Devin Hughes has written up a response to Dr. John Lott’s testimony before the House Judiciary Subcommittee on the Uvalde mass murder. Hughes represents the gun control organization he founded, GVPedia. Unfortunately, his claims are inaccurate. Let’s go through each of his claims in order.
Universal Background Check laws, and particularly licensing laws, would have stopped several mass shootings in the past decade. In his spoken testimony, Lott argued that “There isn’t one mass public mass shooting this century that would’ve been stopped if such a law had been in effect and had been perfectly enforced.” Yet a report issued in 2019 found: “of the 27 deadliest shootings over the last six years where we could identify how the firearms were obtained, a federal licensing requirement may have prevented the shooter from acquiring the firearms used in 52% of the incidents. In these 14 incidents, 169 individuals were fatally shot and 131 individuals were shot and injured.” Specific incidents that may have been stopped by universal background checks include the 2019 Midland-Odessa mass shooting, the 2016 Dallas mass shooting, and several others that are detailed in this GVPedia factsheet.
Mass public shootings are murders in a public place where four or more people are murdered and the FBI excludes “shootings that resulted from gang or drug violence.” For further information about how the FBI defines “public places” and some of the academic research that uses this definition see here. Take the cases that Hughes offers as evidence that background checks would stop mass public shootings.
The 2016 Dallas police murderer was Micah Xavier Johnson. People dishonorably discharged are banned from buying guns, but despite some initial claims, numerous news reports indicate that Johnson was ultimately honorably discharged. That isn’t too surprising, given that he left the service for the offense of stealing one pair of a woman’s panties. For example, from the Chicago Tribune:
“Johnson originally faced removal from the Army altogether, said Texas-based defense attorney Bradford Glendening, which was “highly unusual” since sexual harassment cases typically wind up with a soldier receiving counseling. The case ended in September 2014, when Johnson signed paperwork agreeing to receive a “less than honorable” discharge from the Army, Glendening said. But Johnson wasn’t discharged until April 2015, and Glendening said last week he was told that Johnson received an honorable discharge.” Associated Press, “When Army career ended in disgrace, Dallas gunman was ostracized,” Chicago Tribune, July 15, 2016.
Hughes mentions a protective order against Johnson in his factsheet, but the discussion is misleading. The protective order was “sought” while he was in the military by the female soldier from whom he took the underwear. There is no evidence that the military granted the protective order, let alone that it was in effect when he tried to buy his gun.
The 2019 Midland-Odessa murderer was Seth Ator. The FBI defines a mass killing as four or more people murdered in one place. This attack occurred in many locations as Ator drove along various highways and roads, and in no single place was more than one person murdered. The attacks occurred miles apart from each other.
In his “factsheet,” Hughes mentions several other cases, but none of these are what he claims them to be.
The 2015 Charleston AME Church shooter was Dylann Roof. Again, Hughes gets the facts wrong. While the FBI originally claimed that Dylann Roof was charged with felony drug possession, the FBI later corrected that error. Being charged with a felony that could result in a prison term of at least two years is sufficient to result in a person being prohibited from buying a gun (see copy of relevant part of criminal code below). He can also be banned from buying a gun if he is an addict or admits to being an addict. But he was charged with a misdemeanor and one arrest for possession isn’t enough to establish that the person is a drug addict.
2013 Santa Monica College shooter. Five people died from that mass murderer, though two died in a home, not a public place. The other three murders in public were spread out as the murderer drove around town, with two murders in one location and one additional murder in another. So there is no place where four or more people were murdered in the same public place. There are claims that the California Department of Justice advised Zawahri in an October 2011 letter that he was ineligible to purchase a firearm, but no explanation is provided by Hughes or anyone else for this denial, so it is possible that was in error.
Hughes mentions a mass shooting that occurred in a residence, so it wasn’t a mass public shooting, but no other identifying details are provided for this case.
