Objective: To compare the predictive validity of different diagnostic classifications for disruptive behavioral disorder (DBD) in youths aged 9 to 15 years of age: the DSM-IV categorical diagnosis, the number of DSM-IV DBD symptoms and Achenbach’s taxonomy (Child Behaviors Checklist scores for social problems, rule-breaking behaviors, aggressive behaviors, and attention problems). The incremental validity, after including clinical structured interviews with parents and children into models that contained CBCL scores, is reported. Method: Longitudinal study in a Spanish high-risk population, comparing each classification in 1 and 2-year follow-ups. Both predictors and criteria pertain to measures of the same class of behaviors: clinical information regarding impairment at school and community. Results: Predictive accuracy was dependent on the specific outcomes and the children’s age. The CBCL scores obtained the best predictive results. Dimensional approaches (the number of DSM-IV DBD symptoms and CBCL) obtained higher R 2 values than categorical DSM-IV diagnosis. Assessment based on clinical interviews with family significantly increased the predictive validity of Achenbach’s taxonomy. Conclusions: These results provide guidance for detecting early disruptive behaviors and planning measures for the prevention of subsequent antisocial conduct. This work may provide direction to future validation studies of classifications and diagnostic systems for behavioral adjustment in youths.