Student performance should be among the criteria used to evaluate teachers. Some advantages to comparing teachers on the basis of their students' learning are that such comparisons can provide (1) benchmarks by which to judge the validity of the criteria and measures of teaching effectiveness, (2) the evidence needed to match certain kinds of students with teachers whose style elicits maximum performance, and (3) the evidence needed to match team teachers with the areas in which they are most effective. In order for teachers to be evaluated fairly by such systems, the evaluation must be individualized to the students, the class, the particular subject taught, and the constraints of the environment in which teachers work. Teachers must be involved in designing the criteria rather than only making cosmetic changes in existing evaluation forms. Because of wide variations in goals and resources between state and local school districts, the evaluation should also be a localized procedure, occurring at the district level rather than the state or national level. (IS)