COMPLIANT SINNERS, OBSTINATE SAINTS: HOW...

COMPLIANT SINNERS, OBSTINATE SAINTS: HOW POWERAND SELF-FOCUS DETERMINE THE EFFECTIVENESS OFSOCIAL INFLUENCES IN ETHICAL DECISION MAKING

MARKO PITESAGrenoble Ecole de Management

STEFAN THAULondon Business School

In this research, we examine when and why organizational environments influencehow employees respond to moral issues. Past research has proposed that socialinfluences in organizations affect employees’ ethical decision making, but has notexplained when and why some individuals are affected by an organizational environ-ment and some disregard it. To address this problem, we drew on research on powerto propose that power makes people more self-focused, which, in turn, makes themmore likely to act upon their preferences and ignore (un)ethical social influences.Using both experimental and field methods, we tested our model across the three mainparadigms of social influence: informational influence (Studies 1 and 2), normativeinfluence (Study 3), and compliance (Study 4). Results offer converging evidence forour theory.

Members of organizations are often faced withmoral issues, defined as situations in which “aperson’s actions, when freely performed, may harmor benefit others” (Jones, 1991: 367). An employeemight be tempted to exaggerate, rather than accu-rately report, time spent on a project to receive alarger bonus; another employee may have to decidewhether to deceive the other party in a negotiationto secure a more favorable outcome. While decid-ing on moral issues, employees are exposed to so-cial influences that can impact their ethical deci-sion making. These social influences range fromexplicit organizational ethical standards to others’(un)ethical behavior as an indirect indicator ofacceptable conduct in their organization (for a re-view, see Kish-Gephart, Harrison, and Treviño[2010]). Both ethical decision-making theory(Sonenshein, 2007; Treviño, 1986) and relatedstudies (Gino, Ayal, & Ariely, 2009; Pierce & Sny-der, 2008; Treviño, Butterfield, & McCabe, 1998)suggest that employees’ ethical decision makingwill be consistent with the social influence they areexposed to. Those working in an organization withan ethical code of conduct are expected to make

more ethical decisions (Treviño et al., 1998),whereas those working in organizations in whichunethical behavior is widespread are more likely tomake unethical decisions themselves (Pierce &Snyder, 2008).

Although social influence has a powerful impacton individual decision making (Cialdini & Gold-stein, 2004; Turner, 1991), not all employees areswayed by their context when responding to moralissues. For example, news reports tell stories aboutcheating employees within organizations whosemembers typically do not cheat (Weisman, 2007).Similarly, there are employees who make ethicaldecisions while their coworkers follow their self-interest (Russell, 2011). This suggests that someemployees ignore their social context and, for bet-ter or worse, follow their own preferences for eth-ical or unethical conduct. This phenomenon re-mains unexplained by past theories on ethicaldecision making (Jones, 1991; Sonenshein, 2007;Treviño, 1986). In this research, we propose andempirically test a theoretical model that explainsboth when and why employees disregard their or-ganizational environment and instead follow theirown (un)ethical preferences.

Our model starts with the observation that deci-sion makers are embedded in hierarchically or-dered groups, in which they have different levels of

The authors thank Philip Culbertson, Gillian Ku,Olivier Trendel, and Carolina Werle for their helpfulcomments on previous versions of this article.

635

� Academy of Management Journal2013, Vol. 56, No. 3, 635–658.http://dx.doi.org/10.5465/amj.2011.0891

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

social power. We expect powerful decision makersto be more likely to disregard social influence intheir organization, whether ethical or unethical,and to act more consistently with their (un)ethicalpreferences. The assumptions of our model arebased on the situated focus theory of power (Gui-note, 2007b), which suggests that the experience ofpower leads to an increase in private self-focus,meaning that as a consequence of power, peopledirect more attention to themselves and their per-sonal preferences. Consequently, we expect the in-teractive effect of social influence and power onethical decision making to be mediated by privateself-focus. We test our model in a series of experi-mental and field studies, including different mea-sures and manipulations of social influence,power, self-focus, ethical preferences, and (un)eth-ical decisions and behaviors.

By developing and testing a theoretical modelthat predicts both when and why organizationalenvironments influence how employees make eth-ical decisions, our research contributes to an un-derstanding of employee responses to (un)ethicalsocial influences in organizations, such as organi-zational ethical culture (Treviño et al., 1998), peerpressure from unethical coworkers (Ashforth &Anand, 2003), and reliance on others’ responses tointerpret complex ethical issues in organizations(Sonenshein, 2007). Our studies also extend pastresearch in social psychology on power and socialinfluence (e.g., Briñol, Petty, Valle, Rucker, &Becerra, 2007; Galinsky, Magee, Gruenfeld, Whit-son, & Liljenquist, 2008; Guinote, 2008; See, Mor-rison, Rothman, & Soll, 2011; Tost, Gino, & Larrick,2011) by being the first to test private self-focus asthe explanation for the effect of power on individ-uals’ susceptibility to social influence. Finally, ourwork may also offer practitioners better under-standing of how to manage (un)ethical social pro-cesses in organizations.

ETHICAL DECISION MAKING INORGANIZATIONAL CONTEXTS

Aspects of organizational environments, includ-ing ethical codes, organizational culture, and co-worker behavior, affect how employees respond tomoral issues by exerting social influence (Kish-Gephart et al., 2010; Treviño, 1986). Social influ-ence occurs when people change their thoughts,feelings, or behaviors in response to their environ-ment (Allport, 1954; Turner, 1991). People canhave different motives for making decisions in

keeping with their environment, and these can bebroadly classified as obeying direct requests (com-pliance); copying others because their behavior ispresumed to be the right response in a situation(informational conformity); and doing what othersdo in order to fit in (normative conformity [Cialdini& Goldstein, 2004]). We provide a comprehensivetest of our model across all three types of socialinfluence, which we describe in more detail below.

Compliance occurs when a purposeful requestproduces a behavioral response in the intendeddirection (Cialdini & Goldstein, 2004). The influ-ence that elicits compliance may come in manyforms: It may be “explicit, as in the direct solicita-tion of funds in a door-to-door campaign for char-itable donations, or it may be implicit, as in apolitical advertisement that touts the qualities of acandidate without directly asking for a vote” (Cial-dini & Trost, 1998: 168). Many organizations relyon their codes of conduct (McCabe, Treviño, & But-terfield, 1996; Schwartz, 2001), their incentive sys-tems (James, 2000), or the values they explicitlypropagate (Treviño et al., 1998) to ensure compli-ance with ethical standards. For instance, Treviñoet al. (1998) found that explicit standards of ethicalbehavior (i.e., ethical culture) elicit compliance onthe part of employees, leading to lower levels ofunethical behavior in organizations.

In contrast to compliance, conformity does notentail a direct request to behave in a particular way.Rather, an individual is internally motivated to im-itate others’ decisions (Cialdini & Goldstein, 2004;Crutchfield, 1955). For example, when an individ-ual is uncertain about the appropriate decision in agiven situation, and as a result conforms to thedecisions of others, informational influence is saidto occur (Deutsch & Gerard, 1955). Sonenshein(2007) pointed out that many ethical issues in or-ganizations involve a great deal of equivocality,making it likely for individuals to use social inputto construct their interpretation of ethical issuesand decide on the appropriate response. For in-stance, studies by Tang and Farn (2005) showedthat informational influence (awareness of others’decisions) explained why some individuals weremore willing to engage in software piracy.

Finally, employees may also conform becausethey are motivated to gain social approval (i.e.,normative influence [Deutsch & Gerard, 1955]).Cialdini, Petrova, and Goldstein noted that “honestemployees can be converted into wrongdoers invarious ways, but the process often begins withpeer pressure” (2004: 70). Models describing the

636 JuneAcademy of Management Journal

causes and consequences of organizational corrup-tion also highlight the need for approval as themotivator by which corrupt organizational environ-ments stimulate unethical behavior (Ashforth &Anand, 2003; Pinto, Leana, & Pil, 2008). Gino et al.(2009) provided some empirical support for thesearguments, finding that the level of unethical be-havior in a group increases when a group memberbehaves unethically.

Despite the evidence for the influence of theirorganizational environment on ethical decisionmaking, some individuals resist social influenceand instead make decisions that are consistent withtheir preferences. For instance, an employeemight not comply with explicit organizational pol-icies promoting ethical conduct and instead followhis or her tendency to deceive and manipulateother people for personal gain. Arguably, not allemployees’ ethical decision making is affected by(un)ethical social influences in organizations. Yetthe current body of work on ethical decision mak-ing does not explain when and why employeesmake decisions that are consistent with social in-fluence and when they make decisions that areconsistent with their preferences. This explanatoryproblem limits understanding of a wide range ofethically relevant phenomena in organizations. Forinstance, scholars know that ethical culture maypromote ethical conduct (Treviño et al., 1998), butdo not know when and why it does so. It is alsoknown that unethical coworker conduct may leademployees to behave more unethically (Ashforth &Anand, 2003; Pierce & Snyder, 2008) but not whenand why this actually happens. In the followingsection, we develop a theoretical model explaininghow decision makers’ social power and self-focuscan help answer these questions. Figure 1 showsthe theoretical model.

POWER AND SELF-FOCUS IN ETHICALDECISION MAKING

Social power can be defined as an individual’s“capacity to modify others’ states by providing orwithholding resources or administering punish-ment” (Keltner, Gruenfeld, & Anderson, 2003: 265).There are various bases of power, including differ-ences in roles, expertise, and connections to pow-erful others (Brass, 1984; French & Raven, 1959),and the definition we use allows for any possiblesource of power. Finally, this definition takes intoaccount that power is a contextual phenomenon—aperson may experience power in one particular

relationship, group, or situation but not in others(Anderson & Berdahl, 2002; Emerson, 1962; Thi-baut & Kelley, 1959).

Theory and research suggest that the experienceof power has consequences for people’s attention toexternal stimuli. The argument of the situated fo-cus theory of power (Guinote, 2007b) is that be-cause the powerful are less dependent on others fortheir outcomes, their attention is more focused onthemselves (see also Fiske, 1993; Fiske & Dépret,1996). In contrast, the powerless focus their atten-tion more on contextual stimuli because they areobjectively more dependent on external circum-stances and need to increase the predictability oftheir environment. The greater attention to the selfshould lead the powerful to become more aware oftheir own personal thoughts and feelings. For in-stance, Weick and Guinote (2008) found that powermakes individuals more sensitive to their own sub-jective experiences when they are forming attitudesand judgments.

The attentional state of being aware of personalthoughts and feelings has been previously concep-tualized as private self-focus (Fenigstein, 1979: 76).A heightened private self-focus makes people moreaware of their own preferences, leading to a greaterlikelihood that people will act upon them (From-ing, Walker, & Lopyan, 1982; Scheier & Carver,1980). For example, when an employee works in anorganization with a strong ethical culture, self-fo-cus should lead to lower levels of compliance, asthe employee would pay more attention to his orher preferences rather than to the environment.Consistently with this argument, Froming et al.(1982) found that participants with experimentallyheightened private self-focus were less likely to

FIGURE 1Overview of the Theoretical Model

Power

SocialInfluence

Self-focus

(Un)EthicalDecisions

H2

H1

H2

2013 637Pitesa and Thau

comply when asked to deliver shocks to a confed-erate. Therefore, if a decision maker experiencespower in a given situation, the heightened privateself-focus caused by power should make the pow-erful more likely to act consistently with their pref-erences (even if this means ignoring their socialcontext). We predict:

Hypothesis 1. Power weakens the effect of so-cial influences in ethical decision making.

