My psych professor in the summer talked about this and laughed at all the fringe lunatics that claim without a doubt that violence on TV and in games causes violence in people.
See, there's two kinds of relationships between two variables: causal and correlational. With causal, you can definitively say that A caused a change in B. With correlational, you can only say that when you change A, there appears to be an effect on B, but there are extraneous variables so you can't make any conclusion without being laughed at by everyone else. You actually get mad props in the scientific community for admitting that you couldn't come up with anything.
Maybe violent kids are attracted to violent video games and play them more than other people? HAHA THE TABLES HAVE TURNED.