One 1957-1958 study of GGI with juvenile delinquents failed to provide strong and convincing evidence of GGI effectiveness. A second, 1974 study did provide convincing evidence for the effectiveness of a community treatment program involving GGI over incarceration for individuals who otherwise would have been incarcerated and a modest superiority of GGI over probation for those who would otherwise have been assigned to probation. Some evaluations of school-based prevention programs that have been cited in support of the efficacy of GGI are methodologically flawed and may be misleading. A 1982-1983 evaluation using better design and controls suggested that exposure to GGI may have increased delinquent behavior among students exposed to it. Because of inconsistent and conflicting results, it is suggested that GGI be applied only in experimental contexts and that research should involve near replications to determine how and why observed effects occur. 7 tables and 60 references. (Author abstract modified)