Hypothesis 2. Private self-focus mediates themoderating effect of power on the effect ofsocial influences in ethical decision making.

It is relevant to note an underlying assumption ofour theoretical model. In making our predictions,we assume that individuals will have certain pref-erences, ethical or unethical, when they respond tomoral issues. This assumption is relevant becausehaving preferences on a decision issue can explainwhy power sometimes leads to more and some-times to less susceptibility to social influence (cf.Galinsky et al., 2008: 1462). Past studies have doc-umented more situation-consistent behavior as aresult of power when individuals had no personalpreferences in the given situation and were insteadmotivated to act in accordance with situational de-mands (Guinote, 2007a, 2008). In contrast, studieshave documented less susceptibility to social influ-ence as a function of power in situations that weredesigned in such a way that the powerful firstformed some personal preferences (e.g., an attitudeon an issue or an assessment of a problem) andwere then exposed to others’ input (Briñol et al.,2007; Galinsky et al., 2008; See et al., 2011; Tost etal., 2011).

In the ethical decision-making context that isrelevant to our research question, theory and stud-ies in moral psychology suggest that when con-fronted with moral issues, individuals always havecertain (un)ethical preferences affecting their deci-sion making. These preferences are automaticallyactivated (Cushman, Young, & Hauser, 2006; Haidt,2001) and are even present irrespective of and priorto conscious responses and deliberation (Haidt,2001; Haidt & Hersh, 2001). If it is true that peoplehave certain preferences when they decide onmoral issues, then powerful decision makersshould be less swayed by social influences andinstead follow their own preferences, as the in-creased self-focus brought about by power shouldmake the ethical preferences more salient in thedecision makers’ minds.

OVERVIEW OF THE PRESENT RESEARCH

We tested our predictions in four studies usingboth experimental and field study methods. As ourtheory predicts that, because of its effect on self-focus, power leads to a weaker effect of social in-fluence in ethical decision making, we provide aconservative and comprehensive test of our theoryby testing it across all three main types of socialinfluence (informational social influence, norma-tive social influence, and compliance). Studies 1through 3 used experiments to test whether power(either primed or structurally manipulated) altersthe effect of social influences in ethical decisionmaking in the presence of informational social in-fluence (Studies 1 and 2) and normative social in-fluence (Study 3). In Study 4, we surveyed manag-ers to test whether their subjective sense of powermoderates the effect of social influences in ethicaldecision making when social influence is based ondirect pressures on an individual’s behavior (organ-izational compliance pressures). Finally, we testedthe hypothesized role of private self-focus in allstudies, either by measuring the mediator, privateself-focus (Studies 1, 3, and 4) or by manipulating it(Study 2; see Spencer, Zanna, and Fong [2005]).

STUDY 1: METHODS

Study 1 examined whether power reduces theeffect of social influences in ethical decision mak-ing when accuracy is the salient motive for confor-mity (informational social influence). Followingresearch indicating that ambiguous situations facil-itate informational social influence (Sherif, 1936),we asked participants in Study 1 to respond to anethical dilemma that had no clear solution. In thetreatment experimental condition, participantswere given (bogus) information about others’ deci-sions, which indicated that everyone else preferredone particular solution to the ambiguous ethicaldilemma. The apparent consensus in this ambigu-ous situation served to create pressure to conformto others’ behavior (Festinger, 1954; Sherif, 1936).In the control experimental condition, participantsdid not receive information about others’ decisions.

As a second experimental factor, participants’sense of power was manipulated via a primingprocedure prior to their responding to the ethicaldilemma. Some participants were primed for eitherlow power or high power, and some were notprimed as to sense of power (control). We expectedthat power would not affect participants’ decisions


when they did not receive information about oth-ers’ decisions. However, we expected that powerwould significantly influence participants’ re-sponses when they were exposed to social influ-ence in such a way that feeling powerful wouldlead to less conformity and that feeling powerlesswould lead to more conformity, compared to thecontrol condition. Thus, we expected both low andhigh power (compared to the control) to moderatethe effect of social influence in such a way that lowpower strengthened the effect of social influencewhile high power weakened the effect of socialinfluence on ethical decisions. Finally, we alsomeasured participants’ private self-focus to exam-ine whether it mediated the moderating effect ofhigh and low power on the effect of social influencein ethical decision making.

Participants

Participants were 256 students who volunteeredto participate in this experiment at the end of dif-ferent class sessions. The average age was 23.16(s.d. � 1.50), and 55.6 percent were female. Wefollowed the recommendations for classroom datacollection (Loyd, Kern, & Thompson, 2005) andasked participants to individually respond to apackage containing the experimental materials us-ing pen and paper. The packages were randomlyordered and distributed in advance based on athree (power: high, low, control) by two (socialinfluence vs. no social influence) between-personsdesign.

Procedure and Materials

Power manipulation. Participants’ sense ofpower was manipulated using a priming procedure(Galinsky, Gruenfeld, & Magee, 2003). Participantswrote a narrative essay about an incident in theirlives, ostensibly as part of a study focusing on pastexperiences. In the high power condition, partici-pants recalled an incident in which they had powerover another individual or individuals; in the lowpower condition, they recalled an incident inwhich someone else had power over them; in thecontrol condition, participants recalled a trip to agrocery store.

Private self-focus measure. We measured pri-vate self-focus with the three-item private self-awareness scale developed by Govern and Marsch(2001). Participants rated each item (e.g., “Rightnow, I am conscious of my inner feelings”; “Right

now, I am aware of my innermost thoughts”) on ascale ranging from 1 (“strongly disagree”) to 7(“strongly agree”; � � .86).

Ethical decision-making scenario. After com-pleting the power manipulation and the self-focusmeasure, participants read an ethical dilemmaadapted from Flynn and Wiltermuth (2010). Thedilemma had no one correct response, as it in-volved a conflict of different moral values (Ba-daracco, 1997). Participants read:

You manage a small company that is trying to securean additional round of venture-capital financing.The firm employs five people, each of whom has anirreplaceable set of skills. If any of the five were toleave, the company would struggle to secure addi-tional financing. One of the principal employees,whom you consider a friend, has recently informedyou that he has received an extremely appealingoffer from another company that is much morelikely to succeed. The employee must make a deci-sion in the next two days. Out of respect for you, thisemployee has told you that he will go to the othercompany only if you offer your blessing.

We selected this dilemma from among similar ones(see Flynn & Wiltermuth, 2010) after a pretest withan independent sample (n � 51) showed that re-sponses to this dilemma grouped closest to thecenter value between the two options, indicating anambiguous decision-making task and, therefore, afertile ground for informational conformity in eth-ical decision making (Festinger, 1954; Sherif, 1936;Sonenshein, 2007).

Social influence manipulation. Social influencewas manipulated by providing different answersheets for participants in the social influence ver-sus no social influence conditions (Epley & Gilo-vich, 1999). For both conditions, the answer sheetcontained room for responses from ten partici-pants. For the no social influence condition, theanswer sheet contained no previous responses andso participants’ responses to the ethical decision-making scenario were the first to be entered on theanswer sheet. In the social influence condition, par-ticipants entered their responses on an answersheet on which bogus answers from seven otherparticipants were already entered. In contrast to theresults of the pretest, which showed that peopletend to select the center value of 4, the bogus re-sponses unmistakably tended toward higher val-ues, with all seven values above 4, thus suggestingthat other participants were overwhelmingly notinclined to discourage the valuable employee fromleaving in this ethical dilemma. Such an apparent


social consensus serves as a powerful indicator ofvalid choices in ambiguous situations (Festinger,1954; Sherif, 1936). However, because people canfeel surprised when confronted with a majorityview that differs drastically from their initial posi-tion (Baker & Petty, 1994), the answers included allthree higher values (5, 6, and 7), with the meanvalue of 6, to make the bogus responses seem morerealistic.

Ethical decision-making measure. After readingthe ethical decision-making scenario, participantsused the answer sheets described above to indicatewhether they would discourage this employee fromleaving (1 � “definitely would discourage” to 7 �“definitely would not discourage”).

Finally, participants responded to several ques-tions probing them for suspicion and were thendebriefed.

STUDY 1: RESULTS

Four participants were excluded from the analy-ses because they expressed having doubted the au-thenticity of the bogus responses. All analyses wereconducted on the remaining 252 participants. Ex-clusion of these cases did not alter any of the re-sults. Finally, we conducted a preliminary analysisto examine whether gender influenced the re-sponses, and we found no effect of gender, in thisor any subsequent studies, so the data were col-lapsed across gender in all analyses.

Manipulation Checks

Group means are presented in Figure 2. In keep-ing with the pretest, participants in the no socialinfluence condition tended to respond close to thecentral value of 4 (mean � 3.92, s.d. � 1.03): con-firming that the ethical dilemma presented an am-biguous decision-making task, as intended. Tocheck the effectiveness of the social influence ma-nipulation, we followed past research (e.g., Griffin& Buehler, 1993; McFerran, Dahl, Fitzsimons, &Morales, 2010) and examined the degree to whichparticipants changed their “behavior to match theresponses of others” (Cialdini & Goldstein, 2004:613). Confirming the effectiveness of the social in-fluence manipulation, the responses in the socialinfluence condition were significantly higher(mean � 5.18, s.d. � 0.90, t[250] � 10.35, p � .001),indicating that participants conformed to the bogusresponses.

To examine whether power itself affected partic-ipants’ responses in this ethical dilemma (insteadof just moderating the effect of social influence), weexamined the effect of power within each conditionof social influence. We expected a significant effectof power when participants were exposed to socialinfluence (indicating that power changed how theyresponded to social influence), but no effect whenthey were not exposed to social influence (indicat-ing that power itself did not affect their preferencesin this ethical dilemma). In keeping with our ex-pectations, the interaction between power and so-cial influence was significant (F[2, 246] � 7.80, p �.001, �p

2 � .060), and analysis of simple effectsshowed a significant effect of power in the socialinfluence condition (F[2, 246] � 9.27, p � .001, �p

2

� .070), but not in the no social influence condition(F[2, 246] � 0.82, p � .441, �p

2 � .007).

Hypothesis 1 Test

Hypothesis 1 predicted that high (low) powerwould weaken (strengthen) the effect of social in-fluence in ethical decision making. Thus, whenexposed to social influence, the powerful shouldrespond farther from others’ responses than controlparticipants; in contrast, the powerless should re-spond closer to others’ responses than controlparticipants.

To test Hypothesis 1, we conducted simple com-parisons between the three levels of power withinthe social influence condition (see right portion ofFigure 2). The simple comparisons showed a sig-nificant difference between the high-power condi-tion (mean � 4.73, s.d. � 0.77) and the controlcondition (mean � 5.19, s.d. � 0.99) within thesocial influence condition (F[1, 246] � 4.89, p �.028, �p

2 � .019). Therefore, when exposed to so-cial influence, high-power participants, comparedto control participants, conformed less to others’responses.

An opposite effect occurred as a result of the lowpower prime: The difference between the low-power condition (mean � 5.62, s.d. � 0.70) and thecontrol condition (mean � 5.19, s.d. � 0.99) withinthe social influence condition was significant (F[1,246] � 4.39, p � .037, �p

2 � .017), in such a waythat participants primed with low power selectedvalues closer to others’ responses than did partici-pants in the control condition (indicating greaterconformity for low-power participants).


Hypothesis 2 Test

Hypothesis 2 states that private self-focus medi-ates the moderating effect of power on the effect ofsocial influences in ethical decision making. In thecontext of Study 1, this means that private self-focus should mediate the interaction between bothhigh and low power (each compared to control) andsocial influence on ethical decisions.

