THE TEACHING ECONOMIST - William A. McEachern 
| Home | About The Teaching Economist | Contact the Editor | Support |
Issue 6, Spring 1993
William A. McEachern, Editor
Student Evaluations
Over the years, I have heard few good words about student evaluations of instructors. Are they a valid reflection of teacher effectiveness? There have been more than 1,300 studies of student evaluations. (For a review of some major studies see H.W. Marsh, "Student Evaluations of University Teaching: Research Findings, Methodological Issues, and Directions for Future Research," International Journal of Education Research, Vol. 11 (1987), pp. 253-388.) With that many studies, one could find support for nearly any position, but the broad brush of research tends to support the view that student ratings are a relatively reliable measure of teacher effectiveness.
For example, the classes in which students give the instructor higher ratings tend to be the classes where the students learn more. And an instructor's self-evaluation tends to be significantly correlated with student ratings. Student ratings are also correlated with ratings by faculty colleagues and by alumni. This last piece of evidence challenges the popular view that students appreciate quality teaching only after they graduate and are seasoned by the real world.
What biases are there in student ratings? Two issues that don't seem to matter are the gender and age of the instructor. But there is some bias. Regular faculty tend to receive higher rating than graduate teaching assistants. Students can be influenced more by style (e.g., enthusiasm) than by substance. Ratings tend to be better when the student has had a prior interest in the subject or is taking the course as an elective. Students tend to give higher ratings in courses where they have had to work harder. And students give somewhat higher ratings if they expect a higher grade, thought the bias here is not as strong as most think.
Ratings tend to be higher if students are required to sign their names, if the instructor is present during the evaluation, and if the announced purpose of the ratings is for personnel decisions rather than for self-improvement only. Finally, instructors who receive feedback from student ratings administered during the first half of the term do substantially better in ratings administered at the end of the term than do instructors without the benefit of the first rating.
Student ratings are far from perfect, but, if interpreted with care and used in conjunction with other information, they are probably more reliable and more useful than many of us think. The problem is that they are often relied upon as the primary, if not the sole, source of information about teaching. Nobody would claim that student ratings are the whole story.