ATTITUDES AND SOCIAL COGNITION

The Dark Side of Creativity: Original Thinkers Can Be More Dishonest

Francesca Gino Dan Ariely
Harvard University Duke University

Creativity is a common aspiration for individuals, organizations, and societies. Here, however, we test
whether creativity increases dishonesty. We propose that a creative personality and a creative mindset
promote individuals’ ability to justify their behavior, which, in turn, leads to unethical behavior. In 5
studies, we show that participants with creative personalities tended to cheat more than less creative
individuals and that dispositional creativity is a better predictor of unethical behavior than intelligence
(Experiment 1). In addition, we find that participants who were primed to think creatively were more
likely to behave dishonestly than those in a control condition (Experiment 2) and that greater ability to
justify their dishonest behavior explained the link between creativity and increased dishonesty (Exper-
iments 3 and 4). Finally, we demonstrate that dispositional creativity moderates the influence of
temporarily priming creativity on dishonest behavior (Experiment 5). The results provide evidence for an
association between creativity and dishonesty, thus highlighting a dark side of creativity.

Keywords: creativity, ethics, morality, moral flexibility, unethical behavior

Evil always turns up in this world through some genius or other. creativity of products and individuals’ accomplishments (e.g.,
—Denis Diderot (1713–1784) Amabile, 1983; Baer, Kaufman, & Gentile, 2004; Kaufman, Baer,
Cole, & Sexton, 2008; Plucker & Renzulli, 1999); other work has
The ability to generate novel ideas and think creatively about explored the cognitive and motivational processes that lead to
problems has long been considered an important skill for individuals creative ideas (e.g., Friedman & Förster, 2001; Hirt, McDonald, &
as well as for organizations and societies. Creative thinking allows Melton, 1996; Smith, Ward, & Finke, 1995; Sternberg, 1999a) and
individuals to solve problems effectively (Mumford & Gustafson, the contextual factors that influence creative problem solving (e.g.,
1988) and to remain flexible (Flach, 1990) so that they can cope with Amabile, 1979; Förster, Friedman, & Liberman, 2004; Galinsky &
the opportunities and changes in their day-to-day lives (Runco, 2004). Moskowitz, 2000; Maddux & Galinsky, 2009; Markman, Lind-
At a more macro level, societies use new inventions, original scien- berg, Kray, & Galinsky, 2007); and still other research has exam-
tific findings, and novel social programs to advance, and organiza- ined the relationship between individuals’ personality and their
tions need them to adapt to changing environments and to succeed in creativity (Kershner & Ledger, 1985; for reviews, see Feist, 1998,
the marketplace (Goldenberg & Mazursky, 2001; Goldenberg, Ma- 1999; Simonton, 2000, 2003).
zursky, & Solomon, 1999; Oldham & Cummings, 1996). The impor- Despite their varying focus, these approaches share a basic prem-
tant role of creativity in human progress and adaptation is likely one ise: Because creativity improves problem solving and opens doors to
reason why scholars across disciplines long have been interested in new solutions and opportunities, creativity should be stimulated. But
understanding how creative thinking occurs and how it can be fos- is creativity always beneficial? Although the positive aspects of cre-
tered (Simonton, 2003). ativity have been praised and tested empirically (Goldenberg & Ma-
Creativity research in psychology has been conducted from zursky, 2001; Sternberg, 1999a, 1999b), it is possible that creative
different perspectives. Some work has focused on evaluating the thinking may also have a hidden cost in the form of increased
dishonesty when used to resolve ethical dilemmas.
In the current research, we test for this possibility and propose
that creativity has a dark side when applied to ethical behavior. We
This article was published Online First November 28, 2011.
Francesca Gino, Harvard Business School, Harvard University; Dan conduct five studies to examine whether there is a positive and
Ariely, Fuqua School of Business, Duke University. reliable relationship between creativity and dishonesty, and to
We are grateful to Max Bazerman and Ting Zhang for their insightful investigate the psychological mechanisms explaining this link.
feedback on drafts of the article. We also greatly appreciate the support and
facilities of the Computer Lab for Experimental Research at the Harvard
Business School and the Center for Decision Research at the University of Creativity and Dishonest Behavior
North Carolina at Chapel Hill.
Correspondence concerning this article should be addressed to Francesca Creativity is defined as the ability to produce ideas that are both
Gino, Baker Library, 447, Soldiers Field Road, Harvard Business School, novel (i.e., original, unexpected) and appropriate (i.e., useful,
Boston, MA 02163. E-mail: fgino@hbs.edu adaptive to task constraints; Amabile, 1983, 1988). Over the past
Journal of Personality and Social Psychology, 2012, Vol. 102, No. 3, 445– 459
© 2011 American Psychological Association 0022-3514/12/$12.00 DOI: 10.1037/a0026406

445
446 GINO AND ARIELY

several decades, researchers have explored many of the psycho- Thus, when individuals can easily generate justifications to
logical factors that are considered vital to the creative process and reinterpret or rationalize unethical actions they are tempted to
have identified two main components underlying creative perfor- engage in, they will be more likely to behave dishonestly for
mance: divergent thinking (Guilford, 1968, 1982) and cognitive monetary gains compared to situations in which justifying the
flexibility (Spiro & Jehng, 1990). Divergent thinking refers to the behavior is rather difficult. As noted by Kunda (1990), people
ability of individuals to develop original ideas and to envision reach the conclusions they want to reach, “but their ability to do so
multiple solutions to a given problem. It involves thinking “with- is constrained by their ability to construct seemingly reasonable
out boundaries” or “outside the box” (Thompson, 2008, p. 226). justifications for these conclusions” (p. 480). Greater creativity,
Cognitive flexibility, by contrast, describes the ability of individ- we suggest, facilitates this self-serving justification process. More
uals to restructure knowledge in multiple different ways depending specifically, we propose that when people are motivated to behave
on changing situational demands (i.e., the complexity of the situ- dishonestly so as to benefit financially in a given situation (or to
ation). advance their self-interest in other forms), divergent thinking is
Typically operating together, divergent thinking and cognitive likely to help them develop original ways to bypass moral rules.
flexibility help people find creative solutions to difficult problems, Similarly, cognitive flexibility is likely to help them reinterpret
which may be interpreted from different points of view. One such available information regarding their own behavior in a self-
context is provided by ethical dilemmas. Ethical dilemmas often serving way. Indeed, as prior research has suggested, creative
require people to weigh two opposing forces: the desire to maxi- people are able to perceive and describe what remains hidden from
mize self-interest and the desire to maintain a positive view of the view of others (Carson, Peterson, & Higgins, 2003), and they
oneself (Mead, Baumeister, Gino, Schweitzer, & Ariely, 2009). are also able to develop original ideas and to envision multiple
Recent research has suggested that individuals tend to resolve this solutions to a given problem (Guilford, 1968, 1982).
tension through self-serving rationalizations: They behave dishon- As an example, consider a person’s process of figuring out
estly enough to profit from their unethical behavior but honestly which tax deductions he or she is comfortable with and which lie
enough to maintain a positive self-concept as honest human beings beyond his or her ethically acceptable boundaries. A person who is
(Gino, Ayal, & Ariely, 2009; Mazar, Amir, & Ariely, 2008). When highly creative or has been asked to think creativity about this task
facing the opportunity to behave dishonestly, in fact, most people may be more likely to identify original steps to follow and to
cheat, if only by a little bit, but not as much as they possibly could justify misreporting on taxes in novel ways. As this example
(Ayal & Gino, 2011; Gino et al., 2009). This “minor” cheating can illustrates, greater creativity may promote dishonesty in two ways.
be justified through various means. For instance, one might reason First, it can help individuals find creative loopholes to solve
that other people would cheat under the same circumstances or that difficult tasks they are facing, even if that entails crossing ethical
a little cheating will not hurt anyone. boundaries.1 Second, creativity may help individuals generate var-
Such self-serving justifications can help individuals convince ious credible reasons to justify their own actions before engaging
themselves that their behavior is in fact morally appropriate and, as in them— even when those actions are unethical. In this article, we
a result, that there is no need to negatively update their moral focus on this second direct consequence of creativity. In our
self-image. As a result, any situation in which there is room to studies, we employ tasks in which participants have the opportu-
justify potential dishonest or self-interested behavior is likely to nity to behave dishonestly and are tempted to do so. We do not
promote dishonesty (Schweitzer & Hsee, 2002; Shalvi, Dana, study whether participants cheat in more or less creative ways
Handgraaf, & De Dreu, 2011). This tendency to behave in a given a set of rules to complete the tasks. Rather, we focus on
self-interested manner when the behavior in question can be jus- whether creativity influences the way people who are tempted to
tified was demonstrated years ago in a compelling study by Sny- behave dishonestly justify their unethical actions.
der, Kleck, Strenta, and Mentzer (1979). In the study, participants In short, we expect creativity to be positively associated with
had to choose one of two rooms in which they would watch a dishonest behavior when people face ethical dilemmas (Hypothesis
movie with another person and then answer a short survey. In one 1). In addition, we expect this relationship to hold both in the case
room, the other person was physically handicapped; in the other of dispositional creativity and in the case of primes that temporar-
room, the person was not. When the movie shown was the same in ily trigger a creative mindset (Hypothesis 2). Several studies have
both rooms, participants were more likely to choose the room demonstrated that simple primes can automatically activate certain
where they would sit with the handicapped person rather than the goals and mindsets (Chartrand & Bargh, 1996; Schaller, 2003),
room where they would sit with the nonhandicapped partner. which, in turn, influence perception and behavior without explicit
However, when different movies were projected in the two rooms, conscious awareness (Bargh, 1990; Bargh & Chartrand, 1999). For
most people chose to avoid the stigmatized person because they instance, Fitzsimons, Chartrand, and Fitzsimons (2008) found that
could easily justify their choices. participants primed with Apple logos (i.e., the logos of a company
Similarly, Schweitzer and Hsee (2002) conducted a negotiation commonly associated with creativity and innovation) behaved
study in which sellers of a car provided a buyer with a mileage
estimate from a range of possible values and could lie about the 1
For instance, in the field of professional legal services, lawyers who are
estimate. The results indicated that sellers lied to a greater extent
creative or who are paid to think creatively often end up exploiting the
when the provided range was wide rather than narrow; they could loopholes and ambiguities of the law on behalf of clients, and their
justify the lie by using their increased uncertainty about the true “creative compliance” with regulatory requirements undermines the pur-
mileage. Sellers processed the information about the car’s mileage pose and effectiveness of existing regulations (McBarnet, 1988; McBarnet
in a self-serving manner, allowing them to gain financially. & Whelan, 1991).
 CREATIVITY AND DISHONESTY 447

