Background: The connection between childhood violence exposure and antisocial behavior in adolescence has received much attention and has important implications for understanding and preventing criminal behavior. However, there are a limited number of well-designed prospective studies that can suggest a causal relationship, and little is known about the magnitude of the relationship. Methods: This meta-analysis provides a quantitative comparison of 18 studies (N = 18,245) assessing the relationship between childhood (before age 12) violence exposure and adolescent antisocial behavior. An overall effect size (Cohen's d) was calculated for each study, an average for the 18 studies, and averages for subsets of analyses within studies. Results: Results indicated a small effect from prospective studies (d = 0.31) and a large effect from cross-sectional studies (d = 0.88). The effect for victimization (d = 0.61) was larger than for witnessing violence (d = 0.15). Conclusions: Effect size varied across studies employing different methodologies, populations, and conceptualizations of violence exposure and antisocial behavior. These findings do not support a simple, direct link from early violence exposure to antisocial behavior but suggest that many factors influence this relationship. (Contains 2 tables.)