We used moderated path analysis to test thisprediction (Edwards & Lambert, 2007). To providesupport for the type of mediated moderation thatwe predict in Hypothesis 2, the following must beestablished. First, high (low) power should have apositive (negative) effect on private self-focus. Sec-ond, private-self focus should weaken the effect ofsocial influence on participants’ decisions. Finally,the product of these two paths should be significant(see Grant and Berry [2011] for a similar analyticalapproach).

Model 1 in Table 1 shows that high and lowpower (compared to control) both have a significanteffect on private self-focus. The direction of theseeffects is consistent with our predictions: Highpower increases private self-focus (b � 0.53, s.e.� 0.16, t[246] � 3.42, p � .001), and low power

reduces it (b � �0.62, s.e. � 0.16, t[246] � �3.98,p � .001). Moreover, model 3 shows that the inter-action between private self-focus and social influ-ence significantly predicts participants’ ethical de-cisions. The nature of this interaction is consistentwith our predictions: Social influence had a weakereffect on participants’ decisions when private self-focus was one standard deviation above the mean(5.08; b � 0.47, s.e. � 0.12, t[244] � 4.08, p � .001)than it did when it was one standard deviationbelow the mean (2.86; b � 0.83, s.e. � 0.12, t[244]� 7.04, p � .001). Next, we tested the significanceof the product of the path from high (low) powerto private self-focus and the path from the privateself-focus by social influence interaction to ethicaldecisions. Using the coefficients from the prioranalyses, we applied the bootstrap method to con-struct bias-corrected confidence intervals of the in-direct effects based on 10,000 random replace-ments from the full sample. The 95% bias-corrected confidence interval for the size of theindirect effect excluded zero (indicating a signifi-cant indirect effect [Shrout & Bolger, 2002]) for boththe indirect effect of the low power by social influ-ence interaction (0.023–0.218) as well as the high

FIGURE 2Study 1, Mean Scores with Standard Errors by Group

Mean of “Others’” Responses

No Influence Social Influence

7

6

5

4

3

2

1

PowerManipulation:

Low powerControlHigh power

Social Influence Manipulation

MeanDilemma

Responses


power by social influence interaction (�0.188 to�0.021). In sum, our results support the predictionthat private self-focus mediated the moderating ef-fects of high and low power on the effect of socialinfluence on participants’ decisions in this ethicaldilemma.

STUDY 1: DISCUSSION

The results of Study 1 provide support for ourmodel. Conformity to social influence when re-sponding to an ethical dilemma varied as a func-tion of power in such a way that a heightened senseof power led to less conformity and a sense ofpowerlessness led to more conformity, compared tothe control condition. Specifically, when exposedto social influence, high-power participants, ascompared to control participants, tended to followothers’ responses less. In contrast, those primedwith low power, compared to participants notprimed with power, tended to follow others’ re-sponses more. These findings also show that thiseffect holds along the continuum of power: Thepowerful are relatively less swayed by social influ-ences, whereas those lacking power are particularlysusceptible to social influences when making ethi-cal decisions. Finally, Study 1 shows that this phe-nomenon occurs because of differences in privateself-focus. Mediation analyses showed that private

self-focus mediated the moderating effect of lowand high power on the effect of social influences inethical decision making.

STUDY 2: METHODS

The primary goal of Study 2 was to construc-tively replicate the findings of Study 1 and to testthe proposed mechanism, private self-focus, bymanipulating the psychological process instead ofmeasuring it. This testing strategy manipulatesboth the independent variable and the hypothe-sized psychological mechanism in a fully crosseddesign permitting observation of whether the twofactors interact in such a way that the effect of theindependent variable is only significant when thepsychological mechanism is not independently ac-tivated (Spencer et al., 2005). This pattern of resultsprovides experimental evidence of the hypothe-sized psychological mechanism that is more con-servative than tests relying on a measurement of themediating variable (and thus essentially constitut-ing correlational evidence). Therefore, followingthe power manipulation, we manipulated partici-pants’ private-self focus by asking them to eitheruse self-relevant words or neutral words in an es-say-writing task. Participants were then presentedwith the same ambiguous ethical dilemma used inStudy 1, and all participants were exposed to social

TABLE 1Study 1, Hypothesis 2 Test: Regression Analysis Resultsa

Variablesb

Model 1: Private Self-FocusModel 2: Ethical Decisions,

Step 1Model 3: Ethical Decisions,

Step 2

b LLCI ULCI b LLCI ULCI b LLCI ULCI

Constant 4.00** 3.78 4.22 4.56** 4.36 4.76 4.58** 4.38 4.77Low powerb �0.62** �0.92 �0.31 0.14 �0.14 0.43 �0.02 �0.30 0.26High powerb 0.53** 0.22 0.84 �0.17 �0.45 0.12 �0.05 �0.33 0.23Social influencec 0.06 �0.15 0.28 0.63** 0.43 0.83 0.65** 0.46 0.84Low power � social influence �0.15 �0.46 0.16 0.29* 0.01 0.57 0.15 �0.13 0.43High power � social influence �0.03 �0.34 0.27 �0.29* �0.57 �0.01 �0.21 �0.49 0.07Self-focusd �0.22** �0.34 �0.11Self-focus � social influence �0.16** �0.27 �0.05

R2 .19** .35** .41**�R2 .06**

a LLCI � 95% confidence interval lower limit; ULCI � 95% confidence interval upper limit.b Of theoretical interest in this research was the effect of high (low) power as compared to the control condition, so dummy coding was

used for the two variables.c The social influence manipulation was effect-coded.d Mean-centered.

* p � .05** p � .01


influence (the same bogus responses used inStudy 1).

We expected that high power, compared to lowpower, would reduce conformity, but only whenparticipants’ private self-focus was not experimen-tally heightened. If power reduces conformity inethical decision making because it makes peoplemore self-focused, then it should have no effectwhen the level of private self-focus is already high.By contrast, for participants whose private self-focus is not experimentally heightened, high power(compared to low power) should lead to greaterself-focus and therefore to less conformity.

We also used a different manipulation of powerin Study 2. Participants were assigned to either ahigh-power position (“manager”) or a low-powerposition (“subordinate”). The reason that we ma-nipulated power structurally (by assigning partici-pants to different roles) in Study 2, instead ofpriming it, was to increase confidence in the gen-eralizability of findings to different instantiationsof power.

Participants

Ninety-one students participated in a lab studyin exchange for money. The average age was 19.32(s.d. � 1.10), and 62.6 percent were female. Partic-ipants met individually with an experimenter, whotold them they would participate in several unre-lated studies. They were randomly assigned to con-ditions of a two (power: high vs. low) by two (self-focus prime vs. control) between-persons design.


Power manipulation. Power was manipulatedby announcing that in an upcoming task, partici-pants would act as either managers supervising andmaking decisions that would affect their subordi-nates, or as subordinates with no decision-makingauthority. To strengthen the manipulation, partici-pants were provided with badges, folders, and bo-gus task reports, all indicating their position (eithermanager or subordinate) and emphasizing their de-cision-making authority (e.g., “You are in charge”vs. “You follow orders”) in the upcoming task.

Private self-focus manipulation. Following thepower manipulation, participants were told theywould engage in a couple of short tasks while wait-ing for other participants to join them for the grouptask. First, they were asked to write a story usingeither self-related words (e.g., “I,” “mine”) or neu-

tral words of similar length and frequency of usage(e.g., “a,” “some”). This procedure was developedby Fenigstein and Levine (1984) and has been suc-cessfully used in past research to manipulate pri-vate self-focus (e.g., Fransen, Fennis, Pruyn, &Vohs, 2011; Goukens, Dewitte, & Warlop, 2009).

Ethical decision-making scenario and mea-sure. Next, participants responded to the same eth-ical dilemma used in Study 1 by indicatingwhether they would discourage a valuable em-ployee from leaving their company. All partici-pants were provided with an answer sheet onwhich fake answers, ostensibly from seven otherparticipants, had been entered. Thus, the answersheets were identical to those used in Study 1 whenparticipants were exposed to social influence. Atthe end of the experiment, participants wereprobed for suspicion and debriefed.

STUDY 2: RESULTS

Hypothesis 1 Test

Group means are presented in Figure 3. Powerhad a significant effect on participants’ responsesto the ethical dilemma (F[1, 87] � 16.31, p � .001,�p

2 � .158), so that participants in the high-powercondition (mean � 4.78, s.d. � 0.79) conformedless to others’ responses than did participants inthe low-power condition (mean � 5.39, s.d.� 0.74).

Hypothesis 2 Test

The private self-focus manipulation had a signif-icant main effect on conformity (F[1, 87] � 9.85,p � .002, �p

2 � .102), so that a greater privateself-focus reduced conformity: Participants whoused self-related words (mean � 4.85, s.d. � 0.76)conformed less than did participants who usedneutral words (mean � 5.33, s.d. � 0.83).

Supporting Hypothesis 2, power and self-focusinteracted significantly (F[1, 87] � 4.12, p � .045,�p

2 � .045), in such a way that high power de-creased conformity when participants’ self-focuswas not subsequently experimentally heightened(neutral words used: F[1, 87] � 18.64, p � .001, �p

2

� .176) but did not have an effect when partici-pants’ self-focus was subsequently experimentallyheightened (self-related words used: F[1, 87] �2.04, p � .157, �p

2 � .023) (see Figure 3). Specifi-cally, the difference between the low-power condi-tion (mean � 5.78, s.d. � 0.52) and the high-power


condition (mean � 4.86, s.d. � 0.83) was signifi-cant within the control condition, but it was notsignificant within the private self-focus prime con-dition (low power: mean � 5.00, s.d. � 0.74; highpower: mean � 4.70, s.d. � 0.76). These resultsdemonstrate that the effect of power on conformitydepends on private self-focus—power leads to dif-ferent levels of conformity only in as much as it canproduce different levels of private self-focus. Con-sequently, when private self-focus is high withinboth groups, differences in power have no effect onethical decision making.

Interpreted differently, the pattern of this inter-action shows that the difference between partici-pants who used self-oriented words and those whoused neutral words was greater in the low-powercondition (F[1,87] � 13.51, p � .001, �p

2 � .134)than in the high-power condition (F[1,87] � 0.71, p� .401, �p

2 � .001), where the effect was not sig-nificant. This suggests that the powerful were al-ready self-focused, so the self-focus manipulationdid not affect their conformity. By contrast, theself-focus manipulation reduced conformity amongparticipants in the low power condition, who were

initially less self-focused, and only after a subse-quent self-focus manipulation (but not after usingneutral words) did their focus increase and there-fore their conformity level drop.

STUDY 2: DISCUSSION

Study 2 provides further evidence that privateself-focus explains the effect of power on suscepti-bility to social influence in ethical decision mak-ing. Consistently with the findings of Study 1,power reduced the effect of social influences inethical decision making, but this effect was elimi-nated when private self-focus was subsequentlyprimed. This result shows that power reduces theeffect of social influences in ethical decision mak-ing only in as much as it can increase private self-focus. By manipulating the psychological mecha-nism and showing that the effect of the powermanipulation depended on the manipulation ofprivate self-focus, Study 2 provides experimentalevidence of the psychological mechanism that con-stitutes a particularly conservative test of our the-ory (Spencer et al., 2005).

FIGURE 3Study 2, Mean Scores with Standard Errors by Group

PowerManipulation:

SubordinateManager

Mean of “Others’” Responses

7

6

5

4

3

2

1Control Self-Focus Prime

Self-Focus Manipulation

MeanDilemma

Responses


In addition, in Study 2, power was manipulatedstructurally by assigning participants to differentroles in a group task. Despite the fact that a struc-tural manipulation of power was used, the effect ofpower was similar to that observed in Study 1, thusincreasing the confidence in the generalizability ofour findings.