more creatively on subsequent tasks compared to participants same 7-point scale). Finally, respondents identified their depart-
primed with IBM logos and control participants. Finally, we sug- ment within the company and indicated how much creativity they
gest that creativity promotes dishonesty by increasing people’s thought was required on their job (on a 7-point scale: 1 ⫽ not at
ability to self-justify their bad deeds (Hypothesis 3). That is, we all, 7 ⫽ very much). Three managers in the executive office also
propose that creativity increases moral flexibility, which we define provided ratings for the level of creativity required in each depart-
as individuals’ ability to justify their immoral actions by generat- ment (using a 10-point scale: 1 ⫽ not at all, 10 ⫽ very much).
ing multiple and diverse reasons that these actions can be judged We computed the z scores for all the measures included in the
as ethically appropriate. study. As Table 1 shows, the creativity required on the job (as
judged by both employees and managers) was positively correlated
Overview of the Present Research with employees’ self-reported dishonesty. These results provide
some preliminary evidence for the hypothesized association be-
We test our main hypotheses in a series of studies. First, as a tween creativity and dishonesty.
pilot study, we collect field data to examine whether people in jobs
that require high levels of creativity are more morally flexible than
Experiment 1: Effects of a Creative Personality
others. Next, we conduct five laboratory studies in which partici-
pants have the opportunity to behave dishonestly by overstating Our first study tests the hypothesis that individuals who natu-
their performance and, as a result, earn more money. In Experi- rally have a more creative personality are also more likely to
ment 1, we measure creativity as an individual difference and behave dishonestly. In addition, the study examines whether cre-
examine whether this personality trait is associated with increased ativity as an individual difference is a better predictor of dishon-
dishonest behavior. In Experiment 2, we prime cognitions associ- esty than another dispositional factor often linked to creativity:
ated with creativity and examine whether they temporarily pro- intelligence.
mote dishonesty. In Experiments 3 and 4, we explore the mecha- Whereas prior personality research has found a negative rela-
nism explaining the link between creativity and dishonesty by tionship between intelligence and academic cheating (Hethering-
focusing on people’s ability to justify unethical behavior. Finally, ton & Feldman, 1964; Johnson & Gormly, 1971), Sternberg (2001)
in Experiment 5, we examine whether individual differences in proposed that there is a dialectical relationship between creativity
creativity moderate the effect of priming a creative mindset on and intelligence. In his view, intelligence is a necessary condition
dishonesty. for creativity, which depends both on generation of novel ideas and
Across all our studies, we consistently find that greater creativ- critical analysis of them. If Sternberg’s proposed positive relation-
ity promotes dishonesty by increasing individuals’ ability to justify ship between intelligence and creativity does in fact exist, one
their unethical actions, both when considering measures of creative might also wonder whether it is intelligence and not creativity that
personality and when temporarily activating a creative mindset. leads to dishonesty. Experiment 1 jointly tested the links between
intelligence, creativity, and dishonest behavior.
Pilot Study: Creativity and Dishonesty in the Field
Method
We started our investigation of the relationship between creativ-
ity and dishonesty by collecting data from an advertising agency Participants. Ninety-seven students from local universities
located in the Southern United States. Ninety-nine employees (40 in the Southeastern United States (45 male; Mage ⫽ 21 years, SD ⫽
male; Mage ⫽ 33.48 years, SD ⫽ 8.16) across 17 different depart- 3.59) enrolled in the study for payment. Participants were paid a $2
ments within the same company responded to a short online show-up fee and were paid $3 for completing the online portion of
survey. the study; they then could earn an additional $20 based on the
First, respondents indicated how likely they would be to engage choices they made throughout the study.
in each of eight ethically questionable behaviors (e.g., “Take home Procedure. A week before the laboratory portion of the
office supplies from work,” “Inflate your business expense report”; study, participants completed an online survey that included dis-
␣ ⫽ .78) on a 7-point scale (1 ⫽ not likely, 7 ⫽ very likely). Next, positional measures of both intelligence and creativity. On the day
they read two scenarios describing a person who has the opportu- they showed up at the laboratory, participants were told the study
nity to behave dishonestly (from Gino, Norton, & Ariely, 2010; see included three different tasks testing their problem-solving abili-
Appendix A) and then indicated how likely they would be to ties, general knowledge, and perceptual skills, which had been
behave unethically if they were in the actor’s shoes (using the combined for convenience. The three tasks were a perceptual task,

Table 1
Descriptive Statistics (on Raw Measures) and Correlations (on z Scores of Each Measure), Pilot Study

Ratings type M SD 1 2 3 4

1. Ratings provided by managers for the level of creativity required in each department 7.10 1.94 .53ⴱⴱⴱ .30ⴱⴱ .46ⴱⴱⴱ
2. Ratings provided by employees for the level of creativity required in their job 4.84 1.55 .20ⴱ .24ⴱ
3. Ratings provided by employees for their likelihood to behave dishonestly across the eight behaviors 2.58 1.05 .37ⴱⴱⴱ
4. Ratings provided by employees for their likelihood to behave dishonestly across the two dilemmas 3.69 1.52
ⴱ ⴱⴱ ⴱⴱⴱ
p ⬍ .05. p ⬍ .01. p ⬍ .001.
448 GINO AND ARIELY

a problem-solving task, and a multiple-choice task. We random- and the participants had to identify which of the two triangles
ized the order in which these tasks were presented to participants. (right or left) contained more dots by clicking either on a button
Each task provided participants with the opportunity to cheat. labeled “more on left” or on a button labeled “more on right.” Each
We used multiple measures of creativity and intelligence as well trial included a square with a different number of dots in the left
as various tasks to assess cheating to test the robustness of our and right triangles.
proposed relationship between creativity as an individual differ- The instructions participants received explained how the task
ence and dishonesty. worked and gave them an example. The instructions informed
Dispositional measures. participants that their task was “to indicate whether there were
Creativity. We used three measures to assess participants’ more dots on the right side of the square or on the left side of the
creativity, all of which have been shown to robustly predict cre- square” in each round. They were also informed that a dot might
ative performance (e.g., Barron & Harrington, 1981; Feist, 1999; sometimes be on the box’s diagonal line.
Gough, 1979; Tierney & Farmer, 2002; Zhou & Oldham, 2001). Importantly, the payout in each trial was determined by the
The first measure was Gough’s (1979) Creative Personality Scale. following rule: For each left decision (“more on left”), participants
This measure asked participants to choose adjectives that best earned 0.5 cents, whereas for each right decision (“more on right”),
described them from a list of 30 adjectives. The scoring key was they earned 10 times as much (i.e., 5 cents). Using this payment
such that participants received a point every time they checked an structure, on every trial where there were more dots on the left, the
adjective related to creative personality (e.g., insightful, inventive, task presented a conflict between providing an accurate answer
original, resourceful, unconventional). (indicating left) and profit maximization (indicating right). Thus,
The second measure consisted of Hocevar’s (1980) Creative this payment structure triggered a motivation to find more dots on
Behavior Inventory. This inventory includes a list of 77 activities the right side, given that participants received the payoff simply on
and accomplishments that are considered to be creative (e.g., the basis of their responses (“more on the left” or “more on the
painted an original picture, wrote an original computer program— right”) and not on the basis of accuracy.
excluding school or university work). For each item, participants To make sure participants understood the task, they first played
indicated the frequency of the behavior in their adolescent and a few practice trials with no payment. Once the task was clear,
adult life. The scoring rule was to sum up each participant’s ratings participants played 200 trials (which were based on two blocks of
for the activities included in the inventory. 100 identical trials) on which they earned real money. On each
Finally, the third measure of creative personality was a five-item trial, they received feedback about their earnings on that trial and
scale assessing an individual’s creative cognitive style (Kirton, on their cumulative earnings up to that point.
1976). The scale included items such as “I have a lot of creative In 50 of the trials (out of each block of 100), it was clear that one
ideas,” and “I prefer tasks that enable me to think creatively” (␣ ⫽ triangle had more dots than the other, whereas in the remaining 50
.82). Participants indicated the extent to which they agreed with trials, it was somewhat ambiguous whether there was a larger
each item on a 7-point scale (1 ⫽ strongly disagree, 7 ⫽ strongly number of dots in the left or right triangles (see Figure 1 for
agree). The scoring rule was to average each participant’s ratings examples). We refer to these trials as “ambiguous,” and we focus
across the items. on them in our analysis because these are the trials that allowed for
Intelligence. As for intelligence, we used two different mea- self-serving interpretation of the position of the dots. In each
sures. The first measure was the Cognitive Reflection Test, which ambiguous trial, the participants could benefit from cheating by
consisted of three questions testing individuals’ reliance on logic creatively misinterpreting the ambiguous information they were
versus intuition; the questions are correlated with IQ (Frederick, asked to evaluate. That is, participants could intentionally misrep-
2005). Each question presents an easy “intuitive” answer that is resent their actual perception of these ambiguous trials and report
actually incorrect. For instance, one of the questions asked the “more on the right” simply because they realized that by doing so
following: “A bat and a ball cost $1.10 in total. The bat costs $1.00 they would earn a higher payoff. Thus, we use these ambiguous
more than the ball. How much does the ball cost?” One might trials to measure dishonesty.
intuitively but incorrectly say, “$0.10.” A person who is more Given the structure of this task, participants could earn a max-
thorough might respond that the ball actually costs $0.05 [0.05 ⫹ imum of $10 on this perceptual task (by always pressing the “more
(1 ⫹ 0.05) ⫽ 1.1]. Those with higher IQs tend to notice that the
intuitive answer contains inconsistencies that deserve a further
(and more time-consuming) examination.
The second measure used was the Mill Hill Vocabulary Scale,
which assesses verbal intelligence (Raven, Raven, & Court, 1998).
In this task, participants were presented with a series of 10 words
(e.g., dwindle, palliate); for each word, they were asked to choose
which of six answer options was closest in meaning to the target
word.
Tasks.
Perception task. In the visual perception task (developed by
Gino et al., 2010), participants were presented with a square that
was divided into two triangles by a diagonal line. In each trial, a
total of 20 dots appeared inside the square for 1 s and then
disappeared. The dots were distributed between the two triangles, Figure 1. Example of a perceptual task, Experiment 1.
 CREATIVITY AND DISHONESTY 449

