Randomly assigned experimental and control groups were used to test the experimental intervention's effectiveness compared to existing institutional services. Experimental subjects received training in consequential thinking, drug and alcohol refusal, drug and alcohol avoidance, social and problemsolving, and self-control skills. A behavioral role- play instrument, the Adolescent Problem Situation Inventory, was used to assess subjects' skill levels before and immediately after the experimental training. Following the 10-week intervention, experimental subjects had significantly higher consequential thinking, drug and alcohol refusal, drug and alcohol avoidance, social and problemsolving, and self-control skills than did the control subjects. There were no effects of sex, race, offense type, or pretreatment drug use on post-test skill level, suggesting the intervention was effective with a variety of delinquent subgroups. 2 figures, 16 tables, 144 references. (Author abstract modified)