models. Organiza-tional Research Methods, 12: 418-435.

Landis, R. S., Beal, D. J., & Tesluk, P. E. 2000. A compar-ison of approaches to forming composite measuresin structural equation models. Organizational Re-search Methods, 3: 186-207.

Likert, R. 1961. New patterns of management. NewYork: McGraw-Hill.

Lindell, M. K., & Brandt, C. J. 1997. Measuring interrateragreement for ratings of a single target. Applied Psy-chological Measurement, 21: 271-278.

Mental Health Advisory Team (MHAT-IV). 2006. Oper-ation Iraq Freedom 05-07. Office of the SurgeonGeneral, U.S. Army Medical Command. http://www.armymedicine.army.mil.

Mayer, D. M., Kuenzi, M., Greenbaum, R., Bardes, M., &Salvador, R. 2009. How low does ethical leadership

flow? Test of a trickle-down model. OrganizationalRehavior and Human Decision Processes, 108:1-13.

Neuhert, M., Carlson, D. S., Kacmar, K. M., Roberts, J., &Chonko, L. 2009. The virtuous influence of ethicalleadership behavior: Evidence from the field. Jour-nal ofRusiness Ethics, 90: 157-170.

Page, S. E. 2007. The difference: How the power ofdiversity creates better groups, firms, schools, andsocieties. Princeton, NJ: Princeton University Press.

Roethlisberger, F. J., & Dickson, W. J. 1939. Human rela-tions. In Management and the worker: 551-568.Cambridge, MA: Harvard College.

Schein, E. H. 1985. Organizational culture and leader-ship (1st ed.). San Francisco: Jossey-Bass.

Schein, E. H. 2010. Organizational culture and leader-ship (4th ed.). Hoboken, NJ: Jossey-Bass.

Simon, H. A. 1973. The organization of complex systems.In H. H. Pattee (Ed.), Hierarchy theory: 1-27. NewYork: George Braziller.

Snijders, T. A. B., & Bosker, R. J. 1999. Multilevel anal-ysis: An introduction to basic and advanced mul-tilevel modeling. London: Sage.

Treviño, L. K. 1986. Ethical decision making in organi-zations: A person-situation interactionist model.Academy of Management Review, 11: 601-617.

Treviño, L. K. 1990. A cultural perspective on changingand developing organizational ethics. Research inOrganizational Change and Development, 4: 195-230.

Treviño, L. K., Butterfield, K. D., & McCahe, D. M. 1998.The ethical context in organizations: Influences onemployee attitudes and behaviors. Rusiness EthicsQuarterly, 8: 447-476.

Treviño, L. K., & Youngblood, S. A. 1990. Bad apples in badbarrels: A causal analysis of ethical decision-makingbehavior. Journal of Applied Psychology, 75: 378-385.

Victor, B., & Cullen, J. B. 1988. The organizational basesof ethical work climates. Administrative ScienceQuarterly, 33: 101-125.

Waldman, D. A., & Yammarino, F. J. 1999. CEO charis-matic leadership: Levels-of-management and levels-of-analysis effects. Academy of Management Re-view, 24: 266-285.

Walker, L. J., & Hennig, K. H. 2004. Differing conceptionsof moral exemplarity: Just, hrave, and caring. Jour-nal of Personality and Social Psychology, 86: 629-647.

Walumhwa, F. O., Mayer, D. M., Wang, P., Wang, H.,Workman, K., & Christensen, A. L. 2011. Linkingethical leadership to employee performance: The


roles of leader-member exchange, self-efficacy, andorganizational identification. Organizational Be-havior and Human Decision Processes, 115: 204-215.

Walumbwa, F. O., & Schaubroeck, J. 2009. Leader per-sonality traits and employee voice behavior: Medi-ating roles of ethical leadership and work grouppsychological safety. Journal of Applied Psychol-ogy, 94; 1275-1286.

Weiss, H. M. 1977. Subordinate imitation of supervisorybehavior; The role of modeling in organizational so-cialization. Organizational Behavior and HumanPerformance, 19; 89-105.

Yammarino, F. J. 1994. Indirect leadership; Transforma-tional leadership at a distance. In B. M. Bass & B. ).Avolio (Eds.), Improving organizational effective-ness through transformational leadership: 26-47.Thousand Oaks, CA: Sage.

Yammarino, F. J., & Dansereau, F. 2008. Multi-level na-ture of and multi-level approaches to leadership.Leadership Quarterly, 19; 135-141.

Yang, J., Zhang, Z. X., & Tsui, A. S. 2010. Middle managerleadership and frontline employee performance; By-pass, cascading, and moderating effects. Journal ofManagerial Studies, 47; 654-678.

Yukl, C. 2002. Leadership in organizations (5th ed.).Upper Saddle River, NJ; Prentice Hall.

Zohar, D., & Luria, G. 2005. A multilevel model of safetyclimate; Cross-level relationships between organiza-tion and group-level climates. Journal of AppliedPsychology, 90; 616-628.

John M. Schaubroeck ([email protected]) is theJohn A. Hannah Distinguished Professor of Psychologyand Management at Michigan State University. He re-ceived his Ph.D. from Purdue University. His researchinterests relate to primarily to leadership processes, em-ployee well-being, and ethical behavior.

Sean T. Hannah ([email protected]) is a colonel inthe U.S. Army and the director of the Center for the ArmyProfession and Ethic, at West Point. He received a Ph.D.in organizational behavior with a focus on leadership

from the University of Nebraska-Lincoln. His researchfocuses on leadership development; character develop-ment; leader identity, self-efficacy, and courage; andleadership in extreme and complex contexts.

Bruce J. Avolio ([email protected]) received hisPh.D. from the University of Akron. He is the Marion B.Ingersoll Professor and the director of the Center forLeadership & Strategic Thinking in the Foster School ofBusiness, University of Washington. His current researchfocuses on examining how to accelerate leadership de-velopment across levels and determining the value ofleadership.

Steve W. J. Kozlowski ([email protected]) is a professorof psychology in the Organizational Psychology Programat Michigan State University. He received his Ph.D. fromThe Pennsylvania State University. His meta-theoreticalwork focuses on multilevel phenomena and the dynam-ics of emergence. Substantive research centers on indi-vidual and team learning, team effectiveness, leadership,and use of simulation to create synthetic experience.

Robert G. Lord ([email protected]) is a distinguishedprofessor of psychology at the University of Akron. Hereceived his Ph.D. from Carnegie-Mellon University. Hisresearch interests include leadership, motivation, infor-mation, and ethical behavior.

Linda K. Treviño ([email protected]) is DistinguishedProfessor of Organizational Behavior and Ethics in theSmeal College of Business, The Pennsylvania State Uni-versity. She received a Ph.D. from Texas A&M Univer-sity. Her current research focuses on behavioral ethics inorganizational context with an emphasis on values, eth-ical leadership, ethical culture, internal reporting andwhistle-blowing.

Nikolaos Dimotakis ([email protected]) is an assis-tant professor at Ceorgia State University. He receivedhis Ph.D. from Michigan State University. His researchfocuses on affective processes, well-being andmotivation.

Ann C. Peng ([email protected]) is a doctoral student inmanagement at Michigan State University. She receivedher M.Phil, in management at Lingnan University ofHong Kong. Her research interests include leadership,employee well-being, and proactive work behavior.

Copyright of Academy of Management Journal is the property of Academy of Management and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

Documents

Embedding Ethical Leadership Within and Across Organization Levels