on the right” button). Upon completion of this task, participants cheat for more money. For this question, participants indicated the
reported their performance as indicated on the computer on a extent to which overreporting performance on the tasks would be
collection slip, which they were to hand to the experimenter at the morally wrong and unethical (␣ ⫽ .73) using a 7-point scale (1 ⫽
end of the study so they could be paid.2 not at all, 4 ⫽ somewhat, 7 ⫽ extremely).
Problem-solving task. For the problem-solving task, each On average, participants earned $11.24 (SD ⫽ 1.26) in this pilot
participant received two sheets of paper. The first was a worksheet study, which lasted about 45– 60 min. They reported the payment
containing 20 matrices, each based on a set of 12 three-digit to be fair (M ⫽ 4.20, SD ⫽ 0.98) and indicated that overstating
numbers (e.g., 5.78; see Mazar et al., 2008). The second sheet was performance on the tasks included in the experiment would be
a collection slip on which participants were asked to report their unethical (M ⫽ 5.11, SD ⫽ 0.76)—a rating significantly higher
performance. In this task, participants had 5 min to find two than the scale midpoint, t(45) ⫽ 9.46, p ⬍ .001. Together, these
numbers per matrix that added up to 10, but this duration was not results suggest that any evidence of dishonesty in Experiment 1
sufficient for anyone to solve all 20 matrices. For each pair of would occur in a context in which participants perceive their
numbers correctly identified, participants received $0.25 (for a payment as fair and construe the tasks as ethical dilemmas.
maximum payment of $5). After the 5 min had passed, participants
were told to fold their worksheets and to place them in a recycling Results and Discussion
box positioned in a corner of the room; next, they were asked to As Table 2 shows, the three measures of creative personality
write down their performance scores on their collection slips. were significantly and positively correlated with one another, as
There were no identifiers on the worksheets, thus allowing partic- were the two measures of intelligence. The measures of creative
ipants to feel anonymous as they reported their performance on the personality were also positively and significantly correlated with
task. However, we changed the last two digits in one of the the level of dishonesty on each of the three tasks included in the
matrices on the worksheet and in the example provided on the back study.3 However, we did not find evidence of a link between
of the collection slip so that we could compare actual to reported creativity and intelligence or a link between intelligence and
performance. dishonesty.
Multiple-choice task. This task consisted of a general knowl- Next, we computed a z score for each of our measures and
edge quiz with 50 multiple-choice questions of varying difficulty averaged the individual scores to create one composite measure for
(e.g., “How far can a kangaroo jump?” “What is the capital of creative personality, intelligence, and extent of dishonesty.4 We
Italy?”). Participants received $.10 per correct answer (for a max- used this aggregate measures in a regression analysis testing
imum payment of $5). The experimenter told them to circle their whether dispositional creativity predicted dishonesty while con-
answers on their question sheet and explained that they would trolling for intelligence. This analysis revealed that creativity as an
transfer their answers to a bubble sheet after finishing. When individual difference was positively and significantly associated
participants finished the quiz, the experimenter told them that, by with dishonesty, B ⫽ 0.48, SE ⫽ 0.084, ␤ ⫽ .51, t ⫽ 5.73, p ⬍
mistake, she had photocopied bubble sheets that already had the .001, whereas intelligence was not, B ⫽ 0.059, SE ⫽ 0.086, ␤ ⫽
correct answers lightly marked on them. She then asked the par- .062, t ⫽ 0.695, p ⫽ .49. These results provide support for our
ticipants to use these pre-marked bubble sheets, recycle the test hypothesis that dispositional creativity promotes dishonesty.
sheets with their original responses, and submit the bubble sheets
for payment. From these instructions, it was clear that participants Experiment 2: Creative Mindset and Dishonesty
could use the pre-marked bubble sheets to cheat when transferring
their responses. However, as in the previous task, we were able to In Experiment 2, we examine whether activating a creative
determine the extent of cheating by including a unique ID on both mindset temporarily promotes dishonest behavior in the same way
the test sheet and the bubble sheet. a creative personality does.
Pilot study. We recruited a nonoverlapping group of partic-
ipants (N ⫽ 46; Mage ⫽ 21.26 years, SD ⫽ 1.84) and conducted a 2
Although we use this task to assess cheating, choosing an answer that
pilot study to examine whether participants would perceive the ensures higher levels of payment could reflect motivated perception (Bal-
payment used in Experiment 1 as fair. In addition, we tested cetis & Dunning, 2006). That is, participants may perceptually disambig-
whether participants would consider cheating on the tasks em- uate what they see on the screen to serve their goal to earn more money.
ployed in Experiment 1 as unethical and morally wrong. We asked We acknowledge this limitation, which is in part alleviated in this study by
participants in this pilot study to follow the same procedure as in the use of multiple tasks to assess dishonest behavior.
3
the laboratory portion of Experiment 1 (without the online survey), For the visual perception task, we conducted further analyses to ex-
with only one difference. This time participants did not have the amine cheating on unambiguous trials. We found that it was highly
opportunity to cheat in any of the three tasks. We used the same correlated with cheating on ambiguous trials (r ⫽ .91, p ⬍ .001) and was
payment structure as in the laboratory portion of Experiment 1 moderately correlated with creative personality (Gough’s, 1979, Creative
Personality Scale: r ⫽ .27, p ⬍ .01; creative cognitive style: r ⫽ .19, p ⫽
(i.e., $2 show-up fee, plus additional payment based on perfor-
.069; Hocevar’s, 1980, Creative Behavior Inventory: r ⫽ .19, p ⫽ .065).
mance throughout the study). Upon completion of the study, we
However, cheating on unambiguous trials was not correlated with intelli-
asked them two questions. First, we asked them to evaluate gence (Cognitive Reflection Test score: r ⫽ .04, p ⫽ .72; Mill Hill
whether the payment structure used in the study was fair compared Vocabulary Scale score: r ⫽ –.02, p ⫽ .87).
to other studies they may have participated in (1 ⫽ not fair at all, 4
We also created z scores for participants’ actual performance on the
4 ⫽ fair, 7 ⫽ more than fair). Second, we asked them to imagine problem-solving task and the multiple-choice task and found that they were
having the opportunity to self-report performance in each of the not correlated with dispositional creativity (r ⫽ .15, p ⫽ .15; and r ⫽ –.06,
three tasks included in the study they had just completed and thus p ⫽ .59, respectively).
450 GINO AND ARIELY

Table 2
Descriptive Statistics and Correlations (Using Raw Scores for Each Measure), Experiment 1

Variable Measure M SD 1 2 3 4 5 6 7

Creative personality 1. Gough’s Creative Personality Scale 11.24 2.81

2. Creative cognitive style 3.74 0.75 .54ⴱⴱⴱ
3. Hocevar’s Creative Behavior Inventory 147.29 40.74 .45ⴱⴱⴱ .54ⴱⴱⴱ

Intelligence 4. Cognitive Reflection Test score 1.22 1.10 .003 ⫺.04 ⫺.04
5. Mill Hill Vocabulary Scale score 6.49 1.67 .001 ⫺.08 .02 .30ⴱⴱ

Tasks 6. Cheating level on problem-solving task 2.92 4.54 .53ⴱⴱⴱ .35ⴱⴱ .42ⴱⴱⴱ .04 .04
7. Cheating level on multiple-choice task 3.33 5.86 .31ⴱⴱ .25ⴱ .25ⴱ .02 ⫺.11 .62ⴱⴱ
8. Cheating level on perceptual taska 71.95 23.38 .33ⴱⴱ .23ⴱ .25ⴱ .10 .08 .34ⴱⴱ .26ⴱ

Note. The dark rectangle depicts the relationship between the creativity measures and dishonesty. The dark square depicts the relationship between the
intelligence measures and dishonesty.
a
Cheating on the perceptual task is calculated as the number of times that participants chose “more on the right” in ambiguous trials.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⬍ .001.

