Teaching

To Stop Exam Cheats, Economists Say, Try Assigning Seats

October 14, 2015

Keith Morris, Writer Pictures via AP Images
Steven Levitt, the "Freakonomics" co-author, collaborated on a study of cheating in college classrooms. The takeaway? There is magic in assigned seating.
Think seating charts in the classroom are needed only in elementary school? According to a new study, randomly assigned seats are also the most immediate way to prevent cheating among college students.

The study was set in an introductory science course at an unnamed "top American university" in 2012, as 242 students prepared to take their final examination. They were allowed to sit wherever they wanted upon entering the room, as they had for all previous tests. At the last minute, though, the instructor gave them a seating chart. The experiment led the researchers to conclude that at least a tenth of the students had cheated on the previous midterm. But the final exam, given with assigned seating and three additional proctors, produced nothing but clean consciences. Evidence of copying all but vanished.

The findings are the work of Steven D. Levitt, a professor of economics at the University of Chicago who is a co-author of the best seller Freakonomics, and Ming-Jen Lin, a professor of economics at National Taiwan University, in a study published by the National Bureau of Economic Research. Mr. Levitt and Mr. Lin developed an algorithm to detect evidence of copying during exams after a professor, frustrated by several complaints about cheating, asked the researchers to use his classroom as a test case.

The experiment led the researchers to conclude that at least a tenth of the students had cheated on the previous midterm.
Mr. Levitt and Mr. Lin studied the matching multiple-choice answers of student pairs who sat next to each other during midterm exams. The researchers found an increase of 1.1 shared incorrect answers when those students were allowed to sit where they pleased — nearly twice as many as would be expected to happen by chance. The economists' conclusion: At least 10 percent of students had cheated on their midterms. (They ruled out the possibility that student pairs with high rates of shared mistakes had simply studied together for the final exam. At the test, they recorded the seats students chose before the new seating chart was handed out. Once separated, the students who had initially wanted to sit together did not show a pattern of more shared incorrect answers between them. A high number of shared answers indicates possible cheating.)

After the study was complete, the professor who had opened up his classroom tried to use Mr. Levitt’s and Mr. Lin's findings to deliver some discipline. He sent the names of the 12 students deemed most suspicious to the dean’s office for investigation. According to the paper, a hearing was scheduled, and four students admitted to cheating even before it was convened. But the investigation was then canceled. (Mr. Levitt and Mr. Lin attribute that decision to "pressure from parents.") The professor withheld grades of the dozen suspected students, but the university took no larger disciplinary action.

Can Cheating Be Curtailed?

The study draws a conclusion that many professors may have reached on their own: There is magic in assigned seating. Random pairings, combined with an increase of instructors, eliminated all signs of copying during the final exam.

Randomizing might be the best way to fight cheating when it comes to large lecture halls and multiple-choice tests, says Susan D. Blum, a professor of anthropology at the University of Notre Dame and author of My Word! Plagiarism and College Culture (Cornell University Press, 2009). "The bigger the class, the less students feel a personal investment," she says.

Many students engage in premeditated cheating, and "if you can’t sit next to the people you planned to," random seating "would break up that kind of premeditation," Ms. Blum says. Still, that doesn’t completely eliminate opportunistic cheating — sneaking a glance at an unsuspecting classmate’s paper.

But an atmosphere of trust is essential. "Faculty don’t want to see themselves as opposing students or suspecting them of misdeeds," Ms. Blum says.

In a smaller classroom, trust can go a long way, says Linda K. Trevino, a professor of organizational behavior and ethics at Pennsylvania State University’s Smeal College of Business. In her classes of 25 to 40 students, Ms. Trevino gives frequent quizzes along with an honor code that students must sign. Then she leaves the classroom during the tests. While the code doesn’t eliminate cheating altogether, "many people respond in a positive way to that kind of trust," she says. "You teach students what it’s like to live in an environment where there are high expectations for their conduct."

Over all, rates of cheating have been high and stable since the 1960s, and there is no clear way to arrest the trend, says James M. Lang, a professor of English at Assumption College, in Massachusetts, and author of Cheating Lessons: Learning From Academic Dishonesty. But colleges should still try to pay attention to learning environments and "modify course design and classroom practices to reduce cheating," says Mr. Lang, who contributes articles to The Chronicle’s Advice section. For example, "if students only have two or three high-stakes tests," the pressure to earn high marks increases, thus increasing the likelihood of cheating.

While there is no good evidence that severe punishment reduces cheating, colleges must make sure that "faculty feel supported when they report cases to the administration," Mr. Lang says. "The decision about what kind of punishment is appropriate should come from the faculty member," and it depends on the type of cheating and the individual case. "It’s not ‘one size fits all’ punishment," he says.

Correction (10/14/2015, 10:50 a.m.): The original version of this article referred to Ming-Jen Lin on second reference as "Ms. Lin." He is in fact Mr. Lin. The text has been corrected.