Identifying and Profiling Scholastic Cheaters:
Their Personality, Cognitive Ability, and Motivation
Kevin M. Williams, Craig Nathanson, and Delroy L. Paulhus
University of British Columbia
Despite much research, skepticism remains over the possibility of profiling scholastic cheaters. However,
several relevant predictor variables and newer diagnostic tools have been overlooked. We remedy this deficit
with a series of three studies. Study 1 was a large-scale survey of a broad range of personality predictors of
self-reported cheating. Significant predictors included the Dark Triad (Machiavellianism, narcissism, psy-
chopathy) as well as low agreeableness and low conscientiousness. Only psychopathy remained significant in
a multiple regression. Study 2 replicated this pattern using a naturalistic, behavioral indicator of cheating,
namely, plagiarism as indexed by the Internet service Turn-It-In. Poor verbal ability was also an independent
predictor. Study 3 examined possible motivational mediators of the association between psychopathy and
cheating. Unrestrained achievement and moral inhibition were successful mediators whereas fear of punish-
ment was not. Practical implications for researchers and educators are discussed.
Keywords: antisocial, psychopathy, plagiarism, scholastic cheating
Student cheating remains a disconcerting problem for educators.
In a typical survey, two thirds of college students report having
cheated at some point during their schooling (e.g., Stern & Hav-
licek, 1986; Cizek, 1999). If anything, the problem appears to have
worsened in recent years (Josephson Institute of Ethics, 2008) with
lifetime cheating rates as high as 80% in some student samples
(Robinson, Amburgey, Swank, & Faulker, 2004). One contributor,
the escalating access to the Internet has greatly facilitated plagia-
rism—especially among computer-savvy students (Ma, Wan, &
Lu, 2008; Underwood & Szabo, 2003).
Some investigators argue that situational factors are paramount
in the explanation of cheating behavior (see Murdock, Miller, &
Goetzinger, 2007). Other researchers have sought to profile pre-
dispositions, that is, identify the best individual difference predic-
tors of cheating. To date, the predictors garnering the most support
are poor scholastic attitudes and poor academic preparation (e.g.,
Cizek, 1999; Whitley & Keith-Spiegel, 2002). Those same reviews
were pessimistic about the value of personality and cognitive
ability measures in predicting cheating.
The present report comprises three studies designed to challenge
that pessimism. Our challenge is based on two weaknesses in
previous research: First is the omission of key personality predic-
tors. Second is the failure to exploit objective measures of cheating
in a meaningful setting. Before detailing our research, a brief
review of that earlier research is warranted.
Measurement of Scholastic Cheating
In their taxonomy of academic dishonesty, Whitley and Keith-
Spiegel (2002) listed copying, plagiarism, facilitation, misrepre-
sentation, and sabotage. For each variety of cheating, specific
methods are in common use for measurement and detection. In
scholastic settings, there is a clear trend favoring the use of
high-tech to replace traditional low-tech methods (Cizek, 1999). In
research settings, a third category of cheating assessment may be
added to this list—laboratory cheating.
Each of these approaches
to cheating measurement involves advantages and disadvantages.
In the case of multiple-choice exams, college instructors have
traditionally employed low-tech methods such as direct observa-
tion of exam copying or passing answer keys. In the case of
plagiarism, instructors have often relied on their ability to detect a
familiar source or writing quality unlikely to have been generated
by the student. Such methods are rarely used in empirical research.
As new technologies such as the Internet, cellular phones, and
personal digital assistants (PDAs) became available, so too did
new methods for engaging in scholastic cheating. Fortunately, the
instructor’s arsenal of cheating detection methods has also bene-
fited from technological innovation. Of particular importance is
the availability of new computer software.
For example, software for the detection of multiple-choice an-
swer copying includes several commercially available programs;
others—notably, Signum (Harpp, Hogan, & Jennings, 1996) and
S-Check (Wesolowsky, 2000)—are freely available from their
authors. These programs conduct a pairwise comparison of stu-
dents’ responses to multiple-choice tests to search for excessive
overlap in the answer patterns. For each possible pair of students,
an index of similarity is calculated: Those with suspiciously high
overlap (i.e., those that are identified as obvious outliers among the
distribution of similarity scores) are flagged as potential cheating
pairs (Frary & Tideman, 1997; Harpp & Hogan, 1993; Harpp et al.,
1996; Wesolowsky, 2000). The validity of these methods is cor-
Although many such measures are available (Nicol & Paunonen,
2002), they will not be detailed here. This form of cheating is not scholastic
in the sense of being motivated by higher grades.
Kevin M. Williams, Craig Nathanson, and Delroy L. Paulhus, Depart-
ment of Psychology, University of British Columbia.
Correspondence concerning this article should be addressed to Delroy L.
Paulhus, Department of Psychology, 3519 Kenny Building, University of
British Columbia, Vancouver, BC, Canada V6T 1Z4. E-mail:
Journal of Experimental Psychology: Applied © 2010 American Psychological Association
2010, Vol. 16, No. 3, 293–307 1076-898X/10/$12.00 DOI: 10.1037/a0020773
roborated by the fact that flagged pairs of students are almost
invariably found to have been seated adjacent to each other
(Nathanson, Paulhus, & Williams, 2006).
For the detection of essay plagiarism, another category of an-
ticheating software is available. The most widely used program,
Turn-It-In, is accessed via a commercial website (iParadigms,
LLC, 2004). It is now the standard plagiarism screen for major
academic institutions across the globe (Dahl, 2007; Jones, 2008).
The program algorithm compares the text of a submitted paper
against the continually updated entries in its comprehensive data-
base. Items in this database range from previously submitted
student papers to academic and professional articles, as well as
current and previous Internet web pages. The program notes
strings of (at least) seven consecutive words that match previous
papers and calculates an overall percentage score of plagiarized
text. The output includes a copy of the essay with a different color
code for each source and the exact citation.
The Turn-It-In program operates on the same principle as the
more crude method of inserting essay text into an Internet search
engine such as Google (McCullough & Holmberg, 2005). In either
case, instructors are able to assess plagiarism rates more objec-
tively and efficiently than could be achieved with instructor judg-
ment alone. Drawbacks include the fact that some programs are
costly; others are complicated to use. Some, including Turn-It-In,
have triggered legal challenges.
Self-reports. To estimate cheating rates and their correlates in
large samples, the most efficient method is to collect self-reports
(e.g., Robinson et al., 2004; Underwood & Szabo, 2003). In the
same survey, one could inquire about a wide variety of cheating
behaviors. Questions could also cover a substantial time period
(e.g., how many times did you cheat during high-school?). Be-
cause cumulative self-reports are more reliable than single, or even
multiple behavioral measures, they have more statistical power for
evaluating individual difference correlates. This traditional survey
technique is also the least expensive and labor-intensive (Paulhus
& Vazire, 2007).
The obvious concern is the credibility of self-reports (Paulhus,
1991). Questions about a specific recent test (“Did you cheat on
the psychology midterm exam?”) are of dubious value because
students may fear repercussions for admitting the offense. Ques-
tions about past cheating (e.g., during high-school) may be of more
value because of the time interval and the lack of possible reper-
cussions. It is interesting that high values such as the 80% lifetime
cheating rate found by Robinson et al. (2004) suggest that impres-
sion management is not a serious concern in anonymous surveys of
Summary. Low-tech methods for cheating measurement have
rarely been used in research. Self-report and behavioral measures
are widely used but each has pros and cons. Depending on the
purpose of the research, either one may be appropriate. In the
research presented below, we exploited both methods.
Research on Demographic Predictors
Research on demographic predictors of cheating has also raised
complexities. One consistent finding is that men are more likely
than women to report having cheated (e.g., Jensen, Arnett, Feld-
man, & Cauffman, 2001; Lobel & Levanon, 1988; Newstead,
Franklyn-Stokes, & Armstead, 1996; Szabo & Underwood, 2004).
Yet concrete measures do not confirm such a sex difference
(Culwin, 2006; McCabe, Trevino, & Butterfield, 2001; Nathanson
et al., 2006; Whitley, Nelson, & Jones, 1999). It is unclear whether
this difference is the result of men overreporting their actual
cheating, women underreporting, or both.
Differences in cheating across college majors have been re-
ported in a handful of studies. Business students report higher rates
than nonbusiness students (McCabe, Butterfield, & Trevino,
2006). Students in science and engineering report higher levels of
cheating than those with arts majors (Marsden, Carroll, & Neill,
2005; Newstead et al., 1996). Given the higher rate of males in
science and engineering, however, it is not clear whether gender or
major is the ultimate source. Whitley and colleagues (1999) argue
that major is more important: In a meta-analysis, they showed that
women in science and engineering cheat virtually as much as their
Even fewer studies have examined cultural differences in scho-
lastic cheating. Hayes and Introna (2005) reported that, compared
to students from the United Kingdom, East Asian students held
more tolerant attitudes toward scholastic cheating. However,
Nathanson and colleagues (2006) found no behavioral differences
between East Asian and European students in behavioral indicators
of cheating. Altogether, then, the literature gives little indication of
demographic differences in actual cheating behavior.
Research on Personality Predictors
Comprehensive reviews of research on cheating predictors have
downplayed the value of personality predictors (Cizek, 1999;
Whitley & Keith-Spiegel, 2002). However, a number of personal-
ity variables have not yet been given sufficient attention. The
reason may simply be that standard measures of these variables
have only recently become widely used. Among the overlooked
variables are several with obvious relevance to cheating: narcis-
sism, psychopathy, and the Big Five personality dimensions of
Agreeableness, Conscientiousness, and Openness to Experience.
For possible inclusion in our research, we will address each in
The Dark Triad. The constructs of narcissism, Machiavel-
lianism, and psychopathy are commonly referred to as the Dark
Triad of personality (Paulhus & Williams, 2002). Narcissists are
characterized by grandiosity, entitlement, and a sense of superior-
ity over others (Raskin & Terry, 1988). Such individuals are
arrogant, self-centered, self-enhancing (Morf & Rhodewalt, 2001)
and ultimately, interpersonally aversive (Paulhus, 1998). Most
relevant to cheating, we suspect, is the sense of entitlement
(Emmons, 1987). Narcissists feel entitled to recognition for their
intellectual superiority even when their academic accomplish-
ments are mediocre. Therefore, attaining the plaudits they deserve
may require cheating.
Individuals high in Machiavellianism are characterized by cyn-
icism, amorality, and a belief in the utility of manipulating others
(Christie & Geis, 1970). A wealth of evidence confirms that these
individuals exploit a range of duplicitous tactics to achieve their
goals (see Jones & Paulhus, 2009; McHoskey, 2001). All these
tendencies increase the likelihood of indulging in scholastic cheat-
ing. However, the few studies exploring this possibility have
revealed only weak links, at best (Cizek, 1999; Flynn, Reichard, &
294 WILLIAMS, NATHANSON, AND PAULHUS
Psychopathy is characterized by the four key features of erratic
lifestyle, manipulation, callousness, and antisocial tendencies
(Hare, 2003; Williams, Paulhus, & Hare, 2007). All four suggest
that psychopaths are more likely to cheat than are nonpsychopaths.
