This study attempts to identify the specific demographic and sociological characteristics related to improvement for 48 male alcoholic patients who participated in two contrasting one-month inpatient treatment programs. Of the 48 subjects, 27 participated in a broad spectrum behavior therapy program (experimental group). The remaining 21 subjects participated in a traditional group psychotherapy program (comparison group). To explore the relationship between the characteristics of the subjects who completed each treatment program and therapeutic outcome, a total change score was obtained for each subject. Within each program each of 18 variables found useful in discriminating between outcome categorizations in similar investigations was then correlated with total change. The outstanding finding was the variance among predictor variables having prognostic significance within each treatment program. For the comparison group, married subjects with an occupational skill, a higher IQ, and fewer arrests during the three months prior to treatment demonstrated greater total change than unmarried subjects, unskilled workers, subjects with lower IQ's, and subjects reporting a greater number of arrests. For the experimental group, contrasting findings emerged. Fine discriminations in terms of subject characteristics must be made in order to adequately assign individuals to treatment regimens. (Author/JLL)