Shots in the Dark

Gun control's shaky empirical foundation


In November 1988 The New England Journal of Medicine published a study that noted Seattle's homicide rate was higher than Vancouver's and attributed the difference to stricter gun control in Vancouver. Although the study had serious flaws, including the failure to take into account important demographic differences between the two cities, it received generous coverage in two major newspapers known for their sympathy to gun control.

The Washington Post covered the report in a 600-word, staff-written story on page A4 under the headline "Impact of Gun Control Indicated in Medical Study." The New York Times story ("Gun Curbs Linked to Homicide Rate") was about the same length, although it was by a stringer and appeared deeper in the A section.

The Times made up for those lapses with an editorial about the study later that month. Under the headline "Guns Do Kill People," it said "the study appears to buttress common-sense wisdom about public safety [i.e., our position on gun control]."

This month, when a government-appointed panel of experts announced that their comprehensive review of the relevant scientific literature (including the Seattle/Vancouver study) had failed to find evidence that gun control works, The Washington Post gave the story about 200 words in its "Findings" column. The New York Times (D.C. edition) ran fewer than 150 words of an A.P. story on the bottom of page A23, under a tiny headline that gave no indication of the report's conclusions.

So far the Times has not run an editorial conceding that the research review—which by a strictly scientific or journalistic reckoning ought to carry considerably more weight than a single inconclusive study—"appears to undermine common-sense wisdom about public safety." I'm guessing it won't, even though the report was commissioned by the U.S. Centers for Disease Control and Prevention (CDC), usually a gun control booster.

It's natural, of course, to highlight information that fits one's preconceptions while downplaying information than doesn't. With that in mind, it's important to note that the CDC panel's review, in addition to criticizing studies purportedly showing that gun control reduces violence, finds fault with economist John Lott's research on the crime-deterring benefits of allowing people to carry concealed firearms.

The panelists considered 51 published studies examining seven different kinds of laws, including bans on specific firearms, restrictions on who may own a firearm, and waiting periods for gun purchases. They "found insufficient evidence to determine the effectiveness of any of the firearms laws or combinations of laws."

In other words, after more than half a century of local, state, and federal gun control legislation, we still don't know whether these laws do what they're supposed to do. The report's most consistent finding was inconsistent findings: Sometimes gun control is associated with reduced violence, and sometimes it's associated with increased violence.

The world is messy, and it can be difficult to control for all the relevant variables when you're trying to determine the impact of a particular law. Not surprisingly, the CDC panel calls for more and better research, and it cautions that "insufficient evidence to determine effectiveness should not be interpreted as evidence of ineffectiveness."

But it's scandalous that politicians have been legislating in the dark all these years, promising that the gun control solution du jour would save lives when there was no evidence to back up such claims. If gun control laws have any positive effect at all, it must be pretty modest to have escaped documentation so far.

How could it be otherwise? The typical gun control measure is laughably inadequate to accomplish its ostensible goal.

Regarding criminal background checks, for instance, the CDC panel notes that "denial of an application does not always stop applicants from acquiring firearms through other means." Assuming that a buyer with a disqualifying record is seeking a gun to use in a crime, there are plenty of sources where no questions are asked. According to the report, Americans own some 200 million guns, with around 10 million changing hands every year; retail sales account for less than half of these transactions.

Even if gun availability could be dramatically reduced, any restrictions would disproportionately affect law-abiding people. Criminals have no compunction about breaking the law, and they're highly motivated to obtain the tools of their trade, so anything short of a magical spell that makes all guns disappear is not likely to have a noticeable impact on violence.