Psychopathy is strongly and consistently associated with a wide
range of misconduct in nonoffenders (alcohol and drug abuse,
bullying, antiauthority abuse, driving offenses, criminal behavior;
Williams & Paulhus, 2004). We predict that this link will extend to
scholastic cheating. To date, only one study has investigated the
relation between psychopathy and a behavioral cheating behavior
(Nathanson et al., 2006): Cheating was predictable from two
self-report measures—the Self-Report Psychopathy scale
(Paulhus, Neumann, & Hare, in press) and the Psychopathic Per-
sonality Inventory (Lilienfeld & Andrews, 1996).
Note that our use of the term psychopathy does not imply
clinical or forensic levels. Accumulating research suggests that the
construct tapped by self-report psychopathy questionnaires is con-
ceptually identical to that tapped by interview methods in clinical/
forensic samples (Lebreton, Binning, & Adorno, 2005). Of course,
the mean scores in college students are substantially lower than
those in clinical/forensic samples (Forth, Brown, Hart, & Hare,
1996; Paulhus, Nathanson, & Williams, in press). Nonetheless,
roughly 3% still qualified for a clinical diagnosis of psychopathy.
Terms such as “nonoffender psychopathy” or “subclinical psy-
chopathy” are avoided in this paper, in order to minimize the
assumption that nonoffender/subclinical psychopathy is qualita-
tively different from clinical/forensic psychopathy.
The Big Five. The Big Five personality traits—Extraversion,
Agreeableness, Conscientiousness, Emotional Stability, and Open-
ness to Experience—are now widely viewed as the fundamental
dimensions of personality (Goldberg, 1994; Costa & McCrae,
1992). Extraversion is characterized as the tendency to be sociable,
talkative, energetic, and sensation-seeking. Agreeableness in-
volves cooperating with others, and maintaining harmony. Con-
scientiousness entails ambition, responsibility, and orderliness.
Emotionally stable individuals are anxiety-free, well-adjusted, and
resilient to stress. Finally, openness entails independent thinking,
along with esthetic and intellectual interests.
Given the consensus on their importance, it is surprising how
few studies of scholastic cheating have included the Big Five traits.
Of the five, only extraversion and stability (vs. neuroticism) have
received any attention. Those two factors were studied in depth by
Eysenck (e.g., Eysenck, 1970) long before the Big Five became
prominent as an organizational unit.
It is unfortunate the studies on extraversion and cheating have
yielded equivocal results. Cizek (1999) reported that, in three out
of four studies, extraversion showed a small significant positive
correlation with cheating. However, Jackson and colleagues re-
cently obtained a negative, albeit weak, association between ex-
traversion and cheating (Jackson, Levine, Furnham, & Burr, 2002).
Similarly, studies of stability have shown weak (though consis-
tently positive) correlations with cheating (Cizek, 1999; Jackson et
The three other Big Five factors have yet to be studied in the
context of cheating. Low agreeableness (i.e., disagreeableness)
seems especially relevant to cheating, given that its central features
include confrontation and lack of cooperation (Costa & McCrae,
1992). Current understanding of openness to experience suggests
no obvious association with cheating.
The Big Five variable with the closest conceptual connection to
cheating is (low) conscientiousness. This trait seems particularly
relevant given its contribution to academic preparedness, the
broader concept noted earlier. The published research is minimal
but some argue that dishonesty has clear conceptual links with
conscientiousness (e.g., Emler, 1999; Murphy, 2000). In a study
conducted before the Big Five labels became popular, Hethering-
ton and Feldman (1964) showed that students low in trait respon-
sibility were found to be more likely to cheat. Prudence, another
construct related to conscientiousness, has been linked (negatively)
to self-reported cheating (Kisamore, Stone, & Jawahar, 2007). A
wealth of research in industrial settings has shown that those
scoring low on conscientious-related traits engage in a persistent
pattern of dishonest behaviors such as theft, absenteeism, and
bogus claims of worker compensation (Hogan & Hogan, 1989).
Such behaviors may be seen as the workplace equivalent of aca-
Overview of the Present Research
Three studies were conducted to investigate possible links be-
tween scholastic cheating and the overlooked personality variables
noted above. Study 1 examined the role of three demographic and
eight personality predictors—including the Dark Triad and the Big
Five—in a large-scale survey of self-reported cheating behavior.
Study 2 sought to replicate these findings using a behavioral
indicator of plagiarism, namely, scores recorded by the Turn-It-In
program. Also included was a measure of verbal ability to control
for potential overlap between psychopathy and cognitive abilities.
In Study 3, motivational mechanisms underlying the personality-
misconduct link were evaluated via mediation analysis.
Study 1: Predictors of Scholastic Cheating
The primary goal of Study 1 was to fill in the above-mentioned
gaps in the research on personality correlates of scholastic cheat-
ing. Based on the literature reviewed above, scholastic cheating
should be associated with all of the Dark Triad variables (Hypoth-
esis 1.1) with psychopathy as the strongest predictor (Hypothesis
1.2). Of the Big Five, low scores on agreeableness and conscien-
tiousness should predict cheating (Hypothesis 1.3). Based on the
above research, the self-reported cheating rates should be higher in
male than in female students (Hypothesis 1.4) but no ethnic dif-
ferences are expected (Hypothesis 1.5).
Participants. Two-hundred and 49 students in second-year
undergraduate psychology classes at the University of British
Columbia participated in the study for course credit. 70% were
female; the majority of students were of either European (41.4%)
or East Asian (32.5%) ethnicity.
Measures and procedure. Participants enrolled by respond-
ing to an advertisement to participate in a study examining “Per-
sonality and Background Factors.” They picked up a take-home
questionnaire package that included several personality scales, as
well as a variety of misconduct scales embedded in a large-scale
survey. Instructions on the cover page cautioned against including
any personally identifying information (e.g., name, student num-
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS
ber). Given the sensitive nature of some of the questionnaire items,
this procedure was necessary to encourage honest and accurate
responses. Students returned their unmarked and sealed question-
naire package in a bin outside the lab; inside, they signed in to
receive their credit.
Personality questionnaires. Questionnaires were selected for
conceptual relevance and reputable psychometric properties, as
detailed below. Unless otherwise specified below, all items were
presented in Likert format: 1 ⫽“Strongly disagree” to 5 ⫽
Narcissism was assessed with the Narcissistic Personality In-
ventory (NPI; Raskin & Terry, 1988). The NPI contains 40 forced-
choice items such as “I like to be the center of attention.” versus
“I like to blend in with the crowd.” Currently considered the
standard measure of subclinical narcissism, the NPI has well-
established psychometric properties (Morf & Rhodewalt, 2001).
One point was assigned for each narcissistic response.
Machiavellianism was assessed with the 20-item Mach-IV
(Christie & Geis, 1970). Items include “Most people are basically
good and kind” and “It is hard to get ahead without cutting corners
here and there.” The Mach-IV is the most widely used measure of
Machiavellianism, and is psychometrically robust (for the latest
review, see Jones & Paulhus, 2009). In this dataset, items were
scored on a 6-point Likert scale ranging from ⫺3 (disagree
strongly) to ⫹3 (agree strongly).
Psychopathy was measured using the 64-item Self-Report Psy-
chopathy scale (SRP-III; Paulhus et al., in press). The SRP is
patterned after the Psychopathy Checklist-Revised (PCL-R; Hare,
2003), the current gold standard for assessing psychopathy in
forensic and clinical settings. The SRP has generated coherent
results in psychometric studies covering areas such as concurrent
and convergent/discriminant validity (Hicklin & Widiger, 2005),
including correlations with measures of general misconduct
(Camilleri, Quinsey, & Tapscott, 2009; Williams et al., 2007).
Example items include “I have attacked someone with the goal of
hurting them” and “I like to have sex with people I hardly know.”
Total scores on this measure tend to behave similarly to Lilienfeld
and Andrew’s (1996) Psychopathic Personality Inventory (e.g.,
Nathanson et al., 2006).
The 44-item Big Five Inventory (BFI; John & Srivastava, 1999)
was used to assess the Big Five factors of personality. Example
items (and the Big Five trait they assess) include “is talkative”
(extraversion), “is considerate and kind to almost everyone”
(agreeableness), “is a reliable worker” (conscientiousness), “re-
mains calm in tense situations” (stability), and “has an active
imagination” (openness). Substantial evidence has accumulated for
the validity of all five factors (John & Srivastava, 1999).
Scholastic cheating. The two items used to assess cheating
were: “I have cheated on school tests” and “I have handed in a
school essay that I copied from someone else.” Both specifically
referred to high-school to preclude concerns about admitting to
cheating at our university. A cheating index was calculated as the
mean of these two items. To preclude item overlap, the item with
similar content (“Only losers don’t cheat on tests”) was removed
from the psychopathy scale.
Results and Discussion
High-school cheating rates were estimated by coding any stu-
dent with a nonzero score on the two-item index as a cheater. A
substantial 73% of students admitted to cheating at least once in
high school. That value approximated the median of those values
cited in the literature (64%; Cizek, 1999). The reported rate for
plagiarism was 38.8%. Apparently, cheating tendencies among
college students continue at disturbingly high levels.
Demographics. Hypotheses regarding demographics were
supported. Consistent with Hypothesis 1.4, males reported higher
cheating rates than females: t(243) ⫽3.37, p⬍.01; d⫽.35. This
difference has been remarkably consistent across a spate of studies
(Cizek, 1999). Given that the pattern of personality analyses was
similar across gender, however, we only report results for the
pooled sample. Consistent with Hypothesis 1.5, no ethnic differ-
ences were found.
Personality predictors. For purposes other than estimating
high-school rates, self-reported cheating was indexed with a con-
tinuous measure—the mean of the two self-report items. The item
mean was 2.12 (SD ⫽.99) on a 5-point scale. Reliability of the
composite was .73. Analyses with the individual cheating items
showed similar though weaker patterns.
Note from Table 1 that the alpha reliability estimates for the
personality scales were sound, ranging from .78 to .89. Also
displayed in Table 1 are the correlations among the Big Five and
Study 1: Intercorrelations and Descriptive Statistics for Personality Measures and Self-Reported Cheating
1. Psychopathy (.89) .53
2. Machiavellianism (.78) .26
⫺.09 ⫺.01 .39
3. Narcissism (.87) .48
4. Extraversion (.86) .05 .21
5. Agreeableness (.81) .31
.16 .02 ⫺.23
6. Conscientiousness (.82) .25
7. Stability (.83) ⫺.02 .08 [.10]
8. Openness (.78) .07 [.08]
9. Self-reported cheating (.73)
Note. N ⫽228. Values in parentheses are alpha reliabilities. Values in square brackets are disattenuated for unreliability in the criterion.
Indicates significance at p⬍.05, two-tailed.
