Throughout the past decade, numerous states have passed legislation to prohibit the sale of violent video games to children, usually in conjunction with an argument that exposure to violent media increases violent behavior. However, the link between video games and violence is not yet fully understood. This study uses propensity score matching as a method to more adequately address the underlying issue of causality. Using a sample of 6567 8th grade students, these analyses test whether there is a causal link between playing violent video games and violence, non-violent deviance and substance use. Results indicate a substantial decrease in the relationship between video games and these outcomes when a matched sample is used. This suggests that the strength of evidence supporting a relationship has likely been overestimated using other methodologies.