THIS STUDY TESTS THE HYPOTHESIS THAT COLLEGE INSTRUCTOR BEHAVIOR CAN BE IDENTIFIED AND MEASURED BY THE USE OF A GRAPHIC RATING SCALE WHEN THE RATER RESPONDS TO COMMON BIPOLAR ADJECTIVES. THE BEHAVIORS WILL YIELD, WHEN SUBJECTED TO FACTOR ANALYSIS, CLUSTERS OF TRAITS WHICH WOULD IDENTIFY SUBGROUPS OF COLLEGE INSTRUCTORS. STUDENT RATINGS OF COLLEGE INSTRUCTOR BEHAVIOR WERE OBTAINED THROUGH THE USE OF A SCALE COMPOSED OF 12 BIPOLAR ADJECTIVES FROM THE OSGOOD SEMANTIC DIFFERENTIAL, 13 SINGLE ADJECTIVES FROM OTHER RESEARCH STUDIES, AND AN OVERALL GLOBAL RATING OF INSTRUCTION. THE SCALE WAS ADMINISTERED TO UNDERGRADUATE LIBERAL ARTS CLASSES OF THE SAME INSTRUCTORS ON TWO OCCASIONS SEPARATED BY A TIME INTERVAL OF 15 WEEKS. THE TOTAL SAMPLE OF 7,060 STUDENTS RATED 104 INSTRUCTORS IN HUMANITIES, NATURAL SCIENCES, AND SOCIAL SCIENCES AT THE UNIVERSITY OF PITTSBURGH. IT WAS CONCLUDED THAT STUDENTS DISTRIBUTED THEIR JUDGMENTS OF INSTRUCTORS IN A MARKEDLY RELIABLE MANNER, BUT THE VARIANCE OBSERVED DID NOT SIGNIFICANTLY DISCRIMINATE BETWEEN INSTRUCTORS ACCORDING TO ACADEMIC DIVISION NOR DID IT RELATE IN ANY APPRECIABLE DEGREE TO GLOBAL ESTIMATES OF EFFECTIVENESS. THE DISCRIMINATIONS EXPECTED WERE NOT WITHIN THE COMPETENCE OF THE OSGOOD SCALE OR OF THE SINGLE ADJECTIVES LISTED. THIS DOCUMENT IS ALSO AVAILABLE FROM UNIVERSITY MICROFILMS, ANN ARBOR, MICHIGAN, (MICROFILM NUMBER 67-4558), FOR MF $3.00, HC $7.60. (HW)