The statistical approach estimated a generalized regression model in which an indicator of initial victimization is excluded. Data were obtained from the first five waves of the National Youth Survey (NYS). The results showed that assault, robbery, vandalism, and larceny victimization incidents among adolescents and young adults are disproportionately concentrated among relatively few victims. Risk is correlated across year-to-year time periods. Repeat incident victims have a high probability of remaining victims in the following year. Their likelihood of becoming a nonvictim of robbery, larceny, and vandalism increases over time; there was no such pattern observed for assault. The results suggest the need for the development of more detailed specifications of both state dependence and heterogeneity processes. 6 tables, 18 notes, 35 notes, and 2 appendixes