Characteristics that seem to separate effective and ineffective teacher evaluation systems provide a framework for viewing existing evaluation systems and determining necessary changes. Nine commonalities represent the best practices: (1) participants share an attitude that the purpose of an evaluation system is to provide assistance in the improvement of instruction; (2) the requirements placed on the participants in a system completely relect the actual purpose of the system; (3) teacher evaluation is separate from teaching evaluation; (4) some form of goal setting exists between the teacher and the supervisor as the basic supervisory activity; (5) a narrow focus on the teaching act is used for the instructional interaction between supervisors and teachers; (6) preconferences occur prior to observation as suggested by the clinical supervision model; (7) student evaluation and artifacts (study guides, tests, homework assignments) are used in addition to classroom observation; (8) different requirements are made for tenured and nontenured teachers; and, (9) a complete training program is offered for both supervisors and teachers. (MLF)