The article first presents results of correctional intervention effectiveness based on 32 literature reviews and meta-analyses that collectively addressed 20 separate programmatic approaches such as educational training or group counseling. It then examines findings that show the importance of nonprogrammatic factors such as staff characteristics, staff/client interactions, and setting. The article next introduces a proposed new direction for research, i.e., the study of programs as composites of programmatic approaches and nonprogrammatic factors. This new direction is needed to better understand why some interventions work and others fail. Finally, the article presents a "global approach" framework for implementing the proposed new research, and it discusses practical issues associated with it. The "global approach" is designed to examine operations, staff, offenders, and setting factors simultaneously. This strategy is based on the premise that programs are functional entities and on the corollary that it is not sufficient to examine only one area. 38 notes and 55 references