The think tank’s incredibly extensive review of the evidence on gun control might have some staying power in the national conversation on this topic. I’m still sorting through it myself, as it’s packed with detail and countless references to previous studies, but I wanted to recommend it to fellow gun nerds and point out a few things that readers — as well as those who’ve seen articles about the review elsewhere — should keep in mind.
Here’s how the authors themselves characterize the findings:
We reviewed thousands of studies to identify all available evidence for the effects of 13 gun policies on eight outcomes. After excluding studies that did not meet our criteria for establishing a law’s effects, we found little persuasive evidence for the effects of most policies on most outcomes.
For six of the 13 policies, either we found no studies examining the effects on any of the outcomes we considered or the evidence was inconclusive. However, we found some evidence that seven policies [background checks, assault-weapon bans, child-access-prevention laws, concealed carry, minimum-age requirements, mental-illness prohibitions, and Stand Your Ground laws] affect one or more of four of the outcomes.
So, first of all: Thousands of studies have been done, though only a small proportion of them (62) were strong enough to be included in RAND’s review. It’s simply not true that no one is studying gun policy today thanks to a “ban” on gun research by the Centers for Disease Control, as one often hears from the left; the real problem is that this is an incredibly difficult topic because so many confounding variables can affect violence, and the review is limited to studies that can plausibly claim to identify causation, not just correlation.
It may be the case, of course, that we could do more or bigger studies if the government increased funding, as RAND advocates. I would not personally object to more government-funded research on gun policy — though I would keep the CDC itself away given its notorious anti-gun bias.
Further, if you start digging into the areas where RAND found evidence that a policy had an effect — with the strength of the evidence ranging from “limited” (which can mean just one study that hasn’t yet been contradicted by another) to “supportive” (three studies using at least two different data sets), and areas with conflicting studies deemed “inconclusive” — you find that even there the work is often pretty muddled.
The overall table they provide, for example, cites “moderate” evidence that background checks reduce “suicide,” as well as “supportive” evidence that child-access-prevention laws do. Those levels of support, however, apply only to the reduction of gun suicides. Since people without access to guns might kill themselves through other means instead, this finding is not all that helpful by itself. And in both cases, if you click through to their full analysis, you see there’s only “limited” evidence of an effect on total suicides.
For the record, I support both policies, albeit with some caveats, and I’m dubious that non-gun suicides will totally make up for a decline in gun suicides. (Also, the substitution issue is discussed in a lot more detail in RAND’s full report.) But studies that don’t even include non-gun suicides are simply not serious.
What about “moderate” evidence for an effect of background checks on violent crime? That’s only for dealer checks, which are already federal law. Evidence for an effect from private-seller checks, which are what we’re actually debating today, is “inconclusive.”
You find a similar thing when you look into the claim of “limited” evidence that right-to-carry laws increase violent crime. That’s just one study of violent crime in general, which RAND evaluated in isolation despite having tons of studies on specific types of violent crime. Decades’ worth of evidence is “inconclusive” as to whether whether right-to-carry affects total homicides, firearm homicides, robberies, assaults, and rapes. But RAND’s ultimate conclusion is that there’s “limited” evidence that the laws increase violent crime, to the delight of WonkBlog.
This is a formidable review, including a 380-page report and an interactive website, that I plan to learn a lot from. But it needs to be read very carefully.