A latent variable modeling approach to identifying subtypes of serious and violent female juvenile offenders.

Females have recently become an important population in research related to serious and violent juvenile offending. Although a small body of research exists on girls in the deep end of the system, very few studies have examined the degree of heterogeneity within high-risk female samples. This study applied latent class analysis (LCA) to identify subgroups of female juvenile offenders based on their self-report of offending profiles (N=133). Results supported a three-class solution with subgroups characterized by patterns of 'violent and delinquent', 'delinquency only', and 'low' offending patterns. The LCA solution was replicated in an independent sample of high-risk females. The 'violent and delinquent' class was characterized by significantly higher rates of DSM-IV diagnoses for internalizing disorders, affect dysregulation, exposure to violence (within the home, school and neighborhood), and familial histories of criminality. Implications for future research, policy and clinical practice are discussed. (c) 2007 Wiley-Liss, Inc.