STUDY 3: METHODS

In Study 3, we attempted to replicate construc-tively the findings of Studies 1 and 2 in a normativesocial influence situation. Although the differenttypes of social influence are often intertwined, usu-ally it is possible to ascertain the dominant motivedriving conformity (Cialdini & Goldstein, 2004;Cialdini & Trost, 1998). In ambiguous situationssuch as those the participants faced in Studies 1and 2, accuracy-motivated conformity often playsthe key role; in situations that are less ambiguous,the motive to “fit in” with others tends to be moreimportant. In the classic Asch (1955) studies, peo-ple conformed not because they thought the con-sensus of the group was more accurate (as evidentfrom their answers when responding alone) butbecause it is easier to align oneself with a majorityopinion than to face the social consequences of notdoing so (Crutchfield, 1955; Janes & Olson, 2000).The motive to fit in was not present in Studies 1and 2 because participants had no exposure toother participants except for the evidence of others’responses. However, it is possible that differentgoals associated with different types of social influ-ences alter the effect of power (Overbeck & Park,2001, 2006). To account for this possibility, inStudy 3 we tested the hypothesized relationshipsin a situation in which normative social influencewas salient.

Specifically, we primed participants with eitherhigh or low power. We then manipulated norma-tive social influence by exposing participants toexplicit normative social standards prior to theirdecision and then announcing a follow-up discus-sion with their peers, which served to strengthenthe motive to fit in with the group. By contrast,participants in the control condition were not ex-posed to social influence. Subsequently, all partic-ipants were given an opportunity to deceive theiropponent in a negotiation.

We expected that high power, compared to lowpower, would weaken the effect of normative socialinfluence on participants’ decisions to behave un-ethically and would make participants follow their

own preferences instead. Since the task involved aclear ethical versus unethical decision, we wereable to infer participants’ preferences by measuringtheir scores on a Machiavellianism scale. Machia-vellianism is a personality construct emphasizingpreferences for self-interested behavior (Christie &Geis, 1970), and it has been previously found topredict unethical decisions (Kish-Gephart et al.,2010). Therefore, those who are more (less) Machi-avellian should be more (less) willing to deceivein negotiations—unless they conform to socialpressure.

Thus, another way of measuring the degree towhich a person was swayed by social influence isto examine to what degree the person acted consis-tently with what would have been that person’sresponse if no social influence was present (Cial-dini & Trost, 1998; Madden, 1960). If an individualwould have chosen to deceive, but does not do sodue to others’ pressure, social influence occurred.Social influence can therefore be documented bothas the degree to which an individual conforms tosocial pressures as well as the degree to which he orshe follows his/her own preferences in the face ofsocial pressures. In Study 3, we tested our theoryby measuring social influence from both these per-spectives—as both individuals’ tendency to followtheir own preferences and their tendency to followsocial influence. Following our theory, we ex-pected power to increase the degree to which par-ticipants’ own preferences (measured as Machia-vellianism) affect their willingness to deceive innegotiation and at the same time expected power todecrease the degree to which social influence af-fects their willingness to deceive. We expected pri-vate self-focus to mediate these interactive effects.

Participants

Two hundred sixty students volunteered to par-ticipate in an experiment described to them as on“decision making in an organizational context.”The average age was 21.59 (s.d. � 1.36), and 65.8percent were female. We asked participants to in-dividually respond to a package containing the ex-perimental materials, using pen and paper. Thepackages were randomly ordered and distributed inadvance based on a two (power: high vs. low) bytwo (social influence: ethical organizational cli-mate vs. no social influence) between-personsdesign.



Ethical preferences. To examine to what degreeparticipants followed their own preferences whenmaking (un)ethical decisions, we measured indi-vidual differences in Machiavellianism using the20-item Mach IV Scale (Christie & Geis, 1970). Re-spondents rated each item (e.g., “It is hard to getahead without cutting corners here and there”;“The biggest difference between most criminalsand other people is that criminals are stupidenough to get caught”) on a scale ranging from 1,“disagree strongly,” to 7, “agree strongly” (� � .75).After reverse-scoring the appropriate items, we av-eraged the items for each participant so that higherscores reflected a stronger preference for Machia-vellian behavior.

Power manipulation. After participants had re-sponded to the Machiavellianism scale, we manip-ulated power using an essay-writing priming pro-cedure similar to that in Study 1. Participantsrecalled their experiences of either having power(high-power condition) or being powerless (low-power condition), ostensibly as part of a sepa-rate study.

Private self-focus measure. Private self-focuswas measured using the same scale as in Study 1(Govern & Marsch, 2001) (� � .77).

Ethical decision-making task and measure. Afterresponding to the private self-focus measure, partici-pants engaged in a simulated negotiations taskadapted from Schweitzer, DeChurch, and Gibson(2005: 2132). This task is based on the principles ofthe prisoner’s dilemma. All participants were as-signed to the same role and read the following:

You were recently promoted to head the CentralAmerican tour division, one of the most importantdivisions of the Roving Tours Company. This divi-sion has a big impact on the rest of the company andyou are being reminded to make sure you increaseprofits as much as possible. In order to increaseprofits, it is necessary to maximize the number oftours Roving Tours runs. However, a competingcompany, Wandering Tours, also operates at thesame Central American location. If both companiesincrease their tours to this location, this will makethe location less exotic. Therefore, the ideal out-come would be if your competitor ran the least tourspossible, while you ran the maximum number oftours. You received an e-mail from WanderingTours, describing the need to “cut back” in this area,and asking how many tours you plan to schedule forthe upcoming dry season (between 1 and 7). Yourcompetitor offered to match your number.

Participants indicated, on a seven-point scale, thenumber of tours that Roving Tours would actuallyrun the following season, and, on a separate seven-point scale, how many tours they would report toWandering Tours that they intended to run. Partic-ipants were provided with a classical prisoner’sdilemma payoff table (Schweitzer et al., 2005) andwere told that another participant, acting as theircompetitor, would be presented only with the fig-ure they announced and that their competitorwould base his or her decision on this (potentiallymisleading) figure. Therefore, consistently with theprinciples of prisoner’s dilemma, participants hadan incentive to run as many tours as possible, andthey also had an incentive to try to make the otherparty run as few tours as possible. To do this,however, they had to misrepresent their plannednumber of tours in the hope that the other partywould match the misrepresented number. To giveparticipants an additional incentive to misrepre-sent the number of tours they would run, they weretold that the person who reached the agreementmost favorable to Roving Tours would receive 50euros. The specific payoff was irrelevant in thisstudy, as the only variable of interests was thedifference between actual and stated number oftours, an act of purposeful deception and thereforean unethical decision (Strudler, 1995). In reality,there was no other party, and the recipient of the 50euros was randomly selected.

Social influence manipulation. To manipulatenormative social influence, we adapted an ethicalclimate manipulation designed by Aquino andBecker (2005). Ethical climate, or “typical organi-zational practices and procedures that have ethicalcontent” (Victor & Cullen, 1988: 101), constitutes asocial behavioral standard that influences individ-uals’ ethical behavior through normative socialpressures. The relevant paragraph was included atthe end of participants’ negotiation task instruc-tions. Participants in the ethical climate conditionread as follows:

Roving Tours prides itself on being fair and honestin its business dealings. Consequently, it goesagainst the standard practices of Roving Tours towithhold facts or “stretch” the truth in situationslike this one. In fact, you know of several managerswho have been ostracized by peers for not negotiat-ing completely honestly.

For the control condition (no social influence), thisparagraph was omitted. To strengthen the manipu-lation, all participants were informed that they


would discuss their decisions with their peers, alsorepresenting Roving Tours, following the negotia-tion task. Expectations of an immediate discussionwhen others’ views have been shown to strengthenthe effect of normative social influence (Cialdini,Levy, Herman, & Evenbeck, 1973; Wood & Quinn,2003).

The effectiveness of the ethical climate manipu-lation was checked by asking participants to indi-cate (1 � “not important” to 7 � “very important”)how important it was for Roving Tours to conductits business ethically. In addition, participants in-dicated (1 � “definitely believed” to 7 � “defi-nitely did not believe”), whether they believed theywould have to discuss their choices with otherparticipants while they were making the decision.

Control variables. Research suggests that powercan influence mood (Keltner et al., 2003), whichmight influence both how individuals negotiate(Bazerman, Curhan, Moore, & Valley, 2000) andhow they respond to moral issues (Wheatley &Haidt, 2005). Therefore, to control for possible ef-fects of mood, we asked participants to indicate(1 � “fully disagree” to 7 � “fully agree”) how theyfelt during the experiment (“happy,” “cheerful,”“satisfied,” “sad,” “unhappy,” “dejected”; � � .89).

In addition, we controlled for participants’ nego-tiating experience because experienced negotiatorsmight rely more readily on deception (Murnighan,Babcock, Thompson, & Pillutla, 1993). To accountfor this possibility, we asked participants to indi-cate how much negotiating experience they had(1 � “none” to 7 � “a great deal”).

Finally, participants were probed for suspicionand debriefed.

STUDY 3: RESULTS

Manipulation Checks

The ethical climate manipulation was successful;participants in the ethical climate condition re-ported that it was more important for Roving Toursto conduct its business ethically (mean � 6.28, s.d.� 0.92) than did participants in the no social influ-ence condition (mean � 4.41, s.d. � 0.96), t[258] �16.10, p � .001). In addition, all participants se-lected a value above the scale’s midpoint whenanswering whether they anticipated a discussionwith their peers regarding their negotiation deci-sions. Finally, an analysis of simple effects foundthat the power manipulation did not influence par-ticipants’ responses within the no social influence

condition (F[1, 256] � 0.69, p � .406, �p2 � .002),

excluding the possibility that the power manipula-tion itself influenced participants’ propensity todeceive.

Hypothesis 1 Test

Using regression analysis, we tested whetherpower (1) weakened the effect of social influences(ethical climate) and (2) strengthened the effect ofindividuals’ own preferences (Machiavellianism)on negotiation decisions. The results, presented inTable 2 (model 2), show that the power signifi-cantly moderated the effect of ethical climate (b �0.16, s.e. � 0.07, t[252] � 2.20, p � .028) as well asthe effect of Machiavellianism (b � 0.20, s.e.� 0.10, t[252] � 2.11, p � .036). We analyzedconditional effects to interpret the interactions.First, supporting the prediction that power reducesthe strength of social influences in ethical decisionmaking, the effect of ethical climate was weaker inthe high-power condition (b � �0.28, s.e. � 0.10,t[252] � �2.75, p � .006), than in the low-powercondition (b � �0.59, s.e. � 0.10, t[252] � �5.90,p � .001). Second, supporting the prediction thatpower strengthens the effect of individuals’ ownpreferences in ethical decision making, the effect ofMachiavellianism was stronger in the high-powercondition (b � 0.47, s.e. � 0.14, t[252] � 3.43, p �.001) than in the low-power condition (b � .07,s.e. � 0.14, t[252] � 0.48, p � .628).

Hypothesis 2 Test

To test whether private self-focus mediates themoderating effect of power on the effect of socialinfluences and ethical dispositions, it is necessaryto show that private self-focus simultaneously me-diates both the moderating effect of power on theeffect of ethical climate and the moderating effectof power on the effect of Machiavellianism on par-ticipants’ (un)ethical decisions.

