Robert G. Mogull, a statistics professor in the college of business administration, uses his expertise in scientific probability to detect students cheating on exams, publishing a study that gives statistical guidance concerning academic dishonesty.

In a previous semester, Mogull had discovered evidence that two students in his second semester business statistics course, MIS 131, cheated. The two students in question missed 26 identical out of a possible 100 questions.

As a result, Mogull, who typically does studies and projections concerning poverty, decided that he wanted to calculate the probability that the students had been cheating.

“I thought to myself, ‘You ought to be able to determine the likelihood that they cheated.’ I wanted to see what kind of confidence level I could achieve in determining that in fact they had cheated and collaborated,” he said. “It wasn’t an attempt to nail them because the evidence was already quite clear.”

Mogull did not even write the paper until after the students had been confronted and punished.

The method itself, which has been incorrectly identified in news reports as a “secret formula,” involves calculating both the likelihood of a person missing a particular question and that multiple students miss identical questions.

According to Mogull’s paper, which was published in the Journal of College Teaching and Learning, a case of cheating can be established if the probability of students cheating on the exam is calculated to be better than 90 percent.

“Ninety percent is kind of the minimal confidence level that goes along with the social sciences, but it is kind of arbitrary,” Mogull said. “It is something that a reader, another professor might ask. So I offered that as a rough guide because it is the minimal threshold commonly used in social sciences and business.”

In his real life case, Mogull calculated that there was only a 0.00000000000000000004 chance that the students had not cheated. “It was as close to absolute certain as you can get,” Mogull said.

While this detection method really is only applicable to multiple-choice exams, Mogull said that, while difficult, it is possible to detect cheating on other types of tests.

“You would have to use a different method for an essay-type exam. You couldn’t use the probability method, unless you wanted to try and use word analysis,” Mogull said.

Word analysis is a method of detection that involves tracking phrases and words common to multiple documents, similar to an Internet search engine.Many students are opposed to cheating, albeit for different reasons.

“I don’t like to copy or be copied. There are very few instances that I can think of in which I let people even copy my homework,” Ana Maligat, a senior apparel design major, said. “In fact, I don’t think I have done it since high school.”

“You don’t learn anything by cheating,” Aaron Pypers, a senior physics major said.

However, both Maligat and Pypers had muted reactions to the method.

“I think that there are too many discrepancies,” Maligat said.”The probability of error seems too large, and I don’t think that it is morally OK to bust someone for cheating on just the numbers.”

“It seems like a lot of effort,” Pypers said. “If you don’t trust your students to not cheat, you should change the way that you are giving the exam, like moving desks apart or giving multiple forms, things like that.”

Regardless, Maligat said that she is very careful about her work.

“I always cover my answers on tests and I’m careful about where I look during the test,” Maligat said. “I basically just try to stay focused on my exam.”