Sunday, June 10, 2012

Do More Guns Increase the Homicide Rate? A Look at Worldwide Data

Gun violence has been a hot-button topic in American society for many years. You cannot turn on the news or read a newspaper without hearing about the rise of gun violence, although you shouldn't tell the fear-mongering media that the murder rate is at a thirty-year low. Gun control advocates, such as those at the Brady Campaign to Prevent Gun Violence, want you to think that it's too easy to get guns in this country and "where there are more guns, there are more deaths." A friend of mine recently tried to persuade me that the United States has a significantly higher homicide rate because guns are more prevalent here, whereas Japan's low homicide rate is caused by its strict gun control laws. I can easily counter that Switzerland is an example of where gun ownership is amongst the highest in the world, but where its crime rates are nevertheless really, really low. As such, I began to speculate as to whether more guns lead to more gun violence, particularly from an international standpoint.

I typically become suspicious when people say that "X is the cause of Y." Although there are occasions in which such a scenario occurs, most scenarios make us realize that we live in a multivariate world, which means that there are multiple factors that cause something to occur. However, let's take the Brady Campaign's claims seriously for a moment and see if some sort of correlation can be translated into a causation. How would we go about proving that higher gun ownership leads to higher gun violence, or at least correlation exists between the two? Two data sets would be required: homicide rate and firearm ownership rate. The reason to go with rates, as opposed to the firearm and homicide counts, is because with rates, we can measure the frequency of both occurrences [per 100,000 citizens], which is necessary to account for population differentials. Clearly, the most recent and most thorough data is optimal for such analysis. The UN Office on Drugs and Crime (UNODC) put out its 2011 homicide report, and has data dating back to 1995. However, I was unable to find firearm ownership data that recent. The Small Arms Survey, which is a Swiss-based independent research project focused on small arms, collected data from the year 2007. In terms of country coverage, the Small Arms Survey included more nations than the UNODC. My data collection includes countries that were in both data sets, which included 91 countries (N=91). It would have been nice to have included all nations, but 91 is still a large enough sample size to obtain the sought-after results. I compiled the data and ran a linear regression. The information is below:

What were the results? To determine correlation, one looks at R-squared, which is the coefficient of determination that measures correlation. R-squared is measured from 0 to 1. An R-squared of 0 signifies no correlation whatsoever, where as an R-squared of 1 means a perfect correlation. Upon looking at this data, R-squared is 0.0266. Such a low coefficient can only mean that the correlation is so low that the relation between increased firearm ownership rates and increased homicide rates is virtually non-existent.

To put it into plain English, having more guns around doesn't cause more homicides. Looks like the NRA was correct when they came up with the slogan "Guns don't kill people; people kill people." 

No comments:

Post a Comment