THIS REPORT DESCRIBES THE INITIAL VALIDATION OF PERFORMANCE TESTS OF TEACHER EFFECTIVENESS--USING PUPIL GAINS AS THE CRITERION OF EFFECTIVENESS--AND THE STEPS TAKEN IN RECOGNITION OF THE PROPRIETY OF SUCH MEASURES ONLY IF ALL TEACHERS ARE TEACHING FOR THE SAME OBJECTIVES. AS A FIRST STEP, IT WAS HYPOTHESIZED THAT A VALID PERFORMANCE TEST OF TEACHER EFFECTIVENESS SHOULD DISCRIMINATE BETWEEN TWO EXTREME GROUPS-(1) NONTEACHERS AND (2) SUPERIOR EXPERIENCED TEACHERS-BEFORE IT COULD BE USED FOR ASSESSING TEACHERS WHO DIFFER IN SPECIFIED WAYS-(E.G., THOSE WHO ARE AND ARE NOT INTENSIVELY TRAINED TO BRING ABOUT BEHAVIOR CHANGE IN STUDENTS). SCORES ON STUDENT ACHIEVEMENT MEASURES ON TWO INDUSTRIAL EDUCATION TOPICS WERE ASSESSED FOR RELIABILITY AND INTERCORRELATED WITH MEASURES OF GRADE POINT AVERAGE, INTEREST IN THE SUBJECT MATTER, AND WITH WONDERLIC PERSONNEL TEST SCORES. THE OBJECTIVE HERE WAS TO DETECT VARIABLES THAT COULD POTENTIALLY BE USED TO CONTROL FOR STUDENT DIFFERENCES IN SUCH FACTORS AS "SET," INTELLIGENCE, ETC., IN ASSESSING TEACHER EFFECTIVENESS. KUDER-RICHARDSON RELIABILITY COEFFICIENTS OF .44 AND .78 WERE FOUND FOR THE ACHIEVEMENT TESTS. TEST SCORES CORRELATED .68 WITH GRADE POINT AVERAGE. A "PERPLEXING" FINDING WAS HIGHER TEST SCORES AMONG THOSE EXPRESSING LESS INTEREST IN THE INSTRUCTIONAL TOPIC. PRETEST SCORES WERE MORE HIGHLY CORRELATED WITH POSTTEST SCORES THAN WERE WONDERLIC SCORES. (PAPER PRESENTED AT THE 1966 AMER. EDUC. RES. ASSN. MEETING, CHICAGO, FEBRUARY 17-19, 1966). (AF)