Monday, August 7, 2017


I'm working with the 2016 ANES data on something completely different and came across a question asking respondents:

How many guns do you or anyone else living here own?

For the same of full transparency, I'd answer one. I've had it for years.

The data is fascinating. Nearly two-thirds of 3,468 respondents who answered the question answered zero (actually, 65.7 percent). That means everyone else has a gun, but how many? Like me, more answered one gun than any other response (10.2 percent), followed by two guns (7.3 percent), and so on. Like a lot of surveys, you get weird answers as well. Two people said they own 50 guns and one person reported owning 99 guns. That's a lot of guns, except I'm willing to bet it was a survey response tossed off with little or no thought. I suspect they own a lot, but even to me 99 seems unlikely. Twenty, yeah. Ninety-nine? Not so much.

For fun, let's treat this question as an interval-level variable and correlate it with certain other variables. You know, stuff like age, education, and political ideology. Below are some results.

  • Education: unrelated to how many guns you own
  • Slight positive relationship, older equals more guns. Barely.
  • Same as above for income. There, but not strong.
  • The more you identify with the GOP, the more guns ya own.
  • Ditto to above for being more conservative.
We could go on and on with this, correlating this with a jillion variables available in the data, but I doubt we'd be surprised by the basic results. There was also a slight positive relationship with how many guns you own and how satisfied you are with democracy (r = .04, p<.01) and how fair you felt the 2016 election was (r = .07, p<.001). That's comforting, knowing the gun fans are happy.

From a media standpoint, it's no surprise that there's a small but significant correlation between how many guns you own and watching Fox News hosts Sean Hannity (r=.08, p<.001) and Bill O'Reilly (still on the air in '16, r = .09, p<.001). No relationship with watching Anderson Cooper or Chris Matthews.

A more useful analysis, to be honest, is recoding that gun question so it has two levels -- don't own a gun, or do. Really the number of guns measures those who hunt, in which it's not unusual to have several guns depending on the kind of hunting you're doing at that time.

No comments: