A STUDY OF 10 SCHOOL DISTRICTS TESTED THE HYPOTHESIS THAT CLASHES BETWEEN SCHOOL ADMINISTRATORS AND TEACHERS' UNIONS ARE DUE LARGELY TO THE CONFLICTING PERCEPTIONS OF THE SCHOOL ADMINISTRATOR'S LEADERSHIP BEHAVIOR AS DESCRIBED BY THE ADMINISTRATOR HIMSELF, BY THE SCHOOL BOARD, AND BY MEMBERS OF THE TEACHERS' UNION. ACCORDING TO THE LEADER BEHAVIOR DESCRIPTION QUESTIONNAIRE ADMINISTERED, RESPONDENTS OF ALL THREE GROUPS DEFINED IDEAL AND REAL ADMINISTRATORS IN SIMILAR TERMS ON A SCALE OF 12 LEADER BEHAVIOR DIMENSIONS. BACKGROUND INFORMATION WAS OBTAINED BY INTERVIEWING QUESTIONNAIRE RESPONDENTS, LABOR MEDIATION BOARD MEMBERS, AND UNIVERSITY PROFESSORS. FOLLOWING HALPIN'S SCHEME FOR MEASURING LEADER BEHAVIOR, QUADRANT ANALYSES INDICATED THAT THE SIX LEADER BEHAVIOR DIMENSIONS CONTRIBUTING MOST TO CONFLICT WERE (1) CONSIDERATION, (2) INITIATION OF STRUCTURE, (3) INTEGRATION, (4) DEMAND RECONCILIATION, (5) TOLERANCE OF FREEDOM, AND (6) PRODUCTION EMPHASIS. CONCEPTS STUDIED WERE DEVELOPED FROM A SURVEY OF LITERATURE RELATED TO LEADERSHIP, CONFLICT, AND CONFLICT RESOLUTION. RECOMMENDATIONS INCLUDED--(1) STUDY OF APPLICABILITY OF THE RESEARCH MODEL FOR USE IN EDUCATIONAL ADMINISTRATION, (2) EVALUATION OF PROCEDURES FOR SELECTION OF SCHOOL ADMINISTRATORS WITH RESPECT TO EACH SCHOOL'S ORGANIZATIONAL CLIMATE, AND (3) DEVELOPMENT OF ADEQUATE PROGRAMS FOR TRAINING ADMINISTRATORS, WITH GREATER EMPHASIS ON THE BEHAVIORAL SCIENCES. (JK)