Upozornění: Text přílohy byl získán strojově a nemusí přesně odpovídat originálu. Zejména u strojově nečitelných smluv, kde jsme použili OCR. originál smlouvy stáhnete odsud
CES 5621 Příloha_č.6 Výzvy_clanek_Ar1ely Dlshonest NINA MAZAR, ON AMIR, and DAN ARIELY* People like to think of themsBlves as honest. However, dishonesty pays—and it ofteri pays well. How do people resolve this tensian? This research shows that people behave dishonestly enough to profit but honestiy enough to deludé themselves of their own integrity. A little bit oí dishonesty gives a taste of profit without spoiling a positive seff-view. Two mechanisme allow for such self-concept maintenance: inattention ta moral standards and categorization malleability. Six experiments support the authors’ theory of self-concept maintenance and offar practical applications for curbing dishonesty in everyday life. Keywords: honesty, decision maklng, pollcy, šelf The Dishonesty of Honest People: A Theory of Self-Concept Maintenance It is almost íropossible to open a newspaper or tmu on a selfish human behavior (i.e., homo economicus) is the television without being exposed to a report of dishonest belief that people carry out dishonest acts consciously and behavior of one type oř another. To give a few examples, deliberatively by trading off the expected extemal benefits “wardrobing”—the purchase, use, and then retům of the and costs of the dishonest act (AlLingham and Sandmo ušed clothing—costs the U.S. retail industry an estimated 1972; Becker 1968). According to this perspective, people $16 billion annually (Speigbts and Hilinski 2005); the over- would consider three aspects as they pass a gas station: the all magnitude of fraud in the U.S, property and casualty expected amount of cash they st&nd to gain from robbing insurance industry is estimated to be 10% of total claims the plače, the probability of being caught in the act, and the payments, oř $24 billion annually (Accenture 2i}03); and magnitude of punishment if caught. On the basis of these the "tax gap ” oř the difference between what the Intemal inputs, people reach a decision that maximízes their inter- Revenue Service estimates taxpayers should pay and what ests, Thus, according to this perspective, people are honest they actually pay, exceeds $300 billion annually (more than or dishonest only to the extent that the planned trade-off 15% mmeompl janče rate; Herman 2005). If this evidence is favors a particular action (Hechter 1990; Lewicki 1984). In not disturbing enough, perhaps the largest contribution to addltion to being centra! to economic theory, this extemal dishonesty comes from empíoyee theft and fraud, which cost-benefit view plays an important role in the theory of has been estimated at $600 billion a year in tbe United erime and punishment, which forms the basis for most pol- States slone—an amount almoat twice the market capitali- icy measures aimed at preventing dishonesty and guides zation of General Electric (Association of Čerti fied Fraud punishments against those who exhibit dishonest behavior. Examiners 2006). In summary, this standard extemal cost-benefit perspective generates three hypotheses as to the foTces that are expected WHYARE PEOPLE (DÍS)HONEST? to i n urease the frequency and magnitude of dishonesty: higher magnitude of extemal rewards (Ext-Hi), lower Rooted in the philosophies of Thomas Hobbes, Adam probability of being caught (Ext-H2)5 and lower magnitude Smith, and the standard economic model of ratitmal and of punishment (Ext-H3), •"Nina Mazar is Assistant Professor of Marketing, Joseph L. Rntman From a psychological perspective, and in addition to School of Management University of Toronto (e-mail; nina.inazoi@ financial consideratíons, another set of important inputs to utoronío.ca). Oq Amir is Assistant Professor of Marketing, Rady School of the decision whether to be honest is based on intemal Management, University of Califoruia. San Diego (e-mail: uamirOucsd. rewards. Psychologists show that as part of soeialization, edu). Dan Ariely is Visiting ProfessoT of Marketing, Fnqua School of Busi people intemalize the norms and values of their society ness, Duke University (e-mail: dandan@duke.edu). The authors tbank (Campbell 1964; Henrich et al. 2001), which serve as an Daniel Berger, Anat Bracha, Aimee Drolee, and Hffany Kofiolcharoen for intemal benchmark against which a person compares his of their help in condccting the experiments. as well as Ricunlo E. Puxson fur her behavior. Compliance with the interna! values systém his help in creatíng the matrices. Pierre Chándon served as asRociate editor provides positive, rewards, whereas noncomplíance leads to and Ziv Carmon served as guesl editor for this article. negative rewards (Le., punishments). The most direct evi- © 200ti, American Marketing Association Journal ofMarketing Reaeurvh VaL XLV (December 200S), 633-644 ISSN: 0022-2437 (pritit), 1547-7193 (electronic) 633 NINA MAZAR, ON AMIR, and DAN ARIELY* People like to think of themselves as honest. However, dishonesty pays—and it often pays well. How do people resolve this tension? This research shows that people behave dishonestly enough to profit but honestly enough to delude themselves of their own integrity. A little bit of dishonesty gives a taste of profit without spoiling a positive self-view. Two mechanisms allow for such self-concept maintenance: inattention to moral standards and categorization malleability. Six experiments support the authors’ theory of self-concept maintenance and offer practical applications for curbing dishonesty in everyday life. Keywords: honesty, decision making, policy, šelf The Dishonesty of Honest People: A Theory of Self-Concept Maintenance It is almost impossible to open a newspaper or tum on a selfish human behavior (i.e., homo economicus) is the television without being exposed to a report oť dishonest belief that people carry out dishonest acts consciously and behavior of one type or another. To give a few examples. deliberatively by trading off the expected extemal benefits “wardrobing”—the purchase, use, and then return of the and costs of the dishonest act (Allingham and Sandmo ušed clothing—costs the U.S. retail industry an estimated 1972; Becker 1968). According to this perspective, people $16 billion annually (Speights and Hilinski 2005); the over al 1 magnitude of fraud in the U.S. property and casualty would consider three aspects as they pass a gas station: the insurance industry is estimated to be 10% of total claims expected amount of cash they stand to gain from robbing payments, or $24 billion annually (Accenture 2003); and the plače, the probability of being caught in the act, and the the “tax gap,” or the difference between what the Intemal magnitude of punishment if caught. On the basis of these Revenue Service estimates taxpayers should pay and what inputs. people reach a decision that maximizes their inter- they actually pay, exceeds $300 billion annually (more than ests. Thus, according to this perspective, people are honest 15% noncompliance rate; Herman 2005). If this evidence is or dishonest only to the extern that the planned trade-off not disturbing enough, perhaps the largest contribution to favors a particular action (Hechter 1990; Lewicki 1984). In dishonesty comes from employee theft and fraud, which addition to being centrál to economic theory, this external has been estimated at $600 billion a year in the United cost-benefit view plays an important role in the theory of States alone—an amount almost twice the market capitali- crime and punishment, which forms the basis for most pol zation of General Electric (Association of Certified Fraud icy measures aimed at preventing dishonesty and guides Examiners 2006). punishments against those who exhibit dishonest behavior. In summary, this standard external cost—beneťit perspective WH Y ARE PEOPLE (DfS)HONEST? generates three hypotheses as to the forces that are expected to increase the frequency and magnitude of dishonesty: Rooted in the philosophies of Thomas Hobbes, Adam higher magnitude of external rewards (Ext-H|), lower Smith, and the standard economic model of rational and probability of being caught (Ext-Ho), and lower magnitude of punishment (Ext-H^). *N5na Mazar Ls Assistant Professor of Marketing, Joseph L. Rotman School of Management, University of Toronto (e-mail: nina.mazar® From a psychological perspective. and in addition to utoronto.ca). On Amir is Assislanl Professor of Marketing. Rady School of fmancial consideratíons, another set of important inputs to Management. University of California, San Diego (e-mail: oamir@uesd. the decision whether to be honest is based on internal edu). Dan Ariely is Visiling Professor of Marketing, Fuqua School of Busi rewards. Psychologists show that as part of socialization. ness. Duke University (e-mail: dandun@duke.edu). The authors thank people internalize the norms and values of their society Daniel Berger. Anat Bracha, Aimee Drolee. and TilTany Kosolcharoen for (Campbell 1964; Henrich et al. 2001), which serve as an their help in conducting the experiments, as well as Ricardo E. Paxson for intemal benchmark against which a person compares his of his help in creating the matrices. PietTe Chandon served as associate editor her behavior. Compliance with the internal values systém and Ziv Carmon served as guest editor for this article. provides positive rewards, whereas noncompliance leads to negative rewards (i.e., punishments). The most direct evi- © 2008. American Marketing Association Journal of Marketing Research Vol. XLV (Decembcr 2008). 633-644 ISSN: 0022-2437 (prim). 1547-7193 (electronic) 633 634 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 dence of (he existence of such internal reward mechanisms they derive some financial benefit from behaving dishon- comes from brain imaging studies lhát reveal that acts estly but slili maintain their positive self-concept in terms based on sociál norms, such as altruistic punishment or of being honest. To be more precise, we posit a magnitude sociál cooperation (De Quervain el al. 2004; Rilling et al. range of dishonesty within which people can cheat, but 2002), activate the same primary reward centers in the brain their behaviors, which they would usually consider dishon (i.e., nucleus accumbens and caudale nucleus) as external est, do not bear negatively on their self-concept (i.e., they benefits, such as preferred food, drink, and monetary gains are not forced to update their self-concept).1 Although (Knutson et al. 2001; 0'Doherly el al. 2002). many mechanisms may allow people to find such a compro- mise. we focus on two particular means: categorization and Applied to the coniext of (dis)honesty, we propose that altention devoted to one’s own moral standards. Using these one major way the internal reward systém exerts control mechanisms, people can record their actions (e.g., “I am over behavior is by influencing people’s self-concept—that claiming $x in tax exemptions”) without confronting the is, the way people view and perceive themselves (Aronson moral meaning of their actions (e.g., “I am dishonest”). We 1969; Baumeister 1998: Bem 1972). Indeed. it has been focus on these two mechanisms because they support the shown that people typically value honesty (i.e.. honesty is role of the self-concept in decisions about honesty and part of their internal reward systém), that they háve strong because we believe that they háve a wide set of important beliels in their own morality, and that they want to maintain applications in the markelplace. Although not always mulu- this aspecl of their self-concept (Greenwald 1980; Griffin ally exclusive, we elaboráte on each separately. and Ross 1991; Josephson Institute of Ethics 2006: Sani- tioso. Kunda, and Fong 1990). This means lhal if a person Categorization fails to comply with his or her internal standards for hon esty. he or she will need to negatively update his or her self- We hypothesize that for ccrtain types of actions and mag- concept, which is aversive. Conversely, if a person compiies nitudcs of dishonesty, people can categorize their actions with his or her internal standards, he or she avoids such into more compatible terms and find rationalizations for negative updating and maintains his or her positive self- their actions. As a consequence, people can cheat while view in terms of being an honest person. Notably, this avoiding any negative sclf-signals that might affect their perspective suggests that to maintain their positive self- self-concept and thus avoid negatively updating their self- concepts, people will comply with their internal standards concept altogethcr (Gur and Sackeim 1979). even when doing so involves investments of effort or sacri- ficing financial gains (e.g., Aronson and Carlsmith 1962; Two important aspects of categorization are its relative Harris, Mussen, and Rutherford 1976; Sullivan 1953). In malleability and its limit. First, behaviors with mallcablc our gas station example, this perspective suggests that categorization are those that allow people to reinterpret people who pass by a gas station will be influenced not them in a sclf-scrving manner, and the degrec of malleabil only by the expected amount of cash they stand to gain ity is likely to be determined by their context. For example, from robbing the plače, the probability of being caught, and intuition suggests that it is easier to steal a $. 