Evidence and control of biases in student evaluations of teaching Available to Purchase

This paper shows that class size has inappropriately influenced students’ judgements on evaluations of lecturers in the Faculty of Economics and Administration, University of Malaya. A bias exists whereby courses with small enrolment receive good overall teaching ratings, whereas larger classes have produced poor evaluations. On the other hand, teaching ratings are not affected by instructor characteristics (instructional experience, rank and gender) or other course characteristics (type and level of subject, and time and day course is taught). To improve the construct validity of student ratings as a measure of teaching performance, this paper suggests using results from a regression model to estimate the bias factor associated with class size, and adjusting the overall teaching ratings based on this estimate in order to control for the bias.