296 WILLIAMS, NATHANSON, AND PAULHUS
Dark Traid personality variables: The pattern is similar to that
found in previous studies (e.g., Hicklin & Widiger, 2005; Paulhus
& Williams, 2002; Williams & Paulhus, 2004).
Of special interest are the correlations of the personality
variables with scholastic cheating. Consistent with Hypothesis
1.1, each of the Dark Triad variables exhibited significant
positive associations with scholastic cheating. Consistent with
Hypothesis 1.2, psychopathy showed the strongest correlation
(.58) followed by Machiavellianism (.39) and narcissism (.20;
all p⬍.01). There are plausible mechanisms for each of these
Narcissists are known for their arrogance and sense of entitle-
ment (Emmons, 1987). Expecting to achieve more than others,
they often underperform (Wallace & Baumeister, 2002). Such
ego-threat can lead narcissists to behave in an antisocial fashion
(Twenge & Campbell, 2003). Cheating may be necessary to reaf-
firm their self-perceived superiority. As far as we know, there is no
previous evidence confirming this association empirically.
Given their manipulative tendencies, it is not surprising to find
that Machiavellian individuals have cheated in academic settings.
More surprising, however, is that this association has seen little to
no empirical support in previous research (Cizek, 1999; Flynn et
al., 1987). Those failures may be attributed in part to weakness in
methodology. For example, the Flynn et al. (1987) study used an
inferior measure of Machiavellianism, artificially dichotomized
students into high- and low-Machiavellian groups, and used a
contrived cheating measure. Our improved methodology may have
provided a more powerful test of the expected duplicity of Machi-
Regressions. Given the statistical overlap among the person-
ality constructs, multiple regression analysis was conducted to
determine the unique contribution of the relevant predictors. Re-
sults indicated that, after controlling for the other predictors, only
psychopathy, ␤⫽.50, t(220) ⫽6.71; p⬍.01, remained a
significant predictor of scholastic cheating.
The Big Five. Consistent with Hypothesis 1.3 were the
significant negative correlations with conscientiousness (⫺.28)
and agreeableness (⫺.23; all p⬍.01). Aspects of low consci-
entiousness such as irresponsibility, disorganization, and im-
pulsivity likely contribute to cheating behavior (Hogan &
Hogan, 1989). Because they end up less prepared and have
poorer study skills, they find themselves in desperate straits
(Hogan & Hogan, 1989).
The uncooperativeness inherent in disagreeableness presents a
plausible explanation for its association with cheating. Along with
conscientiousness, however, agreeableness lost significance in a
regression with the Dark Triad members. Presumably, the direct
relevance of the Dark Triad to antisocial behavior confers the
advantage to those three variables in predicting this narrow crite-
Unsuccessful cheating predictors. The remaining Big Five
predictors—emotional stability, extraversion, and openness—failed
to predict self-reported cheating. The results with extraversion and
emotional stability are consistent with the previous reviews (Cizek,
1999; Jackson et al., 2002; Whitley & Keith-Spiegel, 2002). The only
previous study of openness to experience also failed to produce
significant results (Nathanson et al., 2006).
Study 2: Objective Measurement of Scholastic
Our use of self-report measures in Study 1 was appropriate for
exploring new predictors in a large sample survey. However, the
limitations of self-report are well-known (e.g., Paulhus, 1991).
Although the maximization of anonymity helps minimize impres-
sion management, other response biases such as self-deception
may magnify associations via common method variance. Along
with sexual behavior, scholastic cheating constitutes a sensitive
self-disclosure that is vulnerable to underreporting or flat denial
among college students. Alternatives such as peer-evaluations
have considerable value in some measurement contexts (Paulhus
& Vazire, 2007), but are difficult to apply to scholastic cheating.
To provide a behavioral measure of essay plagiarism, the
Internet-based computer program Turn-It-In (iParadigms, LLC,
2004) was used in Study 2. As reviewed above, this program
compares a submitted paper against the constantly updated entries
in its extensive database. Items in the database range from previ-
ously submitted student papers to academic and professional arti-
cles, as well as current and previous Internet web pages. By
examining strings of consecutive words, each paper receives a
percentage score that indicates how much of the paper directly
matches sources in the databank.
One previous behavioral study reported a link between person-
ality and multiple-choice answer copying (Nathanson et al., 2006),
but personality predictors of plagiarism have yet to be studied.
Although both are forms of scholastic cheating, multiple-choice
answer copying and plagiarism may not have the same personality
correlates (Marsden et al., 2005). Consider, for example, that
multiple-choice answer copying is typically spontaneous and un-
planned whereas plagiarism is more deliberate and effortful. Ac-
cordingly, personality traits such as psychopathy and low consci-
entiousness (given their connection with poor impulse-control)
would be more relevant to answer-copying than to plagiarism.
Nonetheless, we hypothesize that psychopathy will again be the
principal predictor of plagiarism.
The role of cognitive ability. The association between (poor)
cognitive ability and cheating has been studied extensively. It
appears that students with poorer academic skills tend to cheat
more—perhaps to compensate for their shortcomings. It is worth
examining the evidence for this argument, which has recently been
summarized by three major reviews.
Whitley and Keith-Spiegel (2002) were pessimistic about any link
between cognitive ability and cheating but Cizek (1999) concluded
that there is a negative association. The most comprehensive review
was recently conducted by Paulhus, Nathanson, and Williams (in
press). Only behavioral indicators of cheating were considered but
measures of ability included various IQ tests, SAT scores, and other
aptitude tests. The results were quite consistent across 13 studies: in
every case, cheating rates were higher in students with lower cogni-
tive ability. The mean effect size was ⫺.26.
The possibility that psychopathic individuals have poorer cog-
nitive ability suggests an alternative explanation for their higher
cheating rates. Psychopaths may just be compensating for their low
ability. The empirical literature, however, does not support the
premise. In several studies, self-report psychopathy scores have
been found to be uncorrelated with measures of general intelli-
gence and knowledge in students (Nathanson et al., 2006; Paulhus
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS
& Williams, 2002), community members (Ishikawa, Raine, Lencz,
Bihrle, & Lacasse, 2001), and patient/offender samples (Crocker et
On occasion, psychopathy is occasionally found to be negatively
correlated with verbal intelligence (e.g., Nathanson et al., 2006).
Given the potential for overlap, it behooves us to disentangle their
roles with respect to scholastic cheating. Accordingly, we included
a measure of verbal intelligence in Study 2. Our decision to
measure verbal (as opposed to some other component of) intelli-
gence was its relevance to plagiarism.
Summary. Study 2 would extend Study 1 by turning to a
behavioral measure of scholastic cheating, namely, plagiarism
scores from Turn-It-In. Using the same set of personality predic-
tors as in Study 1, Study 2 also evaluated the contribution of
We expect the Turn-It-In plagiarism rate to be lower than that of
self-reported cheating (Hypothesis 2.1). Based on previous behav-
ioral research, there should be no sex differences (Hypothesis 2.2)
or ethnic differences in behavioral cheating rates (Hypothesis 2.3).
Among the personality variables, we predict that the Dark Triad
will be significant predictors of behavioral cheating (Hypothesis
2.4), with psychopathy as the strongest personality predictor (Hy-
pothesis 2.5). Based on Study 1, low agreeableness and low
conscientiousness should also predict plagiarism (Hypothesis 2.6).
Poor verbal ability will also be related to plagiarism independently
of psychopathy (Hypothesis 2.7).
Participants. We solicited participants from two sections of
introductory psychology that had an essay requirement. Of the 114
students enrolled in these sections, 107 agreed to participate in a
Seventy-two (67.3%) of them were female,
and the majority of students were of either East Asian (41.0%) or
European (43.0%) ethnicity.
Measures and procedure. Participants completed a battery of
personality scales through an Internet webpage: It prevented stu-
dents from reporting any personally identifying information (e.g.,
name, student number). Instead, students created a random 8-digit
student ID, which was used to obtain their credit at a predeter-
mined location upon completing the survey.
Personality and verbal ability scales. The personality ques-
tionnaires included on the webpage were identical to those used in
Study 1: the Self-Report Psychopathy Scale (SRP-III), Narcissistic
Personality Inventory (NPI), Mach-IV, and Big Five Inventory
(BFI). One minor difference is that, in contrast to Study 1, the
Mach IV scale responses were collected on 5-point (Disagree to
The verbal ability test was based on the Quick Word Test
(QWT; Borgatta & Corsini, 1964), a 100-item power vocabulary
test. In the past, the QWT has shown strong convergent validity
with other standard intelligence tests such as the Wechsler Adult
Intelligence Scales (see Bass, 1974). Internal consistency estimates
on the full test average .91. The QWT items were updated and the
revision, renamed the UBC Word test, has been normed and
validated (Nathanson & Paulhus, 2007). Each item is five letters in
length and respondents must select the best synonym from four
choices. Administration time was set to a maximum of eight
minutes. To control for variation in the number attempted, scores
were calculated as the ratio of correct answers to questions an-
Behavioral cheating measure. Plagiarism scores were based
on two essays assigned to the students by their course instructor.
The first paper required students to summarize a research project
whereas the second paper addressed a personal life experience.
Shortly before the essays were assigned, students were given an
essay outline that informed students that their papers would be
scrutinized by Turn-It-In. The outline also pointed students to
various university websites describing Turn-It-In, proper APA
format guidelines, and the definition of plagiarism.
As detailed earlier, Turn-It-In examines student essays for pla-
giarism by comparing each one to an extensive database of written
works. This process results in each paper receiving a percentage
score that indicates how much of the paper directly matches
sources in the databank. The output displays the percentage of
overlapping text, which is then categorized and color-coded based
on the original text source.
Because Turn-It-In also flagged legitimate overlapping text such
as quotes and citations, it was necessary to have research assistants
further scrutinize the results (this drawback has been rectified in
more recent versions of the program). Discounting instances of
legitimate overlap yielded a genuine proportion plagiarized (the
plagiarism index). Two research assistants showed 100% agree-
ment on the plagiarism index. Note that the Turn-It-In algorithm
has since been improved to discount legitimate text citations.
Results and Discussion
Reliability. Note from Table 2 that reliability estimates for the
personality scales ranged from .71 to .88. The reliability for UBC-
Word test was calculated with an odd-even estimate (.90): This
method is considered appropriate for a speeded test (Crocker &
Algina, 1986). A reliability estimate for the Turn-It-In index was
derived from the correlation of plagiarism scores from the two essays
(r⫽.41). The reliability estimate for the composite was .57.
Operationalizing cheating. Plagiarism was defined as any
nonzero percentage detected by Turn-It-In (after screening). The
mean plagiarism rate was 23% (SD ⫽22.5). A total of 16 students
(15.0%) plagiarized on at least one of their essays. To reduce
skewness, plagiarism scores were transformed into a dichotomous
variable. Students who plagiarized on at least one essay were
assigned a score of 1; all others were assigned a zero. Agreeement
between our two raters was 100%. Similar procedures have been
used previously to deal with the highly skewed distributions com-
mon in cheating studies (e.g., Daly & Horgan, 2007).
