RANDOM VERSUS NONRANDOM ASSIGNMENT IN THE EVALUATION OF TREATMENT FORCOCAINE ABUSERS

Cocaine-abusing patients randomly assigned to day-hospital or inpatien t rehabilitation were compared with patients who self-selected these t reatment settings Co examine differences in substance use and psychoso cial outcomes under experimental and nonexperimental designs. There wa s little evidence of setting or assignment effects or Setting x Assign ment interactions over the 12-month follow-up period. However, Assignm ent x Time interactions were obtained with 2 measures of cocaine use a nd measures,pf family-social and psychiatric problem severity. These i nteractions indicated greater problem severity at intake among the ran domized patients coupled with greater improvements by the 3-month foll ow-up relative to the nonrandomized patients. Findings suggest that ra ndomized studies of treatment for cocaine abuse may produce somewhat l arger estimates of improvement than what is observed in more typical t reatment situations.