Discriminant Analysis and Classification of Teaching Effectiveness Using Student Ratings: The Search for Doctor Fox.

This study was designed to: (1) derive a set of weights for student-faculty rating item scores that maximize differences in mean rating scores for groups of students known to have experienced real differences in instructional effectiveness; and (2) attempt to cross validate the new scoring methods in a second study of students who rated lectures under the same conditions. Participants in the first study were 207 undergraduate and graduate students who were enrolled in general studies sections, while participants in the second study were 213 students enrolled in 12 sections of an undergraduate psychology course. The four types of faculty lectures were: (1) high information-high enthusiasm; (2) high information-low enthusiasm; (3) low information-high enthusiasm; (4) low information-low enthusiasm. In the first study, sections of the same class were randomly divided, while in the second study intact sections of the same class constituted the study group. Lecturers were randomly assigned to student groups. Students in all groups viewed one lecture presentation, rated the presentation using an 18-item questionnaire, and were tested on the material. The use of discrimination analysis to develop student-rating scales that are valid with respect to faculty enthusiasm is supported by this study. This study also indicated that empirically based selection and weighting of items does not improve the validity of student rating scores in detecting real differences in information giving. Current practices in the evaluation of teaching effectiveness are limited almost entirely to proxy measures, namely, student ratings. (SK)