10 pencil from the magnitude of punishment if caught but also by the way a friend than to steal S. 10 out of the frienďs wallet to buy a the act of robbing the store might make them perceive pencil because the former sccnario offers more possibilitics themselves. to categorize the action in terms that are compatible with friendship (e.g., my friend took a pencil from me once; this The utility derived from behaving in line with the self- is what friends do). This thought experiment suggests not concept could conceivably be just another part of the cost- only that a higher degree of categorization malleability benefil analysis (i.e., adding another variable to account for facilitatcs dishonesty (stealing) but also that some actions this utility). However, even if we consider this utility just are inherently less malleablc and thcreforc cannot be cate- another input, it probably cannot be manifested as a simple gorized successfully in compatible terms (Dana, Weber, and constant, because the influence of dishonest behavior on the Kuang 2005; for a discussion of the idea that a medium, self-concept will most likely depend on the particular such as a pcn. can disguise the finál outcome of an action. action, its symbolic value, its contexl, and its plasticity. In such as stealing, see Hsee et al. 2003). In other words, as the following sections, we characterize these elemenls in a the categorization malleability inereases, so does the mag theory of self-concept maintenance and test the implica- nitude of dishonesty to which a person can cornmit without tions of this theory in a set of six experiments. influencing his or her self-concept (Baumeister 1998; Pina e Cunha and Cabral-Cardoso 2006; Schweitzer and THE THEORY OF SELF-CONCEPT MAINTENANCE Hsee 2002). People are oflen torn between two competing motiva- The second important aspcct of the categorization tions: gaining from cheating versus maintaining a positive process pertains to its inherent limit. The ability to categor self-concept as honest (Aronson 1969; Harris, Mussen, and ize behaviors in ways other than as dishonest or immoral Rutherford 1976). For example, if people cheat, they could can be ineredibly useful for the šelf, but it is difficult to gain financially but al the expense of an honest self- imagine that this mechanism is without limits. Instead, it concept. In contrast, if they také the high road. they might may be possiblc to "streteh" the truth and the bounds of forgo financial benefits but maintain their honest self- mental representations only up to a ccrtain point (what concept. This seems to be a win-lose situation, such that choosing one palh involves sacrificing the other. 'Our self-concept maintenance theory is based on how people dcťine honesty and dishonesty for themselves. regardless of whether their defini- In this work. we suggest that people typically solve this tion matchcs the objective definition. motivational dilemma adaptively by finding a balance or equilibrium between the two motivating forces, such that The Dishonesty of Honest People 635 Piaget [1950] calls assimilation and accommodalion). If we EXPERIMENT 1: INCREASING ATTENTION TO assume that the categorization process has such built-in STANDARDS FOR HONESTY THROUGH REL/G/OUS limils, we should conceptualize categorization as effective only up to a threshold, beyond which people can no longer REMINDERS avoid the obvious moral valence of their behavior. The generál setup of all our experiments involves a Attention to Standards multiple-question task. in which participants are paid according to their performance. We compare the perform The other mechanism lhal we address in the current work ance of respondents in the control condilions, in which they is the attention people pay to their own standards of con- háve no opportunity to be dishonest, with that of respon duct. This idea is related to Duval and Wicklunďs (1972) dents in the “cheating” conditions, in which they háve such theory of objective self-awareness and Langeťs (1989) con- an opportunity. In Experiment 1, we test the prediction lhal cept of mindlessness. We hypothesize that when people increasing people’s attention to their standards for honesty attend to their own moral standards (are mindful of them), will make them more honest by contrasting the magnitude any dishonest action is more likely to be reflected in their of dishonesty in a condition in which they are reminded of self-concept (they will update their self-concept as a conse- their own standards for honesty with a condition in which quence of their actions), which in turn will cause them to they are not. adhere to a stricter delineation of honest and dishonest behavior. However, when people are inattentive to their own On the face of it, the idea that any reminder can decrease moral standards (are mindless of them), their actions are not dishonesty seems strange; after all. people should know that evalualed relative to their standards, their self-concept is it is wrong to be dishonest, even without such reminders. less likely to be updated, and, therefore, their behavior is However, from the self-concept maintenance perspective, likely to diverge from their standards. Thus, the attention- the question is not whether people know that it is wrong to to-standards mechanism predicts that when moral standards behave dishonestly but whether they think of these stan are more accessible. people will need to confront the mean- dards and compare their behavior with them in the moment ing of their actions more readily and therefore be more hon of temptation. In other words, if a mere reminder of hon est (for ways to increase accessibility, see Bateson, Nettle, esty standards has an effecl, we can assert that people do and Roberts 2006; Bering, McLeod, and Shackelford 2005; not naturally attend to these standards. In Experiment 1, we Diener and Wallbom 1976; Haley and Fessler 2005). In this implement this reminder through a sirnple recall task. sense, greater attention to standards may be modeled as a tighter range for the magnitude of dishonest actions that Method does not trigger updating of the self-concept or as a lower threshold up to which people can be dishonest without Two hundrcd twenty-nine students participatcd in this influencing their self-concept. experiment, which consisted of a two-task paradigm as part of a broader experimental session with multiplc. unrelated Categorization and Attention to Standards paper-and-pcneil tasks that appeared together in a booklet. In the first task, we asked respondents to write down cither Whereas the categorization mechanism depends heavily the names of ten books they had read in high school (no on stimuli and actions (i.e., degree of malleability and mag- moral reminder) or the Ten Commandmcnts (moral nitude of dishonesty), the attention-to-standards mechanism reminder). They had two minutes to complcte this task. The relies on internal awareness or salience. From this perspec idea of the Ten Commandmcnts recall task was that inde tive, these two mechanisms are distinct; the former focuses pendent of pcop!c’s religion, of whether people believed in on the outside world, and the latter focuses on the inside God. or of whether they knew any of the commandmcnts. world. However, they are related in lhal they both involve knowing that the Ten Commandmcnts are about moral rules attention, are sensitive to manipulations, and are related to would be enough to increase attention to their own moral the dynamics of acceptable boundaries of behavior. standards and thus increase the likelihood of behavior con- sistent with these standards (for a discussion of reminders Thus, although the dishonesty that both self-concept of God in the contcxt of generosity, see Shariff and Noren- maintenance mechanisms allow stems from different zayan 2007). The sccond, ostensibly separate task consisted sources, they both tap the same basic concept. Moreover, in of two sheets of páper: a test shect and an answer shcct. The many reál-world cases, these mechanisms may be so inler- test sheet consisted of 20 matrices, each based on a set of related that it would be difficult to distinguish whether the 12 three-digit numbers. Participants had four minutes to source of this type of dishonesty comes from the environ- find two numbers per matrix that added up to 10 (see Fig ment (categorization) or the individual (attention to stan uře 1). We seleeted this type of task because it is a search dards). In summary, the theory of self-concept maintenance task, and though it can také some time to find the right that considers both external and internal reward Systems suggests the following hypolheses: Figuře 1 Ext&lnt-H|: Dishonesty increases as attention to standards for A SAMPLE MATRIX OF THE ADDING-TO-10 TASK honesty decreases. 1.69 1.82 2.91 Ext&lnt-H2: Dishonesty increases as categorization malleabil ity increases. 4.67 4.81 3.05 Ext&lnt-Hv Given the opportunity to be dishonest. people are 5.82 5.06 4.28 dishonest up to a certain level that does not force them to update their self-concept. 6.36 5.19 4.57 636 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 answer, when it is foiind. the respondents could unambigu- izenry overclaims on taxes, the very act of overclaiming is ously evaluate whether they had solved the i|uestion cor- generálly accepted and therefore not necessarily considered rectly (assuming that they could add two numbers to 10 immoral. However, if this interpretation accounted for our without error), without the need for a solution sheet and the findings, increasing people’s attention to morality (Ten possibility of a hindsight bias (Fischhoff and Beyth 1975). Commandments/recycle condition) would not háve Moreover, we ušed this task on the basis of a pretest that decreased the magnitude of dishonesty. Therefore, we inter- showed that participants did not view this task as one that preted these findings as providing initial support for the reflected their malh ability or intelligence. The answer self-concept maintenance theory. sheet was ušed to report the total number of correctly solved matrices. We promised that at the end of the session. Notě also that, on average, participants remembered only two randomly selected participants would earn $10 foreach 4.3 of the Ten Commandments, and we found no significanl correctly solved matrix. correlation between the number of commandments recalled and the number of matrices the participants claimed to háve In the two control conditions (after the ten books and Ten solved correctly (r = -.14, p = .29). If we use the number of Commandments recall task, respectively), at the end of the commandments remembered as a proxy for religiosity, the four-minute matrix task, participants continued to the next lack of relalionship between religiosity and the magnitude task in the booklet. At the end of the entire experimental of dishonesty suggests that the efficacy of the Ten Com session, the experimenler verified their answers on the mandments is based on increased attention to internal hon matrix task and wrote down the number of correctly solved esty standards. leading to a lower tolerance for dishonesty matrices on the answer sheet in the booklet. In the two (i.e., decreased self-concept maintenance threshold). recycle conditions (after the ten books and Ten Command- ments recall task, respectively), at the end of the four- Finally, it is worth contrasting these results with people’s minute matrix task, participants indicated the total number lay theories about such siluations. A separate set of students of correctly solved matrices on the answer sheet and then (n = 75) correctly anticipated lhal participants would cheat tore out the originál test sheet from the booklet and placed when given the opportunity to do so, but they anticipated it in their belongings (to recycle later), thus providing lbem that the level of cheating would be higher than what it with an opportunity to cheat. The entire experiment repre- really was (Mpred Books/recycie = 9.5), and they anticipated sented a 2 (type of reminder) x 2 (ability to cheat) between- that reminding participants of the Ten Commandments subjects design. would not significantly decrease cheating (MpredXei1 Commandments/recycle — 2.8, t(73) = 1.61, p = . 1 1). The COntraSt Results and Discussion of the predicted results with the actual behavior we found suggests that participants understand the economic motiva The results of Experiment 1 confirmed our prediclions. tion for overclaiming, but they overestirnate its influence on The type of reminder had no effect on participants’ per behavior and underestimate the effect of the self-concept in formance in the two control conditions (Muooks/controi = 3.1 regulating honesty. versus M-fen Commandments/comrol = 3.1, F( 1, 225) = .012./? = .91), which suggests lhal the type of reminder did not influ- EXPERIMENT2: INCREASING ATTENTION TO ence ability or motivation. Following the book recall task. STANDARDS TOR HONESTY THROUGH however, respondents cheated when they were given the COMMITMENT REMINDERS opportunity to do so (MBooks/recvc|e = 4.2), but they did not cheat after the Ten Commandments recall task (MTei1 Another type of reminder, an honor eode, refers to a pro Coinmandments/recycle = 2.8; F(1, 225) = 5.24, p - .023). creat- ceduře that asks participants to sign a statement in which ing a significanl interaction between type of reminder and they declare their commitment to honesty bcforc tuking part ability to cheat (F(3, 225) = 4.52. p = .036). Nolably, the in a task (Dickerson ct al. 1992; McCabc and Trevino 1993, level of cheating remained far below the maximum. On 1997). Although many explanations háve been proposed for average, participants cheated only b.1% of the possible the effectivcness of honor codes ušed by many academie magnitude. Most important, and in line with our notion of institutions (McCabc, Trevino, and Butterfield 2002; sec self-concept maintenance, reminding participants of stan http://www.academicintegri.ty.org), the self-concept mainte dards for morality eliminated cheating completely: In the nance idea may shed light on the internal proccss under- Ten Commandments/recycle condition. participants' per lying its success. In addition to manipulating the awarcness formance was undistinguishable from those in the control of honesty standards through commitment reminders at the conditions (F( 1,225) = .49, p = .48). point of temptation, Experiment 2 represents an extension of Experiment 1 by manipulating the financial incentives We designed Experiment I to focus on the attention-to- for performance (i.e., external benefits); in doing so, it also standards mechanism (Ext&Int-H!)