As in Study 1, we used moderated path analysisto test this prediction (Edwards & Lambert, 2007).To support Hypothesis 2, power should be found tohave a positive effect on private self-focus. In ad-dition, private self-focus should weaken the effectof social influence (strengthen the effect of Machi-avellianism) on negotiation decisions. Finally, theproduct of these two paths should be significant.

Consistently with expectations, model 1 in Table2 shows that power positively affected private self-focus (b � 0.31, s.e. � 0.06, t[252] � 5.28, p � .001).


Moreover, model 3 in Table 2 shows that privateself-focus weakens the effect of ethical climate andstrengthens the effect of Machiavellianism. Specif-ically, the interaction between private self-focusand ethical climate was significant (b � 0.27, s.e.� 0.08, t[249] � 3.43, p � .001), with ethical cli-mate having a weaker effect when private self-focuswas one standard deviation above the mean (5.10; b� �0.21, s.e. � 0.10, t[249] � �2.08, p � .038) thanit did when private self-focus was one standarddeviation below the mean (3.12; b � �0.65, s.e.� 0.10, t[249] � �6.30, p � .001). The interactionbetween private self-focus and Machiavellianismwas also significant (b � 0.27, s.e. � 0.09, t[249] �3.02, p � .003), with Machiavellianism having astronger effect when private self-focus was onestandard deviation above the mean (5.10; b � 0.53,s.e. � 0.12, t[249] � 4.34, p � .001) than when itwas one standard deviation below the mean (3.12; b� �0.03, s.e. � 0.14, t[249] � �0.26, p � .796).

Next, we tested the significance of the product ofthe path from power to private self-focus and thepath from the private self-focus by ethical climateinteraction (the private self-focus by Machiavel-lianism interaction) to deception. We used the co-efficients from the prior analyses and applied thebootstrap method to construct bias-corrected confi-dence intervals of the indirect effects based on10,000 random replacements from the full sample.The 95% bias-corrected confidence interval for the

size of the indirect effect excluded zero (indicatinga significant indirect effect) for both the indirecteffect of the power by ethical climate interaction(0.025–0.160) as well as the power by Machiavel-lianism interaction (0.016–0.193). Therefore, pri-vate self-focus mediated the moderating role ofpower in the effect of ethical climate, as well asthe effect of Machiavellianism, on participants’(un)ethical decisions in the negotiation task.

STUDY 3: DISCUSSION

Replicating the pattern of results in Studies 1 and2, Study 3 showed that power determines the effec-tiveness of social influences in ethical decisionmaking. Study 3 tested the moderating effect ofpower in a different social influence paradigm (nor-mative social influences), further increasing confi-dence in the generalizability of our findings. Par-ticipants whose sense of power was experimentallyincreased, relative to participants whose sense ofpower was experimentally decreased, were less af-fected by social influence and relied more on theirown preferences (measured as Machiavellianism)when making ethical decisions. Finally, mediationanalyses showed that power weakens the effect ofsocial influences because it increases private self-focus, making individuals focus less on their ontheir environment and more on their ownpreferences.

TABLE 2Study 3, Hypotheses 1 and 2 Test: Regression Analysis Resultsa

Variables

Model 1: Private Self-Focus Model 2: Deception, Step 1 Model 3: Deception, Step 2


Constant 4.11** 3.99 4.23 2.57** 2.43 2.71 2.55** 2.41 2.68Mood 0.03 �0.06 0.12 �0.04 �0.15 0.06 �0.04 �0.15 0.06Negotiation experience 0.03 �0.08 0.15 0.08 �0.06 0.22 0.05 �0.08 0.19Power 0.31** 0.19 0.43 0.06 �0.08 0.20 0.06 �0.09 0.20Ethical climate 0.06 �0.05 0.18 �0.43** �0.57 �0.29 �0.43** �0.57 �0.30Machiavellianism �0.01 �0.16 0.16 0.27** 0.08 0.46 0.25** 0.06 0.44Power � ethical climate �0.01 �0.12 0.11 0.16* 0.02 0.30 0.09 �0.05 0.24Power � Machiavellianism �0.01 �0.17 0.15 0.20* 0.01 0.40 0.14 �0.06 0.34Self-focus 0.03 �0.11 0.18Self-focus � ethical climate 0.27** 0.11 0.42Self-focus � Machiavellianism 0.27** 0.09 0.44

R2 .11** .19** .24**�R2 .05**

a LLCI � 95% confidence interval lower limit; ULCI � 95% confidence interval upper limit. All categorical variables were effect-codedand all continuous variables were mean-centered for this analysis.

* p � .05** p � .01


STUDY 4: METHODS

Study 4 tested whether power weakens the effectof social influence when social influence is basedon purposeful pressures on behavior, rather thanon a decision maker’s internal motivation to con-form. Cialdini and Trost (1998) noted that compli-ance behavior often depends on a number of real-world factors, so we sought to complement ourexperimental findings with a field study in order toincrease the ecological validity of our research. Weassessed organizational compliance pressures byasking a sample of managers about the ethical cul-ture in their organization (Treviño et al., 1998).Ethical culture consists of “formal and informalorganizational systems aimed at behavioral con-trol” (Kish-Gephart et al., 2010: 7) and is thus con-sistent with the notion of compliance as acquies-cence to an intentional influence (Cialdini & Trost,1998).

In Study 4, we measured a variety of unethicalbehaviors to examine the behavioral consequencesof power in ethical decision making. Finally, wealso measured power, ethical preferences, and pri-vate self-focus in different ways. As in previousstudies, we expected power to weaken the effect ofsocial influence on unethical behaviors (and thusmake individuals more likely to rely on their ownpreferences), and we expected private self-focus tomediate these interactive effects.

Sample and Procedure

We approached individuals at an internationalairport and told them that we were conducting astudy concerning managers’ experiences at work.Of those who identified themselves as managers,30.38 percent (n � 312) agreed to participate in thestudy. Because of the sensitive nature of the re-search topic, they were assured of complete ano-nymity and were not asked to provide any personaldata apart from demographic information. In addi-tion, the order of the measures was randomized foreach participant to minimize order effects (Podsa-koff, MacKenzie, Lee, & Podsakoff, 2003). The av-erage age of respondents was 43 years (s.d.� 10.25), and they had an average 18.80 years ofworking experience (s.d. � 10.36). Of these, 12.2percent were first-level supervisor or at a lowerorganizational level; 27.2 percent were midlevelmanagers; 35.9 percent were upper-level managers;and 24.7 percent were executive-level managers.Women comprised 37.5 percent of the respondents.

Measures

Organizational ethical compliance pressures.We operationalized organizational compliancepressures mandating ethical behavior by measuringperceived organizational ethical culture (Treviño etal., 1998). We used the 14-item ethical environmentsubscale of the ethical climate measure (Treviño etal., 1998) to assess the degree to which respon-dents’ organizations punished unethical behaviorand rewarded ethical behavior, the degree to whichtheir leaders served as ethical role models, andmore. Respondents indicated their agreement withstatements describing their organizations (e.g.,“Unethical behavior is punished in this organiza-tion” and “Management in this organization disci-plines unethical behavior when it occurs”) on ascale ranging from 1, “completely false,” to 7 “com-pletely true” (� � .92).

Ethical preferences. We measured individuals’preference to behave ethically or unethically usingthe five-item internalization subscale of Aquinoand Reed’s (2002) moral identity scale, which mea-sures how iimportant moral identity is to the self.The internalization subscale examines whethermoral trait associations are embedded in a person’sself-concept. Past research has shown that the in-ternalization subscale predicts ethical decisionsand behavior (Aquino & Reed, 2002; Aquino, Reed,Thau, & Freeman, 2007). Participants indicated (1� “disagree strongly” to 7 � “agree strongly”), howimportant traits (e.g., “caring,” “compassionate,”“fair,” etc.) were to them (e.g., “It would make mefeel good to be a person who has these characteris-tics”). Items were averaged to form a single scale (�� .78).

Sense of power. Respondents’ sense of powerwas measured using the sense of power scale (An-derson & Galinsky, 2006). Respondents indicatedtheir agreement (1 � “strongly disagree” to 7 �“strongly agree”) with eight items (e.g., “In my re-lationships with others, I think I have a great deal ofpower” and “If I want to, I get to make the deci-sions”; � � .88).

Private self-focus. We used the private self-con-sciousness scale to measure respondents’ disposi-tional tendency to focus their attention on self(Fenigstein, Scheier, & Buss, 1975). Respondentsindicated whether the ten behaviors described inthe scale (e.g., “I’m generally attentive to my innerfeelings” and “I’m constantly examining my mo-tives”) were characteristic of them (1 � “extremely


uncharacteristic” to 7 � “extremely characteristic”;� � .72).

(Un)ethical behavior. We used seven items frompast research describing behaviors that violate or-ganizational ethical norms (Bennett & Robinson,2000; Treviño & Weaver, 2001): “Taken propertyfrom work without permission,” “Falsified a re-ceipt to get reimbursed for more money than youspent on business expenses,” “Taken an additionalor longer break than is acceptable,” “Intentionallyworked slower than you could have worked,” “Dis-cussed confidential company information with anunauthorized person,” “Used an illegal drug orconsumed alcohol on the job,” and “Dragged outwork in order to get overtime” (� � .82). The mea-sure asked respondents to indicate the frequencywith which they engaged in these behaviors (1 �“never” to 7 � “daily”); therefore, higher valuesreflected a greater frequency of unethical behaviorsat work.

Pretest of unethical behavior items. To confirmthat the seven items we used describe behaviorsthat can be considered as unethical, we conducteda pretest using an independent sample of employedadults (n � 94). We asked the participants to indi-cate (1 � “definitely ethical,” 7 � “definitely un-ethical”) how ethical the behaviors described ineach item were. Following ethical decision-makingresearch, we defined unethical behavior as any ac-tion that is “either illegal or morally unacceptableto the larger community” (Jones, 1991: 367). All theitems received a mean score of 5 (“somewhat un-ethical”) or higher (average mean score was 5.84,s.d. � 0.89), confirming that the items describeunethical behaviors and adding to the ecologicalvalidity of our measure.

Control variables. Although respondents wereassured of complete anonymity, we controlled forparticipants’ tendency to respond in a socially de-sirable manner using the impression management

subscale from the Balance Inventory of DesirableResponding (Paulhus, 1984). Respondents indi-cated (1 � “not true,” 7 � “very true”) their agree-ment with statements (e.g., “Once in a while I laughat a dirty joke”), and we averaged responses to forma single continuous scale (Stöber, Dette, & Musch,2002) (� � .76).

The means, standard deviations, internal consis-tency of the measures, and correlations among vari-ables are presented in Table 3.

STUDY 4: RESULTS

Hypothesis 1 Test

Regression analysis was used to test Hypothesis1. As summarized in Table 4 (model 2), the inter-action between power and ethical culture is signif-icant (b � 0.08, s.e. � 0.04, t[304] � 2.17, p � .031).Ethical culture significantly decreased unethicalbehaviors when power was one standard deviationbelow the mean (3.51; b � �0.13, s.e. � 0.05, t[304]� �2.33, p � .021), but it had no effect when powerwas one standard deviation above the mean (5.55;b � 0.04, s.e. � 0.05, t[304] � 0.73, p � .463).Therefore, power significantly weakened compli-ance with organizational ethical culture.

In addition, the interaction between power andmoral identity is also significant (b � �0.10, s.e.� 0.04, t[304] � �2.42, p � .017), with moralidentity having no effect on unethical behaviorswhen power was one standard deviation below themean (3.51; b � 0.01, s.e. � 0.07, t[304] � 0.15, p �.887) but significantly decreasing unethical behav-iors when power was one standard deviation abovethe mean (5.55; b � 0.20, s.e. � 0.06, t[304] ��3.21, p � .001). Therefore, power significantlystrengthened individuals’ reliance on their ownpreferences when they were engaging in unethicalbehavior.