The National Instant Check System (NICS) for background checks is accurate. In his spoken testimony, Lott contended that 99% of initial background check denials are mistakes. This is egregiously false. Both the FBI and Office of Inspector General find that such denials for purchasing a firearm are 99.3-99.8% accurate.
The US Department of Justice IG report is titled “Audit of the Handling of Firearms Purchase Denials Through the National Instant Criminal Background Check System.” While the entire sample of cases is readily available for evaluating the error rate, the IG report didn’t even pick a random sample. Instead, they picked a very tiny “judgmentally selected” sample. I have pointed this out before to Hughes, but this is not how you do statistics. Here are two quotes from the IG report.
“We judgmentally selected 447 denied transactions and found that only one transaction was incorrect, resulting in a 99.8 percent accuracy rate.”
“Using the judgmentally selected 447 of the 373,900 FBI denied firearm applications (standard and delayed denials) for FY 2008 through FY 2012″
Here are two previous pieces on these errors in the New York Times and the New York Daily News. As Reagan Dunn, the first national coordinator for Project Safe Neighborhoods, a Justice Department program started in 2001 to ensure gun laws are enforced, noted: “This incredibly high rate of false positives imposes a real burden on the most vulnerable people.”
For example, in the fiscal year 2017, there were 112,090 denials, and 12,710 were referred from the national BATF office to the BATF field offices. So 89% of the cases were dropped at that stage. Unfortunately, the Government Accountability Office (GAO) doesn’t provide a detailed breakdown of the cases at the other stages of review. Still, they tell us how many cases are deemed sufficiently strong to prosecute at the end of the process. As of June 2018, US Attorney’s Offices prosecuted 12 cases — just 0.011% of denials resulted in convictions. People in the US DOJ have often told me that these are very easy cases to prosecute. Was the person a felon? Did he state on the 4473 that he wasn’t a felon? If the answers are “yes” and “no,” they could be easily prosecuted. If one talks to people in enforcement, they will tell you that they would prosecute them if they were real cases. There are so few prosecutions (even adding in the few additional state prosecution) because these weren’t real cases.
Hughes has been informed about these points multiple times, but he just keeps citing the DOJ IG report.
The US has a substantially higher mass shooting rate than other high-income countries, regardless of the mass shooting definition used. In his spoken testimony, Lott argued that the US mass shooting rate is lower than the rest of the world. To inflate the international share of mass shootings, Lott’s research includes group attacks by uniformed soldiers, paramilitary groups, terrorist organizations, and massacres by large rebel groups. There are more than 1,000 of these international cases that Lott misclassified as mass shootings. One of these misclassifications is an attack on an entire village on the Uganda-Kenya border by 300 Pokot raiders that resulted in many deaths, the burning of 200 houses, and theft of 300 head of cattle. As Dr. Adam Lankford of the University of Alabama has noted, including such cases is highly misleading and at odds with Lott’s own assertions about his research.
These claims are simply false. You can see the published research available here and here. We followed the FBI/DHS definitions of mass public shootings using the University of Maryland Global Terrorism Database. We excluded any categories that they list murders as involving government-sponsored or war-related shootings. We also report excluding terrorist-related attacks, though there are a number of “terrorist” type attacks in the U.S., such as Fort Hood, San Bernadino, or Chattanooga. But no one argues that we should exclude these attacks from the U.S. count. Still, excluding these attacks from other countries and not from the United States doesn’t dramatically change the results.
Mass shootings do not overwhelmingly occur in gun-free zones, with credible studies finding that only 12-13% of such shootings occur in areas that ban firearms. In his spoken and written testimony, Lott stated that 94% of mass shootings occur in gun-free zones. This conflicts with Lott’s own more recent estimate that 96% of mass shootings occur in gun-free zones. Before the past several years, Lott argued that 98% of mass shootings occurred in gun-free zones. Regardless of which figure is used, it is false. Lott’s research on gun-free zones is marred by massive data errors. When studying the period from 1977-1997, Lott treats every mass shooting fatality as an individual mass shooting, greatly distorting his numbers. When GVPedia revealed this error in 2019, Lott corrected it without public update, but then reintroduced it in his most recent research.
