I do believe murders also positively correlate with ice cream sales.

I remember asking my first stats teacher. Something like

"So sometimes it's obvious that there is a 3rd reason why two things are strongly correlated so we know it's absurd to assume that one causes the other. But what about when it's not so easy to tell?. When many people believe one does cause the other. ? (Even in scientific literature). How do we know for sure what is actually causal related and what just looks related?".

Short answer. More research. And if possible. Controlled experiment using the sciencetific method. Not always possible.

--

She then went on to explain the research on video game violence and behavior violence in kids who play such games.

This is where things get complicated.

Violent kids (behavioral traits) seem to be encouraged to be more physically violent by engaging with violent games/media.

But kids who aren't already aggressive or violent don't get that effect.

Kids with violent tendencies are more attracted to violent video games and play them more.

Many behavior and social problems are like this.

One type of engagement increases a specific type of behavior that further increases the engagement that perpetuates the behavior.

Things become cycles.

For instance. Depression and anxiety may encourage people to self medicate with recreational drugs. Recreational drugs have long term effects, many of which are increases in anxiety and depression. Which then encourages the person even more to seek out drugs to cope.

Studying phenomena in the world that can't be investigated with controlled lab experiments is actually very complex and they often have to gather a lot more data and use more sophisticated statistics.