This morning a lot of websites are touting a new study in Pediatrics that claims various forms of gun control reduce “pediatric” firearm deaths. By this the researchers mean deaths among all Americans aged 21 and under. (As I’ve pointed out before, also in the context of gun research in Pediatrics, even when you limit the age range to 18, you’re mainly talking about older teens, not young kids.)
Normally, if you want to know what effect a law has, you look to see what changed before and after it went into effect. Do states that enact the law experience different trends relative to states that don’t? This study doesn’t do that; though it combines five years’ worth of data, it just compares gun-death rates across states with various scores from the anti-gun Brady Campaign. It’s “cross-sectional,” in other words. It doesn’t tell us whether the laws actually do anything, or if states with fewer gun deaths are more likely to enact the laws to begin with.
You can add assorted “control variables” to make these comparisons a bit better, and the authors do, but at the end of the day this is not compelling. It’s not rigorous enough to even be included in the RAND Corporation gun-study review I wrote about last year, which quite sensibly excluded cross-sectional research entirely.
In addition, because they’re focusing on “gun deaths,” the researchers don’t take account of substitution effects; it’s possible for a law to reduce gun deaths without reducing total deaths, if people switch to other methods of suicide and homicide. And on top of that, of the three specific laws they analyze beyond the overall Brady scores, one of them (microstamping/ballistic fingerprinting) has been passed in just two states and another (background checks for ammo purchases) in three.
Just to troll, though, I’ll point out that the study included gun ownership as a control variable, and the result for that variable is statistically insignificant, meaning we can’t be confident that gun ownership is correlated with gun deaths at all. If anything, in fact, the result suggests that having a gun-ownership rate above the national median reduces gun deaths by 4 percent, coincidentally the same reduction the researchers claim for a ten-point higher Brady score.
In all seriousness, this shouldn’t budge anyone’s priors on gun control in the slightest.