STUDENT EVALUATIONS OF TEACHING: THE GENERALIZABILITY OF CLASS MEANS

The reliability of student evaluations of instruction has been investigated intensively for over a decade, but there is still considerable doubt about the dependability of such measurements, in part because of ambiguity in the definition of reliability. This paper uses generalizability theory (Cronbach, Gleser, Nanda, & Rajaratnam, 1972) to present a more comprehensive view of the dependability of student evaluations of instruction. The central issue is the dependability of the mean rating of an instructor over a class of students and a set of items. The discussion that follows has two purposes. First, it is intended to illustrate the application of generalizability theory to a design commonly employed in the collection of evaluation data. Second it provides a detailed analysis of the dependability of student evaluations of college teaching.