To Catch a Cheat
EIME student TJ Bliss receives funding for a study on cheating.
It takes work to catch a cheat. Based on his proposal to research statistical methods to detect cheating, EIME graduate student TJ Bliss received the Joe E. Covington Award for Research on Testing for Licensure. This award, sponsored by the National Conference of Bar Examiners, gave Bliss $6,000 towards his study of statistical methods to detect cheating.
Bliss’s work focuses on data related to cheating and numerical methods of detecting it. “One common way is to look for similar response patterns in students, especially incorrect answers. If two students’ answers mirror each other, statistically, it’s more likely that someone’s cheating,” Bliss explained.
Recently, however, with many tests being administered on computers, measuring response time has become a popular way to look for potential cheaters. A computerized test monitors the amount of time spent on each individual question, and test taking irregularities become apparent. “If a student spends very little time on difficult questions and gets them right, it’s something to look into,” said Bliss, “especially if they spent longer amounts of time on easier questions or got easier questions wrong.”
Bliss added that use of statistics alone does not confirm cheating. Even if the numbers suggest cheating, other factors must be taken into consideration. “If two students have the same test results but one took the test in Texas and the other in Canada, we can be pretty sure they weren’t cheating, despite the statistics,” said Bliss. “If we find that they were sitting next to each other, that’s a different story. The statistics flag irregularities, but they don’t prove anything.”
Bliss’ work will produce a literature review of the gathered statistical information, which he plans to publish in a measurement journal. “It’s all happened really fast,” said Bliss. “But things are going well. The award was a pleasant surprise for me.”
19 March 2012