Method objects as performing atypical functions (i.e., the box of tacks can
be used as a stand; Duncker, 1945; Glucksberg & Weisberg, 1966;
Participants. One-hundred eleven undergraduate and gradu-
Maddux & Galinsky, 2009).
ate students from local universities in the Southeastern United
Problem-solving task. Next, participants completed the same
States (52 male; Mage ⫽ 23.27 years, SD ⫽ 3.32) participated in
problem-solving task with the 20 matrices used in Experiment 1.
the study for payment. They were paid a $4 show-up fee and could
We assessed cheating on this task by computing the difference
earn an additional $10 based on their reported performance.
Procedure. Participants engaged in three presumably unre- between participants’ self-reported and actual performance.
lated tasks: a creativity prime (our manipulation) followed by a Final questionnaire. Finally, participants filled out a ques-
2-min filler task, a creativity task (used as a manipulation check), tionnaire including demographic questions and postexperimental
and the matrix task employed in Experiment 1 (used to assess questions assessing their awareness of the priming (Bargh &
dishonest behavior). Chartrand, 2000). Participants were excluded from the data if they
Creativity prime. Participants were randomly assigned to indicated awareness of the priming (e.g., “Something to do with
one of two priming conditions: creative mindset versus control. creativity, originality, novelty”; “Words like creativity, originality
Previous research has successfully employed priming to activate a used often”) or the purpose of the experiment. No participant was
creative mindset (e.g., Fitzsimons et al., 2008; Sassenberg & disqualified under these exclusion criteria.
Moskowitz, 2005). We employed a scrambled sentence test—a
frequently used method (Bargh & Chartrand, 1999; Chartrand &
Bargh, 1996) for manipulating respondents’ momentary mindset.
All participants were asked to construct grammatically correct
four-word sentences (e.g., “The sky is blue”) from a set of five
randomly positioned words (e.g., sky, is, the, why, blue). For the
participants in the creative-mindset condition, 12 of the 20 sen-
tences included words related to creativity (creative, original,
inventiveness, novel, new, innovative, invention, creativity, inge-
nious, imagination, originality, and ideas), whereas for the partic-
ipants in the control condition, no words related to creativity were
included. This priming task was followed by a 2-min filler task to
distract participants.
Manipulation check. We measured creativity triggered by
the prime using the Duncker Candle Problem (Duncker, 1945).
Participants were shown a picture containing several objects on a
table: a candle, a pack of matches, and a box of tacks, all of which
were next to a cardboard wall (see Figure 2). Participants were
given 3 min “to figure out, using only the objects on the table, how
to attach the candle to the wall so that the candle burns properly
and does not drip wax on the table or the floor.” The correct
solution consists of emptying the box of tacks, tacking it to the
wall, and placing the candle inside, so that the box of tacks is used
as a candleholder. In this task, finding the correct solution is a
measure of insight creativity because it involves the ability to see Figure 2. Duncker Candle Problem, Experiment 2.
 CREATIVITY AND DISHONESTY 451

Results and Discussion As their next task, participants were asked to roll a six-sided die
anonymously (a task adapted from Shalvi et al., 2011). For this
Manipulation check. The percentage of participants who task, they would earn money based on the reported outcome.
correctly solved the candle task was larger in the creative-mindset Before the beginning of the study, the experimenter placed a
condition than in the control condition (47.3% vs. 26.8%), ␹2(1, plastic cup with a die in it on each desk. We introduced our second
N ⫽ 111) ⫽ 5.00, p ⬍ .05, suggesting that our priming manipu- manipulation in this task. In the low-justification condition, par-
lation was effective. ticipants were instructed to privately roll the die in the cup only
Cheating on the problem-solving task. The average number once (by shaking the cup) and then report their outcome on the
of matrices by which participants overstated their performance was collection slip they had received. In the high-justification condi-
greater in the creative-mindset condition (M ⫽ 2.71, SD ⫽ 3.15) tion, participants instead were asked to roll the die a first time and
than in the control condition (M ⫽ 1.09, SD ⫽ 1.98), t(109) ⫽ then to roll it again a few more times to make sure the die was
3.25, p ⬍ .01.5 Furthermore, the percentage of participants who legitimate. After making sure the die was legitimate, participants
overstated their performance was also higher (49% vs. 27%), ␹2(1, had to report the outcome of the first roll. In both conditions, the
N ⫽ 111) ⫽ 5.87, p ⬍ .05. bonus payment for this task varied with the outcome of the die roll:
Taken together, these results demonstrate that even when acti- Participants received $1 if the outcome was 1, $2 if it was 2, and
vated temporarily, a creative mindset promotes dishonesty. so on, up to $6 if the outcome of the die roll was 6. This task gave
participants the opportunity to lie by reporting an outcome higher
Experiment 3: The Power of Justifications than the one actually obtained on the first die roll.
Having the possibility to roll the die multiple times gives par-
So far, we have demonstrated that creativity promotes dishon- ticipants room to justify their potential lies (Shalvi et al., 2011).
esty when people face an ethical dilemma and are motivated to We used this justification manipulation to find evidence for the
behave unethically, both when creativity is measured as an indi- mechanism linking creativity and dishonest behavior through mod-
vidual difference and when it is temporarily activated through eration. We predicted that rolling the die multiple times would
priming. In Experiment 3, we investigate the psychological mech- moderate the relationship between a creative mindset and dishon-
anism that may explain this relationship. We proposed that cre- esty. Specifically, in the low-justification condition, we expected
ativity promotes dishonesty by increasing people’s ability to jus- greater lying in the creative-mindset condition than in the control
tify their unethical actions. condition. However, we expected this difference to become less
In Experiment 3, we test this hypothesis by manipulating a prominent in the high-justification condition because individuals
feature of the task participants completed so that they would have in a creative mindset already have the ability to generate reasons
more or less room to justify dishonesty. The less room the task they can use to rationalize their unethical behavior.
provides for justifying cheating, the more moral flexibility indi-
viduals will need if they are to behave dishonestly on that task Results and Discussion
without feeling too guilty about their actions. We expect creativity
Manipulation check. We used the number of correct re-
to be particularly conducive to cheating on those tasks because it
sponses on the RAT as our manipulation check to determine
enhances moral flexibility. By contrast, on tasks that provide room
whether participants primed to think creatively were more likely to
for justification by design, creativity may be less “beneficial” for
complete this creative task successfully. Participants in the
dishonesty because moral flexibility is not needed to justify cheat-
creative-mindset condition performed better on the RAT (M ⫽
ing.
8.41, SD ⫽ 2.92) compared to those in the control condition (M ⫽
6.91, SD ⫽ 2.94), t(143) ⫽ 3.07, p ⫽ .003. This result suggests
Method that our priming was effective.
Die roll outcome. A 2 (prime) ⫻ 2 (justification) between-
Participants. One-hundred forty-five individuals from a city
subjects analysis of variance using the self-reported outcome of the
in the Northeastern United States (57 male; 112 students; Mage ⫽
die roll as the dependent measure revealed a significant interaction,
22.41 years, SD ⫽ 2.81) participated in a series of unrelated
F(1, 141) ⫽ 4.05, p ⫽ .046, ␩2p ⫽ .03 (see Figure 3). For
studies for $20. Our study was the first one that the participants
participants in the low-justification condition, a creative mindset
completed.
led to higher self-reported outcomes on the die roll compared to a
Design and procedure. The study employed two between-
neutral mindset, F(1, 141) ⫽ 13.58, p ⬍ .001. However, for
subjects factors: prime (control vs. creative) and room for justifi-
participants in the high-justification condition, the difference in
cation (high vs. low). As their first task, participants completed the
reported outcomes between the creative-mindset and the neutral-
sentence scramble task used in Experiment 2 and were thus ex-
mindset condition did not reach significance, F(1, 141) ⫽ 1.49,
posed to either a creative prime or a control prime. Next, they
p ⫽ .225. This analysis also revealed a main effect for both the
completed the Remote Association Task (RAT; Mednick,
prime manipulation, F(1, 141) ⫽ 12.83, p ⬍ .001, ␩2p ⫽ .08, and
1962)—a measure commonly used to assess creativity by assess-
the justification manipulation, F(1, 141) ⫽ 5.50, p ⫽ .02, ␩2p ⫽
ing people’s ability to identify associations between words that are
.04.
normally associated. In this task, participants are asked to find a
word that is logically linked to all of three words provided. For
instance, “cold” is the common word linking the words “sore– 5
Participants’ actual performance on the problem-solving task did not
shoulder–seat.” Participants were given 5 min to solve 17 RAT differ across conditions (Mcreative-mindset ⫽ 7.31, SD ⫽ 2.05 vs. Mcontrol ⫽
items (see Appendix B). 7.20, SD ⫽ 2.24), t(109) ⬍ 1, p ⫽ .78.
452 GINO AND ARIELY

6 Narcissistic Personality Inventory developed by Ames, Rose, and

Cameron (2006). This is a forced-choice measure, which includes
5.5
items such as “I really like to be the center of attention,” and “I
Self-reported outcome of die roll

