Resolves once we have better sociological and psychological data and there's a decent consensus.

Some argue that playacting at violence can get them into a habit of resolving disputes that way, or that seeing graphic scenes desensitizing them and makes them less averse to causing such things themselves, or that movie hereos using violence glorifies it. Others argue that perhaps violence in media could deceases a person's chance of committing violence, by satisfying their violent urges in a fictional setting, or letting them see how horrible the aftermath is. Or maybe it just doesn't matter.

If there's a difference between specifics, I'll resolve based on the overall effect. e.g. if video games increase violence but movies have no effect, I'll resolve to "increases violence". If movies increase the chance but video games decrease it in about equal measure, I'll resolve to "no effect".

This is about the causal effect, not correlation. If people who are violent in real life are more likely to enjoy violent games because that's what they're used to, that doesn't matter.