Some Math About Guns

Trying to get a clearer picture of gun violence in the US. Image courtesy of Mike Maguire via Flickr CC.

Turns out it can be really difficult to understand our collective relationship to guns, gun violence, and gun control. What seems to be obvious to some, runs completely counter to others. This was illustrated nowhere better than in the recent report out of the RAND corporation on gun policy. It studies all sorts of relationships between our attitudes about guns and our impressions of the state of gun violence in the US. An article in Vox gives a really thorough summary of the RAND report, and leaves one with two major impressions: (1) we don’t have nearly enough research on gun policy, and (2) despite the fact that there should be plenty of data about this stuff, opinions about what makes us safe seem to be totally subjective.

When faced with something like the incredibly politically divisive debate around gun violence happening in the US now – and always – it’s helpful to quantify.

Mark Reid, who writes the matching learning and statistics blog Inductio Ex Machina, recently posted some data and plots relating gun ownership to gun violence. He sourced the data form Wikipedia and wrote the plots using the statistical computing software R. A quick glance at the plot below shows that the US owns a whole lot of guns and has a whole lot of gun violence.

Gun deaths per capita versus gun ownership in OECD countries compiled by Mark Reid for Inductio Ex Machina.

Reid’s post also includes several other plots, some that incorporate the non-OECD countries, and some that differentiate between gun deaths and gun homicides. The comments section of Reid’s post is also full of alternative questions prompted by the data – like what’s up with Switzerland? – and lots of useful links for similar analyses.

Based on the same data set, Kyle Kinsburg, who writes the blog Aphyr (pronounced “AY-fur”), recently published several plots relating gun death, gun ownership and economic inequality. In particular, he compares gun homicides to Gini index, which produces a linear looking relationship. This isn’t exactly news, we’ve known for a long time that income inequality is correlated to violent crime for all sorts of reasons. Kingburg does point out that there is something interesting to be observed here about gun homicides and prevalence of guns, namely, prevalence of guns doesn’t tell the complete story. For example, Brazil and Argentina have the same prevalence of guns, but Brazil has nearly 10 times more violent crime.

Graph of gun deaths versus the Gini index complied by Kyle Kingsbury for Aphyr.

The R code for Kingbury’s plots are available on his blog, and the data for gun ownership and gun deaths is available on Wikipedia or as .csv files on Reid’s blog. As Reid points out, and I feel obliged to reiterate, this isn’t a rigorous analysis, but it’s cool that we have the tools and technology to get a reasonably quick quantified sense of the problem.

If this sort of data interests you, last year the podcast Science Vs did an episode about guns that includes a good analysis of the data surrounding guns. It is definitely worth a listen, and it draws attention to the relationship between gun ownership and suicide.