5 insist upon getting the respect that is due to me.” We included a

measure for narcissism because prior research has demonstrated
4.5 that people who think they are more creative than others also tend
4
One roll (low to be narcissists (Goncalo, Flynn, & Kim, 2010). Thus, it is
justification)
important to show that the link between dispositional creativity
3.5 Multiple rolls and dishonest behavior that we observed in Experiment 1 holds
(high justification)
even when controlling for narcissism. The survey also included the
3
RAT (the same 17 items as in Experiment 3). Second, differently
2.5 from Experiment 3, this study only included the justification
manipulation and not the priming one.
2 Consistent with the results of Experiment 3, we expected that
Neutral mindset Creative mindset
having room to justify potential lies would moderate the effects of
Figure 3. Self-reported outcome of die roll by condition, Experiment 3. dispositional creativity on dishonesty. In particular, we predicted
that the relationship between creativity as an individual difference
and dishonesty would be stronger in the low-justification condition
Together, these results provide evidence consistent with the than in the high-justification condition.
hypothesis that creativity promotes dishonesty by increasing indi-
viduals’ ability to generate reasons to justify their unethical be-
havior. Results and Discussion
Creative personality and creative performance. We first
Experiment 4: Justifications and Creative Personality examined whether participants who scored high on dispositional
The results of Experiment 3 suggest that moral flexibility ex- creativity also performed better on the RAT. As shown in Table 3,
plains the relationship between a creative mindset and increased this was in fact the case. Next, we conducted a regression analysis
dishonesty. In Experiment 4, we examine whether justifications using a composite z score for the measures of creative personality
play a similar role in explaining the link between a creative as predictor of RAT performance while controlling for narcissism.
personality and increased dishonesty. Consistent with the correlations reported in Table 3, we found that
a creative personality predicted creative performance on the RAT,
B ⫽ 1.95, SE ⫽ 0.41, ␤ ⫽ .37, t ⫽ 4.70, p ⬍ .001. Narcissism did
Method
not, B ⫽ 0.03, SE ⫽ 0.32, ␤ ⫽ .007, t ⫽ 0.095, p ⫽ .93.
Participants. One-hundred fifty-nine individuals from a city Die roll outcome. Next, we examined whether our justifica-
in the Southeastern United States (89 male; 111 students; Mage ⫽ tion manipulation moderated the relationship between disposi-
25.38 years, SD ⫽ 4.62) participated in the study. They were paid tional creativity and higher self-reported outcomes on the die roll
$4 for completing an online survey and could earn up to $6 during (which are suggestive of lying). We tested this hypothesis using
the laboratory component of the study (in addition to potential the moderated regression procedures recommended by Aiken and
earnings related to other studies occurring in the same session). West (1991). In our regression analysis, we controlled for narcis-
Participants were randomly assigned to one of two conditions: sism. As shown in Table 4, we found a significant interaction
high versus low justification. between the composite measure of a creativity personality and the
Procedure. The study employed the same procedure as in justification manipulation in predicting participants’ self-reported
Experiment 3, with two main differences. First, participants com- outcome on the die roll (␤ ⫽ –.20, p ⬍ .05). To interpret the form
pleted an online survey a week prior to the laboratory session. The of the interaction, we plotted the simple slopes for the relationship
survey included the same three measures assessing creativity em- between dispositional creativity and self-reported outcome on the
ployed in Experiment 1, and a measure assessing narcissism as a die roll at each level of our justification manipulation (see Figure
personality trait. We measured narcissism by using the 16-item 4). When participants rolled the die one time only (low-

Table 3
Descriptive Statistics and Correlations (on Raw Measures), Experiment 4

Measure M SD 1 2 3 4 5
ⴱⴱⴱ ⴱⴱⴱ
1. Gough’s Creative Personality Scale 2.18 1.62 .44 .33 .06 .19ⴱ
2. Creative cognitive style 5.01 1.14 .43ⴱⴱⴱ .23ⴱⴱ .45ⴱⴱⴱ
3. Hocevar’s Creative Behavior Inventory 147.01 50.35 .36ⴱⴱⴱ .22ⴱⴱ
4. Narcissism 4.30 3.74 .11
5. Performance on the RAT 8.25 4.13

Note. RAT ⫽ Remote Association Task.

ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⬍ .001.
 CREATIVITY AND DISHONESTY 453

Table 4
Regression Analysis Predicting Self-Reported Outcome on the Die Roll, Experiment 4

Measure B SE ␤ t

Narcissism ⫺0.026 0.107 ⫺.018 ⫺0.243

Justifications 0.568 0.205 .197 2.78ⴱⴱ
Creative personality 1.046 0.186 .561 5.61ⴱⴱⴱ
Creative Personality ⫻ Justifications ⫺0.539 0.265 ⫺.201 ⫺2.04ⴱ
R2 ⫽ .23
F(4, 154) ⫽ 11.31ⴱⴱⴱ
ⴱ ⴱⴱ ⴱⴱⴱ
p ⬍ .05. p ⬍ .01. p ⬍ .001.

justification condition), dispositional creativity was associated Design and procedure. We used the same design and pro-
with higher self-reported outcomes on the die roll (␤ ⫽ .56, p ⬍ cedure as in Experiment 2, with only one difference. In addition to
.001). When participants rolled the die multiple times (high- attending the session in the laboratory, participants completed an
justification condition), this association was still significant but online survey with the three measures of creative personality that
was not as strong (␤ ⫽ .27, p ⫽ .01). we used in Experiments 1 and 4.
These results suggest that a creative personality promotes dis-
honest behavior by increasing the ability of individuals to justify Results and Discussion
their (potential) unethical actions.
Means, standard deviations, and correlations for the three cre-
ativity variables appear in Table 5. We created individual z scores
Experiment 5: Doubling on Creativity
for each of these measures and then averaged them into an aggre-
Our first four studies demonstrated a robust relationship be- gate measure for dispositional creativity.
tween creativity and dishonest behavior, when creativity was both We tested whether a creative personality moderates the effect of
assessed as an individual difference and experimentally manipu- priming a creative mindset on both creative performance and on
lated. In addition, through moderation, we found that moral flex- dishonest behavior, following the moderated regression procedures
ibility explains this link. We designed a final study to test whether recommended by Aiken and West (1991). Table 6 displays the
dispositional creativity moderates the effect of priming a creative results of our regression analyses. As we predicted, there was a
mindset on dishonest behavior. We predicted that dispositional significant interaction between our prime manipulation and dispo-
creativity would moderate the impact of primed creativity on sitional creativity in predicting creative performance, B ⫽ –1.31,
dishonesty such that those who are more dispositionally creative SE ⫽ 0.61, Wald ⫽ 4.65, p ⫽ .031; the amount of cheating (i.e.,
are less influenced by the creativity prime compared to those who the percentage of participants who cheated), B ⫽ –1.12, SE ⫽
are less dispositionally creative. 0.56, Wald ⫽ 3.94, p ⫽ .047; and the extent of cheating (i.e., the
degree to which participants cheated), B ⫽ –1.67, SE ⫽ 0.84, t ⫽
Method –1.99, p ⫽ .049 (see Figure 5).
When participants scored low on (the aggregated measure of)
Participants. One-hundred eight students from local univer- dispositional creativity, a creative mindset was associated with
sities in the Southeastern United States (53 male; Mage ⫽ 21.94 higher levels of creative performance (␤ ⫽ .55, p ⫽ .001) and
years, SD ⫽ 3.33) participated in the study for pay. They were paid greater cheating (␤ ⫽ .52, p ⫽ .001 for the decision to cheat, and
a $2 show-up fee and were paid $4 for completing an online ␤ ⫽ .48, p ⫽ .003 for the extent of cheating, respectively). By
survey; they could earn an additional $10 based on their perfor- contrast, when individuals scored high on dispositional creativity,
mance in the study. a creative mindset was no longer associated with higher creative
performance (␤ ⫽ .05, p ⫽ .73) or with greater cheating (␤ ⫽ .03,
p ⫽ .87 for the decision to cheat, and ␤ ⫽ –.02, p ⫽ .89 for the
5 extent of cheating, respectively).
Self-reported outcome of die roll

These results show that dispositional creativity moderates the

4.5 relationship between priming a creative mindset and creative per-
4 formance as well as the relationship between priming a creative
One roll (low mindset and dishonest behavior.
3.5 justification)

3 Multiple rolls General Discussion

(high justification)
2.5 Over the last three decades, an increasing number of studies
have highlighted the importance of creativity for individuals, or-
2
Low creativity High creativity ganizations, and societies. The majority of this work has stressed
the potential and real benefits of creative thinking. For instance,
Figure 4. Simple slopes for self-reported outcome of die roll, Experiment 4. research has shown that creative products generate an average
454 GINO AND ARIELY

Table 5
Descriptive Statistics and Correlations for Variables Assessing Creativity (on Raw Measures),
Experiment 5

Measure M SD 1 2 3

1. Gough’s Creative Personality Scale 10.08 3.77 .46ⴱⴱⴱ .35ⴱⴱⴱ

2. Creative cognitive style 3.56 0.76 .33ⴱⴱ
3. Hocevar’s Creative Behavior Inventory 87.78 54.12
ⴱⴱ ⴱⴱⴱ
p ⬍ .01. p ⬍ .001.

