The study found significant and stable levels of differentiation between offenders and victims of violence; and this differentiation was predictable with explanatory variables. The evidence of stability over time indicates that theory must account for why some individuals remain primarily offenders while others tend to remain victims. Further, the data indicate that some explanatory variables are significantly associated with the differentiation between offenders and victims. The article also suggests possibilities for future research using Osgood and Schreck’s statistical method in defining differentiation in offending versus victimization at the individual level. The research involved two waves of the public-use version of the Add Health, which provides data on a variety of health and social issues for a nationally representative sample of adolescents who attend school in the United States between grades 7 and 12. The authors used the statistical approach that Osgood and Scheck (2007) developed for analyzing specialization in violent versus property offending. This approach was used to analyze tendencies to gravitate toward violent offending compared with a tendency to become a victim of violence. In this procedure, the differentiation into victim and offender roles was an individual-level latent variable while controlling for confounding between the likelihood that individuals would take either the offender or victim role in violent acts, as well as their overall number of encounters with violence as either an offender or victim. 5 tables, 78 references, and 1 appendix