- but one aspecl of the tests the external cost-bencfit hypothesis that dishonesty results—the finding that the magnitude of dishonesty was increases as the expected magnitude of reward from the dis limited and well below the maximum possible level in the honest act increases (Ext-H|). two recycle conditions—suggested that the categorization mechanism (Ext&Int-H2) could háve been at work its well. Method A possible alternativě interpretation of the books/recycle Two hundred seven students parlicipated in Experiment condition is that over their lifetime. participants developed 2. Using the same matrix task. we manipulated two faclors standards for moral behavior according to which overclaim- between participants: the amount earned per correctly ing by a few questions on a test or in an experimental set- solved matrix ($.50 and $2, paid to each participant) and ting was not considered dishonest. If so, these participants the attention to standards (control, recycle, and recycle + could háve been completely honest from their point of view. honor code). Siinilarly. in a country in which a substantial part of the cit- The Dishonesty of Honest People 637 In the two control conditions, at the end of five minutes, magnitude of cheating; cheating was slightly more common participants handed both the test and the answer sheets to (by approximately 1.16 questions), lhough not significantly the experimenter, who verified Iheir answers and wrote so, in the $.50 condition (F(l. 201) = 2.1, p = .15). Thus, down the number of correctly solved matrices on the we did not find support for Ext-Hj. A possible interpreta answer sheet. In the two recycle conditions, participants tion of this decrease in dishonesty with increased incentives indicated the total number of correctly solved matrices on is that the magnitude of dishonesty and its effect on the the answer sheet, folded the originál test sheet, and pluced categorization mechanism depended on both the number of it in their belongings (to recycle later), thus providing them questions answered dishonestly (which increased by 2.8 in an opportunity to cheat. Only after lhal did they hand the lbe $.50 condition and 1.7 in the $2 condition) and the answer sheet to the experimenter. The recycle + honor code amount of money inaccurately claimed (which increased by condition was simiiar to the recycle condition except that at S1.4 in the $.50 condition and $3.5 in the $2 condition). If the top of the test sheet, there was an additional statement categorization malleability was affected by a mix of these that read, “I understand that this shorl survey falls under two factors, we would háve expected the number of ques MIT’s [Yale’s] honor systém." Participants printed and tions that participants reported as correctly solved to signed their names below the statement. Thus, the honor decrease with greater incentives (at leasl as long as the code statement appeared on the same sheet as the matrices, extemal incentives were not too high). and this sheet was recycled before participants submiiled their answer sheets. In addition. to provide a test for Ext- Most important for Experiment 2, we found that the two H|, we manipulated the payment per correctly solved recycle + honor code conditions ($.50 and $2: Mrecyc,e +. matrix ($.50 and $2) and contrasted performance levels honor code = 3.0) eliminated cheating insofar as the perform between these two incentive levels. ance in these conditions was undistinguishable from the two control conditions ($.50 and $2: Mcontro| = 3.3; F(l, Results and Discussion 201) = .19, p = .66) but significantly different from the two recycle conditions ($.50 and $2: Mre:cyc|e = 5.5; F(l. 201) = Figuře 2 depicts the results. An overall analysis of vari 19.69. p < .001). The latter result is notable given that the ance (ANOVA) revealed a highly significant effect of the two recycle + honor code conditions were procedurally attention-to-standards manipulation (F(2, 201) = I 1.94, /; < simiiar to the two recycle conditions. Moreover, the two .001), no significant effect of the level of incentive manipu instimtions in which we conducted this experiment did not lation (F(l, 201) = .99, /> = .32), and no significant inter- háve an honor code systém al the time, and therefore, action (F(2, 201) = .58, p = .56). When given the opportu objeclively. the honor code had no iinplications of external nity, respondents in the two recycle conditions ($.50 and punishment. When we replicated the experiment in an insti- S2) cheated (Mr t.|e = 5.5) reiative to those in the two con tulion that had a slrict honor code, the results were identi- trol conditions ($.50 and $2: MC0n„.O| = 3.3; F(l, 201) = cal, suggesting that it is not the honor code per se and its 15.99, p < .001), but again, the level of cheating fell far implied external punishment but rather the reminder of below the maximum (i.e., 20); participants cheated only morality that was at play. 13.5% of the possible average magnitude. In line with our findings in Experiment I, this latter result supports the idea Again, we asked a separate sel of students (n = 82) at the that we were also observing the workings of the categoriza institutions without an honor code systém to predict the tion mechanism. results, and lhough they predicted that the increased pay ment would marginally increase dishonesty (Mpred_$2 = 6.8 Between the two levels of incentives ($.50 and $2 condi tions), we did not find a particularly large difference in the versus Mpred_$.5() = 6.4; F(l, 80) = 3.3, p = .07), in essence Figuře 2 predicting Ext-H], they did not anticipate that the honor code would significantly decrease dishonesty (Mpred_recy|ce + EXPERIMENT 2: NUMBER OF MATRICES REPORTED SOLVED honor code = 6.2 versus Mpred_recycle — 6.9, F( 1, 80) — .74. p — .39). The contrast of the predicted results with the actual Control Recycle Recycle + HC behavior suggests that people understand the economic motivation for overclaiming, that they overestimate its Notes: Mean number of “solved” matrices in the control condition (no influence on behavior, and that they underestimate the ability to cheat) and the recycle and recycle + honor code (ilC) conditions effect of the self-concept in regulating honesty. In addition, (ability to cheat). Tile payment scheme was either $.50 or $2 per correct the finding that predictors did not expect the honor code to answer. Error bars are based on standard errors of the means. decrease dishonesty suggests that they did not perceive the honor code manipulation as having impiications of extemal punishment. EXPERIMENT3: INCREASING CATEGORIZATION MALLEABILITY Making people mindful by increasing their attention to their honesty standards can curb dishonesty. but the theory of self-concept maintenance also implies that increasing the malleability to interpret one’s actions should increase the magnitude of dishonesty (Schweitzer and Hsee 2002). To test this hypothesis, in Experiment 3, we manipulate whether the opportunity for dishonest behavior occurs in terms of money or in terms of an intermediary medium (tokens). We posil that introducing a medium (Hsee et al. 638 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 2003) will offer participants more room for interpretation EXPERIMENT 4: RECOGNIZ/NG ACTIONS BUT NOT of their actions, making the moral implicalions of dishon UPDATING THE SELF-CONCEPT esty less accessible and thus making it easier for partici pants to cheat at higher magnitudes. Our account of self-concept maintenance suggests that by engaging only in a relutively low level of cheating, par Method ticipants slayed within the threshold of acceptable magni tudes of dishonesty and thus benefited from being dishonest Four hundred fifty students partícipated in Experiment 3. without receiving a negative self-signal (i.e., their self- Participants had five minutes to complcte the matrix task concept remained unaffected). To achieve this balance, we and were promised $.50 for each correctly solved matrix. posil that participants recorded their actions correctly (i.e., We ušed three betwecn-subjects conditions: the same con they knew that they were overclaiming). but the categoriza trol and recycle conditions as in Experiment 2 and a recy tion and/or attention-to-standards mechanisms prevented cle + token condition. The latter condition was simiiar to this factual knowledge from being morally evaluated. Thus, the recycle condition, except participants knew that each people did not necessarily confront the true meaning or correctly solved matrix would earn them one token, which implications of their actions (e.g., “I am dishonest”). We they would exchange for $.50 a few seeonds later. When the test this prediction (Ext&Int-H^) in Experiment 4. Five minutes clapsed, participants in the recycle + token condition recycled their test sheet and submitted only their To test the hypothesis that people are aware of their answer sheet to an experimenter, who gave them the corrc- actions bul do not update their self-concepts, we manipu- sponding amount of tokens. Participants then went to a sec- lated participants’ ability to cheat on the matrix task and ond experimenter, who exchanged the tokens for money measured their prediclions about their performance on a (this experimenter also paicl the participants in the other second matrix task that did not allow cheating. If partici conditions). We counterbalanced the roles of the two panls in a recycling condition did not recognize that they experimenters. overclaimed, they would base their predictions on their exaggeraled (i.e., dishonest) performance in the first matrix Results and Discussio/t task. Therefore, their predictions would be higher than the predictions of those who could not cheat on the first task. Simiiar to our previous findings, participants in the recy However, if participants who overclaimed were cognizant cle condition solved significantly more questions than par of their exaggeraled claims, their predictions for a situation ticipants in the control condition (Mrccj,c|c = 6.2 versus that does not allow cheating would be attenuated and, theo- Mcontral = 3.5; F( 1. 447) = 34.26. p < .001), which suggests retically, would not differ from their counterparts’ in the that they cheated. In addition, participants’ magnitude of control condition. In addition, to test whether dishonest cheating was well below the maximum—only 16.5% of the behavior influenced people’s self-concept, we asked partici possible average magnitude. Most important, and in line pants about their honesty after they completed the first with Ext&lnt-H->, introducing tokens as the medium of matrix task. If participants in the recycling condition (who immediate exchange further increased the magnitude of dis were cheating) had lower opinions about themselves in honesty (Mrecy|e + tokcn = 9.4) such that it was significantly terms of honesty than those in the control condition (who larger than it was in the recycle condition (F(l, 447) = were not cheating), this would mean that they had updated 47.62, p < .001)—presumably without any changes in the their self-concept. However, if cheating did not influence probability of being caught or the severity of the their opinions about themselves, this would suggesl that punishment. they had not fully accounted for their dishonest behaviors and, consequently, that they had not paid a price for their Our findings support the idea that claiming more tokens dishonesty in terms of their self-concept. instead of claiming more money offered more categoriza tion malleability such that people could interpret their dis Method honesty in a more self-serving manner, thus reducing the negative self-signal they otherwise would háve received. In Forty-four students partícipated in this experiment, terms of our current account, the recycle + token condition which consisted of a four-task paradigm. administered in increased the threshold for the acceptable magnitude of dis the following order: a matrix task. a personality test, a pre honesty. The finding that a medium could be such an diction task. and a second matrix task. In the first matrix impressive facilitator of dishonesty may explain the incom- task. we repeated the same control and recycle conditions parably excessive contribution of employee theft and fraud from Experiment 2. Participants randomly assigned to (e.g., stealing office supplies and merchandise, putting either of these two conditions had five minutes to complete inappropriate expenses on expense accounts) to dishonesty the task and received $.50 per correctly solved matrix. The in the marketplace, as we reported previously. only difference from Experiment 2 was that we asked all participants (not just those in the recycle condition) to Finally, it is worth pointing out that our results differ report on the answer sheet the total number of matrices they from what a separate set of students (n = 59) predicted we had correctly solved. (Participants in the control condition would find. The predictors correctly anticipated that partici then submitted both the test and the answer sheets to the pants would cheat when given the opportunity to do so experimenter, who verified each of their answers on the test (Mrrerecycle > recycle+). 5 ; — ~ r- 4The goal of the recycle+ condition was to guarantee participants that c\ m ihu lI IlIlIlIlIlHl their individual actions of taking money from the jar would not be observ- able. Therefore. it was impossible to measure how much money each nu ni respondent took in this condition. We could reeord only the sum of money ,, ;; — ; missing at the end of each session. For the purpose of statistical analysis, 21 24 27 30 33 36 39 42 45 48 we assigned the average amount taken per recycle-t- session to each partici- panl in that session. Number of Questions Notes: Frequency distribution of number of “solved” questions in the control condition (no ability to cheat) and two cheating conditions: no- recycle and recycle. The values on the y-axis represent the percentage of participants having “solved” a particular number of questions; the values on the x-axis represent ±l ranges around the displayed number (e.g., 21 = participants having solved 20. 21, or 22 questions). 