Duke's Cheating Problem
Posted by: Louis Lavelle on August 12, 2011
When Duke Professor Dan Ariely suspected some of his students were cheating, he did what any behavioral economist would do: He let his students run an experiment. The results of the experiment show many students are still tempted to cheat, but a reminder of a strong, direct honor code could be a deterrent, to a point. Ariely’s experiment is also interesting because it took place at a university that has already had a cheating scandal and takes cheating seriously.
Ariely, a professor at Duke’s Fuqua School of Business, where disciplinary action was taken against 34 MBA students after cheating was discovered on a take-home final in 2007, teaches a class on behavioral economics to undergrads at Duke’s Trinity College of Arts & Sciences. Last semester, he came to believe that some of the 500 students in that class were glancing at other students’ papers during weekly quizzes based on the similarity of their answers, reports the Chronicle of Higher Education’s Wired Blog.
Those suspicions led Ariely to discuss his own research on dishonesty and cheating with his class. After the discussion, two of his students decided to conduct the assigned research project on the class, using their own classmates as unwitting research subjects in a provocative experiment, he wrote on his blog.
The two students sent the class an email from a fictitious student and included a link to a website said to contain the answers to last year’s final exam. Half of the emails also included a postscript: “P.S. I don’t know if this is cheating or not, but here’s a section of the University’s Honor Code that might be pertinent. Use your own judgment: ‘Obtaining documents that grant an unfair advantage to an individual is not allowed.’”
The link, which contained no answers, allowed the researchers to track who clicked on it using Google Analytics.
In all, about 69 percent of students who received emails not containing honor code language clicked on the link, and presumably were sorely disappointed. Of students who received emails with the honor code language, 41 percent still clicked on the link.
"This is a tempting email," Ariely said in an interview with Bloomberg Businessweek. "It's hard not to open it. It doesn't reflect cheating, but it reflects falling into temptation."
In his blog, Ariely wrote that students who didn't receive emails with the honor code language thanked the sender for the link, while students who did receive honor code language were "upset and offended."
"I wish the numbers would've been lower. I expected there to be a difference between the honor code and non-honor code [groups], but I hoped it wouldn't be too big," he said. "Forty [percent] is still too high."
Encouragingly, so many students emailed or approached Ariely with concerns about cheating that the two students running the experiment were forced to reveal it before disclosing their findings, he said.
The day after the final exam, he surveyed students asking them to anonymously report their own cheating and the suspected cheating of their peers. Students estimated 30 percent to 45 percent of their peers cheated, while few admitted to it themselves. Ariely said he didn't think many of his students actually cheated on the final exam, given the 70 percent average.
On his blog, Ariely writes that even if few students actually cheat, the perception that many do "can become an incredibly damaging social norm" because students may "feel that it is socially acceptable to cheat and feel pressured to cheat in order to stay on top."
Ariely said he could have done more as a professor to change that. Not addressing cheating on the quizzes "created a culture that this is OK to do," he said. "I created an environment where I was not strict about it or stopped it. I didn't think about the long-term consequences and the local culture deteriorated."
The experiment in Ariely's class suggests that honor codes do work, especially if students are reminded of the code at precisely the right time: when temptation presents itself. But the fact that two out of three Duke students involved in the experiment were willing to click on a link that promised them a big advantage on the final shows just how limited the impact of such codes can be, absent a culture where dishonesty is not tolerated.
News of the Duke experiment comes less than a month after a professor at NYU's Stern School of Business revealed widespread plagiarism in his class, a level of academic dishonesty that occurred despite another strong ethics code that prohibits "plagiarism, misrepresentation, and falsification of data." Apparently zero tolerance is the only cheating policy that works.
--Kiah Lau Haslett
