New device detects cheating on mulitple-choice exams

By Jennifer Kawai
Online Forty-Niner
Contributing Writer

Student cheating has become such an easy task that a California State University, Sacramento professor, Robert G. Mogull, created a technique that will confirm instructor's suspicions of students who cheat on multiple-choice Scantron tests.

The technique came into production after two of Mogull's students missed the same questions on four different exams. After an item analysis was completed, which shows the number of Scantrons graded, the average number of correct answers and the frequency of incorrect answers per question, Mogull found that two of his students had the same grade on all of their semester exams.

He was sure the students had cheated, and the probability device he created only confirmed his suspicion. Mogull calculated the probability of this instance happening at random.

Needless to say, the business statistics professor failed both students for the course. Mogull's technique called "A Device to Detect Student Cheating," was published in the September 2004 issue of the Journal of College Teaching and Learning.

Although Mogull was dealing with only two students, the technique could be used for cases in which more than two students are involved. The device works by first calculating the probability of missing a specific exam question. The next step is to calculate the probability that two random students would miss the same question. Then, the final probability that both students would miss the same questions is calculated by finding the product of the individual probabilities of any two students missing the same questions. The more students that are involved and the more questions that are missed, there is a smaller probability that the students worked independently, and a higher probability that, according to Mogull, the students "collaborated, colluded, worked jointly, shared, copied answers or cheated."

Mogull's method can be used by instructors who suspect cheating. With a few calculations, the probability can be computed, and students could possibly end up finding themselves with a failing grade.

"A lot of people wouldn't like it," said Krista Minnich, a psychology major at CSULB. "I don't usually cheat on tests, so I personally wouldn't mind the new technique."

The CSU policy for cheating defines it as copying, discussion of answers unless allowed by the instructor, giving or receiving copies of exams without instructor permission, displaying "cheat sheets" on tests that are supposed to be from self-knowledge and allowing someone else besides the officially enrolled student to take an exam. Any of these actions could result in a student failing the course or being expelled from the CSU system.

Faculty who suspect a student of cheating must have reasonable evidence and then arrange an informal school conference to inform the student. An instructor can give a student an""I" for incomplete if the student cannot be reached. Further information on cheating can be read in the Schedule of Classes for spring 2005.

"Finals time is a period more prone to cheating but hopefully students will understand the consequences and not be inclined to do so," said Steve Katz, director of judicial affairs at CSULB. While only the most severe cases are sent to the judicial affairs office, Katz generally refers to about a dozen cases a year.

According to Mogull, students are the best at cheating — his advice to other instructors is to use different forms of a test for each test given, to reduce the temptation to glimpse or copy off of someone else's test.

"Plagiarism and cheating is a problem in universities," said Katie Gibson, journalism and communications studies department professor. "In order to preserve the integrity of a college degree, this issue needs to be dealt with."