In my data are questions asking a national sample of U.S. adults if they own a dog and if they own a gun. They're throwaway questions, not really meant for any serious analysis. So of course I put them up to (semi) serious analysis.
Here's the main finding -- if you own a gun, you're likely to also own a dog. Among those who own a gun, 59.6 percent also have a dog. Among those who don't own a gun, 41.9 percent have a dog. Despite what the NRA thinks, not everyone is packing. About a third in the national survey say they own a gun. A little under half own a dog. Fido is more popular than Beretta.
What separates those who own a dog (but no gun) from those who own a dog and a gun? The obvious stuff. Owners of both are more conservative, more likely Republicans, are older, of higher income, more likely to be white, more likely to watch Fox News, and more likely to feel financially insecure. Education plays no role, and neither does a measure of neuroticism. I don't have a good measure of rural versus urban, but I'm sure that would play a role as well.
For fun I tossed these into a regression model to predict owning a gun. This allows various factors to statistically control for one another to see which variables win out. The result?
- Factors that reduce the likelihood of owning a gun: education, being female. That's it. Those two.
- Factors that increase the likelihood: income, being white, being Republican, and uncertainty about your financial future. Oh, and owning a dog. You got it. After this and other controls, owning a dog still predicts owning a gun. I should point out if we flip the analysis and make owning a dog the dependent variable and plop owning a gun among the list of independent (predictor) variables, it still is significant.
If I were really interested, and I'm not, I'd dig deeper and look at other likely variables. But let's face it. I'm doing this instead of real work.