return that is significantly higher than that of “common” products moderates the effects of activating a creative mindset on dishonest
(Horibe, 2001) and that investments in creativity and innovation behavior. Priming a creative mindset is the most beneficial means
positively impact organizational performance (Lev, 2004). Cre- of enhancing creative thinking for individuals who score low
ativity is also beneficial at the individual level, as it helps us (rather than high) on dispositional creativity.
manage our daily lives and find creative solutions to both ordinary
and difficult problems. Theoretical Contributions
This article casts a shadow on the widespread view that creativ-
ity always leads to “good.” In five studies, we demonstrated that The present research contributes to the creativity literature by
creativity might also produce negative effects by leading individ- offering new insights on the potential dark side of creative think-
uals to more frequently engage in dishonest behavior. An initial ing. Prior work has identified several variables that significantly
pilot study conducted in the field demonstrated that employees promote or inhibit creative performance and has argued for the
who are in positions that require creativity are more likely to be importance of enhancing these factors with the primary goal of
morally flexible and to engage in wrongdoing in the workplace. increasing creativity. Here, we highlight the potential unintended
Experiment 1 found a significant relationship between creative consequences of creativity. Greater creativity helps individuals
personality and dishonesty, and demonstrated that creativity is a solve difficult tasks across many domains, but creative sparks may
better predictor of dishonest behavior than intelligence. Experi- lead individuals to take unethical routes when searching for solu-
ment 2 showed that participants primed with a creative mindset tions to problems and tasks.
were more likely to cheat than were participants in a control Our work also contributes to research on moral psychology and
condition. Experiments 3 and 4 explored the mechanism explain- ethical decision making. Our findings are consistent with studies
ing this link and demonstrated that participants who were primed highlighting the importance of psychological factors in driving
to think creatively (Experiment 3) or who were highly creative people’s dishonesty. An emerging literature has begun to identify
(Experiment 4) were more likely to behave dishonestly because of when these often-subtle factors influence decisions to behave
their greater ability to justify their dishonest behavior. Finally, in unethically, both consciously and unconsciously (Chugh, Bazer-
Experiment 5, we both assessed creativity through an individual man, & Banaji, 2005; Gino & Pierce, 2009; Jordan & Monin,
difference measure and manipulated it experimentally through 2008; Mazar et al., 2008; Monin, Sawyer, & Marquez, 2008; Shu,
priming. This final study demonstrated that a creative personality Gino, & Bazerman, 2011; Tenbrunsel & Messick, 2004). Here, we

Table 6
Moderated Regression Analyses, Experiment 5

Decision to cheat on the problem-solving

Performance on the candle task task

Measure B SE Wald B SE Wald

ⴱⴱ
Mindset (1 ⫽ creative, 0 ⫽ neutral) 1.44 0.461 9.79 1.18 0.417 7.96ⴱⴱ
Creative personality 1.16 0.497 5.43ⴱ 0.93 0.432 4.62ⴱ
Creative Personality ⫻ Mindset ⫺1.31 0.606 4.65ⴱ ⫺1.12 0.563 3.94ⴱ
Nagelkerke R2 ⫽ .179 Nagelkerke R2 ⫽ .149

Extent of cheating

B SE ␤ t

Mindset (1 ⫽ creative, 0 ⫽ neutral) 1.53 0.64 .23 2.41ⴱ

Creative personality 1.24 0.62 .28 2.00ⴱ
Creative Personality ⫻ Mindset ⫺1.67 0.84 ⫺.28 ⫺1.99ⴱ
R2 ⫽ .09
F(3, 104) ⫽ 3.43ⴱ
ⴱ ⴱⴱ
p ⬍ .05. p ⬍ .01.
 CREATIVITY AND DISHONESTY 455

5 Limitations and Directions for Future Research

4 These contributions must be qualified in light of several impor-
Low
Extent of cheating

dispositional tant limitations of our research, which could be addressed by future

3 creativity work. First, in our studies we created situations in which partici-
High pants were tempted to cheat. We used tasks that likely triggered a
2 dispositional
creativity conflict between participants’ short-term desire to earn undeserved
1 money by cheating and their long-term desire to be ethical and thus
not cheat. People commonly use self-control resources to resolve
0 this type of conflict (Mead et al., 2009). Future research could
Neutral mindset Creative mindset
investigate whether creativity would lead people to act in ways that
Figure 5. Simple slopes for the extent of cheating, Experiment 5.
satisfy their selfish, short-term desires rather than their higher
goals when facing other types of self-control dilemmas, such as
eating a slice of chocolate cake when trying to lose weight. As in
the studies presented here, creativity may help individuals generate
extend this body of work by showing that greater creativity can a variety of reasons to justify such self-serving behaviors.
lead to greater dishonesty by increasing individuals’ ability to Future research could also further examine whether creativity
justify their immoral actions. influences individuals’ motivations to behave dishonestly as well
Our research speaks to existing work on elastic justification as their ability to justify self-serving actions. Although our work
(Hsee, 1995, 1996). This research finds that unjustifiable factors focused on the effects of the latter, one could examine whether
(those we wish to take into consideration but know we should not) dispositional creativity and creative primes also increase individ-
influence individuals’ judgments more when there is ambiguity in uals’ desire to cheat. For instance, creativity may lead people to
the justifiable factors (those we believe we should take into con- think of more and diverse ways they could benefit from the
sideration) than when there is not. Our studies demonstrate that monetary gains from cheating, thus making cheating itself more
this type of self-justification process occurs more easily when tempting. Our studies did not specifically test for a link between
people have a creative personality or when they are in a creative creativity and the desire to cheat, but they do provide evidence that
mindset. is inconsistent with this alternative explanation. Indeed, the results
The findings of the present research can be evaluated with of Experiments 3 and 4 demonstrate that creativity increases
respect to work on moral credentials. According to moral creden- individuals’ moral flexibility, thus increasing their ability to be-
tialing theory (Monin & Miller, 2001), when individuals become have dishonestly. Thanks to greater creativity, people have more
aware of their own moral deeds, they are more likely to act and diverse reasons to justify their own unethical behavior.
immorally on subsequent endeavors because they feel as if they Future work could also examine how creativity may lead to
earned “moral credits.” Similarly, being able to generate several different consequences when people face ethical dilemmas if they
are not tempted by the possibility of behaving dishonestly to earn
original justifications for one’s own unethical actions thanks to
more money. In our studies, we used tasks that would heighten
creativity may lead people to feel licensed to cheat.
participants’ desire to cheat. It is possible, however, that individ-
Finally, the present work plants a first step in a research
uals may have different goals when facing ethical dilemmas. For
domain that is highly relevant to the increasingly changing,
instance, they may be motivated to behave consistently with their
innovative, and competitive world of the 21st century. As
moral principles and standards. Depending on the salient goal
innovation has increased, this century already has weathered a
driving individual behavior, creativity may result in actions that
series of accounting scandals and the collapse of several billion-
are prosocial rather than unethical, or actions that show moral
dollar companies, resulting in dramatic changes to the business
courage. For instance, a negotiator may decide to disclose infor-
landscape. Similarly, over the last decade, we have increasingly mation that affects how much his counterpart thinks the issue
witnessed cases of academic dishonesty by both students and being negotiated is worth, even if hiding the information would
teachers, as well as scandals of scientific cheating. Dishonesty lead to a better deal for him. In this case, creativity would help the
and innovation are two of the topics most widely written about negotiators generate many different reasons why disclosing infor-
in the popular press. Yet, to date, the relationship between mation is the proper action to take. Examining such possibility
creativity and dishonest behavior has not been studied empiri- would further our understanding of the role of creativity in ethical
cally. We believe that understanding this relationship has sev- decision making.
eral important implications for education, business, and policy. Finally, future research could investigate the boundary condi-
We are often surprised to learn that successful and ingenious tions of the effects observed in our studies and examine how
decision makers in these contexts have crossed ethical bound- people and organizations can foster creativity and can benefit from
aries. The results from the current article indicate that, in fact, individuals’ creative sparks while avoiding unintended evil solu-
people who are creative or work in environments that promote tions. For instance, this research could manipulate the saliency of
creative thinking may be the most at risk when they face ethical ethical standards (e.g., by using an ethics code), the strength of
dilemmas. Our work offers the first empirical demonstration of ethical norms, or the identity of moral exemplars, and examine
the association between creativity and dishonesty as well as whether these factors could be effectively used to combat the
evidence of the important role of justifications in explaining potential dark consequences of creativity. Similarly, future work
this relationship. could study the effects of creativity in groups and could manipu-
456 GINO AND ARIELY

