Although college faculty may hesitate to evaluate each others' teaching, it is a sensible alternative to evaluation by students or by administrators. Peer evaluations are not only useful for improving teaching, but also affect promotion and tenure decisions. When faculty are rated by observation, repeated visits and rater training are required in order to insure reliability. A qualitative approach, involving descriptions of classroom instruction based on the observer's perceptions, is preferred. Rating scales may be useful in assessing the appropriateness of course objectives, value of instructional materials, student achievement, and teacher's knowledge of subject matter. One college uses a Committee on Teaching to evaluate teaching performance. In addition to reviewing materials presented by the instructor and describing classroom observations, the committee interviews both former and current students. Another approach to teacher evaluation, which is used by very few institutions, is a variable weight approach. Faculty members select, in advance, an area of their performance--research, teaching, or service--which would be given extra weight during that evaluation year. (GDC)