The appropriate questions in teacher evaluation in higher education are: who should evaluate, for what purposes, and using what means. Frequently, evaluation systems provide information for administrative decisions on promotion/tenure and merit. Increased reliance on student ratings of teachers has the disadvantage of distorting factors that may influence ratings, such as class size, lecture or discussion structure, and years teaching. When used as a measure of student satisfaction and supplemented with other sources, student ratings may be justified to judge some, but not all, aspects of teaching. Evaluating teacher effectiveness in terms of student achievement depends on whether the responsibility of learning rests with the pupil or the teacher. In addition to lack of agreement between faculty self-ratings and ratings by others, teachers tend to give themselves better ratings than their students do. Colleague evaluation by direct observation is best used formatively to improve teaching. Indirect review of syllabi, tests, and materials is generally useful with other evidence. The problems of defining effective teaching make evaluation particularly difficult. Evaluation will remain judgmental, but the best systems seem to use a variety of sources of evidence and explicate as clearly as possible how judgments will be made. (CM)