Studies on the frequency distribution of victimizations have significantly advanced our understanding of risk factors associated with repeat victimizations, yet limited attention has been given to the existence of an excessive number of non-victims in victimization data. Using the Korean Youth Panel Survey, this study tests whether including an immunity effect into statistical models better explicates adolescent victimization risk factors and estimates the determinants of youth repeat victimization in South Korea. In this study, we find that accounting for an immunity effect identifies not only predictors of immunity but also better exposes risk factors for victimization. The results from the zero-inflated models show that social attachment and contextual factors play a critical role in deciding both risk and immunity, while self-control, self-esteem, and the lifestyles of students have limited influence on school victimization.