The resulting proportion of cheaters was 15% (SD ⫽38).
Consistent with Hypothesis 2.1, rates of Turn-It-In plagiarism
were much lower than the self-report cheating rates from Study 1.
Consistent with Hypotheses 2.2 and 2.3, plagiarism rates did not
differ according to ethnic background [European vs. East Asian;
However, students were not specifically advised that Turn-It-In results
would be used in our research. This procedure was critical to the study, and
was approved by the course instructor as well as the ethical review boards
at both the departmental and university level. The instructor was free to use
the Turn-It-In results to penalize students at her discretion, but did not.
Similar procedures have been used by other researchers of behavioral
plagiarism (Daly & Horgan, 2007).
298 WILLIAMS, NATHANSON, AND PAULHUS
t(72) ⫽⫺.44, p⬎.05; d⫽⫺.06] or gender: [t(105) ⫽⫺.24, p⬎
Predictors of cheating. Having established personality asso-
ciations in Study 1, we used one-tailed tests of significance for the
parallel Study 2 tests. Note from Table 2 that Dark Triad correlates
of Turn-It-In plagiarism were all significant, supporting Hypoth-
esis 2.4. The pattern of correlates was comparable to that with
self-reported scholastic cheating in Study 1. Supporting Hypothe-
sis 2.5, psychopathy was the strongest predictor. Hypothesis 2.6
was partially supported in that agreeableness but not conscien-
tiousness was a significant predictor.
Consistent with Hypothesis 2.7, low verbal ability was also a
significant predictor. To examine the possibility of poor verbal
ability as an alternative explanation for the psychopathy-
plagiarism link, partial correlations were conducted. Specifically,
the correlation between psychopathy and plagiarism was recalcu-
lated, controlling for verbal ability. This partial correlation (.21,
p⬍.01) was virtually identical to the original correlation (.22, p⬍
.01). The lack of a significant change may be traced to the fact that
psychopathy and verbal ability were almost completely orthogonal
(r⫽⫺.04, p⬎.05). This orthogonality of psychopathy and
cognitive ability is a consistent finding in both clinical samples
(e.g., Hare, 2003) and nonclinical samples (Paulhus & Williams,
Behavioral indicators. One major advance of Study 2 was the
use of a behavioral indicator of cheating. Nonetheless, personality
correlates of cheating were similar in Studies 1 and 2. This
consistency suggests that both methods detect cheating in mean-
ingful ways. The pattern of correlates echoed a previous study
using a behavioral indicator of multiple-choice answer copying
(Nathanson et al., 2006).
Together, Studies 1 and 2 have established a robust personality
predictor of scholastic cheating, thereby addressing the skepticism
of some commentators (e.g., Whitley & Keith-Spiegel, 2002). The
impact of psychopathy was strong and consistent across self-report
and behavioral assessments of scholastic cheating. An obvious
next step is to explore the psychological mechanisms by which the
psychopathy-scholastic cheating link operates. In Study 3, the
motivational mediators of this link are explored, in an attempt to
understand why psychopathic individuals engage in scholastic
Study 3: Psychological Mediators of Scholastic
To identify possible mediators, we searched the literature for
motivations students have offered for engaging in or avoiding
scholastic cheating (e.g., Anderman, Griesenger, & Westerfield,
1998; Cizek, 1999; Rettinger, Jordan, & Peschiera, 2004). Three
categories appear repeatedly in the literature. One is a motivation
for cheating, namely, unrestrained achievement motivation: That
is, some students strive to attain academic success without regard
to fairness. A common motivation reported for not cheating is fear
of punishment: Most students are concerned with repercussions
such as suspension or expulsion from school. Another deterrent to
cheating may be labeled moral inhibition: That is, students who
consider themselves honest and principled are less likely to cheat.
The first of these three may be considered an approach or incentive
motivation, whereas the final two are avoidance or deterrence
motivations (i.e., avoiding punishment and guilt, respectively).
There is reason to believe that all three of these motivations are
linked to psychopathy. First, unrestrained achievement maps onto
the unmitigated agency quadrant of the interpersonal circumplex
(i.e., high dominance and low nurturance)—the same quadrant that
houses psychopathy (Jones & Paulhus, 2010; Salekin, Trobst, &
Krioukova, 2001). Second, insensitivity to punishment was asso-
ciated with psychopathy as far back as the earliest laboratory
research (Hare, 1966). Finally, the impoverished moral identity in
psychopaths is also evident from the scientific literature (O’Kane,
Fawcett, & Blackburn, 1996; Trevethan & Walker, 1989). In short,
their links to both psychopathy and cheating suggest that all three
motivations (unrestrained achievement, fear of punishment, moral
Because the plagiarism scores were non-normal, we repeated the
analyses involving demographic variables and psychopathy with chi-
square tests of independence and Mann–Whitney tests, respectively. The
same results were obtained.
Study 2: Intercorrelations and Descriptive Statistics for Personality Measures and Plagiarism
1234 5 6 789 10
1. Psychopathy (.88) .49
.03 ⫺.04 ⫺.14 .22
2. Machiavellianism (.77) .23
⫺.08 ⫺.13 .01 .14
3. Narcissism (.81) .36
⫺.06 .19 .17 ⫺.10 .12
4. Extraversion (.88) .11 .13 .24
.19 ⫺.04 .08 [.11]
5. Agreeableness (.77) .22
6. Conscientiousness (.78) .22
.14 .05 ⫺.06 [⫺.08]
7. Stability (.80) .14 .21
8. Openness (.71) .38
9. Verbal ability (.90) ⫺.14
10. Turn-It-In plagiarism (.57)
Note. N ⫽107. Values in parentheses are alpha reliabilities. Values in square brackets are disattenuated for unreliability in the criterion.
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS
inhibition) are viable candidates to be mediators of psychopathic
cheating. Their role will be evaluated statistically via mediation
The present study. The primary goal of Study 3 was to
determine which of the three motivational factors (unrestrained
achievement, fear of punishment, moral inhibition) could explain
the psychopathy-cheating link. As a first step, a principal compo-
nents analysis was conducted to organize and simplify a wide
range of motivations for academic cheating.
Each of these motivational factors was then evaluated as a
psychological mediator using the most recent analytic methods
(see Chaplin, 2007). Mediation was determined to occur if the link
between psychopathy scores and cheating outcomes could be
explained by an indirect path via one of the motivations. Essen-
tially, the impact of the mediator corresponds to the product of (a)
the path between the predictor and the mediator and (b) the path
between the mediator and the outcome. Significance tests for
mediation were conducted using the bootstrap procedures devel-
oped by Shrout and Bolger (2002) and programmed by Preacher
and Hayes (2004).
In sum, we hypothesize that psychopathy will correlate signif-
icantly with each of the motivations for cheating (Hypothesis 3.1).
Each of the motivations for cheating will correlate significantly
with cheating (Hypothesis 3.2). Each motivation will provide
partial mediation of the link between psychopathy and cheating
Participants. Two-hundred and 23 students enrolled in under-
graduate psychology classes participated for course credit. One-
hundred and 41 (63.2%) were female, and the majority were of
either East Asian (44.4%) or European (28.3%) ethnicity. Because
gender and ethnic differences were minimal, the analyses were
based on the pooled sample.
Measures and procedure. The data collection procedure was
similar to Study 2. Students participated by responding to an
advertisement listed on the department’s Internet-based research
participation system. They completed a battery of personality
scales on a lab webpage. The procedures were designed to maxi-
mize anonymity by advising participants not to report any person-
ally identifying information (e.g., name, student number). Instead,
they selected a random 8-digit student ID, which was later used to
obtain a course credit of one percent.
Personality and cheating questionnaires. Unless otherwise
specified, all items are scored with a five-point Likert scale (1 ⫽
“Strongly disagree” to 5 ⫽“Strongly agree”). Again, the Self-
Report Psychopathy Scale (SRP-III; Paulhus et al., in press) was
used to assess psychopathy (alpha reliability ⫽.89).
Cheating behavior was measured with admission items from the
Self-Report Cheating Scale (Paulhus, Williams, & Nathanson,
2004). Twenty-six items assess misconduct behaviors such as
“Brought hidden notes to a school test” and “Copied someone
else’s answers on a school test without them knowing.” Eighteen
of the items specifically assess cheating behaviors, whereas the
remaining eight were fillers measuring general misconduct. When
combined to generate an overall self-report cheating score, the
alpha reliability of these 18 items was .85.
Potential mediators of cheating were measured using the moti-
vation items of the Self-Report Cheating Scale (Paulhus et al.,
2004). Based on results from previous studies and reviews (e.g.,
Cizek, 1999), 20 items were generated. Respondents were asked to
rate various factors that have influenced their decision to cheat (or
refuse to cheat) on previous academic tasks, or might influence
their decision to cheat (or refuse to cheat) in the future. Example
items include “I needed to do it to get (or keep) a scholarship,”
“I’m not concerned about the punishments involved if I am
caught,” and “I pride myself in being a good and trustworthy
Factoring the motivations for cheating. To structure a man-
ageable number of distinct motivations for cheating, a principal
components analysis was conducted on the 20 common motiva-
tions for cheating. Maximum Likelihood extraction generated sim-
ilar results, but Principal Axis Factoring results were not as clear.
Given the exploratory nature of this analysis, an oblique rotation
(direct oblimin) was used. The first four eigenvalues were 5.32,
1.91, 1.44, and 1.15. Parallel analysis indicated that a three-factor
solution was appropriate. We used the interpolation tables pro-
vided by Cota, Longman, Holden, Fekken, and Xinaris (1993):
The minimal value for a third eigenvalue was 1.41.
Fortunately, three factors were interpretable: They corresponded
substantially with the three common motivations for cheating
found in the literature. Following common PCA practice, items
with pattern matrix coefficients above .30 were retained. The
pattern matrix is displayed in Table 3. Seven items loaded above
.30 on the first factor: They were combined to form a subscale with
an alpha reliability of .71. These items concern the acceptability of
cheating to gain some academic goal, for example, high grades,
winning a scholarship, or receiving praise. Although most students
seek these goals, only a subset feel that cheating is an appropriate
strategy for obtaining these and other goals. It is this subset of
individuals who are of particular relevance in this context. Ac-
cordingly, the first factor was named “Unrestrained Achievement.”
Four items loaded at least .30 on the second factor and were
combined to form a composite score. The reliability of this sub-
scale was .51. High loading items dealt with concerns about
detection by professors and teaching assistants, and punishment
such as suspension or expulsion from the academic institute.
Accordingly, the second factor was labeled “Fear of Punishment.”
Nine items loaded at least .30 on the third factor and were
combined to form a composite score with an alpha reliability of
.54. This factor involves personal beliefs about one’s own charac-
ter and morals. Some students view themselves as honest individ-
uals who stick to their principles. Presumably, such individuals
would be less likely to engage in scholastic cheating. Conversely,
individuals who neither value these attributes nor feel they possess
them would be more likely to cheat. Other items referred to
excuses about their cheating behavior (e.g., test taking surround-
ings make it too easy to cheat). Example items include “I pride
myself in being a good and trustworthy person,” and “Being honest
and moral is not a high priority for me” (reverse-scored). Accord-
ingly, the third factor was named “Moral Inhibition.”