642 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 a slronger lest would be to determine whether this conclu- trade-off and are unappreciative of the regulative effective- sion also holds across all six experiments. To do so, we ness of the self-concept.6 converted the performance across all the experiments to be proportional, that is, the number of questions reported In principle, the theory we propose can be incorporated solved relative to the maximum possible. Analyzing all con into economic models. Some formalizations related to it ditions across our experiments (n = 1408), we again find appear in recent economic theories of utility rnaximizalion slrici stochastic dominance of the performance distributions based on models of self-signaling (Bodner and Prelec 2001) in conditions that allowed cheating over conditions that did and identity (Bénabou and Tirole 2004, 2006). These mod not ([} = .15, t(1406) = 2.98, p = .003). We obtain similarly els can be adopted to account for self-concept maintenance reliable differences for each quantile of the distributions, by incorporating attention to personál standards for honesty suggesling that the overall mean difference (|3 = .134, (meta-utility function and salience parameter sl, respec t( 1406) = 9.72, p < .001) was indeed caused by a generál tively) and categorization malleability (interpretation func shift in the distribution ralher than a large shift of a smáli tion and probability 1 - X, respectively). These approaches porlion of the distribution. convey a slowly spreading conviction among economists that to study moral and sociál norms. altruism. reciprocity, GENERAL DISCUSSION or antisocial behavior, the underlying psychological moli- valions that vary endogenously with the environment must People in almost every society value honesty and main be understood (see also Gneezy 2005). The data presented tain high beliefs about their own morality; yet examples of herein offer further guidance on the development of such significant dishonesty can be found everywhere in the mar- models. In our minds, the interplay between these formal ketplace. The standard cost-benefit model, which is centra! models and the empirical evidence we provide represents a to legal theory surrounding crime and punishment, assumes fruitful and promising research direction. that dishonest actions are performed by purely selfish. cal- culating people, who only care about external rewards. In Some insighls regarding the functional from in which the contrast, the psychological perspective assumes that people external and internal rewards work togelher emerge from largely care about internal rewards because they want, for the dala, and these findings could also provide worthwhile example, to maintain their self-concept. On the basis of palhs for further investigations in both economics and psy these two exlreme starting poinls, we proposed and tested a chology. For example, the results of Experiment 2 show theory of self-concept maintenance that considers the moti- that increasing external rewards in the form of increasing valion from both external and internal rewards. According benefits (monetary incentive) decreased the level of dishon to this theory, people who think highly of themselves in esty (though insignificantly). This observation matches terms of honesty make use of various mechanisms that findings from another matrix experiment in which we allow them to engage in a limited amount of dishonesty manipulated two factors between 234 participants: the abil- while retaining positive views of themselves. In other ily to cheat (control and recycle) and the amount of pay- words, there is a band of acceptable dishonesty that is lim menl to each participant per correctly solved matrix ($. 10. ited by internal reward consideratíons. In particular, we S.50, $2.50. and S5). In this 2x4 design, we found limited focus on two related but psychologically distinct mecha dishonesty in the $.10 and $.50 conditions but no dishon nisms that influence the size of this band—categorization esty in lbe $2.50 and $5 conditions. Furthermore, the mag- and attention to standards—which we argue háve a wide set nilude of dishonesty was approximately the same for $.10 of important applications in the marketplace. and $.50. Togelher, these observations raise the possibility of a step function-like relationship—constanl, limited Across a set of six experiments we found support for our amount of dishonesty up to a certain level of positive exter theory by demonstraling lhal when people had the ability to nal rewards, beyond which increasing the external rewards cheat, they cheated, bul the magnitude of dishonesty per per could limit categorization malleability, leaving no room for son was relatively low (relative to the possible maximum under-the-radar dishonesty. In this way, dishonesty may amount). We also found that. in generál, people were insen- actual ly decrease with extemal rewards. sitive to the expected extemal costs and benefits associated with the dishonest acts, but they were sensitive to contex- Finally, it is worthwhile noling some of the limitations of tual manipulations related to the self-concept. In particular, our results. The first limitation is related directly to the rela the Level of dishonesty dropped when people paid more tionship between external and intemal rewards. Arguably, at attention to honesty standards and climbed with increased some point at which the extemal rewards become very high. categorization malleability (Dana, Weber, and Kuang 2005). they should lempt the person sufficiently to prevail (because the external reward of being dishonest is much Some of the results provide more direcl evidence for the larger than the internal reward of maintaining a positive self-concept maintenance mechanism (Experiment 4) by self-concept). From that point on, we predicl that behavior showing that even though participants knew that they were would be largely infiuenced by external rewards, as the overclaiming, their actions did not affect their self-concept standard economic perspective predicts (i.e., ultimalely, the in terms of honesty. Notě also that, in contrast. predictors magnitude of dishonesty will increase with increasing. high expected dishonest actions to háve a negative effect on the extemal rewards). self-concept. This misunderstanding of the workings of the self-concept also manifested in respondents’ inability to 6Note that our manipulations in their generál form may bc vicwed as predicl the effects of moral reminders (Ten Commandments and honor code) and mediums (tokens), suggesting that, in priming. In this scnsc. our results may generalize to a much larger class of generál, people expecl others to behave in line with the manipulations that would curtail cheating behavior and may be useful standard economic perspective of an external cost-benefit when. for example. the Ten Commandmcnts or honor codes are not a fea- sible solution, such as purehasíng environments. The Dishonesty of Honest People 643 Another limitation is that our results did not support a gest that increasing people's attention to their own stan sensitivity to others' reported behaviors, implying that, for dards for honesty and deereasing the categorization mal- example, self-estccm or norm compliance consideratíons do leabílily could be effective remedies. However. the means not influence people’s decisions about being dishonest. We by which to incorporate such manipulations into everyday do not imply that such effects are not převalení or perhaps scenarios in which people might be tempted to be dishonest even powerful in the marketplace. For example, it could be (e.g., returning ušed clothes, filling out tax returns or insur- that the sensitivity to others operates slowly toward chang- ance claims), to determine how abstract or concrete these ing a person’s globál internal standards for honesty. rather manipulations must be to be effective (see Hayes and Dun- than having a large influence on the local instances of dis ning 1997), and to discover methods for fighting adaptation honesty, such as those that took plače in our experiments. to these manipulations remain open questions. From a practical perspective. one of the two main ques REFERENCES tions about under-the-radar dishonesty pertains to its mag nitude in the economy. By its very nátuře, the level of dis Accenture (2003), “One-Fourth of Americans Say Iťs Acceptable honesty in the marketplace is diťflcult to measure, but if our to Defraud Insurance Companšes," (February 12), (accessed studies are any indication, it may far exceed the magnitude December I, 2006), [available at httpt/Avww.accenture.com/xd/ of dishonesty committed by “standard, run-of-the-mill” xd.asp?it=enweb&xd=_dyn%5Cdynamicpressrelease_577. eriminals, who consider only the external rewards in their xml], decision. Across the six experiments (exeluding the recycle + token condition), among the 791 participants who Allinghatn. Michael G. and Agnar Sandmo (1972), “Incoine Tax could cheat, we encountered only 5 (.6%) who cheated by Evasion: A Theoretical Analysis,” Journal of Public Economics, the maximal amount (and, thus, presumably engaged in 1 (November), 323-38. extemal cost-benefit trade-off analysis, leading to standard rational dishonesty), whereas most cheated only slightly Aronson, Elliot (1969), "A Theory of Cognitive Dissonance: A (and, thus, presumably engaged in a trade-off of external Current Perspective.” in Advances in Experimental Sociál Psy and internal rewards, leading them to engage in limited dis chology, Vol. 4. Leonard Berkowitz, ed. New York: Academie honesty that flies under the self-concept radar). Further Press. 1-34. more, the total costs incurred as a result of limited dishon esty were much greater than those associated with the --------- and J. Merrill Carlsmith (1962), “Performance maximal dishonesty. Taken at face value, these results sug- Expectancy as a Determinant of Actual Performance,” Journal gest that the effort that society at large applies to deterring ofAbnormal and Sociál Psychology, 65 (3), 178-82. dishonesty—especially standard rational dishonesty— might be misplaced. Association of Certified Fraud Examiners (2006), “2006 ACFE Report to the Nation on Occupational Fraud & Abuse,” Another important applicd speculation involves the (accessed August 1, 2008), [available at http://www.acfe.com/ medium experiment. As society moves away from cash and documents/2006-mn.pdf). electronic exchanges become more prevalent, mediums are rapidly more available in the economy. Again, if we také Bateson. Meiissa, Daniel Netlle, and Gilbert Roberts (2006), our results at face value, particular attention should be paid "Cues of Being Watched Enhance Coopemtion in a Reál-World to dishonesty in these new mediums (e.g., backdating Seldng,” Biology Letters, 2 (June), 412—14. stoeks) because they provide more opporlunities for under- the-radar dishonesty. In addition, we observed that the Baumeister, Roy F. (1998). “The Šelf,” in Handbook of Sociál medium experiment not only allowed people to cheat more Psychology, Daniel T. Gilbert. Susan T. Fiske, and Gardner but also increased the level of maximal cheating. In the Lindzey, eds. New York: McGraw-Hill, 680-740. medium experiment, we observed 24 participants who cheated maximally, which indicated that the tokens not only Becker, Gary S. (1968), “Crime and Punishment: An Economic allowed them to elevate their acceptable magnitude of dis Approach,” Journal of Political Economy, 76 (2), 169-217. honesty but also liberated them from the shackles of their morality altogether. Bern, Daryl J. (1972), “Self-Perception Theory,” in Advances in Experimental Sociál Psychology, Vol. 6, Leonard Berkowitz, ed. When we consider the applied implications of these New York: Academie Press, 1-62. results. we must emphasize that our findings stem from experiments not with eriminals but with students at elite Bénabou, Roland and Jean Tirole (2004), “Willpower and Per universities, people who are likely to play important roles sonál Rules,” Journal of Political Economy, 112 (4), 848-86. in the advancement of the United States and who are a lot more simiiar to the generál public. The prevalence of dis ---------and---------- (2006), “Identity, Dignity and Taboos,” work- honesty among these people and the finding that, on an ing páper, Department of Economics and Woodrow Wilson individual level. they were mostly honest rather than com School. Princeton University. pletely dishonest suggest the generálizability of our results. As Goldstone and Chin (1993) conclude, people seem to be Bering, Jesse M., Katrina McLeod. and Todd K. Shackelford moral relativists in their everyday lives. (2005), “Reasoning About Dead Agents Reveals Possible Adap- tive Trends,” Human Nátuře, 16 (4), 360-81. From a practical perspective, the next question is thus related to approaches for curbing under-the-radar dishon Bodner. Ronit and Dražen Prelec (2001), “Self-Signaiing and esty. The results of the honor code. Ten Commandments, Diagnostic Utility in Everyday Decision Making,” working and token manipulations are promising because they sug páper. Sloan School of Management. Massachusetts Institute of Technology. Campbell, Ernest Q. (1964), “The Internalization of Moral Nonns.” Sociometry, 27 (4), 391-412. Dana, Jason, Roberto A, Wreber, and Jason Xi Kuang (2005), “Exploiting Moral Wiggle Room: Experiments Demonstrating an Illusory Preference for Fairness,” working páper, Department of Psychology, University of Illinois at Urbana-Champaign. De Quervain, Dominique J.-F., Urs Fischbacher, Valerie Treyer, Mclanie Schclthammer, Ulrich Schnyder, Alfred Buck, and Ernst Fehr (2004). 'The Neural Basis of Altruistic Punishment,” Science, 305 (August 27), 1254-58. Dickerson, Chris A., Ruth Thibodeau, Elliot Aronson, and Dayna Miller (1992), “Using Cognitive Dissonance to Encourage 644 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 Watcr Conservation,” Journal ofApplied Sociál Psychology, 22 Sclectively Recruits Nucleus Accumbcns.” Journal of Neuro- (II), 841-54. science, 21 (16), 1-5. Langer, Ellen J. (1989), “Minding Matters: The Conscqucnccs of Dicncr, Edward and Marc Wallbom (1976), “Effccts of Sclf- Mindlessness-Mindfulncss,” in Advances in Experimental Awarcness on An ti normative Behavior,” Journal of Research in Sociál Psychology, Leonard Berkowitz, cd. San Dicgo: Acade Personality, 10 (I). 107-1 I I. mie Press, 137-73. Lewicki, Roy J. (1984), “Lying and Dcccption: A Bchavioral Duval. Thomas S. and Robert A. Wicklund (1972), A Theory of Model,” in Negotiation in Organizations, Max H. Bazerman Objective ŠelfAwareness. New York: Academie Press. and Roy .1. Lewicki. cds. Bcverly Hilis, CA: Sagc Pubiications. Fischhoff, Baruch and Ruth Beyth (1975), "I Knew It Would Hap- 68-90. pen: Remembered Probabilitics of Oncc-Futurc Things,” Orgci- McCabe, Donald L. and Linda Klebe Trevino (1993), “Academie nizational Behavior and Human Performance, 13(1), 1—16. Dishonesty: Honor Codes and Other Contextual Influences.” Gnccz.y, Uri (2005), “Dcccption: The Role of Conscqucnccs,” Journal of Higher Education. 64 (5), 522—38. American Economic Re view, 95 (1), 384—94. --------- and---------- (1997), “Individual and Contextual Influences on Academie Dishonesty: A Multicampus Investigation,” Goldstone, Robert L. and Calvin Chin (1993), "Dishonesty in Research in Higher Education, 38 (3), 379-96. Sclf-Report of Copics Madc: Moral Relativity and the Copy --------- , --------- , and Kenneth D. Butterfield (2002), “Honor Machine,” Basic and Applied Sociál Psychology, 14(1), 19-32. Codes and Other Contextual Influences on Academie Integrity: Greenwald, Anthony G. (1980), "The Totalitarian Ego: Fabrica- A Rcplication and Extension to Modified Honor Code Set- tion and Revision of Personál History,” American Psychologist, tings,” Research in Higher Education, 43 (3), 357-78. Nagin, Daniel S. and Greg Pogarsky (2003), “An Experimental 35 (7), 603-618. Investigation of Deterrence: Cheating, Self-Serving Bias, and Griffm, Dále W. and Lee Ross (1991), “Subjcctive Construal. Impulsivity.” Criminology, 41 (I), 167-94. 