late the level of competitiveness or cooperation within a group to motivation and social cognition: Foundations of social behavior (Vol. 2,
examine whether promoting a climate of cooperation can reduce pp. 93–130). New York, NY: Guilford Press.
the effects of creativity on dishonest behavior. Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of
By calling attention to a previously underexplored relationship, being. American Psychologist, 54, 462– 479. doi:10.1037/0003-
that between creativity and dishonesty, our studies have uncovered 066X.54.7.462
Bargh, J. A., & Chartrand, T. L. (2000). The mind in the middle: A
findings of both theoretical and practical importance. Scholars
practical guide to priming and automaticity research. In H. T. Reis &
need more knowledge regarding both the positive and negative C. M. Judd (Eds.), Handbook of research methods in social and per-
consequences of thinking outside the box before fully embracing sonality psychology (pp. 253–285). New York, NY: Cambridge Univer-
the recommendation to stimulate creativity in organizations and sity Press.
society more broadly. We hope this research will stimulate future Barron, F. B., & Harrington, D. M. (1981). Creativity, intelligence, and
endeavors that can further our understanding of how the process of personality. Annual Review of Psychology, 32, 439 – 476. doi:10.1146/
self-serving justifications, triggered by creativity primes or a cre- annurev.ps.32.020181.002255
ative personality, can lead to dishonest behavior. Carson, S. H., Peterson, J. B., & Higgins, D. M. (2003). Decreased latent
inhibition is associated with increased creative achievement in high-
functioning individuals. Journal of Personality and Social Psychology,
Conclusions 85, 499 –506. doi:10.1037/0022-3514.85.3.499
Chartrand, T. L., & Bargh, J. A. (1996). Automatic activation of social
In the current studies, we found a robust relationship between information processing goals: Nonconscious priming reproduces effects
creativity and dishonesty. This research provides a critical first of explicit conscious instructions. Journal of Personality and Social
step toward understanding how creativity is associated with un- Psychology, 71, 464 – 478. doi:10.1037/0022-3514.71.3.464
ethical behavior—two often-discussed phenomena in our complex Chugh, D., Bazerman, M., & Banaji, M. (2005). Bounded ethicality as a
world. Across five studies, we demonstrated that both a creative psychological barrier to recognizing conflicts of interest. In D. Moore,
personality and an activated creative mindset promote individuals’ D. Cain, G. Loewenstein, & M. Bazerman (Eds.), Conflict of interest:
ability to justify their unethical actions. In turn, this increased Challenges and solutions in business, law, medicine, and public policy
ability to justify potential unethical actions promotes dishonesty. (pp. 74 –95). New York, NY: Cambridge University Press. doi:10.1017/
Our results suggest that there is a link between creativity and CBO9780511610332.006
rationalization. As Mazar et al. (2008) proposed, the ability of Duncker, K. (1945). On problem solving. Psychological Monographs,
58(5), Serial No. 270.
most people to behave dishonestly might be bounded by their
Feist, G. J. (1998). A meta-analysis of personality in scientific and artistic
ability to cheat and at the same time feel that they are behaving as
creativity. Personality and Social Psychology Review, 2, 290 –309. doi:
moral individuals. To the extent that creativity allows people to 10.1207/s15327957pspr0204_5
more easily behave dishonestly and rationalize this behavior, cre- Feist, G. J. (1999). Personality in scientific and artistic creativity. In R. J.
ativity might be a more general driver of this type of dishonesty Sternberg (Ed.), Handbook of creativity (pp. 273–296). New York, NY:
and play a useful role in understanding unethical behavior. Cambridge University Press.
Fitzsimons, G. M., Chartrand, T. L., & Fitzsimons, G. J. (2008). Automatic
effects of brand exposure on motivated behavior: How Apple makes you
References “think different.” Journal of Consumer Research, 35, 21–35. doi:
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and 10.1086/527269
interpreting interactions. Newbury Park, London: Sage. Flach, F. (1990). Disorders of the pathways involved in the creative
Amabile, T. M. (1979). Effects of external evaluation on artistic creativity. process. Creativity Research Journal, 3, 158 –165. doi:10.1080/
Journal of Personality and Social Psychology, 37, 221–233. doi: 10400419009534349
10.1037/0022-3514.37.2.221 Förster, J., Friedman, R., & Liberman, N. (2004). Temporal construal
Amabile, T. M. (1983). The social psychology of creativity: A componen- effects on abstract and concrete thinking: Consequences for insight and
tial conceptualization. Journal of Personality and Social Psychology, 45, creative cognition. Journal of Personality and Social Psychology, 87,
357–376. doi:10.1037/0022-3514.45.2.357 177–189. doi:10.1037/0022-3514.87.2.177
Amabile, T. M. (1988). A model of creativity and innovation in organiza- Frederick, S. (2005). Cognitive reflection and decision making. Journal of Eco-
tions. In B. M. Staw & L. L. Cummings (Eds.), Research in organiza- nomic Perspectives, 19(4), 25–42. doi:10.1257/089533005775196732
tional behavior (Vol. 10, pp. 123–167). Greenwich, CT: JAI Press. Friedman, R. S., & Förster, J. (2001). The effects of promotion and
Ames, D., Rose, P., & Anderson, C. P. (2006). The NPI-16 as a short prevention cues on creativity. Journal of Personality and Social Psy-
measure of narcissism. Journal of Research in Personality, 40, 440 – chology, 81, 1001–1013. doi:10.1037/0022-3514.81.6.1001
450. doi:10.1016/j.jrp.2005.03.002 Galinsky, A. D., & Moskowitz, G. B. (2000). Counterfactuals as behavioral
Ayal, S., & Gino, F. (2011). Honest rationales for dishonest behavior. In primes: Priming the simulation heuristic and considerations of alterna-
M. Mikulincer & P. R. Shaver (Eds.), The social psychology of morality: tives. Journal of Experimental Social Psychology, 36, 384 – 409. doi:
Exploring the causes of good and evil (pp. 149 –166). Washington, DC: 10.1006/jesp.1999.1409
American Psychological Association. Gino, F., Ayal, S., & Ariely, D. (2009). Contagion and differentiation in
Baer, J., Kaufman, J. C., & Gentile, C. A. (2004). Extension of the consensual unethical behavior: The effect of one bad apple on the barrel. Psycho-
assessment technique to nonparallel creative products. Creativity Research logical Science, 20, 393–398. doi:10.1111/j.1467-9280.2009.02306.x
Journal, 16, 113–117. doi:10.1207/s15326934crj1601_11 Gino, F., Norton, M., & Ariely, D. (2010). The counterfeit self: The
Balcetis, E., & Dunning, D. (2006). See what you want to see: Motivational deceptive costs of faking it. Psychological Science, 21, 712–720. doi:
influences on visual perception. Journal of Personality and Social Psy- 10.1177/0956797610366545
chology, 91, 612– 625. doi:10.1037/0022-3514.91.4.612 Gino, F., & Pierce, L. (2009). Dishonesty in the name of equity. Psycho-
Bargh, J. A. (1990). Auto-motives: Preconscious determinant of social logical Science, 20, 1153–1160. doi:10.1111/j.1467-9280.2009.02421.x
interaction. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of Glucksberg, S., & Weisberg, W. R. (1966). Verbal behavior and problem
 CREATIVITY AND DISHONESTY 457

