Correlation Does Not Equal Causation

I read an article recently in a local parenting magazine. The article briefly discussed violent video games and depression in children. This sort of article drives me nuts because it just reports a correlation between two items rather than actually

offering any actual advice on how to proceed. I totally understand that the researchers who originally studied and reported this information shouldn’t offer advice, but if you’re going to report this kind of research in a parenting magazine, offer some kind of response rather than just parrot the researchers and seem to be implying a course of action.

While it’s nice to know that there’s a correlation between two things, what are you actually trying to tell people? Without analyzing the article critically, the author seems to be telling parents not to let their kids play as many violent video games. That isn’t necessarily going to help, though, since depression and playing those video games are just correlated. In order to justifiably tell parents to limit the amount of time their kids play violent video games, a causation needs to be established. You need to show that playing those games CAUSES depression. Perhaps it’s the case that depressed teens are drawn to violent video games. In that case, limiting the amount of time they play games isn’t going to fix the problem and could just hide it.

Knowing that there is a correlation, what do you expect parents or anyone else to do with this information? Maybe parents should just be on the alert so that if they see their children playing violent video games, they consider whether the child may also be depressed. Parents would then need to know the symptoms of depression.

Just because two things are correlated not does necessitate that there is a causal relationship between them. Leaves falling from the trees may be correlated with people buying gifts, but do falling leaves cause people to buy gifts or people buying gifts cause leave to fall? Nope. There are other issues at play.

So two things are correlated. Now what?