There is at present a lack of useful theory to guide policymaking in the juvenile justice system. While the causes of delinquency are important, it is useful to include consideration of the fact that most juveniles discontinue their criminal behavior after some brief experimentation with it. Instead, because there is little research to explain this phenomenon, most policies and programs of the juvenile justice system seem to be predicated on the notion that the unlawful behavior will persist without some kind of intervention by the system. A greater attention placed on classification would lead to the system allocating scarce resources to those delinquents that are less likely to drop out of crime. A study of 932 juveniles indicated that many quit crime while still juveniles. Of those who committed adult felonies, almost all did so in early adulthood. The juvenile Predictive Attribute Analysis (PAA) model can be usefull to researchers interested in studying the 'aging out' phonomenon by pointing out rather specific categories where maximal differences in likelihood of return to a correctional facility occur. Status offenders over 16, for example, are highly unlikely to enter into adult crime and can therefore be safely keep out of the juvenile system with little if any harmful effect. Attention to the 'aging out' phenomenon will serve two important purposes. First, the knowledge will bring about refinement and reformulation of existing theory about the causes of delinquency. Second, policies and programs will result to address the segments most in need of help. Twenty references are included.