Monday, 23 February 2015

No relationship between homicide rate and gun ownership across US states

The Economist recently ran an article about suicide in the United States, which presented a chart showing a strong positive association between suicide rate and gun ownership across US states. This led me to wonder whether a similarly strong association exists between homicide rate and gun ownership. I took data on gun ownership from the same source as The Economist, namely a paper by Okoro et al. (2005) published in the journal Pediatrics. And I took data on homicide rates from the FBI's Uniform Crime Report for 2002 (the same year as the gun ownership data). As the chart below indicates, there is no relationship between the two variables: r = –.05 (p = 0.7, n = 50). I have previously noted that there is also little apparent relationship between homicide rate and gun ownership across countries. 


  1. Maybe control for population density and also limit murders to those committed using a firearm. This paper finds that states with top gun ownership rates have x3 murder rate.

    1. Thanks for the link. When there's no raw association at all, I'm pretty skeptical. North and South Dakotans are apparently armed to the teeth, but they have homicide rates on par with Western Europe. Also, total homicides is the relevant measure, rather than homicides committed with a gun, since people presumably select into alternative methods of killing when guns aren't available.