Using latent class analysis to identify aggressors and victims of peer harassment.

This study used latent class analysis (LCA) to identify and classify individuals into aggressor and victim latent classes. Participants were over 2,000 sixth grade students who completed peer nomination procedures that identified students who had reputations as perpetrators and/or victims of physical, verbal, or relational harassment. Results showed five latent classes. Consistent with previous research, LCA identified latent classes of victims, aggressors, and socially adjusted students. However, rather than a single aggressive-victim subgroup, LCA identified latent classes of highly-victimized aggressive-victims and highly-aggressive aggressive-victims. Comparisons showed differences in mean profiles and classification criteria between LCA and traditional dichotomization approaches. Adjustment outcomes showed that highly-victimized aggressive-victims generally experienced greater negative psychological and social adjustment outcomes than highly-aggressive aggressive-victims. Implications of these findings for better assessment of victim and aggressor subgroups were discussed. Copyright 2008 Wiley-Liss, Inc.