We compared outcomes during and after treatment for mixed substance dependent patients (N=143) randomly assigned to a high-structure, behaviorally-oriented (HSB) or a low-structure, facilitative (LSF) individual counseling style. We hypothesized that patients with different coping characteristics would respond differently to the two styles of counseling. Patients were treated in once-weekly individual HSB or LSF counseling for up to 12 weeks. Outcome measures included patient and counselor ratings of benefit, retention, symptom reduction, and negative urines; follow-up assessments included control of substance use and psychosocial adjustment. While no differences in outcomes during or after treatment were found for the HSB and LSF patients, both groups did improve equally. Contrary to our hypothesis, our coping measures did not predict different outcomes for patients treated in the LSF and HSB styles. Post-hoc analyses, however, revealed that outcomes could be predicted in each style from patterns of pretreatment characteristics, which included measures of coping strategies, psychological characteristics, and treatment readiness. Moreover, the patterns associated with positive outcomes were different for the HSB and LSF patients: high treatment readiness was most important for success in HSB counseling, while low psychiatric severity and positive coping styles were important for the LSF clients. The finding of no HSB-LSF outcome differences calls into question the exclusive emphasis on behavioral treatment approaches by the present-day managed care industry. Also, the traditional approach to matching studies, ie, employing one patient characteristic at a time to predict differential outcomes for particular treatments, may be simplistic. An alternative approach employing multivariate statistical procedures to predict outcomes from several patient characteristics may hold more promise.