Publication: Journal of Economic Education
Volume: Volume 25, No. 2
Issue: Spring 1994
Pages: pp. 121-133
Author(s): Joe Kerkvliet (Oregon State University)
Address (Principal Author):Joe Kerkvliet, Department of Economics, Oregon State University, Ballard Extension Hall 303, Corvallis, OR 97331-3612 (503) 737-2321
Internet Address (Principal Author):
Title: Cheating by Economics Students: A Comparison of Survey Results
Abstract: Principles of economics students are asked about their academic cheating behavior using two different survey instruments. Direct question surveys rely on truthful responses to a sensitive question, while randomized response surveys provide more confidentiality and may encourage more truthful answers. Separate logit models of cheating are estimated using data from both survey types and the results are compared. Statistical results indicate that between 25 and 42 percent of economics undergraduates have cheated on examinations. Some logit results are robust with respect to survey type. The most likely cheater is a heavy drinker who is a resident member of a fraternity or sorority. Males are less likely cheaters and students with college educated parents may cheat more. No statistically significant effects are found for grade point average, total study time, or exercise time. One result is fragile with respect to survey type. Randomized data indicates that a student in the early stages of an academic program is more likely to have cheated. However, direct question data support the opposite conclusion. The results indicate caution in estimating the extent and determinants of illegal or unpopular behavior. A comparison of findings shows that the survey method can affect the sign and statistical significance of parameter estimates and the use of only one method may give misleading conclusions.