, , , , , ,

I apologize up front for a long post; hopefully its substance will make it worthwhile.

Over a year ago now, the Daily Beast published an article in which they compared gun fatality rates with laws protecting gun freedom in the various states in the Union. I know I’m reading it late, but it’s new to me, and I have a thought about it I’d like to share.

In their piece, they did no real comparison. They simply put up the two lists side by side, allowing readers to see a lot of freer states suffering more gun-related deaths per capita and a lot of less free states enjoying fewer such deaths. In fact, they encouraged widespread poor statistical comparison by posting just the top five deadliest states and the bottom five deadliest states — and hiding the full list behind a link. And when it comes to causality, they’re quick to suggest that more restrictions on guns lead to fewer gun-related deaths.

I thought I’d look a little closer at their numbers.

I trusted their reporting of gun-related death rates by state. I accepted that those figures probably included suicides (which account for most gun-related deaths and which, I think, would not be affected by how easily one can get a gun).

I also accepted their reporting of the relative level of gun freedom by state, which they seem to have gotten from a group called the Legal Community against Violence. Those numbers are a simple ranking of the fifty states from the freest (Arizona, at #1) to the least free (California, at #50).

Now because the level of gun freedom was given as a ranking, I thought I should rank the states by gun death rates to compare ranks to ranks rather than comparing ranks to rates. I ranked the states from the gun-deadliest (Louisiana, at #1) to the least gun-deadly (Hawaii, at #50).

First I calculated the statistical correlation between the rank of freedom and the rank of death. I got ρ = 0.56 — a definite correlation. The freer states tend to also be the deadlier ones.

But since one of the basic fallacies of interpreting statistics is assuming that correlation implies causation, I looked at other variables. And I found an interesting one.

It turns out that the correlation between auto death rates (per mile driven) and gun death rates (per capita) by state is much stronger than the correlation between gun freedom and gun deaths, with ρ = 0.76.

So if gun freedom yields gun deaths, then surely bad driving yields more gun deaths, right? Not likely, no matter what you think of road rage. Especially since most gun-related deaths are suicidal.

These data suggest to me a different cause for gun deaths. I suspect that people in some parts of the country tend to have more respect for dangerous machines and the damage they can do than their compatriots in other regions.

So there’s a very high correlation between the damage done with the most dangerous machines that most people own (cars) and the damage done by the second-most dangerous machines that most people own (guns).

And that respect for dangerous machines prompts some such states to pass laws regulating their use. But other safer states tend to prefer liberty and self-control to governmental control. So we have relatively safe Vermont (#37 for gun deadliness) enjoying some of the best gun freedom in the country (at #3), while almost-as-safe California (#33 for gun deadliness) has a lot less freedom (at #50).

In other words, I think there’s one thing that makes New England the safest place to drive a car or own a gun, regardless of the fact the some New Englanders want very little gun freedom (Massachusetts and Connecticut) and others want plenty of it (Vermont and Maine). And the absence of that thing is what makes both free Mississippi and unfree Florida so deadly, whether on the road or at the range.

Some people are just culturally more likely to hurt themselves or others with machines.