Some Evidence on the Variables Associated with Student Evaluations of Teachers

Abstract

Student evaluation of teaching (hereafter SET) scores are of use to academic departments and colleges in the evaluation and rewarding of teaching; they are also of use to the individual instructor in assessing his own performance and in developing the expertise required to become a more effective teacher. Isolating the more important variables which determine the SET score is a problem common to teachers and administrators alike. One purpose of this paper is to provide additional evidence about the variables associated with SET scores. Another purpose is to comment on the use of multiple regression analysis as a tool for adjusting raw scores for factors essentially beyond the control of the instructor when ratings are used to reward teaching performance [3]. Knowledge of these basic relations and the appropriate use of ratings are clouded by problems of model specification and multicollinearity among variables. Some of these problems will be focused upon in the results presented below. The data reported here represent over 5,300 student ratings given to 103 regular instructors and 24 teaching assistants (with full-course responsibilities) in the College of Business Administration, University of Georgia, during the Fall quarter, 1974. The short-form instrument presented in Hildebrand [2] forms the principal component of the questionnaire used.1 The interested reader should consult Hildebrand [2] about instrument reliability, colleague vs. student ratings, and similar matters which he discusses at length.