A structured and reliable diagnostic procedure based on a revised version of the Diagnostic Interview Schedule for Children for children, parents, and teachers was used to assign both DSM-III and DSM-III-R diagnoses to 177 outpatient boys aged 7 to 12 years. Compared to their DSM-III counterparts, DSM-III-R oppositional defiant disorder was 25.5% less prevalent, DSM-III-R dysthymia was 37.8% less prevalent, and DSM-III-R conduct disorder (CD) was 44.3% less prevalent. However, DSM-III-R attention deficit hyperactivity disorder was 14.4% more prevalent than DSM-III attention deficit disorder with hyperactivity. The two definitions of CD were compared to exemplify an empirical approach to diagnostic validation. The DSM-III-R diagnosis of CD appears to be more valid as it is more strongly associated with police contacts, school suspensions, and history of antisocial personality disorder in the biological father, but both CD diagnoses are associated with family histories of criminal convictions.