Social science on the right to bear arms

Doomsday warnings don't hold up

The debate over carrying firearms outside the home often is accompanied by predictions that public carry will lead to more violent crime. Fortunately, such claims are incorrect. The large-scale expansion of legal carry throughout the United States has not resulted in the doomsday scenarios predicted by opponents. This post surveys the pro/con social science evidence presented in the amicus briefs in New York State Rifle & Pistol Association v. Bruen, for which oral argument will be held on Wednesday.

This post is co-authored by Campbell University law professor Gregory Wallace. Professor Wallace and I are two of the five co-authors of the just-published third edition of the textbook Firearms Law and the Second Amendment: Regulation, Rights, and Policy (Aspen, Wolters Kluwer). Professor Wallace was the lead author on Chapter 1, which surveys data and social science on firearms.

In some counties of New York, as in some jurisdictions in California, Massachusetts, Rhode Island, and Delaware, applicants for a concealed carry license may be denied unless they show  some specific, special need to carry a licensed handgun, beyond the general purpose of self-defense. New Jersey and Maryland are even more restrictive, statewide. Hawaii only allows carry by security guards on duty. The rest of the United States has two types of policies. Twenty-one states do not require a permit for concealed carry. Twenty-one other states, plus D.C. and Puerto Rico, issue licenses fairly, to adults who pass background checks (usually, fingerprint-based) and safety training. These laws are called "right to carry" (RTC).

Advocates of New York's restrictive carry regime say that it saves lives. They contend that more lenient concealed carry laws increase the risk of armed violence to the public, to law enforcement, and even to the carrier.

Three amicus briefs in Bruen present empirical evidence that argue for these claims. One brief was filed by 14 social science researchers and public health experts, led by John Donohue at Stanford Law School (Donohue Brief). Another was filed by the Educational Fund to Stop Gun Violence and 35 other organizations and individuals, including public health and policy experts (Educational Fund Brief). A third was submitted by Prosecutors Against Gun Violence, an organization of 74 prosecutors who support efforts to reduce violence involving guns (PAGV Brief).

Three amicus briefs addressing social science evidence also were filed supporting the challenge to New York's prohibitive law. The first was from Dr. William English and the Center for Human Liberty. English is a political economist and public policy professor at Georgetown University (English Brief). Another was filed by attorney Dan Peterson, representing Law Enforcement Groups and State and Local Firearms Rights Groups (Law Enforcement Brief), while a third was filed by the Crime Prevention Research Center, led by John R. Lott, Jr., an economist and leading gun researcher, and Carl Moody, an economics professor at William & Mary University (CPRC Brief). These briefs present empirical evidence showing that nondiscretionary "shall issue" laws (also referred to as "right-to-carry" laws) do not increase violent crime, and challenge the research cited in the Donahue, Educational Fund, and Prosecutors briefs.

Professor Wallace was not an amicus in the Law Enforcement Brief, but he did assist in the research. Professor Kopel one coauthor a law review article with Professor Moody, and presented an empirical study by Moody in an amicus brief in McDonald v. City of Chicago. That study was cited by Justice Alito's opinion for the Court, at note 2. ("Chicago residents now face one of the highest murder rates in the country and rates of other violent crimes that exceed the average in comparable cities. n.2…Brief for International Law Enforcement Educators and Trainers Association et al. as Amici Curiae 17–21, and App. A (providing comparisons of Chicago's rates of assault, murder, and robbery to average crime rates in 24 other large cities).")

After conducting the most thorough judicial review to date of relevant social science on the net public-safety effects of allowing public carriage of guns, Judge Richard Posner in Moore v. Madigan, 702 F.3d 933, 939 (7th Cir. 2012), concluded that such evidence failed to establish a convincing defense of an Illinois statue banning public carry. He explained:

The theoretical and empirical evidence (which overall is inconclusive) is consistent with concluding that a right to carry firearms in public may promote self-defense. Illinois had to provide us with more than merely a rational basis for believing that its uniquely sweeping ban is justified by an increase in public safety. It has failed to meet this burden.

Id. at 942.

Little has changed since Judge Posner's assessment. Social science research has yet to establish a predictive relationship between less restrictive right-to-carry laws and violent crime.

A Rand Corporation metastudy on firearms, published in 2020, examined 29 leading empirical studies from 1998-2019 on the relationship between concealed carry laws and violent crime. Some of the studies raised serious concerns among Rand's methodology review team. Rosanna Smart et al., The Science of Gun Policy: A Critical Synthesis of Research Evidence on the Effects of Gun Policies in the United States. Excluding the studies with significant methodological weaknesses, the Rand metastudy found that there is incon­clusive evidence that state shall issue laws have any statistically significant effects on total homicides, firearm homicides, robberies, assaults, rapes, and mass shootings.

