Compared with studies of client characteristics and treatment process, limited research has been conducted on how program features of drug-dependence treatment programs may affect client outcomes. Of particular interest are those characteristics of programs that may have a clinically significant impact on outcomes and that are amenable to change within programs. Because of the heterogeneity among studies included in this meta-analysis, the data were analyzed in terms of type of outcome variable (drug use and crime), type of design (single-group and treatment-comparison group), and type of treatment (methadone maintenance, therapeutic communities, outpatient drug free, and detoxification). For the more valid treatment-comparison group studies, the weighted mean effect size was 0.29 for drug use outcomes and 0.17 for crime outcomes. Program factors found to be significantly correlated with effect size in one or more modalities were decade of treatment, researcher involvement in treatment delivery, maturity of the program, counselor/client ratio, treatment implementation, treatment exposure, and methadone dosage. 7 tables, 1 figure, a glossary, 8 notes, 49 references, and appended listing of eligibility criteria for studies