Correlation really isn't causation, even when it comes to gun laws
After the Gabby Giffords shooting last January, Richard Florida of the Atlantic produced an interesting graphic purporting to show a correlation between states with one of three gun control laws (assault weapons bans, trigger locks, and safe storage requirements) and lower homicide rates. While he takes great pains to note that the relationship is not causal and merely a positive correlation, the obvious takeaway is that the absence of gun control laws leads to more homicide.
Correlation Between Gun Control Laws and Firearm Deaths
Source: The Atlantic
Florida writes that he finds “substantial negative correlations between firearm deaths and states that ban assault weapons (-.45), require trigger locks (-.42), and mandate safe storage requirements for guns (-.48).” Thus, he asserts, gun control laws are associated with fewer firearm deaths.
Is that finding valid? Who knows? I don’t and you certainly can’t tell from those two data points.
The problem is that we simply can’t tell whether gun control laws cause firearm deaths to go down, whether firearm deaths result in fewer gun control laws, or whether the correlation is purely coincidental. In the second case, it is conceivable that states experiencing more gun deaths pass fewer laws restricting gun ownership because citizens feel safer carrying their own weapons. It is this confusion, though, that makes the simple correlation useless. If states don’t pass gun control laws because they have lots of gun deaths, then of course there is a positive correlation, but the reasoning has reversed. It’s not lax guns laws causing more gun deaths, it’s more gun deaths causing lax gun laws. That belief is certainly subject to debate, but it puts the whole correlation business out of commission.
Now, there are ways to solve this problem statistically. Jens Ludwig at the University of Chicago has a sophisticated paper that I find quite convincing and that looks at a related issue—concealed weapons laws causing more adult homicide. Unfortunately, his work does not fit neatly onto one figure.
You might be inclined to say well, OK, but since the two approaches reach the same conclusion, what’s to worry about? That the two reach the same conclusion is, to me, no more than a coincidence. And making policy choices based on coincidences is a transparently bad idea.
Tomorrow: Stand Your Ground laws lead to miscarriages of justice