You may have heard about the Pornhub report that indicated that people in blue states watch more pornography than people in red states. To some extent, I think it didn’t get more attention because it goes along with people’s prejudices. Now if it had turned out that people in blue states were more into scatological porn, I would have been surprised because I assume (perhaps wrongly) that people in those states tend to be sexually repressed and so tend to get into what I think of as sick and twisted stuff. But the truth is that there is really no reason to think that porn consumption has anything to do with politics. There are various issues.
Jordan Ragusa addressed probably the most important issue in a brief post, Nickelback, Herpes, and Obama’s Vote Share in 2012. The issue is simply that meaningless correlations show up all the time. And that’s especially true when you are looking at a lot of different things. For example, imagine correlating first names with cancer rates. Because you would look at an enormous number of first names, some would—through random chance—be correlated with high cancer rates. The problem would be even worse if you subdivided the kinds of cancer. Being named Jerry makes you twice as likely to get testicular cancer! Um, no.
Seth Masket addresses the problem more directly in his post, Are Democrats Pervs? Some Problems with a State-Level Analysis of Individual-Level Behavior. He’s actually focused on an important issue that I doubt is at play here: the ecological inference problem. This is the problem that we are looking at aggregate data and trying to say something about individuals. In other words: maybe it’s the Republicans in the blue states who are consuming so much porn. Like I said: I don’t think that’s actually the problem here.
Masket doesn’t seem to think this is the problem either. As he wrote, “Chances are that even if there is an individual relationship here, it’s not a direct one.” That is to say that higher levels of porn consumption and Democratic voting patterns correlate to something else. What I original thought of was youth: people (okay: men) consume more pornography when they are young than when they are old. Blue states have younger populations. Social Security and Medicare are the main reasons why red states take more from the government than blue states. So I suspect that population age is most of the effect.
Consider the youngest state in the nation: Utah. It has far more porn consumption than would be indicated by its 2012 voting. On the other side, we have the oldest state: Maine. It has far less porn consumption than the model predicts. And if you go through all the data (Age and Sex Composition: 2010) you will see this pattern over and over. In fact, as far as I can tell, that is the only thing that we see in the data.
But Masket brings up a couple of other interesting possibilities. One is marriage levels. I’m sure that’s part of it. But I suspect that age and marriage are highly correlated, so I doubt this tells us much more than we already knew. He also mentioned poverty. I don’t exactly see how that works. Certainly unemployment could raise the rates of porn consumption. And he mentions high speed internet availability. That could certainly be a factor; you can’t watch porn if you can’t watch porn.
The main thing is simply that correlation is not causation. Age, marital status, and internet access all provide potential mechanisms. Liberalism really doesn’t provide a mechanism except in the sense of a caricature that social conservatives might have of the abortion-on-demand, sodomy-requiring liberal. And it doesn’t really get us anywhere. So when you hear that Democrats are more interested in pornography than Republicans, remember what that means: the Republican base is dying.
H/T: Jonathan Bernstein, Read Stuff, You Should