I thought I’d infuse some (very simple) empiricism into the evolving gun debate. Using data readily available on Wikipedia on gun-related deaths across countries and gun ownership across countries, I simply plotted a simple linear regression using Excel.
(The whole process was very simple. If you ever want to try something like this, just save any Wikipedia entry that has a table of data, then open it w/ Excel. It took me about two minutes to match up the countries, then another minute to make the chart above. So there’s no excuse for not using empirical data!)
Because (gun-related homicide
death rates) data was lacking for many countries, it ended up matching up only 73 countries. I trimmed the sample to limit it to OECD member & observer countries. This is what’s often used as a “peer set” for the United States (countries that we should be on par with on various levels). The result was a dataset of 50 countries. That’s not a massive dataset, but in comparative politics, that’s not too bad (after all, there are only about 200 countries in the world, so this is about a quarter of the total “universe” — plus the sample is made up of mostly “comparable” countries).
Here’s how you read this simple bivariate regression: The red line is a simple regression estimate (you can do bivariate regressions in Excel by choosing the “trendline” option). The little equation (also an option in Excel) gives you the estimation. It’s basically the old y=mx+b equation you learned in basic middle school algebra. The “slope” is the estimated relationship between the two variables. Notice that the slope goes up: more guns, more gun-related homicides
deaths. The R2 (R-squared) number is an estimation of how much of the total variation is explained by the model (the equation). The R2 value is not too high, but still manages to explain about 25% of the variation in the data.
The red dot is the United States, which has the third highest gun-related homicide
deaths in the 50-country sample. The other two are Mexico & South Africa. Notice, however, that even though (individually) Mexico & South Africa have both high gun-related homicide death rates and few guns per capita, the overall relationship among the 50 countries is still clear: more guns, more gun-related homicides deaths.
In particular, it’s useful to note that all the advanced industrial democracies (that is, all the “first world” wealthy countries) are clustered together: few guns, few gun-related homicides
deaths. Why does that matter? Because unless you believe that American criminals are somehow much more resourceful than criminals in Japan, Germany, or Australia, the claim that “if guns were more restricted only criminals would have guns” is patently ridiculous. Despite strong gun restrictions in Japan, Germany, and Australia, criminals there don’t seem to get their hands on enough guns to commit as many murders.
ADDENDUM: A correction using logarithmic scales and regression through the origin is posted here.
CORRECTION: I originally misstated the main variable. I actually didn’t use gun-related deaths, but rather the more specific gun-reated homicides (I had originally used the first, but then corrected myself, but forgot to change the labels on the graph). Using the specific gun-related homicides removes all accidental gun-related deaths & gun-related suicides.