300 WILLIAMS, NATHANSON, AND PAULHUS
Associations among the three motivations were generally small.
The exception was a moderate negative correlation between Un-
restrained Achievement and Moral Inhibition (r⫽⫺.40, p⬍.01).
Intercorrelations. Table 4 presents the intercorrelations
among psychopathy, self-reported cheating and the potential me-
diators. Psychopathy correlated significantly with cheating (r⫽
.55; p⬍.01), after removing overlapping items. The significant
correlations of psychopathy with all three motivations supports
Hypothesis 3.1. Hypothesis 3.2 was partially supported in that
cheating correlated significantly with Moral Inhibition and Unre-
strained Achievement but not with Fear of Punishment.
Mediation analyses. Each of the cheating motivations was
evaluated as a potential mediator of the psychopathy-cheating link.
Using the bootstrap approach (Preacher & Hayes, 2004; Shrout &
Bolger, 2002), 5,000 samples were drawn. This method is consid-
ered more powerful than the traditional Sobel (1982) method,
given that the sampling distribution of the indirect effect is typi-
cally non-normal (Shrout & Bolger, 2002). It also allows for the
simultaneous evaluation of multiple mediators. The latter is im-
portant for our data because the mediators are intercorrelated.
Figure 1 displays the overall mediation model: The impact of
psychopathy on cheating can be seen to drop from .55 to .32 after
Study 3: Pattern Matrix Loadings From Principal Components Analysis of Motivations
Unrestrained achievement Fear of punishment Moral inhibition
Cheat to get a scholarship .73 .03 .26
Cheat to pass a course .71 ⫺.12 .09
Cheat because exam difficult .66 .08 ⫺.09
Cheat because of social pressure .65 .02 ⫺.05
Cheat to compete .64 ⫺.01 ⫺.18
Cheat to get a high grade .62 .07 ⫺.26
Fear of punishment
Punishment is severe .03 .72 ⫺.05
Too many TAs .19 .63 ⫺.16
Punishments are empty threats .38 ⴚ.39 ⫺.23
Not concerned about punishment .29 ⴚ.38 ⫺.18
Cheat because I can ⫺.04 ⫺.09 ⴚ.68
Cheat because no one will know .18 ⫺.17 ⴚ.64
Don’t cheat cause I’m a good person .21 ⫺.06 .58
Cheat because I’m not honest/moral ⫺.02 ⫺.37 ⴚ.57
Cheat without thinking .15 .15 ⴚ.53
Cheat because everyone does it .27 .12 ⴚ.51
Note.N⫽223. Factor extraction was followed by a direct oblimin rotation. Bold entries are the highest loading
for each item.
Study 3: Intercorrelations and Descriptive Statistics for
Psychopathy, Self-Reported Cheating and the
1. Psychopathy (.89) .23
Motivation for cheating
2. Unrestrained achievement (.71) ⫺.02 ⫺.40
3. Fear of Punishment (.51) .01 ⫺.10
4. Moral Inhibition (.54) ⫺.61
5. Self-reported cheating (.85)
Item mean 2.11 2.42 2.30 3.31 1.95
Standard deviation .42 .85 .62 .73 .54
Note.N⫽223. In the text, hypothesized correlations are couched as
one-tailed tests. All items were measured on 5-point scales.
Indicates significance at p⬍.01 (two-tailed).
Figure 1. Analysis of three mediatiors of the relation between psychop-
athy and cheating. All values represent standardized regression coefficients
(betas). The lower path indicates the total effect of psychopathy on cheat-
ing with the indirect effect in parentheses.
denotes statistical significance
at p⬍.01, two-tailed.
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS
the introduction of the three mediators. Analysis revealed with
95% confidence that the total indirect effect (i.e., the difference
between the total and direct effects) of psychopathy on cheating
was significant with a point estimate of .23 and a 95% confidence
interval of .11 to .35. Hence, the overall mediation was significant.
Nonetheless, the direct impact (.32) remained significant ( p⬍.05)
indicating that the three motivations provided only partial media-
tion of the association of psychopathy with cheating.
In support of Hypothesis 3.3, two of the motivations appeared to
be successful and unique mediators: (1) for unrestricted achieve-
ment, the 95% CI of .03 to .09 around the point estimate of .06 did
not include zero and (2) for moral inhibition, the 95% CI (.13; .33)
around the point estimate of .24 did not include zero. By contrast,
the 95% CI (-.22; .28) for the point estimate of fear of punishment
(.003) did include zero.
The impetus for the three studies reported here was the wide-
spread skepticism about the value of individual differences in
predicting scholastic cheating (Cizek, 1999; Whitley & Keith-
Spiegel, 2002). Those two reviews—fully comprehensive at the
time—were published before the advent of several highly relevant
Study 1 addressed this limitation by measuring the Big Five and
Dark Triad traits in a large-scale study of self-reported scholastic
cheating. Study 2 revealed a similar pattern using a behavioral
criterion and a control for intellectual ability. Although traits such
as Machiavellianism, narcissism, disagreeableness, and (low) con-
scientiousness showed some degree of association, psychopathy
was the strongest and most consistent predictor. Indeed, psychop-
athy stood out as a significant predictor in all three studies reported
here. Poor verbal ability also predicted cheating but did not ac-
count for the impact of psychopathy.
This robust link between psychopathy and scholastic cheating is
consistent with a body of research linking psychopathy to a broad
range of misconduct in both offenders and nonoffenders. In of-
fender samples, psychopathy is a notoriously strong correlate of
criminal behavior and recidivism (see Hare, 2003). In nonoffender
samples (e.g., students), psychopathy is typically measured via
self-report, but exhibits a similar pattern of results. For example,
Williams, Paulhus, Nathanson, and colleagues (Nathanson et al.,
2006; Williams & Paulhus, 2004; Williams et al., 2007) have
repeatedly demonstrated associations between psychopathy and a
wide range of misconduct indicators, including concrete behaviors.
This malevolent personality can be traced to an especially volatile
combination of manipulativeness, callous affect, erratic impulsive-
ness, and antisocial tendencies. Only subsets of this synergistic
combination are found in related constructs such as disagreeable-
These broader implications of the psychopathy-cheating link
parallel Blankenship and Whitley’s (2000) supposition about an
underlying cheating personality: This notion arose from their dem-
onstration that scholastic cheaters were also likely to engage in a
wide variety of antisocial behavior including drug use and vio-
lence. The present findings complement that research and further
promote the view of psychopathy as perhaps the single most
destructive personality syndrome. Furthermore, these results pro-
vide further evidence for the viability of psychopathy as a con-
struct with conceptual similarity (if not equivalence) in offender
and nonoffender samples (Lebreton et al., 2005).
Other Individual Difference Predictors of Cheating
Whereas psychopathy demonstrated strong and replicable asso-
ciations with cheating, other personality predictors were less ef-
fective. The identification of weak or null predictors also contrib-
utes to our understanding of cheating behavior. Weak or
moderated predictors require further study whereas consistently
null predictors can safely be excluded from further research.
Narcissism and Machiavellianism. Of the two remaining
Dark Triad constructs, Machiavellianism did show some associa-
tions with cheating—although they were fewer and weaker than
those with psychopathy. Although often predicted, the empirical
association of Machiavellianism with actual cheating behavior has
proved to be surprisingly weak (Christie & Geis, 1970; Cizek,
1999; Flynn et al., 1987). We found interesting that association
remained even after controlling for psychopathy, narcissism, con-
scientiousness, and agreeableness. Lacking the impulsive tendency
of psychopaths, Machiavellians may be more deliberate in their
mischief and more attentive to possible negative consequences
(Jones & Paulhus, 2009).
Finally, narcissism was the least successful predictor of cheating
among the Triad constructs. Regression analyses demonstrated that
any cheating behavior initially attributed to narcissism could be
explained by its overlap with psychopathy and Machiavellianism.
These results fit with previous research. For example, narcissists’
performance motivation is strongly influenced by ego involvement
(Wallace & Baumeister, 2002). Specifically, narcissists’ perfor-
mance motivation is enhanced if an opportunity for self-
enhancement—such as the publicizing of task results—presents
itself. A public posting of grades might have inspired narcissists to
cheat. Apparently, their sense of entitlement and need for recog-
nition was insufficient to provoke cheating in our studies.
The Big Five. We offered hypotheses regarding two of the
Big Five factors. One was that conscientious students would cheat
less. Although this hypothesis was in fact confirmed in Study 1,
the association disappeared when other predictors were included in
the regression equation. Even conscientiousness failed to work in
The rationale for the original hypothesis was that conscientious
students tend to be better prepared academically and, therefore,
have less need to cheat (Hogan & Hogan, 1989). Note however,
that conscientiousness also has a strong ambition component
(Costa & McCrae, 1998). This desire to excel may motivate some
conscientious individuals to cut corners, no matter how well-
prepared they are. In short, conscientiousness combines two com-
ponents that work in opposite directions: The result was a minimal
net effect on cheating. Future research should take advantage of
measures that disentangle these two components (e.g., Jackson,
Paunonen, Fraboni, & Goffin, 1996).
Similarly, initial associations between disagreeableness and
cheating were eliminated after accounting for overlap with the
Triad constructs and conscientiousness. These results are most
likely attributable to the sizable overlap between agreeableness
and the Triad constructs. Of the Big Five traits, only agreeableness
overlaps with each of the Triad, and typically to a substantial
degree (Paulhus & Williams, 2002). It appears that disagreeable-
302 WILLIAMS, NATHANSON, AND PAULHUS
ness alone is not sufficient: Only when it operates in combination
with other unsavory attributes (as in psychopathy) does cheating
occur. Finally, the openness, stability, and extraversion scales were
consistently unrelated to cheating behavior.
Low verbal ability. Another hypothesis concerned the asso-
ciation of poor verbal ability with cheating. Several reviews
(Cizek, 1999; Paulhus et al., in press) have concluded that the
ability-cheating link is a robust one (see also Daly & Horgan,
2007). The underlying principle is that students with poor cogni-
tive ability are less well prepared for tests and essays and therefore
choose to compensate by cheating. With respect to plagiarism, we
assumed that this link should be even clearer when verbal ability
is isolated from more global conceptions of cognitive ability.
