Thursday, January 13, 2011

Rachel Maddow and Correlation vs Causation

On her MSNBC talk show tonight, Rachel Maddow made the point that there were more gun deaths in states with a higher percentage of gun ownership than states with a lower percentage of gun ownership. As you can see in the image below, the two states with the highest gun death rate also have a high percentage of citizens that own guns.

So, gun ownership causes higher gun death rates, right? Gun ownership causes gun-related deaths. Get rid of guns, get rid of deaths. Case closed.

Well, not so fast. I don't have access to Maddow's research but I am assuming that gun ownership refers to the publicly available data regarding registered guns. These citizens have probably passed a gun safety class and passed a criminal background check.

I did a bit of research and I found something interesting. According to, the District of Columbia has the highest rate of gun related deaths per capita. Interestingly enough, it also has the lowest gun ownership rate per capita, according to a survey done by Behavioral Risk Factor Surveillance System in North Carolina. Only 3.8% of D.C. residents own a firearm. While not a state, the district has its own gun laws like each state in the union. In fact, before the U.S. Supreme Court ruled on District of Columbia vs Heller, it was almost impossible to legally own a hand gun in the District of Columbia.

So, if Maddow was being completely honest, her chart would look like this:
Highest Gun Death Rate % Gun Ownership
District of Columbia 3.8%
Louisiana 46.0%
Mississippi 54.0%
Alaska 61.0%
Alabama 57.0%
Nevada 32.0%

Kind of a different picture, huh?

The right question to ask is what percentage of the gun deaths are at the hands of registered gun owners? But, greater gun ownership leads to greater access. Perhaps. How many of the gun-related deaths were by weapons that were legally registered but subsequently stolen and used in the killing. How many of the deaths were committed by individuals who where licensed gun owners? Only then could we begin to move away from strong correlation between gun ownership and high rates of gun-related deaths and into the realm of causation.

Surely there has been a study done on this? Rachel wasn't the first person to ask this question, was she? Fortunately, there was a study done by Harvard professors Matthew Miller, Assistant Professor of Health Policy and Injury Prevention at Harvard School of Public Health, and his colleagues David Hemenway and Deborah Azrael. It was published in the February 2007 issue of Social Science and Medicine.
These results suggest that it is easier for potential homicide perpetrators to obtain a gun in states where guns are more prevalent. “Our findings suggest that in the United States, household firearms may be an important source of guns used to kill children, women and men, both on the street and in their homes,” said Miller.
So, after all of that "household firearms MAY be an important source of guns"? Wow. Solid research Professor. Still no causation, huh?

causation vs correlation

As a disclaimer, I am not a huge fan of guns. Having served in the military, I have seen first-hand what they can do. What I am a huge fan of is the Constitution.


Anonymous said...

I couldn't agree more, Doug. Sadly, it's a short step from "only the government can have guns" to "the government uses guns to tell people how to vote".

Go Constitution,

Anonymous said...

^ oh so you intend to stop the government with your guns?...but then why stop there, because i don't think merely a gun could stop the US govt which is armed with the world's strongest why not arm yourselves with nukes also which might give you a fighting chance to protect your 'constitution'.

Anonymous said...

It absolutely does NOT make sense....gang members and those using these guns in this example are killing each other, not the MASS shootings OUR country is famous for worldwide....