TABLE 3Means, Standard Deviations, Internal Consistencies, and Correlations, Study 4a

Variables Mean s.d. 1 2 3 4 5 6 7

1. Age 43.00 10.252. Impression management 3.48 0.65 .01 (.72)3. Etdical culture 4.15 1.13 �.02 �.04 (.92)4. Moral identity 6.08 0.88 .13 .19 �.01 (.83)5. Sense of power 4.53 1.02 �.01 �.01 �.04 �.02 (.90)6. Private self-focus 3.59 0.63 .01 .02 .02 .03 .18 (.73)7. Unethical behavior 2.05 0.76 �.11 �.14 �.06 �.12 �.03 .03 (.91)

a n � 312. All values greater than |.12| are significant. Cronbach’s coefficient alphas are displayed on the diagonal.


Hypothesis 2 Test

To test Hypothesis 2, we followed the same mod-erated path analysis procedure as in Studies 1 and3. Accordingly, power should have a positive effecton private self-focus. Moreover, private-self focusshould weaken the effect of ethical culture(strengthen the effect of moral identity) on unethi-cal behaviors. Finally, the product of these twopaths should be significant.

Consistently with expectations, model 1 in Table4 shows that power positively affected private self-focus (b � 0.11, s.e. � 0.03, t[304] � 3.13, p � .002).Moreover, model 3 shows that private self-focusweakens the effect of ethical culture and strength-ens the effect of moral identity. Specifically, theinteraction between private self-focus and ethicalculture was significant (b � 0.18, s.e. � 0.06, t[304]� 2.80, p � .006); ethical culture decreased uneth-ical behavior when private self-focus was one stan-dard deviation below the mean (2.96; b � �0.16,s.e. � 0.05, t[304] � �2.98, p � .003), but it had noeffect when private self-focus was one standarddeviation above the mean (4.22; b � 0.07, s.e.� 0.06, t[304] � 1.18, p � .239). In addition, theinteraction between private self-focus and moralidentity was also significant (b � �0.20, s.e. � 0.08,t[304] � 2.58, p � .010); moral identity signifi-cantly decreased unethical behavior when privateself-focus was one standard deviation above the

mean (4.22; b � �0.21, s.e. � 0.07, t[304] � �3.05,p � .003), but it had no effect when private self-focus was one standard deviation below the mean(2.96; b � 0.04, s.e. � 0.07, t[304] � 0.58, p � .565).

Next, we tested the significance of the product ofthe path from power to private self-focus and thepath from the private self-focus by ethical cultureinteraction (the private self-focus by moral identityinteraction) to unethical behavior. We used the co-efficients from the prior analyses and applied thebootstrap method to construct bias-corrected confi-dence intervals of the indirect effects based on10,000 random replacements from the full sample.The 95% bias-corrected confidence interval for thesize of the indirect effect excluded zero (indicatinga significant indirect effect) for both the indirecteffect of the power by ethical culture interaction(0.006–0.042) and the power by moral identity in-teraction (�0.053 to �0.005). Therefore, privateself-focus mediated the moderating role of power inthe effect of ethical culture, as well as the effect ofmoral identity, on respondents’ unethical behavior.

STUDY 4: DISCUSSION

Study 4 demonstrated that power moderates theeffect of social influences on unethical behaviorseven when social influences are based on compli-ance pressures. Power again weakened the effect of

TABLE 4Study 4, Hypotheses 1 and 2 Test: Regression Analysis Resultsa

Variables

Model 1: Private Self-FocusModel 2: Unethical Behavior,

Step 1Model 3: Unethical Behavior,

Step 2


Constant 3.59 3.52 3.66 2.47 2.10 2.83 2.41 2.06 2.77Age 0.01 �0.01 0.01 �0.01* �0.02 �0.01 �0.01* �0.02 �0.01Impression management 0.02 �0.09 0.13 �0.14* �0.27 �0.01 �0.14* �0.27 �0.01Power 0.11** 0.04 0.18 �0.04 �0.12 0.04 �0.06 �0.14 0.02Ethical culture 0.01 �0.05 0.08 �0.04 �0.12 0.03 �0.05 �0.12 0.02Moral identity 0.02 �0.06 0.10 �0.09 �0.19 0.01 �0.09 �0.18 0.02Power � ethical culture 0.05 �0.01 0.11 0.08* 0.01 0.15 0.05 �0.03 0.12Power � moral identity �0.02 �0.09 0.05 �0.10* �0.19 �0.02 �0.06 �0.15 0.03Self-focus 0.06 �0.08 0.19Self-focus � ethical culture �0.20* �0.36 �0.05Self-focus � moral identity 0.18* 0.05 0.30

R2 .05* .08** .12**�R2 .04**

a LLCI � 95% confidence interval lower limit; ULCI � 95% confidence interval upper limit. All continuous variables were mean-centered for the analysis.

* p � .05** p � .01


social influences (ethical culture), making individ-uals more likely to follow their own preferences(moral identity), and private self-focus mediatedthis effect. Study 4 added to the ecological validityof our research by testing the moderating effect ofpower among a diverse sample of managers.

GENERAL DISCUSSION

Four studies tested the idea that power weakensthe effect of social influences, and instead makesindividuals more likely to follow their own prefer-ences when responding to moral issues. This effectwas tested across the three main paradigms of so-cial influence: informational influence (Studies 1and 2), normative influence (Study 3), and compli-ance (Study 4). We operationalized power in differ-ent ways (primed, structurally manipulated, andmeasured), and used different designs to test ourtheory. All four studies offer converging evidencethat power reduces the effect of social influences inethical decision making.

In Study 1, relative to the control group, partici-pants primed with high power conformed less toothers’ responses when responding to an ambigu-ous ethical dilemma. Participants primed with lowpower, in contrast, conformed more than the con-trol group. Private self-focus mediated the moder-ating effect of power on the effect of social influ-ence. Study 2 provided a more rigorous test of thehypothesized psychological mechanism by manip-ulating private self-focus directly. We found thatpower (this time structurally manipulated) reducedthe effect of social influence for participants in theexperimental control condition, but when partici-pants’ private self-focus was experimentallyheightened, the effect of power disappeared, indi-cating that the effect of power depends on its abilityto heighten a decision maker’s private self-focus. InStudy 3, participants primed with high power wereless affected by normative social influences (ethicalclimate) than were those primed with low power,and those primed with high power were insteadmore likely to follow their own preferences (opera-tionalized as Machiavellianism) when decidingwhether to deceive their opponent in negotiations,and private self-focus explained these effects. Fi-nally, using a survey among a sample of managers,Study 4 demonstrated that as the level of managers’subjective sense of power increased, the effect ofcompliance pressures (ethical culture) decreasedand instead the effect of individuals’ own prefer-ences (moral identity) increased with respect to

unethical behavior. Again, private self-focus medi-ated the effects.

Theoretical Implications

This work has implications for several areas ofresearch. First, our research contributes to the eth-ical decision-making literature by identifying self-focus caused by power as an explanation for whenand why organizational environments influenceemployees to make (un)ethical decisions. A largebody of research focuses on understanding ethicaldecision making in organizations (Kish-Gephart etal., 2010), and this research suggests that the char-acteristics of an organizational environmentstrongly influence individuals’ behavior (Gino etal., 2009; Pierce & Snyder, 2008; Treviño, 1986;Treviño et al., 1998). However, not all employeesreact equally to social influences, and past researchhas left unexplained the question of when and whyorganizational environments affect employees.This explanatory problem has precluded a deeperunderstanding of a range of organizational phe-nomena, including ethical climate (Victor & Cullen,1988), ethical culture (Treviño et al., 1998; Treviño,1986) and conformity to (un)ethical colleagues(Gino et al., 2009; Pierce & Snyder, 2008). Ourresearch emphasized the fact that ethical decisionmaking occurs in the context of power differencesand identified variations in power as a parsimoni-ous explanation for when and why organizationalenvironments affect employees’ ethical decisionmaking. We show that power makes people morefocused on their own preferences and thus makesthem less likely to be influenced by characteristicsof their organizational environment when respond-ing to moral issues.

The present research also contributes to socialpsychological research on power by developingand testing, for the first time, an explanation for therole of power in the effectiveness of social influ-ences. We hypothesized and found in four studiesthat power increases private-self focus, which, inturn, determines the likelihood of individuals’ act-ing on their own preferences, at the expense of theeffect of social influences. This finding extendsprevious research on power and social influence(Galinsky et al., 2008; Guinote, 2008) by providinga more comprehensive account of the role of powerin individuals’ susceptibility to social influence. Inso doing, our work allows for more detailed predic-tions of how and when having or not having powermight influence individuals’ susceptibility to so-


cial influence. For instance, in Study 2, we foundthat being powerless, as compared to being in aposition of power, did not increase susceptibility tosocial influence when private self-focus washeightened. Future research can integrate the liter-atures on power and self-focus in a similar mannerto explore additional implications of the role ofprivate self-focus in the interaction between powerand social influence. For example, past researchhas shown that public self-focus, in contrast toprivate self-focus, makes individuals pay greaterattention to standards of conduct (Fenigstein, 1979;Froming & Carver, 1981). It is possible that raisingpublic self-consciousness could alleviate potentialnegative effects of power, such as the tendency todisregard ethical organizational standards (as doc-umented in our Study 4) or ignore useful advice(See et al., 2011; Tost et al., 2011). Future researchis needed to explore such possibilities.

Our research also extends the management liter-ature on power. Management research has gener-ally implied that power has a negative impact onemployees’ behavior toward others (Dubin, 1982;Kipnis, Castell, Gergen, & Mauch, 1976; Kochan,2002). For instance, commenting on the numerousinstances of problematic behaviors in corporationsin recent years, Kochan concluded this: “Powerbecame highly concentrated at the top of organiza-tions, and the adage that ‘power corrupts and abso-lute power corrupts absolutely’ once again hasproven true” (2002: 139). However, our findingsprovide a more nuanced interpretation of the roleof power (cf. Chen, Lee-Chai, & Bargh, 2001; Over-beck & Park, 2001). Power is not a corruptive forcein itself, but it does make employees more self-focused. Self-focus, in turn, weakens the effect ofsocial influences in organizations, thus making thepowerful more likely to follow their own prefer-ences, whether ethical or unethical.

Limitations and Future Research

We offered a parsimonious explanation for whenand why organizational environments affect em-ployees’ ethical decision making. Future researchmay examine more complex theoretical models.For example, it is possible that other factors, suchas employees’ need for affiliation (Cialdini & Trost,1998; McGhee & Teevan, 1967) or the characteris-tics of the source of social influence (Cialdini &Goldstein, 2004; Gino et al., 2009), affect employ-ees’ susceptibility to social influence when they aremaking (un)ethical decisions. Similarly, future re-

search could examine whether documented psy-chological consequences of power other than self-focus, such as self-confidence (Briñol et al., 2007;Tost et al., 2011), explain why powerful peoplemake (un)ethical decisions more consistent withtheir preferences.