solving: Effects of labeling in a functional fixedness problem. Journal of Implications of counterfactual structure for creative generation and
Experimental Psychology, 71, 659 – 664. doi:10.1037/h0023118 analytical problem solving. Personality and Social Psychology Bulletin,
Goldenberg, J., & Mazursky, D. (2001). Creativity templates in new 33, 312–324. doi:10.1177/0146167206296106
products. New York, NY: Cambridge University Press. Mazar, N., Amir, O., & Ariely, D. (2008). The dishonesty of honest people:
Goldenberg, J., Mazursky, D., & Solomon, S. (1999, September 3). Cre- A theory of self-concept maintenance. Journal of Marketing Research,
ative sparks. Science, 285, 1495–1496. doi:10.1126/science.285 45, 633– 644. doi:10.1509/jmkr.45.6.633
.5433.1495 McBarnet, D. (1988). Law, policy, and legal avoidance. Journal of Law &
Goncalo, J. A., Flynn, F. J., & Kim, S. H. (2010). Are two narcissists better Society, 15, 113–121. doi:10.2307/1410078
than one? The link between narcissism, perceived creativity, and cre- McBarnet, D., & Whelan, C. (1991). The elusive spirit of the law: For-
ative performance. Personality and Social Psychology Bulletin, 36, malism and the struggle for legal control. Modern Law Review, 54,
1484 –1495. doi:10.1177/0146167210385109 848 – 873. doi:10.1111/j.1468-2230.1991.tb01854.x
Gough, H. G. (1979). A Creative Personality Scale for the Adjective Check Mead, N. L., Baumeister, R. F., Gino, F., Schweitzer, M. E., & Ariely, D.
List. Journal of Personality and Social Psychology, 37, 1398 –1405. (2009). Too tired to tell the truth: Self-control resource depletion and
doi:10.1037/0022-3514.37.8.1398 dishonesty. Journal of Experimental Social Psychology, 45, 594 –597.
Guilford, J. P. (1968). Intelligence, creativity, and their educational im- doi:10.1016/j.jesp.2009.02.004
plications. New York, NY: Knapp. Mednick, S. A. (1962). The associative basis of the creative process.
Guilford, J. P. (1982). Cognitive psychology’s ambiguities: Some sug- Psychological Review, 69, 220 –232. doi:10.1037/h0048850
gested remedies. Psychological Review, 89, 48 –59. doi:10.1037/0033- Monin, B., & Miller, D. T. (2001). Moral credentials and the expression of
295X.89.1.48 prejudice. Journal of Personality and Social Psychology, 81, 33– 43.
Hetherington, E. M., & Feldman, S. E. (1964). College cheating as a doi:10.1037/0022-3514.81.1.33
function of subject and situational variables. Journal of Educational Monin, B., Sawyer, P., & Marquez, M. (2008). The rejection of moral
Psychology, 55, 212–218. doi:10.1037/h0045337 rebels: Resenting those who do the right thing. Journal of Personality
Hirt, E. R., McDonald, H. E., & Melton, R. J. (1996). Processing goals and and Social Psychology, 95, 76 –93. doi:10.1037/0022-3514.95.1.76
the affect-performance link: Mood as main effect or mood as input? In Mumford, M. D., & Gustafson, S. B. (1988). Creativity syndrome: Inte-
L. L. Martin & A. Tesser (Eds.), Striving and feeling: Interactions gration, application, and innovation. Psychological Bulletin, 103, 27– 43.
among goals, affect, and self-regulation (pp. 303–328). Hillsdale, NJ: doi:10.1037/0033-2909.103.1.27
Erlbaum. Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal
Hocevar, D. (1980). Intelligence, divergent thinking, and creativity. Intel- and contextual factors at work. Academy of Management Journal, 39,
ligence, 4, 25– 40. doi:10.1016/0160-2896(80)90004-5 607– 634. doi:10.2307/256657
Horibe, F. (2001). Creating the innovation culture: Leveraging visionaries, Plucker, J. A., & Renzulli, J. S. (1999). Psychometric approaches to the
dissenters, and other useful troublemakers in your organization. New study of human creativity. In R. J. Sternberg (Ed.), Handbook of cre-
York, NY: Wiley. ativity (pp. 35– 61). New York, NY: Cambridge University Press.
Hsee, C. K. (1995). Elastic justification: How tempting but task-irrelevant Raven, J., Raven, C., & Court, J. H. (1998). Manual for Raven’s Progres-
factors influence decisions. Organizational Behavior and Human Deci- sive Matrices and Vocabulary Scales. Oxford, England: Oxford Psy-
sion Processes, 62, 330 –337. doi:10.1006/obhd.1995.1054 chologists Press.
Hsee, C. K. (1996). Elastic justification: How unjustifiable factors influ- Runco, M. A. (2004). Creativity. Annual Review of Psychology, 55, 657–
ence judgments. Organizational Behavior and Human Decision Pro- 687. doi:10.1146/annurev.psych.55.090902.141502
cesses, 66, 122–129. doi:10.1006/obhd.1996.0043 Sassenberg, K., & Moskowitz, G. B. (2005). Don’t stereotype, think
Johnson, C. D., & Gormly, J. (1971). Achievement, sociability, and task different! Overcoming automatic stereotype activation by mindset prim-
importance in relation to academic cheating. Psychological Reports, 28, ing. Journal of Experimental Social Psychology, 41, 506 –514. doi:
302. doi:10.2466/pr0.1971.28.1.302 10.1016/j.jesp.2004.10.002
Jordan, A. H., & Monin, B. (2008). From sucker to saint: Moralization in Schaller, M. (2003). Ancestral environments and motivated social percep-
response to self-threat. Psychological Science, 19, 809 – 815. doi: tion: Goal-like blasts from the evolutionary past. In S. J. Spencer, S.
10.1111/j.1467-9280.2008.02161.x Fein, M. P. Zanna, & J. M. Olson (Eds.), Motivated social perception
Kaufman, J. C., Baer, J., Cole, J. C., & Sexton, J. D. (2008). A comparison (pp. 215–231). Mahwah, NJ: Erlbaum.
of expert and nonexpert raters using the consensual assessment tech- Schweitzer, M. E., & Hsee, C. K. (2002). Stretching the truth: Elastic
nique. Creativity Research Journal, 20, 171–178. doi:10.1080/ justification and motivated communication of uncertain information.
10400410802059929 The Journal of Risk and Uncertainty, 25, 185–201. doi:10.1023/A:
Kershner, J. R., & Ledger, G. (1985). Effect of sex, intelligence, and style 1020647814263
of thinking on creativity: A comparison of gifted and average IQ Shalvi, S., Dana, J., Handgraaf, M. J. J., & De Dreu, C. K. W. (2011).
children. Journal of Personality and Social Psychology, 48, 1033–1040. Justified ethicality: Observing desired counterfactuals modifies ethical
doi:10.1037/0022-3514.48.4.1033 perceptions and behavior. Organizational Behavior and Human Deci-
Kirton, M. (1976). Adaptors and innovators: A description and measure. sion Processes, 115, 181–190. doi:10.1016/j.obhdp.2011.02.001
Journal of Applied Psychology, 61, 622– 629. doi:10.1037/0021- Shu, L. L., Gino, F., & Bazerman, M. H. (2011). Dishonest deed, clear
9010.61.5.622 conscience: When cheating leads to moral disengagement and motivated
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bul- forgetting. Personality and Social Psychology Bulletin, 37, 330 –349.
letin, 108, 480 – 498. doi:10.1037/0033-2909.108.3.480 doi:10.1177/0146167211398138
Lev, B. (2004, June). Sharpening the intangible edge. Harvard Business Simonton, D. K. (2000). Creativity: Cognitive, personal, developmental,
Review, 109 –116. and social aspects. American Psychologist, 55, 151–158. doi:10.1037/
Maddux, W. W., & Galinsky, A. D. (2009). Cultural borders and mental 0003-066X.55.1.151
barriers: The relationship between living abroad and creativity. Journal Simonton, D. K. (2003). Scientific creativity as constrained stochastic behavior:
of Personality and Social Psychology, 96, 1047–1061. doi:10.1037/ The integration of product, person, and process perspectives. Psychological
a0014861 Bulletin, 129, 475–494. doi:10.1037/0033-2909.129.4.475
Markman, K. D., Lindberg, M. J., Kray, L. J., & Galinsky, A. D. (2007). Smith, S. M., Ward, T. B., & Finke, R. A. (1995). Cognitive processes in
458 GINO AND ARIELY

creative contexts. In S. M. Smith, T. B. Ward, & R. A. Finke (Eds.), The Sternberg, R. J. (2001). What is the common thread of creativity? Its
creative cognition approach (pp. 1–7). Cambridge, MA: MIT Press. dialectical relation to intelligence and wisdom. American Psychologist,
Snyder, M. L., Kleck, R. E., Strenta, A., & Mentzer, S. J. (1979). Avoid- 56, 360 –362. doi:10.1037/0003-066X.56.4.360
ance of the handicapped: An attributional ambiguity analysis. Journal of Tenbrunsel, A. E., & Messick, D. M. (2004). Ethical fading: The role of
Personality and Social Psychology, 37, 2297–2306. doi:10.1037/0022- self-deception in unethical behavior. Social Justice Research, 17, 223–
3514.37.12.2297 236. doi:10.1023/B:SORE.0000027411.35832.53
Spiro, R. J., & Jehng, J. (1990). Cognitive flexibility and hypertext: Theory Thompson, L. (2008). Making the team (3rd ed.). Upper Saddle River, NJ:
and technology for the non-linear and multidimensional traversal of Pearson Education.
complex subject matter. In D. Nix & R. Spiro (Eds.), Cognition, edu- Tierney, P., & Farmer, S. M. (2002). Creative self-efficacy: Its potential
cation, and multimedia (pp. 163–205). Hillsdale, NJ: Erlbaum. antecedents and relationship to creative performance. Academy of Man-
Sternberg, R. J. (Ed.). (1999a). Handbook of creativity. New York, NY: agement Journal, 45, 1137–1148. doi:10.2307/3069429
Cambridge University Press. Zhou, J., & Oldham, G. R. (2001). Enhancing creative performance:
Sternberg, R. J. (1999b). A propulsion model of creative contributions. Review of Effects of expected developmental assessment strategies and creative
General Psychology, 3, 83–100. doi:10.1037/1089-2680.3.2.83 personality. The Journal of Creative Behavior, 35, 151–167.

Appendix A

Scenarios Used in the Pilot Study

1. Steve is the Operations manager of a firm that produces 2. Dale is the Operations manager of a firm that produces health
pesticides and fertilizers for lawns and gardens. A certain toxic food. Their organic fruit beverage has 109 calories per serving. Dale
chemical is going to be banned in a year, and for this reason is knows people are sensitive to crossing the critical threshold of one
extremely cheap now. If Steve buys this chemical, produces, and hundred calories. He could decrease the serving size by 10%. The
distributes his product fast enough, he will be able to make a very label will say each serving has 98 calories, and the fine print will say
nice profit. If you were Steve, how likely is it you would use this each bottle contains 2.2 servings. If you were Dale, how likely is it you
chemical while it is still legal? would cut the serving size to avoid crossing the 100 threshold?

Appendix B

Remote Association Task Items Used in Experiments 3 and 4

The goal in this task is to find a word that is logically linked to “playing– credit–report” (i.e., playing card, credit card, report
all three of the words provided. For example: (1) Manners Round card).
Tennis ⫽ Table; (2) Playing Credit Report ⫽ Card. In the space below, try to solve as many items as you can. Work
Thus, the word “table” is a solution because it links the words as fast as you can without sacrificing accuracy. YOU HAVE FIVE
“manners–round–table” (i.e., table manners, round table, table MINUTES (i.e., 300 seconds) TO WORK ON THIS. Please do not
tennis). The word “card” is a solution because it links the words use any help other than your own knowledge.

(Appendices continue)
 CREATIVITY AND DISHONESTY 459

Table B1

Word 1 Word 2 Word 3 Solution

Blank White Lines Paper

Magic Plush Floor Carpet
Thread Pine Pain Needle
Stop Petty Sneak Thief
Envy Golf Beans Green
Chocolate Fortune Tin Cookie
Barrel Root Belly Beer
Broken Clear Eye Glass
Pure Blue Fall Water
Widow Bite Monkey Spider
Chamber Staff Box Music
Mouse Sharp Blue Cheese
Hall Car Swimming Pool
Square Cardboard Open Box
Ticket Shop Broker Pawn
High Book Sour Note
Gold Stool Tender Bar

Received January 11, 2011

Revision received October 5, 2011
Accepted October 11, 2011 䡲