Computer scientists at Pace University have developed a new way to detect cheating by online test-takers.
The approach, developed by Charles C. Tappert, a professor, and several graduate students, tracks the typing and style choices of users to authenticate their identity.
Mr. Tappert has worked with various students for several years to develop a keystroke biometric system that tracks how long users hold down and release individual keys—their “dwell time”—and their transition from each key to the next.
Others have gone that route, but this year his team added a “stylometric” component, to measure tendencies in users’ writing, which could include types of words used or frequent misspellings.
Identification efforts using this measure have been less accurate than tracking dwell time, but the professor says he will continue to work on it with students this fall.
The project is important, he says, because it could help prevent students from enlisting others to complete their assignments or take exams. “The idea is that if someone else is feeding you the answers, the words you use would be more like theirs than yours,” he says.
The tool could help colleges comply with a provision in the 2008 Higher Education Act calling for distance courses and programs to verify that students enrolled are actually completing their own work.
Mr. Tappert says he and his students plan to present the results of their research at the International Joint Conference on Biometrics, to be held in Washington this October.Return to Top