We also did not systematically examine whetherdifferences in moral issues affect the role of powerin ethical decision making. Moral issues can differin terms of a number of characteristics, such as themagnitude of consequences or proximity of thoseaffected by a decision maker (Jones, 1991). How-ever, we believe that the role of power we docu-mented holds for various types of moral issues.Specifically, social psychological research hasdemonstrated that a heightened self-focus de-creases susceptibility to social influence regardlessof the decision-making issue at hand (Froming etal., 1982; Gibbons & Wright, 1983; Hutton &Baumeister, 1992; Pryor, Gibbons, Wicklund, Fa-zio, & Hood, 1977). Since power reliably increasesself-focus, as documented in our studies and im-plied by other research (Guinote, 2010; Weick &Guinote, 2008), we believe that our findings gener-alize across the different types of moral issues. Nev-ertheless, issue-contingent perspectives on ethicaldecision making (Jones, 1991) do have importantlinks with our research more generally. For in-stance, Jones proposed that when social consensus,or “the degree of social agreement that a proposedact is evil (or good)” (1991: 375) is low, ambiguityin ethical decision making is created. Prior researchshows that ambiguity can make it more likely foremployees to observe and mimic others’ decisions(Sherif, 1936; Sonenshein, 2007), rendering the roleof power more salient. Therefore, characteristics ofmoral issues that render (un)ethical social influ-ences more or less salient should make the role ofpower more or less relevant. Future research isneeded to explore such interplay between charac-teristics of moral issues and power in ethical deci-sion making.

A related point is that we treated ethical andunethical social influences, as well as individuals’ethical and unethical dispositions, in a similarfashion. Our theoretical assumption is that powerreduces the effectiveness of social influence, re-gardless of whether the influence is ethical or un-ethical. Moreover, our theory is agnostic as towhether a decision maker’s preferences are for eth-ical or unethical conduct. Although we find sup-port for our predictions across different types ofsocial influence and different operationalizations


of ethical preferences, it is possible that the effectswe documented differ depending on the particularcombination of type of social influence and indi-vidual preferences. Future research might explorethis possibility by systematically manipulating thecontent of social influence (ethical vs. unethical) aswell as decision makers’ preferences.

Future research could also replicate our findingsusing different measures of (un)ethical decisionmaking and behavior, including more objective in-dicators of ethical and unethical behavior (seeKish-Gephart et al. [2010] for a discussion of differ-ent approaches to the measurement of (un)ethicaldecisions and behavior). Although we found sup-port for our theory using different operationaliza-tions of (un)ethical decisions and behavior, futureresearch could extend our work to more objectiveindicators of unethical behavior (e.g., Detert, Treviño,Burris, & Andiappan, 2007; Gino et al., 2009).

Finally, future work could explore in more detailthe downstream consequences of the interplay be-tween power and social influences in ethical deci-sion making. Our model leads to a number of inter-esting predictions—for instance, consider the caseof an individual with strong unethical preferencesbut who is working in an organization that empha-sizes ethical conduct and enforces its ethical poli-cies diligently. Power should make such an indi-vidual more likely to disregard such the ethicalorganizational context and instead behave unethi-cally. This disregarding, in turn, could ultimatelycome at the expense of this individual’s self-inter-est, as it violates organizational ethical policies,and the employee is likely to be punished. There-fore, following our model, power could in certaincases have the ironic consequence of making em-ployees act against their best interest and harmtheir standing in their organization.

Managerial Implications

Organizations implement various systems to pro-mote ethical employee behavior, ranging from for-mal policies such as incentive and sanctioning sys-tems (James, 2000; McCabe et al., 1996; Schwartz,2001) to more informal systems such as value state-ments and peer socialization (Treviño et al., 1998).These systems have in common that they rely on asocial influence logic to affect ethical decisionmaking. Therefore, our findings can inform organ-izational decision makers of when these systemsare more or less effective. For example, our resultssuggest that people occupying powerful organiza-

tional roles will be the least likely to respond tosocial influence–based systems. Powerful employ-ees may need more incentives (rewards and pun-ishments) to pay attention to these systems than dopeople lacking power. Another practical implica-tion of our results is that organizations will beparticularly challenged to foster ethical conductamong powerful organization members if thesemembers have unethical preferences. This findingunderscores the importance of selecting the “right”people to hold positions of power in organizations.

Our results also imply that employees lackingpower might be particularly susceptible to unethi-cal social influence. Organizations could channeltheir efforts to prevent the diffusion of unethicalbehaviors onto organizational relationships inwhich power differences are salient (e.g., supervi-sors and subordinates). Employees lacking powermay be particularly susceptible and more easilypersuaded to join in unethical conduct by morepowerful organization members (Treviño & Brown,2005). Our findings suggest that organizational in-terventions promoting employees’ private self-fo-cus have the potential to reduce the susceptibilityof employees lacking power. For instance, in Study2, individuals assigned to low-power positions(“subordinates”) were more susceptible to socialinfluence than individuals assigned to high-powerpositions (“managers”). Raising low-power indi-viduals’ private self-focus eliminated low-powerindividuals’ increased susceptibility to social influ-ence. This implies that organizations interested inpreventing unethical behavior should raise em-ployees’ private self-focus by training them to callto mind their personal standards whenever they areconfronted with a moral issue. Research on imple-mentation intentions suggests one way to do thispractically: Through practicing so-called if-thenplans (Gollwitzer, 1999), employees can adopt ahabit of focusing on their personal standards ofethical behavior when they face a moral issueat work.

Conclusion

We have shown that private self-focus caused bypower explains when and why employees disre-gard their organizational environment and insteadfollow their own (un)ethical preferences. Thesefindings are important for understanding variousethical and unethical social processes in organiza-tions—from the effectiveness of ethical organiza-tional policies to the process of organizational cor-


ruption—and thus contribute to an integrated andmore contextualized understanding of ethical deci-sion making.

REFERENCES

Allport, G. W. 1954. The historical background of socialpsychology. In G. Lindzey & E. Aronson (Eds.),Handbook of social psychology, vol. 1: 1–46. NewYork: Random House.

Anderson, C., & Galinsky, A. D. 2006. Power, optimism,and risk taking. European Journal of Social Psy-chology, 36: 511–536.

Anderson, C., & Berdahl, J. L. 2002. The experience ofpower: Examining the effects of power on approachand inhibition tendencies. Journal of Personalityand Social Psychology, 83: 1362–1377.

Aquino, K., & Becker, T. E. 2005. Lying in negotiations:How individual and situational factors influence theuse of neutralization strategies. Journal of Organi-zational Behavior, 26: 661–679.

Aquino, K., & Reed, A. 2002. The self-importance ofmoral identity. Journal of Personality and SocialPsychology, 83: 1423–1440.

Aquino, K., Reed, A. II, Thau, S., & Freeman, D. 2007. Agrotesque and dark beauty: How moral identity andmechanisms of moral disengagement influence cog-nitive and emotional reactions to war. Journal ofExperimental Social Psychology, 43: 385–392.

Asch, S. E. 1955. Opinions and social pressure. Scien-tific American, 193: 31–35.

Ashforth, B. E., & Anand, V. 2003. The normalization ofcorruption in organizations. In B. M. Staw & R. Sut-ton (Eds.), Research in organizational behavior,vol. 25: 1–52. Greenwich, CT: JAI.

Badaracco, J. 1997. Defining moments: When managersmust choose between right and right. Boston: Har-vard Business Press.

Baker, S. M., & Petty, R. E. 1994. Majority and minorityinfluence: Source-position imbalance as a determi-nant of message scrutiny. Journal of Personalityand Social Psychology, 67: 5–19.

Bazerman, M. H., Curhan, J. R., Moore, D. A., & Valley,K. L. 2000. Negotiation. In S. T. Fiske, D. L. Schacter,& C. Zahn-Wexler (Eds.), Annual review of psychol-ogy, vol. 51: 279–314. Palo Alto, CA: Annual Re-views.

Bennett, R. J., & Robinson, S. L. 2000. Development of ameasure of workplace deviance. Journal of AppliedPsychology, 85: 349–360.

Brass, D. J. 1984. Being in the right place: A structuralanalysis of individual influence in an organization.Administrative Science Quarterly, 29: 518–539.

Briñol, P., Petty, R. E., Valle, C., Rucker, D. D., & Becerra,A. 2007. The effects of message recipients’ powerbefore and after persuasion: A self-validation analy-sis. Journal of Personality and Social Psychology,93: 1040–1053.

Chen, S., Lee-Chai, A. Y., & Bargh, J. A. 2001. Relation-ship orientation as a moderator of the effects of so-cial power. Journal of Personality and Social Psy-chology, 80: 173–187.

Christie, R., & Geis, F. 1970. Studies in Machiavellian-ism. New York: Academic Press.

Cialdini, R. B., & Goldstein, N. J. 2004. Social influence:Compliance and conformity. In S. T. Fiske, D. L.Schacter, & C. Zahn-Wexler (Eds.), Annual reviewof psychology, vol. 55: 591–621. Palo Alto, CA: An-nual Reviews.

Cialdini, R. B., Levy, A., Herman, C. P., & Evenbeck, S.1973. Attitudinal politics: The strategy of modera-tion. Journal of Personality and Social Psychology,25: 100–108.

Cialdini, R. B., Petrova, P. K., & Goldstein, N. J. 2004. Thehidden costs of organizational dishonesty. MITSloan Management Review, 45: 67–74.

Cialdini, R. B., & Trost, M. R. 1998. Social norms, con-formity, and compliance. In D. T. Gilbert, S. T. Fiske,& G. Lindzey (Eds.), Handbook of social psychol-ogy, vol. 2: 151–192. Boston: McGraw-Hill.

Crutchfield, R. S. 1955. Conformity and character. Amer-ican Psychologist, 10: 195–198.

Cushman, F., Young, L., & Hauser, M. 2006. The role ofconscious reasoning and intuition in moral judg-ment: Testing three principles of harm. Psychologi-cal Science, 17: 1082–1089.

Detert, J. R., Treviño, L. K., Burris, E. R., & Andiappan, M.2007. Managerial modes of influence and counter-productivity in organizations: A longitudinal busi-ness-unit-level investigation. Journal of AppliedPsychology, 92: 993–1005.

Deutsch, M., & Gerard, H. 1955. A study of normative andinformational social influences upon individualjudgment. Journal of Abnormal and Social Psy-chology, 51: 629–636.

Dubin, R. 1982. Management: Meanings, methods, andmoxie. Academy of Management Review, 7: 372–379.

Edwards, J. R., & Lambert, L. S. 2007. Methods for inte-grating moderation and mediation: A general analyt-


ical framework using moderated path analysis. Psy-chological Methods, 12: 1–22.

Emerson, R. 1962. Power-dependence relations. Ameri-can Sociological Review, 27: 31–41.

Epley, N., & Gilovich, T. 1999. Just going along: Noncon-scious priming and conformity to social pressure.Journal of Experimental Social Psychology, 35:578–589.

Fenigstein, A. 1979. Self-consciousness, self-attention,and social interaction. Journal of Personality andSocial Psychology, 37: 75–86.

Fenigstein, A., & Levine, M. P. 1984. Self-attention, con-cept activation, and the causal self. Journal of Ex-perimental Social Psychology, 20: 231–245.

Fenigstein, A., Scheier, M. F., & Buss, A. H. 1975. Publicand private self-consciousness: Assessment and the-ory. Journal of Consulting and Clinical Psychology,43: 522–527.

Festinger, L. 1954. A theory of social comparison pro-cesses. Human Relations, 7: 117–140.

Fiske, S. T. 1993. Controlling other people: The impact ofpower on stereotyping. American Psychologist, 48:621–628.

Fiske, S. T., & Dépret, E. 1996. Control, interdependenceand power: Understanding social cognition in itssocial context. European Review of Social Psychol-ogy, 7: 31–61.

Flynn, F. J., & Wiltermuth, S. S. 2010. Who’s with me?False consensus, brokerage, and ethical decisionmaking in organizations. Academy of ManagementJournal, 53: 1074–1089.

