Initial interviews were completed from September 1, 1994, through January 31, 1998. The cluster analysis of summary measures of nine baseline alcohol/other drug use and self-reported delinquency variables identified four groups of youths: low-level delinquents and drug users, high-level delinquents, hair-test-identified marijuana and cocaine users, and self-reported drug users. The validity of this typology was assessed by comparing the clusters of youths regarding their education experiences; delinquency referral history; neglect, abuse, and family problem history; close friends' problem behavior; and emotional/psychological functioning and mental health and substance abuse treatment history. The findings show that youths were experiencing overlapping delinquency, alcohol/other drug use, and emotional/psychological problems. The authors advise that intervention efforts which focus on one problem at a time are unlikely to meet with success. The use of cluster analysis to identify subgroups of youths who possibly need more intensive services highlights its usefulness in the efficient use of scarce treatment resources. Such a use of cluster analysis is particularly appropriate for youths who enter the juvenile justice system. 12 tables and 36 references