A study was undertaken to examine the traditional method of evaluating teachers, the multiple-evidence approach used by the University of Georgia's field-based preservice secondary English education program and by the Georgia Department of Education. The study sought to evaluate the instruments used by the university to measure prospective teachers' knowledge, attitudes, and performance, and to use these instruments in a comparative, evaluative study of the Georgia Teaching Field Criterion-Referenced Test in Communicative Arts. Subjects included 40 teacher candidates who had completed the university program, and 40 candidates who in addition had taken the state test as part of the test's tryout. The two groups scored alike on the university's test, establishing the groups' similarity and supporting the test's reliability. No significant relationship was found between the department's test and any other measure of performance, including professional or student ratings. The results indicate that none of the predictor measures is useful for positive prediction of classroom performance, and that these measures have little utility for prediction other than of performance on other measuring instruments. (DF)