0’Doherty, John P., Ralf Deichmann, Hugo D. Critchley, and Ray Sociál Infercnce, and Human Misundcrstanding,” in Advances mond J. Dolan (2002), "Neural Responses During Anticipation in Experimental Sociál Psychology, Vol. 24, Mark P. Zanna, cd. of a Primary Taste Reward,” Neuron, 33 (5), 815-26. New York: Academie Press, 319-59. Piaget, Jean (1950), The Psychology of Intelligence. New York: Gur, Ruben C. and Harold A. Sackcim (1979), “Self-Dcception: A Harcourt Brace & Co. Conccpt in Scarch of a Phenomcnon,” Journal of Personality Pina e Cunha, Miguel and Carlos Cabral-Cardoso (2006), “Shades and Sociál Psychology, 37 (2), 147-69. of Cray: A Liminal Interpretation of Organizational Legality- Halcy, Kevin J. and Daniel M.T. Fcsslcr (2005), “Nobody’s Illegality,” International Public Management Journal, 9 (3). Watching? Subtlc Cues Affect Generosity in an Anonymous 209-225. Economic Game,” Evolution and Human Behavior. 26 (3). Reno, Raymond R„ Robert B. Cialdini. and Carl A. Kallgren (1993). “The Transsituational Influence of Sociál Norms,” Jour 245-56. nal of Personality and Sociál Psychology, 64 (11). 104-112. Harris, Sandra L., Paul H. Mussen, and Eldred Rutherford (1976), Rilling, James K„ David A. Gutman, Thorsten R. Zeli, Giuseppe Pagnoni, Gregory S, Bcrns, and Clinton D. Kilts (2002), "A “Some Cognitive, Bchavioral. and Personality Corrclatcs of Neural Basis for Sociál Coopcralion,” Neuron, (July 18), Maturity of Moral Judgmcnt,” Journal of Genetic Psychology. 395-405. 128 (1), 123-35. Sanitioso, Rasyid, Živa Kunda, and Gcoffrey T. Fong (1990), Haycs, Andrcw F. and David Dunning (1997), “Trait Ambiguity “Motivated Recruitment of Autobiographical Memories,” Jour and Construal Proccsses: Tmplications for Self-Peer Agreement nal of Personality and Sociál Psychology, 59 (2), 229-41. in Personality Judgmcnt.” Journal of Personality and Sociál Schweitzer, Maurice E. and Christopher K. Hsee (2002), “Streteh- Psychology, 72 (3), 664-77, ing lbe Truth: Elastic Justification and Motivated Communica- tion of Uncertain Information,” Journal of Risk and Uncer- Hechter, Michael (1990), “The Attainmcnt of Solidarity in Intcn- tainty, 25 (2), 185-201. tional Communitics,” Rationality and Society, 2 (2), 142-55. Sharšff, Az.im F. and Ara Norenzayan (2007), “God Is Watching Hcnrich, Joscph. Robert Boyd, Samuel Bowles, Colin Camcrer, You: Supematural Agent Concepts Increase Prosocial Behavior in an Anonymous Economic Game,” working páper. Depart Ernst Fchr, Herbert Gintis, and Richard McElreath (2001), 'in ment of Psychology, University of British Columbia. Scarch of Homo Economicus: Behavioral Experiments in 15 Speights, David and Mark Hitinski (2005), "Retům Fraud and Small-Scale Societics,” American Economic Review, 91 (2). Abuse: How to Protéct Profits,” Retailing íssaes Letter, 17 (1), 73-78. Herman, Tom (2005), "Study Suggests Tax Cheating Is on the 1-6. Risc: Most Dctailed Survey in 15 Ycars Finds $250 Billion-Plus Gap; Rampšng Up Audits on Wealthy,” The Wall Street Journal. Sullivan, Harry S. (1953), The Interpersonal Theory of Psychiatry. (March 30), Dl. New York: Norton. Hsee, Christophcr K., Fang Yu, Jiao Zhang, and Yan Zhang (2003), “Medium Maximization,” Journal of Consumer Tesser. Abraham, Murray Millar, and Janet Moore (1988), “Some Research, 30 (I), 1-14. Affective Consequences of Sociál Comparison and Rcflection John, Olivcr P. and Richard W. Robins (1994), “Accuraey and Bias Processes: The Pain and Pleasure of Being Close,” Journal of in Sclf-Perccption: Individual Diffcrcnccs in Sclf-Enhanccment Personality and Sociál Psychology, 54 (1), 49-61. and the Role of Narcissism." Journal of Personality and Socicd Psychology, 66 (1), 206-219. Trivers. Robert (2000), "The Elements of a Scientific Theory of Joscphson Institute of F.thics (2006), “Report Card on the Ethics Self-Deception,” in Evolutionary Perspectives on Human of American Youth,” (October 15), (accessed August I, 2008). Reproductive Behavior, Dori I.eCroy and Peter Moller, eds. [avaiI ablc at http://w wvv.joscphsoni nst itute.org/reporteard/]. New York: New York Academy of Sciences, 114—31. Kallgren, Carl A., Robert B. Cialdini, and Raymond R. Reno (2000), "A Focus Theory of Normative Conduct: When Norms Do and Do Not Affcct Behavior,” Personality and Sociál Psy chology Bulletin, 26 (8), 1002-1012. Knutson, Brian, Charles M. Adams, Gracc W. Fong. and Daniel Hommer (2001), “Anticipation of Increasing Monctary Reward Commentaries and Rejoinder to ‘The Dishonesty of Honest People” Hypermotivation H YPERMOTÍVA TiON SCOTT RICK and GEORGE LOEWENSTEIN* A closer examination of many of the acts of dishonesty in the reál world reveals a striking pattern: Many, if not In their clever and insightful article, Mazar. Amir. and most, appear to be motivated by the desire to avoid (or Ariely (2008) propase that people balance two competing recoup) losses rather than the simple desire for gain. A desires when deciding whether to behave dishonestly: the wide range of evidence suggests that people who find them motivation for personál gain and the desire to maintain a selves “in a hole” and believe that dishonest behavior is the positive self-concept. Their studies focus on the latter fac- only apparent means of escape are more likely to cheat. tor. showing that people behave dishonestly when it pays. steal, and lie. For example, several studies háve found that but only to the extent that they can do so without violating people are more likely to cheat on their taxes when they their perception of themselves as honest. The research is owe than when they are due for a refund (e.g.. Chang and innovalive and important. It has already had an influence on Schultz 1990; Schepanski and Kelsey 1990). policies dealing with conflicts of interest in medicine (Association of American Medical Colleges 2007) and. Prospect theory’s (Kahneman and Tversky 1979) concept even before its own publication. has spawned significant of loss aversion might seem to provide a natural account of follow-up research (Vohs and Schooler 2008). what could be called “hypermotivation”—a visceral statě that leads a person to také actions he or she would normally In our opinion, the main limitation of Mazar, Amir, and deem to be unacceptable. Loss aversion suggests that the Ariely’s article is not in the perspective it presents but motivation to avoid a loss will be greater—approximately rather in what it leaves out. Although it is important to two to three times so—than the motivation to obtain a gain understand the psychology of rational ization, the other fac- of equivalent value, which helps explain why being in a tor that Mazar. Amir, and Ariely recognize but then largely hole produces such strong motivation. However, such sim ignore—námely, the motivation to behave dishonestly—is ple amplification of value does not fully capture the magni arguably the more important side of the dishonesty tude of motivation produced by such situations, which often equalion. has a powerful emotional component—a feeling of intense misery and desperation. Much as miseries, such as hunger The motivation side is especially important, in part and pain, tend to crowd out altruism (Loewenstein 1996). because the propensity to rationalize is itself a function of hypermotivation can cause people to shed. temporarily, the the motivation to do so. Given sufficient motivation, people ethieal constraints of everyday lite. can persuade themselves of almost anything. including why behavior they normally would consider unethical is morally In a remarkable study of the causes of hypermotivation, acceptable. Research on the self-serving faimess bias (for a Cressey (1950) personally interviewed hundreds of incar- summary, see Babcock and Loewenstein 1997) shows that cerated embezzlers and pored through large data sets col- people tend to conflate what is fair with what is in their per lected by other researchers. He found that such erimes were sonál interest, and the same is no doubt true of people’s a response to problems that often began with, as he put it, judgments of what is ethieal. Given a sufficientiy powerful “gambling. drink, and extravagant living” (p. 739). One motivation to commit an act of fraud, in generál, people are prisoner spontaneously came to a simiiar conclusion: more than capable of rationalizing why it does not conflict with their own ethieal precepts. Furthermore, after people Tlíc more I think about it. tlíc more I'm inelined to háve taken the first step toward unethical behavior. a large think that before a person does a thing Iike that he must body of research shows that subsequenl steps into the abyss liuvc doně something previously that tlíc community of immorality become progressively easier (e.g., Lifton wouldiťt approve of. If he’s in an env iron ment and 1986; Milgram 1963). isn't leading a double lifc and docsivt háve anything to liidc, I caif t conccivc of him starting with an embez- *Scott Riek is a posidocloral fellow and lecturer, The Wharton School, zlcment. He has to do something previously. (Cressey University of Pennsylvania (e-mail: srick@ wharton. upenn.edu). George 1953, p. 40) Loewenstein is Herbert A. Simon Professor of Economics and Psychology, Department of Sociál and Decision Sciences, Camegie Mellon University After subjecting his extensive database to an intense (e-mail: gl20@andrew.cmu.edu). Rick thunks the Risk Management and serutiny. which he labeled “negative čase analysis,” and Decision Processes Center at Wharton for funding. Ziv Carmon served as systematically attempting to challenge his own conclusions. guest editor for this commenlary. Cressey (1950, p. 742) proposed the fotlowing: Trustcd persons become trust violators when they con- ceivc of themselves as having a financial problém which is non-shareable, háve the knowiedge or avvarc- © 2008. American Marketing Association Journal of Marketing Research Vol. XLV (Deccmber 2008). 645-653 ISSN: 0022-2437 (prim). 1547-7193 (eleetronic) 645 646 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 ncss that this problém can bc secrctly resolved by vio- people are loss averse. tliey are also powerful ly averse to lation of the position of financial trust, and are able to inequality (Loewenstein, Thompson, and Bazcrman 1989). apply to their own conduct in that situation verbaliza- tions which enable them to acl just their conceptions of Academia is a domain in which reference points are par- themselves as trusted persons with their conceptions of ticularly likely to be defined in terms of the attainments of themselves as users of the entrusted funds or propeny. others. Academia is becoming inereasingly competitive, and the (professional) achievements of others háve never The rational i zations (or “verbalizations”) that Mazar, Amir, been easier to assess (through Online curricula vitae or pro- and Ariely study play a role in Cressey’s framework. How filcs). The increasing intensity of competition within acade ever, they are only the tínal step in a process set into motion mia can be ťelt at all levels. More undergraduale students when people find themselves in troubíc as a result of “non- are entering graduate school with curricula vitae that elicit shareable financial probletns.” jealousy from some of their older peers. and the publication requirements for getting a first job are approaching a level A further difference between Cressey’s (1950) embez that not long ago would háve been sufflcient for tenure at zlers and Mazar, Amir, and Ariely’s cheaters is how the two most institutions. Even journals are becoming inereasingly evaluale their dishonesty in retrospect. Participants in competitive with one another (Huber 2007; Lawrence Mazar, Amir, and Ariely’s Experiment 4 realized that they 2003). With standards ratcheting upward, there is a kind of had cheated, bul they did not believe that they needed to “arms race” in which academics at all levels must produce update the extern to w'hich they view'ed themselves as hon more to achieve the same career gains. Some of this est. In contrast, Cressey’s (1953, p. 120) embezzlers increased productivity is enabled by new technology, such “define themselves as eriminals, find this definition incom- as computers and the Internet, and some comes from people patiblc with their posilions as trusted persons, and usually putting in longer hours. However, some of it. we fear, condemn themselves for their past behavior." Although comes from researchers pushing the envelope of honesty— rationalizations likely preceded the dishonesty observed or worse. both in Mazar, Amir, and Ariely’s (2008) laboratory studies and in Cressey’s real-world cases, Cressey’s findings sug- An unfortunate iinplication of hypermotivation is that as gest that serious acts of dishonesty can be rationalized only competition within a domain increases, dishonesty also for so long. tends to increase in response. Goodstein (1996) feared as much over a decadc ago: DePaulo and colleagues (2004) observe a simiiar pattern in a study in which undergraduale students and nonstudent Throughout most of its recent history, science was con- adults were asked to deseribe the most serious lie they ever strained only by the Iimits of imagination and creativ- told. They found that “the vast majority of serious lies orig- ity of its participants. In the past couple of decades that inate with bad behaviors" (p. 164). Unlike the participants statě of affairs has changed dramatically. Science is in Mazar, Amir, and Ariely’s studies, who lied to achieve a now constrained primarity by the number of research modest amount of additional profit, participants in DePaulo posts, and the amount of research funds available. and colleagues’ study lied to hide extramarital affairs, gam- What had always previously been a purely intellectual bling debts, and other serious transgressions that could competition has now become an intense competition jeopardize careers or marriages if revealed. Moreover. for scarce resources. This change, which is permanent DePaulo and colleagues’ participants reported feeling and irreversible, is likely to háve an undesirable effect disiressed while lelling their lies, and those who were ulti- in the long run on ethieal behavior among scienlisis. mately caught reported feeling guilty and remorseful. Instances of scientific fraud are almost sure to become Again, as in Cressey’s (1950, 1953) studies, it appears that more