The lack of rigorous evaluations and variations in operation of adolescent diversion programs have led to a state of confusion concerning the effects of diversion. In a program at Michigan State University, youths referred to the project from the juvenile court are randomly assigned to undergraduates who are trained and supervised by the project personnel or a control group which receives treatment as usual. The standard 18-week intervention involves a combination of child advocacy and behavioral contracting. Results have been positive but not conclusive. Project youth tend to have fewer and less serious police and court contacts and better school performance when compared with control youth. A major focus of the research has been to identify the conditions under which such a diversion program can be successful. What is done with the adolescent after diversion appears to affect potential reinvolvement with the system. Variations in this and other variables, e.g., where in the system the youth is diverted, who administers the program and the type of youth diverted, are likely to account for the wide variety of results found by diversion programs. Two references are cited in the paper. (Author abstract modified).