Investigating Academic Dishonesty with Concrete Measures

Our recent research revealed two predictors of multiple-choice exam cheating: subclinical psychopathy and poor verbal ability. However, the extent to which individual difference variables predict plagiarism on term papers remains untested. Such research is now possible using the widely used plagiarism detector called Turn-It-In. We collected such objective measures of both plagiarism and multiplechoice cheating in a large undergraduate class (N = 328). The Turn-It-In program identified 20 percent of students as plagiarists. Results confirmed and extended our previous findings: Psychopathy and poor verbal ability predicted both multiplechoice cheating and plagiarism. Interestingly, they also predicted excessive quoting on papers. Recommendations for researchers and educators are discussed. Introduction Academic dishonesty remains a persistent problem for educators, with lifetime cheating rates as high as 80% (Robinson, Amburgey, Swank, & Faulker, 2004). However, recent statistical and technological advances have facilitated the detection of common forms of academic dishonesty. For example, the S-Check program assesses cheating on multiple-choice exams (e.g., Wesolowsky, 2000) and the Turn-It-In program assesses plagiarism on term papers (see Turnitin.com website). Despite claims that individual difference predictors are of little value in explaining academic dishonesty (e.g., Whitley, 1999), results from our own recent research countered these claims (Nathanson, Paulhus, & Williams, in press). Using a concrete behavioral indicator of multiple-choice cheating across two studies, we found two consistent predictors: (1) Deficits in scholastic competence (e.g., poor verbal ability); and subclinical psychopathy, a less extreme variant of its clinical counterpart, (2) characterized by callousness and social deviance (Hare, 1991; Williams & Paulhus, 2004). The present study aimed to replicate and extend our previous work by exploring whether these same associations would hold for the prediction of another form of academic dishonesty, namely, plagiarism. Although Turn-It-In provides an objective indicator of plagiarism, little research has been reported. Using an internet site, we collected myriad individual difference variables, including the Big Five (John & Srivastava, 1999), verbal ability, and the 'Dark Triad' of narcissism, Machiavellianism, and subclinical psychopathy (Paulhus & Williams, 2002). We predicted that subclinical psychopathy and poor verbal ability would again be significant predictors of our two current concrete measures of academic dishonesty, multiple-choice exam cheating and plagiarism. Method Participants. Participants were 328 students enrolled in two sections of an introductory psychology class at a large northwestern university. Sixty-three percent were women. Forty-seven percent were of East Asian heritage, 30% were of European heritage, and the remainder came from other ethnic heritages. Procedure. Students were explicitly warned that the instructor would be watching for exam cheating and plagiarism. Of the 328 students, 107 completed our online battery of standard self-report individual differences measures. Students were not aware of any link between their course work and the personality study. All participants received course credit for participation in the latter. Measuring cheating. To measure cheating on multiple-choice exams, we submitted students' answer sheets to the S-Check program. To minimize false positives, this program conducts pair-wise analyses of all students' responses while controlling for each student's ability, and computes other (potentially) corroborative indices. When plotted on a normal distribution, those pairs suspected of cheating tend to be clear outliers. Moreover, when compared against seating charts, these pairs are invariably seating adjacent (Nathanson et al., in press). Plagiarism. To measure plagiarism, we submitted the students' two term papers to the Turn-It-In program (see Turnitin.com website). This website compares a submitted paper against the constantly updated entries in its comprehensive database. Items in this database range from previously submitted student papers to academic and professional articles. Each paper receives a percentage score that indicates how much of the paper directly matches sources in the databank.