Two issues in the controversy over the use, interpretation, validity, and relevance of student ratings are explored here: the definition of preferred teacher behavior in terms of the rating instrument used, and the relationship between the student's expected grade, grade point average, and the ratings recorded. The objectives of the present research were threefold: (1) to develop a short rating scale and associated diagnostic checklist; (2) to examine the construct validity and reliability of the scale; and (3) to assess the relationships between selected student demographic data and ratings on the scale. After a pilot test, there appeared to be little advantage to using an observed/preferred behavior format, so that format was discontinued in further revision of the scale. A repeated measures analysis of variance indicated that discrepant groups, i.e., those whose actual grade was two or more letter grades away from his/her expected grade, should be omitted from computations to increase the validity of student ratings. Further research was suggested to determine whether demographic variables interact with group discrepancy to influence evaluations of teacher behaviors. (Author/BW)