A system which used pupil outcomes to estimate the predictive validity of assessments of teacher outcomes was field tested in a performance-based teacher education program at Florida State University. Three levels of teaching competency were assessed: (1) competencies associated with cognitive identification of appropriate decisions for implementing a teaching model; (2) competencies for performing model-related behaviors; and (3) competencies associated with inducing pupils to demonstrate model-related classroom behaviors. Five instruments, measuring teacher knowledge and classroom performance and student classroom behavior and outcomes, were used to evaluate program effectiveness and student teachers' readiness to teach in public schools. The field test was conducted in a public school setting with 24 teacher candidates and 120 third grade students. Each candidate delivered six half-hour lessons to five students during a four-week period. In addition to the measurement instruments, videotapes were used by three-person observation teams to assess teacher classroom performance and student classroom performance. Correlation coefficients and multiple regression analysis indicated that pupil outcome scores may be successfully used to estimate the predictive validity of assessment instruments and to aid in the validation of teaching competencies within the teaching model framework. (Author/JAC)