To date, only a limited number of researchers studying children and adolescents who murder have attempted to construct classification systems or typologies to describe these juvenile offenders. Most of the classification systems have been developed post hoc. This has been possile when the researcher has been able to discover enough distinctions or commonalties in a sample of youthful homicide perpetrators to delineate subgroups. Still others have taken existing classification systems designed for other purposes or populations (DSM criteria, FBI typology-of-offender criteria) and examined whether these can successfully account for observed differences among youthful homicide offenders. Although they attack the issue from different angles, researchers using either approach share the belief that it is important for the field to specify underlying assumptions and to study a problem within a clearly stated conceptual framework. When considering the high degree of heterogeneity among preteen homicide offenders, a substantial value should be placed on these efforts.
Before proceeding, however, an argument against the use of classification systems should be discussed. Specifically, considering the relative infancy of the field as a whole, it could be argued that classification systems are based on insufficient data and, therefore, run the risk of prematurely shifting the focus in a direction that may not lead to the clearest under-