�鶹��ý

Recent advances in technology call for a reconsideration of how to deal with the perennial issue of cheating in exams. “In the recent past,” Professor Kang Lee says, “it was easier to see if students were using cell phones or other technology. But advances in AI software such as Chat GPT, and in wearable technology pose unique challenges for proctors and instructors administering exams in large halls.”

At a Focus on Research session attended by OISE faculty, students, postdoctoral fellows, and staff, Lee shared his team’s recent studies on methods that reduce incidents of using unauthorized aids in exams.

Lee is uniquely positioned to undertake ground-breaking research on this subject. For the past three decades, he has focused on developmental and cognitive factors involved in honesty and lying. He holds a Tier 1 Canada Research Chair in developmental neuroscience, and his research has been funded by agencies such as the Social Sciences and Humanities Research Council (SSHRC) and SSHRC and the Natural Sciences and Engineering Research Council (NSERC) in Canada, the National Institutes of Health (NIH) in the US, and the National Natural Science Foundation (NSFC) in China. 

Lee notes that previous work on cheating focusses on associations between academic integrity and specific characteristics and personality traits. “We know that there are specific socio-cognitive tendencies and abilities that are associated with a reduction of cheating.” 

Lee points to the benefit, however, of conducting actual field studies, something that has not previously been attempted when studying cheating. “There is an urgent need to develop scientifically validated solutions to encourage academic integrity,” Lee says. “Otherwise, we are relying on self-reported incidents, which tend to be unreliable.”

In one study, Lee’s team recruited faculty members teaching introductory level undergraduate courses who agreed to allow Lee’s team to plant a fake question on the exam and a corresponding fake answer on the University’s intranet. In some instances, instructors were also asked to remind students of the University’s academic integrity policy prior to the exam.

Students could only obtain this answer by surreptitiously searching online during the exam, inadvertently identifying themselves as cheating in the exam. As this was a double-blind study, individual students were not identified or penalized. Lee and his team, however, were able to compare actual instances of cheating with self-reported instances of cheating and to determine the impact of policy reminders.

Two major findings emerged from this study. First, there was no correlation between instances of cheating and self-reported instances of cheating, underlining the need for this type of study. Second, there was a marked reduction of cheating in the examinations preceded by reminders of university policy. 

Clearly, reminders about academic integrity reduce incidents of cheating. But are some types of reminders more effective than others? To address this question, Lee’s team designed a second study to examine the impact of different types of reminders prior to unproctored exams. They found that a reminder about University policy, while effective, was not nearly as effective as a concrete example of consequences of an incident of cheating. Interestingly, both types of reminders were impactful even when students were not under surveillance during the exam.

This finding prompted Lee to wonder whether reminders are similarly effective in proctored exams, and to initiate a third double-blind, randomized, controlled study. As expected, Lee’s team found that brief mentions of university policy are effective regardless of the exam setting. Surprisingly, however, there were significantly fewer instances of cheating in unproctored exams. Students who were not actively monitored for cheating were less prone to cheating, a phenomenon Lee ascribes to well-established honour codes.

The increasing availability and sophistication of wearable technology, then, does not necessarily mean that there is a need for new methods of proctoring. In fact, subjecting students to surveillance is far less effective at preventing cheating than a school culture in which honour codes are clearly communicated and understood by students. 

“These studies underline the value of putting resources into developing a culture of academic integrity, which is potentially our best tool to curb cheating.”