More guns = more murders, part II
In the last Debunker, we examined the relationship between gun ownership and the murder rate. We found that the most recent statistics from the Federal Bureau of Investigation showed that annual gun purchases increased by more than 60 percent over the past decade, while the murder rate declined by more than 14 percent. We concluded that either “(1) there is no correlation between gun ownership and the murder rate, or (2) gun ownership is negatively correlated with the murder rate (in other words, more guns means fewer murders).”
Some comments online have argued that we did not prove the point: after all, not all murders are committed with firearms. What if murders by guns increased, they ask, while murders committed with weapons other than guns decreased by larger amount, producing the decline in overall murders.
It’s an interesting question. If the dramatic rise in gun purchases coincided with a large enough drop in murders committed by knives, blunt objects, etc. (perhaps deterred by the increased probability that a would-be victim or bystander might have a gun), overall murders might have dropped, while murders by gun remained stable, or even increased (perhaps due to shoot-outs).
There’s nothing a statistics nerd like the Debunker likes better than an interesting question. After looking into this one, however, we must pronounce it:
The FBI’s Uniform Crime Reporting statistics table building tool doesn’t break down murder rates by type of weapon used. However, the FBI’s Crime in the United States appendix, “Supplementary Homicide Data,” includes Expanded Homicide Data tables for Murder Victims by Weapon (including “Total Firearms”) for 2001-2005 and 2006-2010. Combining this data with the U.S. Census Bureau’s population estimates for 2000-2009 and 2010, we can derive the murder rate (per 100,000 population) for murders by gun alone:
Now we can compare annual gun purchases (as measured by the FBI’s National Instant Criminal Background Check System with this murder rate by gun:
The result is that, while gun purchases increased more than 60 percent, the murder rate by guns declined even more than the overall murder rate: The overall murder rate declined some 14 percent, but the murder rate by gun declined more than 20 percent.
A simple linear regression analysis of these figures reveals a coefficient of correlation of about 0.97.
Of course, correlation is not causation, but a coefficient of 0.8 is considered “strong”; a coefficient of 1.0 is perfect correlation; this coefficient—0.97—is incredibly strong. Moreover, the coefficient of determination here is 0.94—meaning that 94 percent of the decline in the murder rate by gun is explained by the increase in gun purchases; only 6 percent is due to other causes.
Conclusion: The argument that increased gun ownership causes increased murders by guns is unfounded. The evidence very strongly shows that increased gun ownership causes murders by gun to fall. In the absence of countervailing evidence, the presumption must be in favor of increasing gun ownership: Restricting the Second Amendment rights of the citizenry means that innocent people will die. The burden of proof is on the proponents of gun control to show otherwise.