Fransen, M. L., Fennis, B. M., Pruyn, A. T. H., & Vohs,K. D. 2011. When fit fosters favoring: The role ofprivate self-focus. Journal of Experimental SocialPsychology, 47: 202–207.

French, J., & Raven, B. H. (Eds.). 1959. The bases ofsocial power. Ann Arbor, MI: Institute for SocialResearch.

Froming, W. J., & Carver, C. S. 1981. Divergent influencesof private and public self-consciousness in a compli-ance paradigm. Journal of Research in Personality,15: 159–171.

Froming, W. J., Walker, G. R., & Lopyan, K. J. 1982. Publicand private self-awareness: When personal attitudesconflict with societal expectations. Journal of Ex-perimental Social Psychology, 18: 476–487.

Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. 2003.From power to action. Journal of Personality andSocial Psychology, 85: 453–466.

Galinsky, A. D., Magee, J. C., Gruenfeld, D. H., Whitson,J. A., & Liljenquist, K. A. 2008. Power reduces thepress of the situation: Implications for creativity,conformity, and dissonance. Journal of Personalityand Social Psychology, 95: 1450–1466.

Gibbons, F. X., & Wright, R. A. 1983. Self-focused atten-tion and reactions to conflicting standards. Journalof Research in Personality, 17: 263–273.

Gino, F., Ayal, S., & Ariely, D. 2009. Contagion anddifferentiation in unethical behavior. PsychologicalScience, 20: 393–398.

Gollwitzer, P. M. 1999. Implementation intentions:Strong effects of simple plans. American Psycholo-gist, 54: 493–503.

Goukens, C., Dewitte, S., & Warlop, L. 2009. Me, myself,and my choices: The influence of private self-aware-ness on choice. Journal of Marketing Research, 46:682–692.

Govern, J. M., & Marsch, L. A. 2001. Development andvalidation of the situational self-awareness scale.Consciousness and Cognition, 10: 366–378.

Grant, A. M., & Berry, J. W. 2011. The necessity of othersis the mother of invention: Intrinsic and prosocialmotivations, perspective taking, and creativity.Academy of Management Journal, 54: 73–96.

Griffin, D., & Buehler, R. 1993. Role of construal pro-cesses in conformity and dissent. Journal of Person-ality and Social Psychology, 65: 657–669.

Guinote, A. 2007a. Power and goal pursuit. Personalityand Social Psychology Bulletin, 33: 1076–1087.

Guinote, A. 2007b. Behaviour variability and the situatedfocus theory of power. European Review of SocialPsychology, 18: 256–295.

Guinote, A. 2008. Power and affordances: When the sit-uation has more power over powerful than power-less individuals. Journal of Personality and SocialPsychology, 95: 237–252.

Guinote, A. 2010. In touch with your feelings: Powerincreases reliance on bodily information. SocialCognition, 28: 110–121.

Haidt, J. 2001. The emotional dog and its rational tail: Asocial intuitionist approach to moral judgment. Psy-chological Review, 108: 814–834.

Haidt, J., & Hersh, M. A. 2001. Sexual morality: Thecultures and emotions of conservatives and liberals.Journal of Applied Social Psychology, 31: 191–221.

Hutton, D. G., & Baumeister, R. F. 1992. Self-awarenessand attitude change: Seeing oneself on the centralroute to persuasion. Personality and Social Psy-chology Bulletin, 18: 68–75.


James, H. S. 2000. Reinforcing ethical decision makingthrough organizational structure. Journal of Busi-ness Ethics, 28: 43–58.

Janes, L. M., & Olson, J. M. 2000. Jeer pressure: Thebehavioral effects of observing ridicule of others.Personality and Social Psychology Bulletin, 26:474–485.

Jones, T. M. 1991. Ethical decision making by individu-als in organizations: An issue-contingent model.Academy of Management Review, 16: 366–395.

Keltner, D., Gruenfeld, D. H., & Anderson, C. 2003.Power, approach, and inhibition. Psychological Re-view, 110: 265–284.

Kipnis, D., Castell, P. J., Gergen, M., & Mauch, D. 1976.Metamorphic effects of power. Journal of AppliedPsychology, 61: 127–135.

Kish-Gephart, J. J., Harrison, D. A., & Treviño, L. K. 2010.Bad apples, bad cases, and bad barrels: Meta-analyticevidence about sources of unethical decisions atwork. Journal of Applied Psychology, 95: 1–31.

Kochan, T. A. 2002. Addressing the crisis in confidencein corporations: Root causes, victims, and strategiesfor reform. Academy of Management Executive, 16:139–141.

Loyd, D. L., Kern, M. C., & Thompson, L. 2005. Classroomresearch: Bridging the ivory divide. Academy ofManagement Learning and Education, 4: 8–21.

Madden, J. M. 1960. Personal preferences and confor-mity. Journal of Social Psychology, 52: 269–277.

McCabe, D., Treviño, L., & Butterfield, K. 1996. The in-fluence of collegiate and corporate codes of conducton ethics-related behavior in the workplace. Busi-ness Ethics Quarterly, 6: 461–476.

McFerran, B., Dahl, D. W., Fitzsimons, G. J., & Morales,A. C. 2010. I’ll have what she’s having: Effects ofsocial influence and body type on the food choices ofothers. Journal of Consumer Research, 36: 915–929.

McGhee, P. E., & Teevan, R. C. 1967. Conformity behaviorand need for affiliation. Journal of Social Psychol-ogy, 72: 117–121.

Murnighan, J. K., Babcock, L., Thompson, L., & Pillutla,M. M. 1993. The information dilemma in negotia-tions: Effects of experience, incentives, and integra-tive potential. International Journal of ConflictManagement, 10: –339313–121.

Overbeck, J. R., & Park, B. 2001. When power does notcorrupt: Superior individuation processes amongpowerful perceivers. Journal of Personality and So-cial Psychology, 81: 549–565.

Overbeck, J., & Park, B. 2006. Powerful perceivers, pow-erless objects: Flexibility of powerholders’ social at-tention. Organizational Behavior and Human De-cision Processes, 99: 227–243.

Paulhus, D. L. 1984. Two-component models of sociallydesirable responding. Journal of Personality andSocial Psychology, 46: 598–609.

Pierce, L., & Snyder, J. 2008. Ethical spillovers in firms:Evidence from vehicle emissions testing. Manage-ment Science, 54: 1891–1903.

Pinto, J., Leana, C. R., & Pil, F. K. 2008. Corrupt organi-zations or organizations of corrupt individuals? Twotypes of organization-level corruption. Academy ofManagement Review, 33: 685–709.

Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsa-koff, N. P. 2003. Common method biases in behav-ioral research: A critical review of the literature andrecommended remedies. Journal of Applied Psy-chology, 88: 879–903.

Pryor, J. B., Gibbons, F. X., Wicklund, R. A., Fazio, R. H.,& Hood, R. 1977. Self-focused attention and self-report validity. Journal of Personality, 45: 513–527.

Russell, J. 2011. Olympus reveals details of accountingscandal. http://www.telegraph.co.uk/finance/newsbysector/mediatechnologyandtelecoms/electronics/8875766/Olympus-reveals-details-of-accounting-scandal.html. Accessed November 8.

Scheier, M. F., & Carver, C. S. 1980. Private and publicself-attention, resistance to change, and dissonancereduction. Journal of Personality and Social Psy-chology, 39: 390–405.

Schwartz, M. 2001. The nature of the relationship be-tween corporate codes of ethics and behaviour. Jour-nal of Business Ethics, 32: 247–262.

Schweitzer, M. E., DeChurch, L. A., & Gibson, D. E. 2005.Conflict frames and the use of deception: Are com-petitive negotiators less ethical? Journal of AppliedSocial Psychology, 35: 2123–2149.

See, K. E., Morrison, E. W., Rothman, N. B., & Soll, J. B.2011. The detrimental effects of power on confi-dence, advice taking, and accuracy. OrganizationalBehavior and Human Decision Processes, 116:272–285.

Sherif, M. 1936. The psychology of social norms. NewYork: Harper.

Shrout, P. E., & Bolger, N. 2002. Mediation in experimen-tal and nonexperimental studies: New proceduresand recommendations. Psychological Methods, 7:422–445.


Sonenshein, S. 2007. The role of construction, intuition,and justification in responding to ethical issues atwork: The sensemaking-intuition model. Academyof Management Review, 32: 1022–1040.

Spencer, S. J., Zanna, M. P., & Fong, G. T. 2005. Estab-lishing a causal chain: Why experiments are oftenmore effective than mediational analyses in examin-ing psychological processes. Journal of Personalityand Social Psychology, 89: 845–851.

Stöber, J., Dette, D. E., & Musch, J. 2002. Comparingcontinuous and dichotomous scoring of the Bal-anced Inventory of Desirable Responding. Journal ofPersonality Assessment, 78: 370–389.

Strudler, A. 1995. On the ethics of deception in negotia-tion. Business Ethics Quarterly, 5: 805–822.

Tang, J. H., & Farn, C. K. 2005. The effect of interpersonalinfluence on softlifting intention and behaviour.Journal of Business Ethics, 56: 149–161.

Thibaut, J. W., & Kelley, H. H. 1959. The social psychol-ogy of groups. New York: Wiley.

Tost, L. P., Gino, F., & Larrick, R. P. 2011. Power, com-petitiveness, and advice taking: Why the powerfuldon’t listen. Organizational Behavior and HumanDecision Processes, 117: 53–65.

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., & Brown, M. E. 2005. The role of leadersin influencing unethical behavior in the workplace.In R. E. Kidwell & C. L. Martin (Eds.), Managingorganizational deviance: 69–87. London: Sage.

Treviño, L. K., Butterfield, K., & McCabe, D. 1998. Theethical context in organizations: Influences on em-ployee attitudes and behaviors. Business EthicsQuarterly, 8: 447–476.

Treviño, L. K., & Weaver, G. R. 2001. Organizationaljustice and ethics program “follow-through”: Influ-

ences on employees’ harmful and helpful behavior.Business Ethics Quarterly, 11: 651–671.

Turner, J. C. 1991. Social influence. Pacific Grove, CA:Thomson Brooks/Cole.

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

Weick, M., & Guinote, A. 2008. When subjective experi-ences matter: Power increases reliance on the ease ofretrieval. Journal of Personality and Social Psy-chology, 94: 956–970.

Weisman, S. R. 2007. Bank’s report says Wolfowitz vio-lated ethics. New York Times, May 15. http://www.nytimes.com/2007/05/15/washington/15wolfowitz.html?scp�5&sq�Wolfowitz&st�cse.

Wheatley, T., & Haidt, J. 2005. Hypnotic disgust makesmoral judgments more severe. Psychological Sci-ence, 16: 780–784.

Wood, W., & Quinn, J. M. 2003. Forewarned and fore-armed? Two meta-analytic syntheses of forewarn-ings of influence appeals. Psychological Bulletin,129: 119–138.

Marko Pitesa ([email protected]) is a Ph.D.candidate in organizational behavior at Grenoble Écolede Management. His research focuses on socially desir-able behavior in organizations, including ethical behav-ior, altruistic behavior, and trust.

Stefan Thau ([email protected]) is an associate profes-sor of organisational behavior at London BusinessSchool. He received his Ph.D. from the University ofGroningen. His research focuses on norm violations, rec-iprocity, trust, and self-regulation.


Copyright of Academy of Management Journal is the property of Academy of Managementand its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

COMPLIANT SINNERS, OBSTINATE SAINTS: HOW...

Documents

Transcript of COMPLIANT SINNERS, OBSTINATE SAINTS: HOW...