The hypothesis was tested and confirmed in Study 2 with a
small but significant effect of verbal ability on plagiarism. We
believe that link between cheating and intelligence is indirect:
Cheating is a method for coping with perceived inadequacies. If
we are right, further research may show that those with poor math
ability are more likely to cheat on math tests. An appropriate
experimental manipulation may confirm this person x situation
Sex differences. The pattern of sex differences in cheating
found in Studies 1 and 3 mimicked those of previous research
(McCabe et al., 2001; Whitley et al., 1999): That is, self-reported
rates of cheating were higher in males than in females. However,
this sex difference disappeared in Study 2 when cheating was
measured in a more concrete fashion. To date, explanations for the
sex difference in self-reported cheating styles have been elusive
and largely speculative (Cizek, 1999). Results involving objective
cheating rates further support the notion that there is no real sex
difference in cheating. Given the confound with academic major,
however, our data cannot tease apart the contributions of gender
Explaining the Psychopathy-Cheating Link
Once the unique association between psychopathy and scholas-
tic cheating had been confirmed in Studies 1 and 2, we explored
the motivational mediators of this link in Study 3. Our mediation
analyses provided a means for quantifying the explanatory power
of these three motivations for cheating.
An unrestrained achievement motivation partially explained this
association. Incentives such as high grades and scholarships seem
to activate dishonesty in these individuals. Callous disregard for
others and lack of impulse control encourage cheating as a means
for achieving success. Indeed, such mechanisms are activated by
psychopathic individuals as methods for achieving all of their life
goals—academic or otherwise (Hare, 2003). It is notable that the
achievement goals shared by most college students trigger cheat-
ing in psychopaths alone.
Also confirmed was a second mediator of psychopathic cheat-
ing—a deficit in moral inhibition. The finding is consistent with
previous demonstrations of links between psychopathy and moral-
ity deficits (Williams et al., 2009). Even if temptations to cheat are
activated, most students avoid acting on them because it compro-
mises their self-image. As the final roadblock to cheating, this
moral identity may be seen as the ultimate deterrent (Acquino &
Douglas, 2003). Psychopaths, however, not only admit to such
deficits, they may well devalue society’s notion of integrity. In
sum, there are both internal (intrinsic) and external (incentive)
factors involved in the thought process underlying psychopathic
Our expectations about the mediating impact of a third motiva-
tion—fear of punishment—were not fulfilled. This failure can be
traced to a lack of association between fear and cheating. We
caution that this finding not be taken to suggest that fear of
punishment has no impact: Indeed, there is a wealth of evidence to
suggest that punishment does in fact deter students from cheating
(Cizek, 1999). Rather than fear per se, it may be that perceived
likelihood of being turned in by a peer is the psychological
mediator (McCabe et al., 2006).
Our use of a behavioral criterion in Study 2 addressed the
limitations of self-report methods, but may have also introduced
other limitations. For example, the eligibility requirements neces-
sary for students in Study 2 led to a relatively small sample size,
which hampered our ability to confirm significant associations. As
expected, the effect sizes (correlations with cheating) were lower
with a behavioral criterion compared to the self-report. For exam-
ple, the .58 correlation of psychopathy in Study 1 fell to .22 in
Study 2. Disattenuation of the criterion variables, however, helped
reduce this difference. Note that common self-report variance in
Study 1 suggests another possible explanation for the high corre-
lation: It may be that psychopaths are more willing to admit their
Another aspect of Study 2 impaired its power to find significant
correlates. The low frequency of plagiarists identified in this
dataset (i.e., roughly 7% to 15%) restricted the range in the
dependent variable and produced a highly skewed measure of
cheating (see Cohen, Cohen, West, & Aiken, 2002). These rates
may seem low compared to previous estimates based on self-report
(Newstead et al., 1996), which are upward of two thirds of students
(Robinson et al., 2004; Stern & Havlicek, 1986). Indeed, our own
self-report estimate in Study 1 was 73%.
Such self-report measures, however, cover a wider scope and
time: Ours, for example, asked whether the student had cheated at
any time in high school. Such self-reports often subsume all
varieties of cheating (e.g., answer copying, plagiarism, using hid-
den notes, etc.). In contrast, our Turn-It-In coverage was restricted
to two discrete opportunities to plagiarize essays in one university
course: Hence our rates—about 7%—represent typical rates of
cheating per opportunity (Lavin, 1965).
This fact highlights one of the trade-offs involved in using
Turn-It-In and similar programs. These programs capture natural-
istic cheating behavior, as opposed to other behavioral methodol-
ogies which, though typically inducing higher frequencies of
cheating, require contrived entrapment scenarios, or are otherwise
unrealistic (e.g., Hoff, 1940; Leveque & Walker, 1970). Further-
more, the essays used in the Study 2 course were designed to
minimally susceptible to cheating: Students were instructed to
write about personal experiences rather than a traditional literature
review or other essay style that could be plagiarized much more
readily. These instructions undoubtedly reduced rates of plagia-
rism even further. This handicap makes the cheating correlations
reported in this study conservative estimates. In that light, our
confirmation of significant associations is especially noteworthy.
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS
Causal inferences. In general, one may be more confident in
inferring causality if (a) the measurement of independent variables
temporally precedes that of dependent variables, (b) it is empiri-
cally demonstrated that changes in the independent variable lead to
changes in the dependent variable, on average, (c) changes in the
dependent variable do not lead to changes in the independent
variable, or (d) other potential causal variables are ruled out
(Cohen et al., 2002).
As with all correlational studies, causal inferences in the present
studies must be qualified. As a general rule, however, it is reason-
able to assume that the trait variables studied here temporally
precede other variables (see review by Bouchard & Loehlin,
2001). Moreover, the plagiarism measure was collected after the
personality measures were collected. It is difficult to argue that
cheating tendencies make people less conscientious or intelligent.
However, caution is warranted in conclusions about the medi-
ation analyses conducted in Study 3. The variables were not
collected in any distinct temporal order. Those mediators—
interpreted as motivations—could be construed as either justifica-
tions (which occurred after the cheating behavior) instead of
motivations (which occur before the cheating behavior).
However, justifications tend to reduce feelings of guilt. There-
fore, in the psychopathy-cheating context, it is unlikely that these
mediators are justifications because psychopathic individuals ex-
perience little guilt (Hare, 2003). Future studies that can establish
precise temporal measurement of these mediators may permit a
clearer explication of this dynamic process.
Future Directions and Recommendations
Behavioral indicators. Our findings have implications for re-
searchers of cheating behavior and educators in general. Both groups
can benefit from the use of concrete, objective criteria such as the
Turn-It-In program used here. Researchers are justifiably concerned
about the biases inherent in self-report measures—especially those
that assess socially undesirable behaviors such as cheating (Paulhus,
1991). Individuals who admit to cheating may also admit to undesir-
able personalities: Spurious correlations are the result. Software indi-
ces are more objective, unobtrusive, and can be used to capture
cheating at naturalistic rates in naturalistic settings.
Nonetheless, the similarity of the results obtained from self-
report (Studies 1 and 3) and computer-based criteria (Study 2)
suggests that both methods have their place as cheating indicators.
Such convergence and replication substantiates our claims about
the personality correlates of self-reported cheating.
Given the option, behavioral outcomes tend to be more convinc-
ing to many behavioral scientists. The success of our research with
programs such as Turn-It-In and S-Check suggests that behavioral
indicators of other forms of misconduct would be ideal in future
studies. However, the logistics of using such measures may prove
difficult, if not impossible, among nonoffender samples. Forensic
measures such as criminal records are unworkable, given that most
students and community members have no offenses. Obtaining
ethical approval and student consent for the use of such measures
would also be complicated. Some researchers have been creative in
their efforts to obtain behavioral indicators of misconduct, includ-
ing the collection of official university records and workplace
reports (e.g., Gustafson & Ritzer, 1995). Again, use of these
measures entails several trade-offs compared to self-report mea-
sures (e.g., sample size, time considerations).
Beyond the Big Five. In future research, we recommend the
exploration of several other individual difference variables. As
noted earlier, Lee and Ashton (2005) have recently expanded the
Five-Factor Model of personality to include a sixth factor—
Honesty-Humility (H-H)—as part of their HEXACO model of
fundamental personality traits. H-H captures characteristics such
as “sincerity, fairness, and modesty versus slyness, pretentious-
ness, and greed” (Lee & Ashton, 2005; p. 1573). H-H demon-
strates strong negative correlations with each of the Dark Triad
(Lee & Ashton, 2005) as well as self-reported scholastic cheating
(Marcus, Lee, & Ashton, 2007). Future research may determine
the independent contributions of H-H and the Triad constructs in
predicting scholastic cheating.
Implications for educators— contending with cheating.
Educators have to deal with cheating at both the abstract and
practical levels. First, they must continually revisit the meaning of
the construct as interpreted by students and test administrators
(Chambliss et al., 2010; Harris, 2001; Murdock & Stephens, 2007).
They are also on the front lines in contending with cheating and,
when it occurs, about documenting the offense (Whitley & Keith-
Spiegel, 2002). The present research supports the interpretation of
a high Turn-It-In score as cheating by linking it to individual
difference variables, namely, psychopathy and poor ability, which
have previously been linked with cheating. The use of such soft-
ware can help overcome some of the problems with traditional
techniques. When suspected for other reasons, confirmation of
plagiarism via computer software is an invaluable tool. In fact,
simply publicizing the fact that such techniques are in use should
reduce the prevalence of cheating on any given exam.
Effecting improvements in students’ cognitive ability and char-
acter is a more challenging goal: To the extent such changes are
even possible, they seem beyond the mandate of the typical edu-
cator. Psychopathic individuals are notoriously unresponsive to
treatment interventions applied by highly trained clinicians, and
sometimes become even more dangerous following treatment
(Rice, Harris, & Cormier, 1992). Instead, a preventative approach
to cheating is more likely to be fruitful. There is no shortage of
useful techniques for preventing cheating, such as alternate exam
forms, clear warnings about the use of cheating detection pro-
grams, banning cell-phones and other electronic devices, random
or assigned seating arrangements, and assigning essays that in-
volve writing about personal experiences that could not be easily
plagiarized from external sources (Cizek, 1999; Gulli, Kohler, &
Patriquin, 2007; Whitley & Keith-Spiegel, 2002).
More generally, educators should benefit from awareness that
the most probable cheaters are those low in scholastic prepared-
ness and high in psychopathy. Attention to the first group requires
redoubling efforts to prevent students from falling behind. Another
approach may be to reduce the degree of competitiveness among
the students. By creating an environment where relative achieve-
ment is de-emphasized, the disadvantaged students would feel less
threatened and less likely to resort to cheating. Such thinking is
hardly new among educators but it might help to acknowledge that
scholastic unpreparedness has its roots in basic traits.
Dealing with those high in psychopathy, on the other hand,
raises more fundamental pedagogical issues. The fact that cheating
is just one in their history of antisocial behaviors suggests that
304 WILLIAMS, NATHANSON, AND PAULHUS
psychopaths top the “most likely to be expelled” list. Yet early
diagnosis and surveillance of such individuals is problematic. It
seems unlikely that school boards and university senates would
approve of mass prescreening of students for psychopathy. Any
attempt to determine probability-of-expulsion in advance suggests
an unsavory “guilty until proven innocent” approach.
Even if prescreening were to be approved, there is no estab-
lished cutoff score for psychopathy in nonoffender populations.