It is simply bizarre that Hughes claims that the CPRC treats every shooting fatality as an individual mass shooting. He provides no evidence to back this claim and simply asserts it. A detailed discussion of the cases and data are available here. The percentage changes over time as mass public shootings occur. Lott referenced that report that was easiest for people to follow.
There is no evidence that mass shooters systematically target gun-free zones. In his spoken testimony, Lott claimed that shooters overwhelmingly want to maximize casualties and therefore chose undefended targets. He cites the 2022 Buffalo shooting that killed ten and wounded three individuals, along with that shooter’s manifesto, as part of this evidence. Yet the Buffalo shooter wrote that there was a “100% guaranteed” chance that he would encounter a civilian with a gun, writing: “This is Buffalo after all so I am expecting some boys to be packing.”
Besides the overwhelming rate that these mass public shootings occur in gun-free zones, examples of mass murderers explicitly discussing attacking places where victims are defenseless are available here.
Lott’s point was these murderers explicitly talk about how they prefer gun-free zones. The Buffalo mass murderer wrote: “Attacking in a weapon-restricted area may decrease the chance of civilian backlash. Schools, courts, or areas where CCW are outlawed or prohibited may be good areas of attack. Areas where CCW permits are low may also fit in this category. Areas with strict gun laws are also great places of attack.” The fact that the Buffalo murderer acknowledges that gang members and criminals might be illegally carrying guns doesn’t negate that he explicitly states he wants to target places where people can’t legally carry guns.
Defensive gun use is not more effective at preventing injury than other means of self-defense. In his spoken testimony, Lott claimed that research overwhelmingly finds that using a gun for self defense is the most effective form of defense. This is false. A 2015 study from Harvard University finds that defensive gun use is no more effective at preventing injury than doing nothing, and is less effective than other defensive strategies.
There is a long list of other research that shows that guns are the safest course of action when you are confronted by a criminal: for example, Gary Kleck and Miriam A. Delone, “Victim Resistance and Offender Weapon Effects in Robbery,” Journal of Quantitative Criminology 9 (1993): 55–81; Lawrence Southwick, “Self- Defense with Guns,” Journal of Criminal Justice 28 (2000): 351–370; and Jongyeon Tark and Gary Kleck, “Resisting Crime,” Criminology 42 (2004): 861–909. See also Lott’s “More Guns, Less Crime,” University of Chicago Press, 3rd edition, 2010.
Evidence indicates that murders did not increase when firearms were banned in other countries. In his written testimony, Lott wrote “Time and again, crime rises when we prevent people from protecting themselves. Indeed, every place in the world that has banned guns has seen an increase in murder.” This is false. Japan is the developed nation that has come closest to completely banning firearms, with several laws enacted from 1946 to 1958 that essentially ban firearms. Since that time, Japan has seen its homicide rate fall more than 75 percent. Further, a 2013 study found that among developed nations, countries with more guns per capita were associated with significantly higher rates of firearm deaths
Below is a detailed post that the CPRC put up in 2016. The entire post that discusses the evidence from other countries is available here. An additional discussion is avalable in Lott’s books such as “More Guns, Less Crime” (“More Guns, Less Crime,” University of Chicago Press, 3rd ediiton, 2010, Chapter 10).
Regarding Japan, the point to make clear is that Japan has had a very low murder rate for as long as data is available. Gun ownership by private citizens was banned or extremely heavily regulated for hundreds of years, with no significant change in the number of people allowed to own guns. Some point to the drop in homicides after the 1958 gun law, but they ignore the 1946 regulations under the Allied Occupation and the 1950 Order that continued “the general prohibition of possession of guns by civilians.” We have consulted Professor Mark Ramseyer at the Harvard Law School, probably the leading US-based expert on the Japanese legal system. He believes that handguns were effectively banned for years long before crime data existed in Japan despite some minor variations in the law over time. Hence, a comparison