Google+
Close

The Corner

The one and only.

The Problem with Social Science



Text  



Do violent video games contribute to violence?

Just as Southington, Conn., steps in with a symbolic burning of video games, social scientists are cuing up to say, “There’s no evidence.”

There’s no evidence that gun control stops mass shootings either, but that doesn’t seem to faze anyone. Note that social science isn’t really required to prove things congenial to cultural liberals. (Changing the standard of evidence is one of the ways liberals imagine themselves to be scientific while conservatives are know-nothings.)

But I was struck by this study, one of whose authors wrote an essay in December claiming video games don’t contribute to violence.

The key problem with past studies suggesting a link, these authors say, is they don’t control for possible confounding factors. It’s not the video games it’s the fatherlessness. It’s the lack of parental support. Or in this case (since fatherlessness was not studied) once you control for “antisocial traits and current depressive symptoms, or by family and peer influences,” video games don’t matter.

Well, maybe so. I’m certainly open to that. But what social scientists who say stuff like this often miss is that “controlling for” the confounding factors may be controlling for the pathways through which something has an effect. I’m guessing fatherless boys are left alone to play a lot more video games, and often choose violent ones because boys like violence and aggression, especially when they feel masters of it.

In the old family-structure debate we referred to this as the “per se” defense. Marriage doesn’t matter “per se” once you control for income, parental support, maternal depression, relationship commitment, parental supervision, warmth, relationship with father. Well yeah. You are controlling for the pathways through which marriage matters.

Plus it’s often not one thing, but many things, that matter, all at once.



Text