Although some researchers have argued that psychopaths form a
distinct group in student samples (Harris, Rice, & Quinsey, 1994),
recent evidence has supported a normal distribution of psychopa-
thy scores—even among offenders (Edens, Marcus, Lilienfeld, &
Poythress, 2006; Lilienfeld & Andrews, 1996; Nathanson et al.,
2006). Either way, the diagnosis of psychopathy in a nonoffender
population is a comparatively more subjective endeavor than that
in a clinical or forensic context. Even if scores were kept confi-
dential, labeling could be extremely harmful to the student. The
surveillance of high scoring individuals would be highly problem-
atic ethically and practically. Indeed, it is possible that such labels
might translate into self-fulfilling prophecies. Furthermore, our
examination of potential mediators, combined with the results of
several forensic studies (see Hare, 2003), suggests that threats of
punishment are likely to go unheeded by psychopathic individuals.
On the whole, our character analysis suggests that the only way to
eliminate cheating among psychopaths is to make it impossible.
Overall, these cheating reduction strategies may be grouped into
two main categories: Altering teaching philosophy and modifying test
administration techniques. The former, which includes reducing the
competitive nature of the classroom environment, may be most ef-
fective for reducing cheating stemming from cognitive ability deficits.
The latter, which includes the use of alternate test forms, should be
most beneficial in eliminating cheating by psychopathic individuals.
Ideally, a combination of philosophical and methodological ap-
proaches may be most effective in abolishing cheating.
Our challenge to previous skepticism about profiling scholastic
justified cheaters appears to have paid off. This series of studies on
key personality variables eventuated in the isolation of subclinical
psychopathy as a powerful predictor. The replication of this asso-
ciation across three studies was essential for confirmation. The
association held up whether self-report or computer-scored behav-
ioral indices of cheating were used as operationalizations. Asso-
ciations also held up when the Big Five personality variables were
partialed out. Had we studied Machiavellianism, narcissism, agree-
ableness, conscientiousness or verbal ability on their own, each
would have yielded a significant link: The unique role of psychop-
athy would not have been so apparent.
In addition, our comparative analyses of the Turn-It-In scores
with self-report cheating provide mutual support for the validity of
each method Although behavioral and self-report measures both
have inadequacies, the converging pattern of correlates with psy-
chological variables raises confidence in both approaches.
Our conclusions may apply to misconduct in other nonoffender
samples. In the business world, for example, it may be that
psychopaths commit other acts of misconduct—such as fraud or
assault—in order to achieve goals such as promotions, wealth, or
power. Although many strive to attain such goals, psychopathic
individuals are most likely to believe that such devious and ag-
gressive tactics are acceptable as means to ambitious ends. It is
also possible that psychopathic individuals’ self-image as tough,
deceitful and callous explains their general tendencies toward
misconduct. Indeed the dynamics uncovered here may apply to all
psychopathic misconduct. A frank analysis may eventuate in suc-
cessful strategies for preventing, or reducing such behavior.
Anderman, E. M., Griesenger, T., & Westerfield, G. (1998). Motivation
and cheating during early adolescence. Journal of Educational Psychol-
ogy, 60, 84–93.
Aquino, K., & Douglas, S. (2003). Identity threat and antisocial behavior:
The moderating effects of individual differences, aggressive modeling,
and hierarchical status. Organizational Behavior and Human Decision
Processes, 90, 195–208.
Bass, B. (1974). Validity of Borgatta and Corsini’s Quick Word Test with
young adults. Psychological Reports, 34, 726.
Blankenship, K. L., & Whitley, B. E., Jr. (2000). Relation of general
deviance to academic dishonesty. Ethics and Behavior, 10, 1–12.
Borgatta, E. F., & Corsini, R. J. (1964). Quick Word Test manual. New
York: Harcourt, Brace, & World, Inc.
Bouchard, T. J., Jr., & Loehlin, J. C. (2001). Genes, personality, and
evolution. Behavior Genetics, 31, 243–273.
Camilleri, J. A., Quinsey, V. L., & Tapscott, J. L. (2009). Assessing the
propensity for sexual coaxing and coercion in relationships: Factor
structure, reliability, and vaidity of the Tactics to Obtain Sex Scale.
Archives of Sexual Behavior, 38, 959–973.
Chambliss, M., Bong, M., Greene, B., Kauffman, D., Loyens, S., & Van
Meter, P. (2010). Building trust by eliminating plagiarism: White paper
from the ad hoc committee on plagiarism. Contemporary Educational
Psychology, 35, 103–107.
Chaplin, W. F. (2007). Moderator and mediator models in personality
research: A basic introduction. In R. W. Robins, R. C. Fraley, & R. F.
Krueger (Eds.), Handbook of research methods in personality psychol-
ogy (pp. 602–632). New York: Guilford.
Christie, R., & Geis, F. L. (1970). Studies in Machiavellianism. New York:
Cizek, G. J. (1999). Cheating on tests: How to do it, detect it, and prevent
it. Mahwah, NJ: Erlbaum.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. A. (2002). Applied multiple
regression/correlation analysis for the behavioral sciences (3rd ed.).
Hillsdale, NJ: Erlbaum.
Costa, P. T., Jr., & McCrae, R. R. (1992). Revised NEO Personality
Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI):
Professional manual. Odessa, FL: PAR.
Costa, P. T., Jr., & McCrae, R. R. (1998). Six approaches to the explication
of facet-level traits: Examples from Conscientiousness. European Jour-
nal of Personality, 12, 117–134.
Cota, A. A., Longman, R. S., Holden, R. R., Fekken, C. G., & Xinaris, S. (1993).
Interpolating 95th percentile eigenvalues from random data: An empirical
example. Educational and Psychological Measurement, 53, 585–596.
Crocker, A. G., Mueser, K. T., Drake, R. E., Clark, R. E., McHugo, G. J.,
Ackerson, T. H., & Alterman, A. I. (2005). Antisocial personality,
psychopathy, and violence in persons with dual disorders: A longitudinal
analysis. Criminal Justice and Behavior, 32, 452–476.
Crocker, L. M., & Algina, J. (1986). Introduction to classical and modern
test theory. New York: Holt, Rinehart, & Winston.
Culwin, F. (2006). An active introduction to academic misconduct and the
measured demographics of misconduct. Assessment and Evaluation in
Higher Education, 31, 167–182.
Dahl, S. (2007). Turn-It-In: The student perspective on using plagiarism
detection software. Active Learning in Higher Education, 8, 173–191.
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS
Daly, C., & Horgan, J. M. (2007). Profiling the plagiarists: An examination
of the factors that lead students to cheat. Journal of Educational Com-
puting Research, 36, 39–50.
Edens, J. F., Marcus, D. K., Lilienfeld, S. O., & Poythress, N. G. (2006).
Psychopathic, not psychopath: Taxometric evidence for the dimensional
structure of psychopathy. Journal of Abnormal Psychology, 115, 131–144.
Emler, N. (1999). Moral character. In V. J. Derlaga, B. A. Winstead, &
W. H. Jones (Eds.), Personality: Contemporary theory and research (pp.
377–404). Chicago: Nelson-Hall.
Emmons, R. A. (1987). Narcissism: Theory and measurement. Journal of
Personality and Social Psychology, 52, 11–17.
Eysenck, H. J. (1970). The structure of human personality. London:
Flynn, S., Reichard, M., & Slane, S. (1987). Cheating as a function of task
outcome and Machiavellianism. Journal of Psychology, 121, 423–427.
Forth, A. E., Brown, S. L., Hart, S. D., & Hare, R. D. (1996). The
assessment of psychopathy in male and female noncriminals: Reliability
and validity. Personality and Individual Differences, 20, 531–543.
Frary, R. B., & Tideman, T. N. (1997). Comparison of two indices of
answer copying and development of a spliced index. Educational and
Psychological Measurement, 57, 20–32.
Goldberg, L. R. (1994). Basic research on personality structure: Implica-
tions of the emerging consensus for applications to selection and clas-
sification. In M. G. Rumsey, C. B. Walker, & J. H. Harris (Eds.)
Personnel selection and classification (pp. 247–259). Hillsdale, NJ:
Gulli, C., Kohler, N., & Patriquin, M. (2007). The great university cheating
scandal. Maclean’s, 120, 32–36.
Gustafson, S. B., & Ritzer, D. R. (1995). The dark side of normal: A
psychopathy-linked pattern called aberrant self-promotion. European
Journal of Personality, 9, 147–183.
Hare, R. D. (1966). Psychopathy and choice of immediate versus delayed
punishment. Journal of Abnormal Psychology, 71, 25–29.
Hare, R. D. (2003). The Hare Psychopathy Checklist – Revised (PCL-R)
(2nd ed.). Toronto: Multi-Health Systems.
Harpp, D. N., & Hogan, J. J. (1993). Crime in the classroom: Detection and
prevention on cheating on multiple-choice exams. Journal of Chemical
Education, 70, 306–311.
Harpp, D. N., Hogan, J. J., & Jennings, J. S. (1996). Crime in the classroom
part II. An update. Journal of Chemical Education, 73, 349–351.
Harris, R. (2001). The plagiarism handbook. Los Angeles: Pyrczak Pub-
Harris, G., Rice, M., & Quinsey, V. (1994). Psychopathy as a taxon:
Evidence that psychopaths are a discrete class. Journal of Consulting
and Clinical Psychology, 62, 387–397.
Hayes, N., & Introna, L. D. (2005). Cultural values, plagiarism, and
fairness: When plagiarism gets in the way of learning. Ethics and
Behavior, 15, 213–231.
Hetherington, E. M., & Feldman, S. E. (1964). College cheating as a
function of subject and situational variables. Journal of Educational
Psychology, 55, 212–218.
Hicklin, J., & Widiger, T. A. (2005). Similarities and differences among antisocial
and psychopathic self-report inventories from the perspective of general per-
sonality functioning. European Journal of Personality, 19, 325–342.
Hoff, A. G. (1940). A study of the honesty and accuracy found in pupil
checking of examination papers. Journal of Educational Research, 34,
Hogan, J., & Hogan, R. (1989). How to measure employee reliability.
Journal of Applied Psychology, 74, 273–279.
iParadigms, L. L. C. (2004). TurnItIn.com informational brochure. Oak-
land, CA: IParadigms, LLC.
Ishikawa, S. S., Raine, A., Lencz, T., Bihrle, S., & Lacasse, L. (2001).
Autonomic stress reactivity and executive functions in successful and
unsuccessful criminal psychopaths from the community. Journal of
Abnormal Psychology, 110, 423–432.
Jackson, C. J., Levine, S. Z., Furnham, A., & Burr, N. (2002). Predictors
of cheating behavior at a university: A lesson from the psychology of
work. Journal of Applied Social Psychology, 32, 1031–1048.
Jackson, D. N., Paunonen, S. V., Fraboni, M., & Goffin, R. D. (1996). A
five-factor versus six-factor model of personality structure. Personality
and Individual Differences, 20, 33–45.
