This report describes the New York City public school system's experiences with decentralization since 1970. The report includes an introductory chapter explaining the conditions that led to the adoption of decentralization; analytical case studies of eight New York City decentralized school districts; and two concluding chapters on findings and implications. The case studies describe the context of decentralization in each district; relationships between school boards and superintendents and between district offices and schools, professional staff, and the community; emergent management styles; and the effects of decentralization on student performance, attendance, and later academic attainment. It is suggested that the New York City experience does not provide an extensive test of decentralization because of legal ambiguities and inadequate implementation. Nevertheless, it is concluded that even with limited decentralization, significant improvements are evident. Among the findings of the study are: 1) student performance improved in many poor, minority districts; 2) the number of curriculum innovations, staff training programs, and programs to link schools and communities increased markedly; 3) more positive changes occurred in districts where superintendents had greater authority and office staffs were more stable; 4) effective superintendents' management styles varied; and 5) decentralization did not enhance parent participation. The report presents a model outlining the elements of district effectiveness, and concludes by identifying some unresolved problems in decentralizing districts. (Author/MJL)