The extent to which school-based management actually altered influence relations between a district office and schools and between principals and teachers is examined in this report. Five innovative initiatives implemented in the Dade County Public Schools--peer evaluation, different teacher roles, block scheduling, school-within-a-school, and teachers as advisors--were evaluated. Methodology involved administrative reviews of individual school-based management pilot schools for 1988, 1989, and 1990, and analysis of proposals to formulate an estimate of impact using the Purdue Teacher Opinionnaire database. Findings indicate that while school-level discretion was allowed in some areas of decision-making, most deviations from regular policy were subject to approval by central office administrators. Within schools, teacher involvement altered teacher-principal relations, which were mitigated by principal supervision of the innovation and limited teacher involvement. Overall, the pilot school-based management project recorded no change in district/school relations and some change in teacher/principal relations. Factors for successful implementation include a supportive superintendent, a systematic training program, a pool of sponsored innovations, and increased knowledge of organizational behavior. The findings call into question the assumptions of increased efficiency through reorganization and increased innovation through autonomy. The appendix presents procedures for teacher peer evaluation. (16 references) (LMI)