coinmon. many of DePaulo and colleagues’ participants were ulti- mately forced to update their self-concept. Whereas recent high-profile instances of data falsifica- tion in the medical Sciences háve received much attention in The feeling of being in a hole not only originates from Science (e.g., Couzin and Schirber 2006; Normile 2006; nonshareable unethical behavior bul also can arise, more Xin 2006). anonymously self-reported data falsification has prosaically, from overiy ambitious goals (Heath, Larrick, rcccntly been documented in fields closer to home as well and Wu 1999). In the lab, Schweitzer, Ordóňez, and Douma (in marketing, see Mason, Bearden, and Richardson 1990; (2004) find that participants who had ambitious goals over- in economics, see List et al. 2001). In addition, Martinson, stated their productivity significantly more often than par Anderson, and De Vries (2005) measured self-reported mis- ticipants who were simply asked to do their best. In the conducl among more than 3000 researchers funded by the classroom, the prospect of falling short of one’s own per National Institutes of Health and found that onc-third formance goals (Murdock and Anderman 2006) or perhaps reported engaging in “questionable research practices” parents’ goals (Pearlin. Yarrow, and Scarr 1967) appears to (e.g., dropping observations or data points front analyses encourage cheating. Likewise, in organizational settings, because of a gut feeling that they were inaccurate). Surely, the desire to meet ambitious profit goals often leads to the moderate amount of self-reported misconduct is a mere questionable accounting practices (Degeorge, Patel, and lower bound on the actual amount of misconduct occurring. Zeckhauser 1999; Jensen 2001; Prentice 2007). CLOSING COMMENTS ACADEMIE, HEAL THYSELF The economist Andrei Shlcifer (2004) explicitly argues Another important reference point that can lead to the against our perspective in an article titled “Does Competi perception of being in a hole is the attainments of others. tion Destroy Ethieal Behavior?” Although he endorses the Research on sociál preferences has shown that as much as premise that competitive situations are more likely to elicit unethical behavior, and indeed offers several examples Commentaries and Rejoinder 647 other than those provided here, he argues against a psycho such registries for the behavioral Sciences because they logical perspective and instead attempts to show that “con- should also offer the addcd benefit of making it more diffi ducl described as unethical and blamed on 'greeď is some- cult for researchers to fudge their data. times a consequence of market competition” (p. 414). However, we believe that he makes a fundamental mistake Some may object to the tise of registries on the grounds in implicilly viewing greed as an individua! difference that they will invariably lower productivity. However, we variable that can be conlrasled with market conditions. In believe that (slightly) reduced productivity could be benefi- contrast to Shleifer, we argue that conditions of extreme ciai. for at least two reasons. One reason is based on the competition Jead to unethical behavior exactly because they vicious circle in which fraud and standards act to influence lead to greed—that is, hypermotivation.1 By raising aspiru- one another: Fraud increases productivity, which in turn tions above what is possible to attain with normál, moral raises standards, which in turn stimulates fraud. Curbing means, competition creates the psychological conditions for research transgressions could break the eyele, reducing pro fraud. ductivity and bringing standards back down to earth. The second benefit is that making it more difficult to publish Shleifer (2004) concludes optimistically, arguing that fudged findings would benefit those who otherwise would competition will lead to economic growth and lhal wealth háve based subsequeni research on those findings. There is tends to promote high ethieal standards. We are more pes- a nonnegligible proportion of findings in our field that is simistic and, we believe, more pragmatie. Competition may difficult to replicate. Whereas some of these failures to promote progress. but it also inevitably creates winners and replicate are due to differences in subject populalions. to losers, and usually more of the latter than the former. The hidden moderators, or to good luck on the part of the initial perceived difference in oulcomes between winners and los researchers or bad luck on the part of those conducting ers (e.g., get the job and the good life versus remain unem- follow-up studies, some are surely due to the research mis- ployed and deprived) has the polenlial to hypermotivate demeanors or felonies of the originál authors. Making it would-be losers to turn themselves into winners through more difficult for researchers to misbehave could reduce unethical behavior. the amount of time spent trying and failing to replicate the unreplicable. How should society respond to the problems caused by hypermotivation? Unfortunately, practical concerns limit REFERENCES the potential for muling the forces of competition, for example, by offering rewards that linearly increase with Ariely, Dan, George Loewenstein, and Dražen Prelec (2003), performance rather than offering all rewards to a single “Coherent Arbdtrariness: Stable Demand Curves Without Stable winner or a smáli number of winners. What, for example, Preferences,” Quarterly Journal of Economics, 118 (1), 73—105. can companies offer rejected job applicants beyond assur- ance that the decision was a difficult one and the obligalory Association of American Medical Colleges (2007), “The Scien- promise to keep their résumé on lile? If making competi tific Basis of Influence and Reciprocity: A Symposium,” spon- tion more humane is impraclical, what can be doně lo curb sored by AAMC and Baylor College of Medicine, Department dishonesty? of Neuroscience and Computational Psychiatry Unit, Washing ton, DC (June 12). We are not quite as pessimistic as Mazar, Amir, and Ariely regarding the importance of factors identified as Babcock. Linda and George Loewenstein (1997), “Explaining important by the standard economic perspective, such as Bargaining Impasse: The Role of Self-Serving Biases,” Journal the probability of getting caught and the magnitude of pun ofEconomic Perspectives, II (I), 109-126. ishment if caught. There is evidence that such factors can be influential (Cox. Cox, and Moschis 1990: Scholz and Chang. Otto H. and Joseph J. Schultz Jr. (1990). “The Income Tax Pinney 1995), particularly when they are made salient Withholding Phenomenon: Evidence from TCMP Data,” Jour (Ariely, Loewenstein, and Prelec 2003. p. 101). However, nal ofthe American Taxation Association, 12 (Fall), 88-93. we do not believe that the key to deterring fraud lies in increasing the probability of getting caught or the severity Couzin. Jennifer and Michael Schirber (2006), “Fraud Upcnds of punishment. Instead. we believe that fostering an atmos- Oral Cancer Field, Casting Doubt on Prevention Trial,” Science, phere of openness and transparency will be most effective, 311 (5760), 448-49. in part by making it more difficult to comniit acts of fraud. In the academie domain, for example, transparency could Cox, Děna, Anthony D. Cox, and George P. Moschis (1990), be promoted by establishing registries lhal make publicly “When Consumer Behavior Goes Bad: An Investigation of available the raw data and complete deseriptions of meth- Adolescent Shoplifting,” Journal of Consumer Research, 17 ods for both published and unpublished studies. Medical (September), 149-59. scientists háve long advocated the crealion of such reg istries to overcome the “filé drawer problém” (Dickersin Cressey, Donald R. (1950). “The Criminal Violation of Financial and Rennie 2003; Simes 1986). We advocate the crealion of Trust,” American Sociological Review, 15 (6), 738-43. •Although we essentially equate hypermotivation witli greed here. oíher --------- (1953), Other People’s Money. Glencoe, IL: The Free conceptualizations of greed would not be consistent with what we call Press. hypermotivation. For example, Wang and Murnighan (2007, p. 2 1 > pro- pose that greed “typically starts with inner desire/temptation, i.e., a poten Degeorge, Francois, Jayendu Patel, and Richard Zeckhauser tial. desirable outcome.” In contrast, we propase that hypermotivation is (1999), “Earnings Management to Exceed Thresholds,” Journal driven hy the desire to avoid or recoup losses rather than the desire to ofBusiness, 72 (l), 1-33. obtain positive outcomes. DePaulo, Bella M.. Matthew E. Ansfield, Susan E. Kirkendol, and Joseph M. Boden (2004), “Serious Lies,” Basic and Applied Sociál Psychology, 26 (2-3), 147-67. Dickersin, Kay and Drummond Rennie (2003), “Registering Clin- ical Trials," Journal of the American Medical Association, 290 (4), 516-23. Goodstcin, David (3996), “Conduct and Misconduct in Science,” Annals of the New York Academy of Sciences, 775 (I), 31—38. Healh, Chip. Richard P. Larrick, and George Wu (1999), “Goals as Reference Points,” Cognitive Psychology, 38 (1), 79-109. 648 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 Huber, Joel (2007), “Journal of Marketing Research in the New Wang, Long and J. Keith Murnighan (2007), “On tlíc Duality of Competitive Journalistic Enviromnent," Journal of Marketing Greed,” working páper, Kellogg School of Management, North Research, 44 (Fcbruary), 1 -3. western University. Jensen, Michael C. (2001), "Corporate Budgeling Is Broken: Leťs Xin, Hao (2006), “Invention of China's Homegrown DSP Chip Fix It,” Harvard Business Review, 79 (6), 94-101. Dismisscd as a Hoax,” Science, 312 (5776), 987. Kahneman, Daniel and Amos Tversky (1979), "Prospect Theory: Homo Economicus’ Soul An Analysis of Decision Under Risk.” Econometrica, 47 (2), 263-92. JOHN R. MONTEROSSO and DANIEL D. LANGLEBEN* Lawrcnce, Peter A. (2003), “The Politics of Publication,” Nátuře, Mazar, Amir, and Ariely's (2008) claiin that microeco- 422 (6929). 259-61. nomic models fail to account for altaiism and other sociál Lifton, Robert J. (1986), The Nazí Doctors. New York: Basic phenomena is not entirely accurate: The neoclassical economisťs perspective is only that behavior can be mod- Books. eled as orderly maximization of utility functions that cap- List, John A., Charles D. Bailey, Patricia J. Euzent, and Thomas L. ture whatever it is that a person cares about (Becker 1991). That said, in recent years, behavioral scientists háve con- Martin (2001), “Academie F.conomists Behaving Badly? A Sur ducted a great deal of important experimental work. vey on Three Areas of Unethical Behavior,” Economic Inquiry, demonstrating ways that human behavior deviates from the 39 (1), 162-70. materiál reward maximizer, “Homo Economicus.” Evolu- Loewenstein, George (1996), “Behavioral Decision Theory and tionary biologists played a founding role, specifying mechanisms by which “selfish genes” produce altruistic Business Bthics: Skewed Tradeofřs Between Šelf and Other,” in organisms (Hamilton 1964; Trivers 1971), and experimen- Codes of Conduct: Behavioral Research into Business Ethics, talists provided abundant confirmation that benevolent and David M. Messick and Ann E. Tenbrunsel. eds. New York: Rus- malevolent sociál motives are potent (Batson, Fultz, and sell Sage Foundation, 214—26. Schoenrade 1987; Camerer and Tlialer 1995) and engage --------- , Leigh Thompson, and Max H. Bazerman (1989), “Sociál much of tlie same neural circuitry as other motivations Utility and Decision Making in Interpersonal Contexts,” Jour (Sanfey et al. 2003; Tabibnia and Lieberman 2007). Mazar. nal of Personality and Sociál Psychology, 57 (3), 426-41. Amir, and Ariely further suggest that dishonesty is con- Martinson, Brian C., Melissa A. Anderson, and Raymond de Vries strained by the unconscious tendency to preserve a ťavor- (2005), “Scientists Behaving Badly,” Nátuře, 435 (7043), able self-image. The logic of this “self-signaling” mecha 737-38. nism is as follows: (I) People value a particular conception Mason. J. Barry, William O. Bearden, and Lynne D. Richardson of their own šelf; they want to possess certain traits and qualities (some of which are "moral”); (2) people infer their (1990), “Perceived Conduct and Professional Ethics Among own traits in much the same way they infer the traits of oth Marketing Faculty." Journal of the Academy of Marketing Sci ers (Bem 1965; Mead f 19341 1962); and thus (3) people’s ence, 18 (3), 185-97. behavior is in part shaped by wanting to provide evidence Mazar, Nina, On Amir, and Dan Ariely (2008), “The Dishonesty (to themselves) that they possess the desirable traits. of Honest People: A Theory of Self-Concept Maintenance,” We agree but notě that neglect of self-signaling in most Journal of Marketing Research, 45 (December), 633-44. economic models is not a mere oversight; there are diťficul- Milgram, Stanley (1963), “Behavioral Study of Obedience,” Jour ties inherent in quantifying self-signaling motivations. As nal ofAbnorrnal and Sociál Psychology, 67 (4), 371-78. Mazar, Amir, and Ariely suggest. people treat new diagnos- tic information about their personál qualities in self-serving Murdock, Tamera B. and Eric M. Anderman (2006). “Motiva- ways, exploiting ambiguity and ignoring evidence when tional Perspectives on Student Cheating: Toward an Intcgratcd possible to allow cheating that “flies below the radar.” Model of Academie Dishonesty,” Educational Psychoiogist, 41 Specifically, they suggest that if a person takés a little extra (3), 129-45. from the proverbial cookie jar, even in the absence of risk of external punishment, there is a cost to bear iu terms of Normile, Dennis (2006), “Panel Discredits Findings of Tokyo potentially aversive diagnostic information indicating a University Team,” Science, 311 (5761), 595. dishonest šelf. What complicates matters is that this cost may be reduced or avoided through diversion of attention Pearlin. Leonard L. Marian R. Yarrow, and Hany A. Scarr (1967), “Unintended Effects of Parental Aspirations: The Čase oťChil- ‘John R. Monterosso is Assistant Professor of Psychology, Department dren’s Cheating,” American Journal of Sociology, 73 (1), of Psychology. University of Southern Calilornia. Los Angeles (e-mail: 73-83. johnrmon@usc.evlu). Daniel D. Langleben is Associale Professor of Psy chiatry, University of Pennsylvania (e-mail: langlebe@upenn.edu). The Prentice, Robert A. (2007), “Ethieal Decision Making: More authors lhank David Seelig for helpíul comments. The authors received Needed Than Good Intentions,” Financial Analysis Journal, 63 financial support from the lollowing sources: NIH DA015746 (DL), NIH (6), 17-30. R01DA021754 (JM), and NIH R01DA023176 (JM). Schepanski. Albert and David Kelsey (1990), “Testing for Frarn- ing Effects in Taxpayer Compliance Decisions,” Journal of the American Taxation Association, 12 (Fall). 