There's a chart being passed around the internet showing a correlation between the percentage of homes with guns and the number of gun deaths per capita. This chart originated during the January 14th episode of the of The Rachel Maddow Show in an editorial regarding the recent shooting in Tucson, Arizona.

This chart is plainly making two related statements: gun deaths are high in states with high gun ownership and low in states with low gun ownership. Regardless of the clear reliance on the post hoc fallacy, there are two glaring issues with this conclusion.

1. Why are they only showing 1/5 of the data?
2. Raw data in a chart is ineffective at showing trends.

Giving an incomplete and unclear look at statistics is usually the sign of data not truly supporting a claim. A simple way to test this is to retrieve the full data set and graph it. Though the web site of the cited source, the Violence Policy Center, does contain the data used in the chart by Rachel Maddow, the rest of the data was not. Thankfully, the web site did cite did cite their sources – the CDC and a paper published in The Journal of the American Academy of Pediatrics. Why the show's researchers didn't bother to dig further, as Rachel would say, is something I just can't guess.

Given the red flags outlined above, it's not surprising that the full data set doesn't support the claims.

As you can see, both gun ownership and gun deaths trend upwards, but the rates are so inconsistent that there is no indication of any direct correlation. Even more curious is the missing data point for Washington, DC – despite both sources providing the data. As both the highest rate of gun deaths and the lowest rate of gun ownership, the absence of DC is a glaring omission which requires explanation.

While this discredits the conclusions made by the Violence Policy Center, things are even worse for Rachel Maddow. As you can see in the video above, press release was cited in the context of the Tucson mass shooting which left 6 people dead. The statements made up until the moment the data appeared on screen clearly leads the viewer into believing that the data is limited to manslaughter. In reality, however, it also includes suicide, accidents and legal intervention¹. If we separate the deaths by intent, it deviates further from the claims it was intended to support.

We can now see that the tendency for gun deaths to trend with ownership is a result of suicides, not homicides. Further, the rate of homicides trends down as the rate of ownership goes up, making it contrary to the conclusion made by Maddow.

It is unfortunate that the Rachel Maddow Show chose to run the limited and, frankly, deceitful data without even using it in the correct context.

¹ Deaths resulting from injuries inflicted by the police or other law-enforcing agents, including military on duty, in the course of arresting or attempting to arrest lawbreakers, suppressing disturbances, maintaining order, and other legal actions. Excludes injuries caused by civil insurrections.

Data Sources

Injury Mortality Reports

http://webappa.cdc.gov/sasweb/ncipc/mortrate10_sy.html

Pediatrics -- Okoro et al. 116 (3): e370 Table 1

http://pediatrics.aappublications.org/cgi/content-nw/full/116/3/e370/T1