Jensen, L. A., Arnett, J. J., Feldman, S. S., & Cauffman, E. (2001). It’s wrong,
but everybody does it: Academic dishonesty among high school and college
students. Contemporary Educational Psychology, 27, 209–228.
John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History,
measurement, and theoretical perspectives. In L. A. Pervin & O. P. John
(Eds.), Handbook of personality: Theory and research (2nd ed.,pp.
102–138). New York: Guilford Press.
Jones, D. N., & Paulhus, D. L. (2009). Machiavellianism. In M. R. Leary
& R. H. Hoyle (Eds.), Handbook of individual differences in social
behavior (pp. 93–108). New York: Guilford Press.
Jones, D. N., & Paulhus, D. L. (2010). Differentiating the Dark Triad
within the interpersonal circumplex. In L. M. Horowitz & S. N. Strack
(Eds.), Handbook of interpersonal theory and research. New York:
Jones, K. O. (2008, June). Practical issues for academics using the TurnItIn
plagiarism detection software. Proceedings of International Conference
on Computer Systems and Technology, Gabrovo, Bulgaria.
Josephson Institute of Ethics. (2008). 2008 Report card on the ethics of
American youth. Retrieved from http://www.josephsoninstitute.org/
Kisamore, J., Stone, T. H., & Jawahar, I. M. (2007). Academic integrity:
The relationship between individual and situational factors on miscon-
duct contemplations. Journal of Business Ethics, 75, 381–394.
Lavin, D. E. (1965). The prediction of academic performance: A theoret-
ical analysis and review of research. New York: Sage.
LeBreton, J. M., Binning, J. F., & Adorno, A. J. (2005). Subclinical
psychopaths. In D. L. Segal & J. C. Thomas (Eds.), Comprehensive
handbook of personality and psychopathology: Vol. 1. Personality and
everyday functioning (pp. 388–411). New York: Wiley.
Lee, K., & Ashton, M. C. (2005). Psychopathy, Machiavellianism, and nar-
cissism in the Five-Factor Model and the HEXACO model of personality
structure. Personality and Individual Differences, 38, 1571–1582.
Leveque, K. L., & Walker, R. E. (1970). Correlates of high school cheating
behavior. Psychology in the Schools, 7, 159–163.
Lilienfeld, S. O., & Andrews, B. P. (1996). Development and preliminary
validation of a self-report measure of psychopathic personality traits in
noncriminal populations. Journal of Personality Assessment, 66, 488–524.
Lobel, T. E., & Levanon, I. (1988). Self-esteem, need for approval, and
cheating behavior in children. Journal of Educational Psychology, 80,
Ma, H. J., Wan, G., & Lu, E. Y. (2008). Digital cheating and plagiarism in
schools. Theory into Practice, 47, 197–203.
Marcus, B., Lee, K., & Ashton, M. C. (2007). Personality dimensions
explaining relationships between integrity tests and counterproductive
behavior: Big Five, or one in addition? Personnel Psychology, 60, 1–34.
Marsden, H., Carroll, M., & Neill, J. T. (2005). Who cheats at university? A
self-report study of dishonest academic behaviours in a sample of Australian
university students. Australian Journal of Psychology, 57, 1–10.
McCabe, D. L., Butterfield, K. D., & Trevino, L. K. (2006). Academic
dishonesty in graduate business programs: Prevalence, causes, and pro-
posed action. Academy of Management Learning and Education, 5,
McCabe, D. L., Trevino, L. K., & Butterfield, K. D. (2001). Cheating in
academic institutions: A decade of research. Ethics and Behavior, 11,
McCullough, M., & Holmberg, M. (2005). Using the Google search engine
306 WILLIAMS, NATHANSON, AND PAULHUS
to detect word-for-word plagiarism in masters theses: A preliminary
study. College Student Journal, 39, 435–441.
McHoskey, J. W. (2001). Machiavellianism and sexuality: On the moder-
ating role of biological sex. Personality and Individual Differences, 31,
Morf, C. C., & Rhodewalt, F. (2001). Unraveling the paradoxes of narcis-
sism: A dynamic self-regulatory processing model. Psychological In-
quiry, 12, 177–196.
Murdock, T. B., Miller, A. D., & Goetzinger, A. (2007). Effects of
classroom context on university students’ judgments about cheating:
Mediating and moderating processes. Social Psychology of Education,
Murdock, T. B., & Stephens, J. B. (2007). Is cheating wrong? Student’s
reasoning about academic dishonesty. In E. A. Anderman & T. B.
Murdock (Eds.), The psychology of academic cheating. San Diego:
Murphy, K. R. (2000). What constructs underlie measures of honesty or
integrity? In R. D. Goffin & E. Helmes (Eds.), Problems and solutions
in human assessment: Honoring Douglas N. Jackson at seventy (pp.
265–283). New York: Kluwer Academic/Plenum Press.
Nathanson, C., & Paulhus, D. L. (2007, June). The UBC Word test. Poster
presented at the 68th annual meeting of the Canadian Psychological
Association, Ottawa, Canada.
Nathanson, C., Paulhus, D. L., & Williams, K. M. (2006). Predictors of a
behavioral measure of scholastic cheating: Personality and competence but
not demographics. Contemporary Educational Psychology, 31, 97–122.
Newstead, S. E., Franklyn-Stokes, A., & Armstead, P. (1996). Individual
differences in student cheating. Journal of Educational Psychology, 88,
Nicol, A. A. M., & Paunonen, S. V. (2002). Validity evidence for the
different item styles of overt honesty measures. Journal of Business and
Psychology, 16, 431–445.
O’Kane, A., Fawcett, D., & Blackburn, R. (1996). Psychopathy and moral
reasoning: Comparison of two classifications. Personality & Individual
Differences, 20, 505–514.
Paulhus, D. L. (1991). Measurement and control of response bias. In J. P.
Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of
personality and social psychological attitudes (pp. 17–59). San Diego:
Paulhus, D. L. (1998). Interpersonal and intrapsychic adaptiveness of trait
self-enhancement: A mixed blessing? Journal of Personality and Social
Psychology, 74, 1197–1208.
Paulhus, D. L., Nathanson, C., & Williams, K. M. (in press). A new look
at the link between cognitive ability and exam cheating. In B. D. Zumbo
(Ed.), Developments in the theories and applications of measurement,
evaluation, and research methodologies across the disciplines (Vol. 1).
Vancouver: University of British Columbia Press.
Paulhus, D. L., Neumann, C. S., & Hare, R. D. (in press). Manual for the
Self-Report Psychopathy scale (SRP). Toronto: Multi-Health Systems.
Paulhus, D. L., & Vazire, S. (2007). The self-report method. In R. W.
Robins, R. C. Fraley, & R. F. Krueger (Eds.), Handbook of research
methods in personality psychology (pp. 224–239). New York: Guilford
Paulhus, D. L., & Williams, K. M. (2002). The Dark Triad of personality:
Narcissism, Machiavellianism, and psychopathy. Journal of Research in
Personality, 36, 556–563.
Paulhus, D. L., Williams, K. M., & Nathanson, C. (2004). The Self-Report
Cheating Scale. Unpublished instrument, University of British Columbia.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for
estimating indirect effects in simple mediation models. Behavior Re-
search Methods, Instruments & Computers, 36, 717–773.
Raskin, R., & Terry, H. (1988). A principal components analysis of the
Narcissistic Personality Inventory and further evidence of its construct
validity. Journal of Personality and Social Psychology, 54, 890–902.
Rettinger, D. A., Jordan, A. E., & Peschiera, F. (2004). Evaluating the
motivations of other students to cheat: A vignette experiment. Research
in Higher Education, 45, 873–890.
Rice, M. E., Harris, G. T., & Cormier, C. A. (1992). An evaluation of a
maximum-security therapeutic community for psychopaths and other
mentally disordered offenders. Law and Human Behavior, 16, 399–412.
Robinson, E., Amburgey, R., Swank, E., & Faulker, C. (2004). Test
cheating in a rural college: Studying the importance of individual and
situational factors. College Student Journal, 38, 380–395.
Salekin, R. T., Trobst, K. K., & Krioukova, M. (2001). Construct validity
of psychopathy in a community sample: A nomological net approach.
Journal of Personality Disorders, 15, 425–441.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-
experimental studies: New procedures and recommendations. Psycho-
logical Methods, 7, 422–445.
Sobel, M. E. (1982). Asymptoptic confidence intervals for indirect effects
in structural equation models. In S. Leinhardt (Ed.), Sociological meth-
odology (pp. 290–312). San Francisco: Jossey-Bass.
Stern, E. B., & Havlicek, L. (1986). Academic misconduct: Results of
faculty and undergraduate student surveys. Journal of Allied Health, 15,
Szabo, A., & Underwood, J. (2004). Cybercheats: Is information and
communication technology fuelling academic dishonesty? Active Learn-
ing in Higher Education, 5, 180–199.
Trevethan, S. D., & Walker, L. J. (1989). Hypothetical versus real-life
moral reasoning among psychopathic and delinquent youth. Develop-
ment & Psychopathology, 1, 91–103.
Twenge, J. M., & Campbell, W. K. (2003). ‘Isn’t it fun to get the respect
that we’re going to deserve?’ Narcissism, social rejection, and aggres-
sion. Personality and Social Psychology Bulletin, 29, 261–272.
Underwood, J., & Szabo, A. (2003). Academic offences and e-learning:
Individual propensities in cheating. British Journal of Educational Tech-
nology, 34, 467–477.
Wallace, H. M., & Baumeister, R. F. (2002). The performance of narcis-
sists rises and falls with perceived opportunity for glory. Journal of
Personality and Social Psychology, 82, 819–834.
Wesolowsky, G. O. (2000). Detecting excessive similarity in answers on
multiple choice exams. Journal of Applied Statistics, 27, 909–921.
Whitley, B. E., Jr., & Keith-Spiegel, P. (2002). Academic dishonesty: An
educator’s guide. Mahwah, NJ: Erlbaum.
Whitley, B. E., Jr., Nelson, A. B., & Jones, C. J. (1999). Gender differences
in cheating attitudes and classroom cheating behavior: A meta-analysis.
Sex Roles, 41, 657–680.
Williams, K. M., Cooper, B. S., Howell, T. M., Yuille, J. C., & Paulhus,
D. L. (2009). Inferring sexually deviant behavior from corresponding
fantasies: The role of personality and pornography consumption. Crim-
inal Justice and Behavior, 36, 130–137.
Williams, K. M., & Paulhus, D. L. (2004). Factor structure of the Self-
Report Psychopathy scale (SRP-II) in non-forensic samples. Personality
and Individual Differences, 37, 765–778.
Williams, K. M., Paulhus, D. L., & Hare, R. D. (2007). Capturing the
four-factor structure of psychopathy in college students via self-report
Journal of Personality Assessment, 88, 205–219.
Received August 29, 2007
Revision received June 28, 2010
Accepted June 29, 2010 䡲
IDENTIFYING AND PROFILING SCHOLASTIC CHEATERS