Any particular teaching performance is limited by time and place and must be regarded as only a small sample of the large collection of activities called teaching. Evaluation is part of teaching, but it may also be a separate activity carried on by a third party. Evaluation methods differ according to the purpose, whether for guidance of preservice teachers, improvement of training programs, or certification. For both teacher education and teacher certification a major problem is that of defining all the kinds of jobs that teachers are expected to fill and adequately describing the social and institutional settings in which the jobs exist. No matter what the main reason for the evaluation of a particular teaching performance, the evaluator must take some account of what the teacher is trying to accomplish. If teaching performance is to be evaluated in detail, some theory of teaching is needed as a guide to how it should be broken down. Adequate justification for the theory must be available in some form. A particularly difficult problem in evaluating teaching performance is that of gathering sufficient accurate data. If teaching performance is to be judged on the basis of pupil learning, great care must be used to eliminate major sources of error in the assessment of learning. Among the many issues related to the problem of evaluating teacher performance are racial bias in teacher selection, merit rating, accountability, differentiated staffing, and community control of the schools. (RT)