60-77, Scholz, John T. and Neil Pinney (1995), "Duty, Fear, and Tax Compliance: The Heuristie Basis of Citizenship Behavior,” American Journal of Political Science, 39 (2), 490-5 12. Schweitzer, Mauriee E.. Lisa Ordóňez. and Bambi Douma (2004), "Goal Sctting as a Motivator of Unethical Behavior,” Academy of Management Journal, 47 (3), 422—32. Shleifer, Andrei (2004), "Does Competition Dcstroy Ethieal Behavior?” American Economic Review, Papers and Proceed- ings, 94 (2), 414-18. Simes. Robert J. (1986), “Publication Bias: The Čase for an Inter national Registry of Clinical Triais,” Journal ofCIinical Oncol- ogy, 4(10), 1529-41. Vohs, Kathlccn D. and Jonathan W. Schooler (2008), "The Valtic of Believing in Free Will: Encouraging a Belief in Determinism Increases Cheating,” Psychological Science, 19 (1), 49-54. Commentaries and Rejoinder 649 and rationalization (what we loosely refer lo as “self- been found to be associated with increased activalion in the deception”). right ventral and lateral prefrontal cortex and decreased response in the limbic systém (Anderson et al. 2004). This Compare self-signal valuation with the valuation of pattern is simiiar to other-deception, with the addition of money: $10,000 is differentially valued, but barring exotic the limbic systém. We speculate that brain activity associ circmnstances, it will always háve some positive value, and ated with labeling self-signals as unwanted and keeping its value can be inferred on the basis of what a person will them from entering awareness may help distinguish self- trade for it. However, as Mazar. Amir, and Ariely rightly ťrom other-deception. The mechanism through which a argue, the value of self-signals depends crucially on the conflict between the existing self-image and an incoming degree of prevailing seif-deception. That superficial moral self-signal is detected may be simiiar to the one involving reminders (i.e., the researchers’ Ten Commandments the medial prefrontal and insular cortices and employed in manipulation) háve such a large impact is, in itself, an error monitoring (Klein el al. 2007). This mechanism may important observation (Shariff and Norenzayan 2007), but be deficient in certain addictive disorders (Forman et al. it also suggests enormous volatility in the price of diagnos- 2004), which also involve gross discrepancies between self- tic information that could complicate the transition from a image and reality. The role of the limbic systém as the clever demonstration to actual modeling of the behavior. modulátor of the prefrontal cognitive and motor control (Goldstein et al. 2007; Schaefer et al. 2002) makes it a can- CAN COGNITIVE NEUROSCIENCE HELP didate for a simiiar role in modulating the degree of aware D/STINGUISH SELF-DECEPTION FROM DECEPTION ness of unwanted self-signals. Thus, we anticipate that neu roscience experiments using functional imaging technology OF OTHERS? would be able to characterize the self-deception and other- deception, as well as other processes underlying the dishon Although Mazar, Amir, and Ariely’s behavioral methods esty of honest people. achieve a high level of external validity, they do not allow individual occurrences of cheating to be recorded. This is The idea that immoral behavior is constrained by motiva understandable; tempting participants and surreptitiously tions related to a desired self-concept is not new. Among capturing individual acts of dishonesty may be incompali- other places, it is well developed in the work of Sigmund ble with the standards of informed consent. However, many Freud, as is the idea of motivated avoidance of undesirable critical questions cannot be addressed by group distribu self-signals (some Freudian “defense" mechanisms, includ- tions (e.g., What distinguishes those who cheat from those ing repression, can bc so characterized). However. the idea who do not?). Moreover, if Mazar, Amir. and Ariely are has largely been ignored by hard-nosed empirical behav correct that deception of others is constrained by self- ioral scientists because there has not been an obvious way signaling and that, in turn, self-signaling is constrained by to move to the type of well-specified predictive models that self-deception (and we believe that they make a compelling made economics the most practically useful of the sociál čase), further progress will depend on finding ways lo study Sciences. The recent interest front economists (Bodner and these underlying processcs. Here, there may be an opportu Prelec 1997) and empiricists, such as Mazar, Amir, and nity for the methods of cognitive neuroscience to be pro- Ariely, may clear a path forward, and neuroimaging may ductively brought to bear. provide the best hopc yet of elucidating the neural mecha nisms of the processes underlying human self-signaling. Cognitive psychology and cognitive neuroscience háve The most exciting implication of Mazar, Amir. and Ariely’s well-developed methods for decomposing complex pro- findings may be that they highlight the long way left to go. cesses (Aguirre and D’Esposito 1999). These methods háve In the meantime, the authors may find themselves in already been applied to the study of deception. intent, and demand the next time U.S. Congress debates the merits of awareness, using functional magnetit' resonance imaging school prayer. (fMRI), which allows changes in brain activity to bc meas- ured on a spatial scale of a few millimeters and a temporal REFRENCES scale of approximately one second (Haynes 200K; Lan gleben 2008; Soon et al. 2008; Spence 2008). Mazar, Amir, Aguirre, G.K. and M. D’Esposiio (1999). “Experimental Design and Ariely’s participants lied. but did they lie to them for Brain fMRI," in Functional MRL C.T.W. Mooncn and P.A. selves, to others, or both? To what extent were the partici Bandettini. eds. New York: Springer, 369-80. pants’ behaviors conscious and motivated? In other words, can sclf-dcception and interpersonal deception (“other- Anderson, M.C.. K.N. Ochsner, B. Kuhl. J. Cooper, E. Robertson, deception”) (Sackeim and Gur 1979) be teased apart by S.W. Gabrieli, G.H. Glovcr, and J.D. Gabrieli (2004), “Neural neuroscience? Previous fMRI studies of deception ušed Systems Underlying the Suppression of Unwanted Memories,” forced-choice protocols that produced motivated, inten- Science, 303 (5655), 232-35. tional, and conscious deception. These studies produced a reproducible pattem of lateral prefrontal and parietal activa- Batson, C., J. Fultz, and P. Schoenrade (1987), “Distress and tion that has been hypothesized to be due to the cognitive Empathy: Two Qualitalively Distinct Vicarious Emotions with elTort involved in inhibiting the “prepotent” tendency Different Motivational Consequences,” Journal of Personality, toward veridical responding (Langleben et al. 2002; Spencc 55 (1), 19-39. et al. 2004). This early hypothesis is likely to be an over- simplification (Langleben et al. 2005; Sip et al. 2008; Becker, Gary S. (1991), A Treatise on the Family. Cambridge, Spence et al. 2008: Ward et al. 2003). MA: Harvard University Press. At the time of this writing, we are not aware oťpublished Bern. D. (1965), “An Experimental Analysis of Self-Persuasion," fMRI studies dedicated to self-deception. Repression (moti Journal ofExperimental Sociál Psychology, I (3), 199-218. vated exclusion of distressing materiál from attention) has Bodner. R. and D. Prelec (1997), “The Diagnostic Value of Actions in a Self-Signaling Model,” mimeo, Massachusetts Institute of Technology (January). 650 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 Camcrer, C. and R. Thaler (1995), “Ultimatums, Dictators, and Ventrolateral Prefrontal Cortex to Vocal Deception,” Neumltn- Munners,” Journal of Economic Perspectives, 9 (2), 209-219. age, 40(4), 1411-18. Tabibnia, G. and M.D. Lieberman (2007), "Fairncss andCoopera- Forman, S.D., G.G. Dougherty, B..I. Cascy, G.J. Sicglc, T.S. tion Are Rcwarding: Evidence from Sociál Cognitive Ncuro- Braver, D.M. Barch. et al. (2004), “Opiáte Addicts Lack Error- science,” Armals of the New York Academy of Sciences, 1118. Dependent Activation of Rostral Anterior Cingulate," Biologi 90-101. emi Psychiatry, 55 (5). 531 -37. Trivers, R.L. (1971), “The Evolution of Reciprocal Altruism,” Quarterly Review of Biology, 46 (I), 35-57. Goldstein, M„ G. Brendel, O. Tucscher, H. Pan, J. Epstein, M. Ward, N.S., D.A. Oaklcy, R.S. Frackowiak. and P.W. Halligan Beutel. et al. (2007), “Neural Substrates of the Interaclion of (2003), “Diffcrcntial Brain Activations During Intcntionally Emotional Stimulus Processing and Motor Inhibitory Control: Simulated and Subjcctivcly Experienced Paralysis,” Cognitive An Emotional Linguistic Go/No-Go fMRI Study,” Neurolmage, Neuropsychiatry, 8 (4), 295-312. 36 (3), 1026-1040. More Ways to Cheat: Expanding the Hamilion, W.D. (1964), “The Genetical Evolutioii of Sociál Scope of Dishonesty Behaviour 1 and II,” Journal ofTheorelical Biology, 7(1), 1 -52. NINA MAZAR, ON AMIR, and DAN ARIELY* Haynes. J.D. (2008), “Detecting Deception from Neuroimaging Signals: A Data-Driven Perspective,” Trends in Cognitive Sci In our article on the dishonesty of honest people (Mazar, ences, 12(4). 126-27. Amir, and Ariely 2008), we proposed a theory of self- concept maintenance that aliows people who consider Klein, T.A., T. Endrass, N. Kathmann. J. Neumann, D.Y. von Cra- themselves honest human beings and háve high beliefs in mon, and M. llllsperger (2007), “Neural Correlates of Error their own morality to cheat “a little bit” without altering Awareness," Neurolmage, 34 (4), 1774—81. their self-image. In other words, they can benefit financially from their dishonesty without paying the cost involved in Langleben, D.D. (2008), “Detection of Deception with fMRI: Are “recognizing” themselves as being dishonest. This concept We There Yet?” Legal and Criminological Psychology, 13 (I). results in a substantial amount of “mostly honest” people. 1-9. Although this is the pessimistic interpretation of our results, the optimistic perspective is that our participants did not --------- , J.W. Loughead, W.B. Bilker. K. Ruparel. A.R. Childress, cheat enough (i.e., as much as standard economics would S.L Busch, and R.C. Gur (2005), "Telling Truth From Lic in predict)! That is, even when there were no external conse Individual Subjects with Fast Evenl-Related fMRI,” Human quences to cheating and even when the amount they could Brain Mapping, 26 (4), 262—72. gain was substantial, they did not cheat much. --------- , L. Schroeder, J.A. Maldjian, R.C. Gur, S. McDonald, J.D. The generál idea behind the theory of self-concept inain- Ragland, C.P. 0’Brien, and A.R. Childress (2002), “Brain tenance is that there is a range of dishonesty within which Activity During Simulated Deception: An Event-Relaled Func people do not need to confront the trne meaning or implica- tional Magnetic Resonance Study,” Neurolmage, 15 (3), tions of their actions. Although our experimental evidence 727-32. supports our theory, we agree with Monterosso and Lan gleben (2008) that a necessary next step is to find ways to Mazar, Nina, On Amir, and Dan Ariely (2008), “The Dishonesty distinguish. at the process level, between dishonesty based of Honest People: A Theory of Self-Concept Maintenance,” on the explicit trade-off of external costs and benefits and Journal of Marketing Research, 45 (December), 633-44. the self-concept maintenance idea we propose (see also Wirtz and Kum 2006). In this regard, we suspect that neu Mead, G.H. ([ 1934] 1962). Minci, Šelf. and Society. Chicago: Uni roimaging research could prove useful (though we are also versity of Chicago Press. obliged to mention that Amos Tversky ušed to say that the main finding of neuroimaging is that people think with Sackeim, H. and R.C. Gur (1979), “Self-Deception, Other- their brains). but achieving this goal would require a differ Deception, and Šelf-Reported Psychopathology,” Journal of ent research methodology than the one we ušed in our stud Consulting and Clinical Psychology, 47 (1), 213-15. ies. In particular, instead of confronting participants with a one-shot game in which there is only one opportunity to be Sanfey, A.G.. J.K. RiIIing. J.A. Aronson, L.E. Nystrom, and J.D. dishonest and participants need to decide whether and to Cohen (2003), “The Neural Basis of Economic Decision- what extent to cheat, a neuroimaging experiment would Making in the Ultimatum Game,” Science, 300 (5626). require multiple decision-making trials, each of which 1755-58. *Nina Mazar is Assistant Professor of Marketing. Joseph L. Rotman Schaefer, S.M.. D.C. Jackson. R.J. Davidson. G.K. Aguirre. D.Y. School of Management, University of Toronto (e-niail: nina.mazar<§> Kimberg, and S.L. Thompson-Schill (2002), “Modulation of utoronto.ca). On Amir is Assistant Professor of Marketing, Rady School of Amygdalar Activity by the Conscious Regulation of Negative Management, University of Califomia, San Diego (e-mail: oamir@ucsd. Emotion,” Journal of Cognitive Neuroscience, 14 (6), 913-21. edu). Dan Ariely is Visiting Professor of Marketing, Fuqua School of Business, Duke University (e-mail: dandan@duke.edu). SharilT, A.F. and A. Norenzayan (2007), "God Is Watching You: Priming God Concepts Increases Prosocial Behavior in an Anonymous Economic Game,” Psychological Science, 18 (9), 803-809. Sip. K.E.. A. Roepstorff, W. McGregor, and C.D. Frith (2008). “Detecting Deception: The Scope and Limits,” Trends in Cogni tive Science, 12 (2), 48-53. Soon, C.S.. M. Brass, H.J. Heinze, and J.D. Haynes (2008), “Unconscious Determinants of Free Decisions in the Human Brain,” Nátuře Neuroscience, I I (5), 543-45. Spence, S. (2008), “Playing DeviPs Advocate: The Čase Against fMRI Lie Detection,” Legal and Criminological Psychology, 13 (1), 11-25. --------- , M.D. Hunter, T.F. Farrow. R.D. Grecn, D.H. Lcung, C.J. Hughes, and V. Ganesan (2004), "A Cognitive Neurobiological Account of Deception: Evidence from Functional Ncuroimag- ing,” Philosophical Transactions of the Royal Society of Lon don, 359 (1451), 1755-62. --------- , C. Kayior-Hughes, T.F. Farrow, and l.D. Wilkinson (2008), “Speaking of Secrets and Lies: The Contribution of Commentaries and Rejoinder 651 would require lhát parlicipants face the tlilemma of benefit- A second way we want to expand our understanding of ing from dishonesty or acting honestly.2 dishonesty is by incorporating the perspective that Rick and Loewenstein (2008) propose. In our basic theory of self- One approach to such a multiple-trial type of experiment concept maintenance, we emphasize the role of two distinct is to use the "dots game” that we recently developed to but interrelated mechanisms: categorization and attention to investigate the dynamics of people’s decisions to cheat over moral standards. We hypothesize that for cerlain types of time (Mazar and Ariely 2008). The proceduře is simple, actions and magnitudes of dishonesty, people can catego involving a Computer-based task in which participants see a rize their actions into more compatible terms and find box divided by a diagonál line with 20 smáli dots inside the rationalizalions for them. As a consequence, people can box. On each trial. the dots are presented for only one sec cheat while avoiding any negative self-signals that might ond, after which they disappear: participants are then asked affect their self-concept and thus avoid negatively updating to indicate whether there were more dots in the left half of their self-concept altogelher (Gur and Sackeim 1979). the box or in the right half of the box. When the response is Whereas the categorization mechanism depends heavily on made. a new trial appears, and this goes on for a while. the external world In terms of the type of exchange, the Although the task explicitly asks participants to indicate on objects in question, and so forth, the attention lo standards which side there are more dots, the payment in this design mechanism relies on salience.