One day at 5:30 a.m., a man with a machete hacked at two doors, leaving a trail of broken glass, to force his way into a DeKalb County, Ga., home. What he planned to accomplish once inside, though, we’ll never know: The homeowner had a gun, confronted the intruder, and fired two shots, killing him.
This is what’s known as a defensive gun use, or DGU. It is abundantly clear that such things happen regularly in this country, which should not surprise us: We have 323 million people, at least as many guns, and plenty of crime, so periodically a person will use a gun to stop a crime. The National Rifle Association’s “Armed Citizen” report, from which this case is drawn, compiles several incidents each week from local news stories.
But exactly how many times do things like this happen each year? Often enough to provide a big potential upside to buying a gun? Often enough that we should worry about various proposed gun-control measures’ reducing the number?
It turns out that it’s a lot harder to count DGUs than it is to count, say, murders, each of which leaves a dead body and, almost without fail, attracts attention from the authorities. And thus the issue has been hotly disputed among researchers for decades.
Perhaps the simplest approach is to conduct a survey asking people whether they’ve used a gun defensively within some preceding time period. Most famously, a survey conducted in 1993 by Gary Kleck and Marc Gertz estimated that more than 1 percent of American adults had used a gun defensively in the past year. This implied something on the order of 2.2 to 2.5 million defensive gun uses annually. Several other surveys conducted around the same time suggest roughly the same thing (including one effort by the Centers for Disease Control that went unpublicized until Kleck stumbled upon the data in 2018, two decades after the questions had been asked).
Especially noteworthy is that while some DGUs can be quite dramatic — the FBI reports that of the 50 “active shooter” incidents in 2016 and 2017, armed civilians stopped four and intervened unsuccessfully in two more (one civilian was wounded, the other scared the attacker off to a new location, where he continued shooting) — most don’t involve firing the gun. Producing a weapon is usually enough to scare off your run-of-the-mill lowlife.
But these are old studies, and it’s not obvious what a survey like this would say today. Crime has fallen drastically since the 1990s, and gun ownership has dipped by some measures as well. However, gun owners today are more likely to own handguns instead of rifles and shotguns, and also far more likely to have concealed-carry permits, thanks to the spread of liberal “shall issue” laws. (In a small but growing number of states, a permit isn’t even required.) Some call it “Gun Culture 2.0”: less focus on hunting, more on self-defense and carrying in public.
And there are serious objections to these surveys aside from their age, the largest of which is that when you’re trying to estimate a very low proportion — which, by all accounts, the share of Americans who use a gun in self-defense each year is — the potential for overestimation is much higher than that for underestimation. For example, if the true number is 0.5 percent, underestimation can result from wrong answers by just 0.5 percent of the respondents, while overestimation can result from wrong answers by 99.5 percent. There are reasons that both groups might answer incorrectly, but the fact that one group is so much bigger than the other makes these risks lopsided.
Further, as the RAND Corporation pointed out in a recent review of the gun-violence literature, if these survey results are accurate, a lot of gunshot injuries — an implausibly high number, arguably — are not resulting in deaths or hospital visits: The Kleck and Gertz estimates “suggest that, while using a firearm for self-defense, U.S. residents likely injured or killed an opponent 207,000 times per year, but only about 100,000 people die or are treated for gunshot injuries in hospitals each year, most of whom either shot themselves or were victims of criminal assaults.”
Kleck has defended his work in great detail for a quarter of a century now, but 2 million is probably too high. Is there a more accurate method?
Some anti-gun researchers have relied instead on the National Crime Victimization Survey (NCVS). As the name suggests, this survey asks Americans about their experiences with crime; it also asks some victims what, if anything, they did to resist. Research with this survey produces drastically lower results: Maybe 100,000 DGUs a year, even fewer in some studies.
But if Kleck-style surveys have a high risk of overstating DGUs, the NCVS obviously produces an underestimate. It doesn’t ask all of the respondents whether they defended themselves against a crime; in order to get to that question, folks have to report (in the course of answering “screener” questions) that they were the victim of crimes or attempted crimes involving personal contact. “Among the potentially excluded respondents,” RAND notes, “are those reporting incidents involving other crimes (e.g., trespassing, commercial crimes), victims of crimes in the included categories . . . who did not report those crimes earlier in the interview, and those reporting incidents that were not completed crimes (e.g., suspected crimes).” On top of that, respondents are simply asked what they did in response to the crime; they are not asked specifically about gun use.
It seems fair to treat the NCVS estimates as a floor and Kleck’s as a ceiling, which gives us a range of about 100,000 to 2 million DGUs per year. With such a wide range, unfortunately, we can’t even say whether defensive gun uses outnumber criminal ones, which the NCVS estimates at around 500,000 per year these days, down from two to three times that in the early to mid 1990s, when Kleck’s surveys took place.
And, of course, none of this should minimize the fact that guns are used quite often as a tool for evil. When it comes to lethal uses of guns, for instance, the criminal ones clearly outnumber the defensive ones: For 2017, the FBI records 299 justifiable firearm homicides by civilians vs. 10,982 murders and non-negligent manslaughters with guns. The former number is underestimated for a variety of technical reasons, but that’s an enormous gap, and there are tens of thousands of gun suicides each year as well on the “con” side of the balance sheet.
So what should gun owners, and those thinking about joining us, make of all this? However imprecise the numbers may be, we can certainly dismiss claims that guns are practically never useful for self-defense. Guns are used in self-defense probably hundreds of thousands of times per year.
Beyond the aggregate data, we must look to our individual circumstances: the crime rates in the neighborhoods we live in, whether we have to travel at night in dangerous areas, whether we’ve been threatened, our mental health, our criminal history and anger issues, our shooting abilities and training, and so on. All else equal, the greater the threats we face, the less our likelihood of misusing a gun ourselves, and the greater our ability to deal with a threat effectively, the more likely it is that the benefits of gun ownership will outweigh the risks. Some of us will quite rationally decide that we’re safer with a gun than without — however infuriating this may be to those who would minimize the value of bearing arms.
This article appears as “Armed Citizen” in the September 30, 2019, print edition of National Review.