-1 is such that it tempts participants to indicate that there are more dots on the right side. That is, a participant receives Rick and Loewenstein (2008) emphasize the role of higher payoffs ($.05) for every trial in which he or she cor motivation as another important faelor. In particular, they rectly or incorrectly identifies that the box has more dots on point to two sources of motivation for dishonesty: competi the right side (simiiar to the situations physicians face) and tion and avoidance or recoup of losses. Motivation can lower payoffs ($.005) when he or she correctly or incor influence people’s dishonesty in two major ways. First, rectly identifies that the box has more dots on the left side. motivation lo gel “out of a hole” can increase people’s By creating this asymmelric payment, the task essenlially propensity for self-serving categorization, thus increasing represents a conflict of interest, such that the main task of self-concept maintenance dishonesty. Second, it can háve a telling the truth sometimes stands in opposition to the direcl effect on dishonesty through standard rational dis financial incentives physicians encounter on a daily basis. honesty. In particular, it can lead to a point al which the perceived extemal costs and benefits of dishonesty trump The results of a few experiments with this paradigm any consideratíons of maintaining an honest self-image; at revealed two main findings: First, participants cheat when this point, people may carry out dishonest acts consciously they háve the incentive and opportunity to do so. More and deliberatively. important, when participants’ behavior is plotted over the course of the entire experiment (100 trials), al some point, Figuře I depicts an extended framework for dishonesty almost all participants switch to a response stratégy in that considers the three different mechanisms of categoriza which they cheat in every trial. This latter observation sug tion, attention. and motivation. Although these mechanisms gests that people who consider themselves honest try to are nol independent from one another, Mazar and Ariely limit their cheating over multiple templations lo cheat, (2006) point out that making the right policy recommenda- which is why they cheat only once in a while in the early stages of the experiment; presumably, this behavior does ,The two mecbanisms are somewhal related. Forexampíe, more salienl not pose any threat to their self-concept. However, if they overstretch the tolerance of their self-concept. they create a moral standards could also inake it more difficult to categorize behavior as new situation in which they must face their own dishonesty, which in turn causes them to give up and cheat all the not pertaining to morality. way—a phenomenon simiiar to the “what-the-hell” effect in the domain of dieting (Baumeister and Heatherton 1996; Figuře 1 Polvy and Herman 1985). AN EXTENDED FRAMEWORK FOR DISHONESTY These results suggest that a neuroimaging experiment could be useful to distinguish between different phases of J f Competition A the process. For example, a functional magnetic resonance imaging (FMRI) technique could deliver further support for V losses the different types of dishonesty we observe in the dots experiment. In particular, it could help determine whether •Nonmalleability Motivation different areas of the brain are involved in the early trials of actions Categorization O Dishonesty (when there is a conflict between the temptation to be dis •Magnitude honest and the benefits of keeping a positive self-concept), of dishonesty Attention the trials in which people switch to a new stratégy, and the Moral 'N finál stage in which people adopt a different stratégy and cheat continuously. Such knowledge could also be crucial C reminders J in discerning the different mechanisms involved in different types of deception as well as in self-deception. -Moreover, care should be taken to convince participants that their dis honesty may go unnoticed. 652 JOURNAL OF MARKETING RESEARCH, DECEMBER 2008 tions to curb dishonesty requires the identification of of office supplies, companies could plače salienl monelary the mechanism that is driving dishonesty in a particular values on these office supplies. In another example targel- situation. ing the accounting profession, Bazerman, Loewenstein, and Moore (2002; see also Bazerman and Loewenstein 2001) If the cause for dishonesty lies mainiy in motivation and suggest passing laws or enforcing standards that bar audi- standard rational dishonesty (i.e., dishonesty stemming tors from offering both Consulting and tax Services to from an explicit analysis of external cost and benefit), the clients to decrease deceplivc audits. benefits of being dishonest must be greater than its costs. The solution then is to shift the imbalance such that the Finally, it is also clear that dishonesty is complex and costs become greater than the benefits. This can be driven by many factors, including cultural norms and the achieved by the standard legal approach of controlling the slrenglhs of a persoifs own moral standards. For example, external costs: increasing either the probability of being in a recent study on the influence of cultural norms and caught or the severity of the punishment. Good examples legal enforcement in controlling corruplion, Fisman and for this approach are the introduction of governmental task Miguel (2007) found that diplomats from high-corruption forces, such as the Department of Justice’s task force on countrics accumulated significantly more unpaid parking intellectual property, which, among other combat strategies. violations. Togelher with other observations, such studies invests in increasing the number of specially trained prose- suggest that societies cannot underinvest in their educa- eutors. The same is true for the Intemal Revenue Service’s tional efforts to strengthen the internalization of moral stan increase in audits and the music industry’s aggressive dards to make them part of their cultural norms. This basic increase in filing lawsuits against individual deceivers. premise requires the considcration of kcy questions, such as the following: How best can this be doně? Is there a critical Recent evidence suggests that the probability of punish age period for the internalization of such standards (as in ment is more important than increasing the severity of the language and visual development)? What should the limits punishment (Nagin and Pogarsky 2003). From this finding. of such efforts be? and Should societies allow all idcologies there might be ways to increase the effectiveness and effi- to participate in the creation of internal standards? In addi ciency of standard measures accordingly. Informing the tion. the considcration of self-concept maintenance also important question of what is the optimal probability for suggests that the theory of optimal punishment (optimal deterrence, Barkan, Zohar. and Erev (1998) suggest that the punishment trades off the benefits of deterrence and the best approach is eliminating the probability of being caught cost of punishing innoccnt people) should be rcconsidercd. altogether—that is, moving to nonprobabilistic punish If the punishment magnitude is determined in a way that ments (see also Erev et al. 2003). According to this per makes the costs slightly higher than the benefits and if spective, a person who expects that driving through a red these costs also include internal standards, the optimal pun light would involve a $500 Fine in 5% of the cases is more ishment will bc lower by that amount. For example, if the likely to drive through it than a person who has the same expected benefit for a particular erime is Y and the internal expected value but with ccrtainty of being caught (i.e., a reward for honesty is X. the standard rational model would definite $25 fine). More important, over time, the person in preseribe a punishment with an expected magnitude of the probabilistic punishment setting is going to discount the —(Y + e), whereas the model that includes internal rewards probability of the punishment further (as long as he or she would preseribe —(Y + e) + X. The complexity is that not is not caught), which in turn will lead to an even greater everyone has the same level of internal standards, and to the tendency for violation. Eliminating the probabilistic com- degree that these are unobservable. it is difficult to assess ponent from all undesirable behaviors is impossible, but it the true level of optimal punishment (though it is possible is clear that there are some cases (e.g., driving through an that someday there will be a test for this). Conversely, signs intersection at a red light) in which this is possible and of repeated erimínal behavior, for example. can be taken as desirable. an indication for a lower level of internalized standards. causing the magnitude of X to be updated as lower. This A less common approach for fighting standard rational type of framework, in which X is an individual variable, dishonesty is to decrease the benefits of dishonesty and to has the potential to help build a theory of repeated punish help people deal with their losses so that they do not feel ment with the same desired principles of optimal punish trapped. This theory implies that measures such as debt ment but with more effectiveness (i.e., a logical guideline consolidation and credit counseling might prove successful. for the magnitude of these policies). If the cause for dishonesty is inattention to moral stan In summary, there is no doubt that dishonesty is převa dards. however, implementing contextual cues at the point lení in daily lite. To increase the effectiveness and effi- oťtemptation might prove much more effective. For exam ciency of measures to prevent dishonesty. it is vital to ple. the Internal Revenue Service could slightly change its understand which of the distinct mechanisms is underlying forms by asking people to sign an honor code statement the behavior in a particular situation. before filling out their tax forms or by making them more personál. Another variation worth trying might be to REFERENCES include a survey that asks taxpaycrs questions such as how much they care about their country, how important honesty Barkan Ráchel, Dov Zohar, and Ido Erev (1998), “Accidents and u'as to their parents, how many people they think lie on Decision Making Under Risk: A Comparison of Four Models,” their taxes, or what the typical profile is of tax evaders. Organizational Behavior and Human Decision Processes, 74 (2), I 18—44. If dishonesty is mainiy driven by categorization, the challenge is to decrease the malleability of actions or to Baumeister, Roy F. and Todd F. Heatherton (E996), “Self- give people less room for various interpretations by increas Regulation Failure: An OverView,” Psychological Inquiry, 7(1). ing transparency. For example, to decrease employee theft 1-15. Commentaries and Rejoinder 653 Bazerman, Max H. and George Loewenstein (2001), “Taking the --------- and---------- (2008), “Temporal Influences on Cheating Bias Out of Bcan Counting,” Harvard Business Review, 79 (I), Behavior: The ‘What the Hcll’ Effect,” working páper, Sloan 28. School of Management, Massachusetts Institute of Technology. --------- ,--------- , and Don A. Moorc (2002). “Why Good Accoun- --------- , On Amir, and Dan Ariely (2008), “The Dishonesty of tants Do Bad Audits,” Harvard Business Review, 80 (11). Honest People: A Theory of Sclf-Concept Maintenance,” Jour 96-102. nal of Marketing Research, 45 (December), 633-44. Erev, Ido, Paul Ingram, Ornit Raz, and Dror Sliany (2003). "On Monterosso, John R. and Daniel D. Langleben (2008), "Homo the Possibility of Gcntlc Rule Enforccment,” working páper. Economicus’ Soul,” Journal of Marketing Research, 45 Department of Tndustria! Enginecring and Management, (December), 648-50. Tcchnion. Nagin, Daniel S. and Grcg Pogarsky (2003), “An Experimental Fisman, Ray and Edward Migucl (2007), “Corruption, Norms, and Investigation of Deterrence: Cheating, Sclf-Scrving Bias, and Legal Enforccment: Evidence from Diplomatic Parking Tick- Impulsivity,” Criminology, 41 (I), 501-527. ets,” Journal of Political Economy, 115 (6), 1020-1048. Polvy, Janet and C. Peter Herman (1985), “Dieting and Binging,” Gur. Ruben C. and Harold A. Sackcim (1979), “Self-Dcccption: A American Psychologist, 40 (2), 193-201. Concept in Seaich of a Phenomenon,” Journal of Personality and Sociál Psychology, 37 (2), 147-69. Rick, Scott and George Loewenstein (2008), “Hypermotivation,” Journal of Marketing Research, 45 (December), 645-48. Mazar. Nina and Dan Ariely (2006), “Dishonesty in Everyday Life and Its Poíicy Implications,” Journal of Public Poiicy & Wirtz, Jochen and Dorecn Kuní (2004), “Consumer Cheating on Marketing, 25 (Spring), 117-26. Service Guarantees,” Journal of the Academy of Marketing Sci ence. 32 (2), 159-75.