The Law Enforcement Brief points out that almost 30 years' experience with increased gun ownership and liberalized carry laws has not led to increased crime. From 1991-2019, the number of civilian-owned firearms in the United States more than doubled, to an estimated 434 million. The number of carry permit-holders increased at least seven-fold to almost 20 million. Forty-two states have adopted either "shall issue" or permitless carry laws. Not one of these states has reverted to a restrictive "may issue" regime or repealed permitless carry. Carry outside the home is the norm throughout most of the country, including many large urban areas, such as the District of Columbia, Chicago, Miami, Philadelphia, San Juan, Seattle, and Houston. During the same period, murder rates and violent crime have plummeted by half. 

Right-to-carry laws and violent crime

The most-cited study supporting New York's very restrictive law is John J. Donohue et al., Right-to-Carry Laws and Violent Crime: A Comprehensive Assessment Using Panel Data and a State-Level Synthetic Control Analysis, 16 J. Empirical Legal Stud. 198 (2019) (Donohue Study). Using panel data and synthetic controls analyses, the study examined right-to-carry laws adopted in states between 1981 and 2007. The Donohue Study is featured prominently in the Donahue, Educational Fund, and PAGV briefs.

The amicus briefs cite the Donahue Study for two main points. First, the study finds that states with more restrictive licensing requirements for concealed carry had a decline in violent crime over the 27-year period that was nearly ten times greater than states that did not have such licensing regimes. Second, the study finds that adoption of more lenient right-to-carry laws led to an immediate, "substantial and statistically significant" increase in violent crime. According to the study's synthetic control analysis, violent crime in these states increased by an average of 13-15 percent by the tenth year after adoption.

The English Brief disputes the Donohue Study, based on English's recent study, The Right to Carry Has Not Increased Crime: Improving an Old Debate Through Better Data on Permit Growth Over Time (English Study), published in July 2021. English used decades of state-level data on the growth of carry permits over time to investigate the relationship between right-to-carry laws and violent crime. English contends that the most direct measure of the real-world effects of right-to-carry laws is how growth in the number  of permits over time affects crime, rather than a simple "before-after" following a one-time change in the law. The English Study finds that "the higher rates of carry made possible by RTC laws have no statistically significant effect on violent crime rates, including murder rates (with or without firearms) for the period of 1977-2014."

The English Brief critiques the Donohue Study's methodology:

• The Donohue Study uses an inferior "binary dummy variable" approach instead of actual carry permit data. In other words, if a state's licensing reform statute went into effect on January 1, 1998, Donahue divides the state's data into two categories: before the 1998 law, and after. English argues that the real-world effects of right-to-carry laws are time-dependent—that there is a time lag between the legal date of such laws and large numbers of permit issuances. For example, by March 1998 there might not be a large number of new licensed carriers in the state, but by January 2001 there could be. So English uses actual permit data, rather than Donohue's more simplistic before/after coding.
• The Donohue Study uses "analytic weights." That is, Donohue assigns greater weight to data from more populous states. English calls Donohue's method "statistically indefensible." For example, Florida, Georgia, Alabama, Mississippi, North Carolina, and South Carolina all provide natural experiments in the effects of changes in carry licensing laws. English weights each of these experiments equally. Donahue gives greater weight to Florida, because it has the largest population. Once Donahue's overweighting of high population states is corrected, even the Donohue Study shows no correlation between right-to-carry laws and increased violent crime.
• The Donohue Study does not compare real states with other real states. Instead, each actual right-to-carry state is compared to a hypothetical (synthetic) state composed of multiple actual states (e.g., California, Michigan, and New York). These synthetic states are stitched together and weighted in different proportions based on purported similarities to the right to carry state. Donahue then compares what happened in the actual state after the RTC law was passed to crime data in the synthetic state. English derides reliance on these "imaginary, stitched-together 'Frankenstein' states." He writes that Donahue's use of synthetic states is compromised by multiple errors in execution, including use of an arbitrarily short analysis period and omission of Washington, D.C., data from the estimating process. Further, Donahue's outcome "lags" do not allow covariates to influence the modeling of the synthetic control. (Covariates are independent variables, such as incarceration rates or poverty rates, that can impact the dependent variable, namely the violent crime rate.) According to the English Brief, "When these issues are corrected, synthetic control analysis indicates that RTC laws have no statistically significant effect on violent crime rates."

The Rand Corporation metastudy also warns about the reliability of the Donohue Study:  "Synthetic control methods are relatively new, and especially when controls are made up of just a few states, as they were in this case, their usefulness for identifying causal effects may be compromised." An in-depth critique of Donahue's methodology is provided by Florida State University criminology professor Gary Kleck's  The Effect of Right-to-Carry Laws on Crime Rates: A Critique of the Research of Donohue et al.

The Donohue, Educational Fund, and PAGV briefs rely on other social science studies with flawed methodologies. For example, as the Law Enforcement Brief points out, the Rand Corporation's 2020 metastudy classifies two studies prominently featured in these briefs as having "serious methodological problems": Michael Siegel et al., Easiness of Legal Access to Concealed Firearm Permits and Homicide Rates in the United States, 107 Am. J. Pub. Health 1923 (2017), and Cassandra K. Crifasi et al., Association Between Firearm Laws and Homicide in Urban Counties, 95 J. Urban Health 383 (2018). The Donohue Brief also features a study from Emma Friedel, another Florida State criminologist, which repeatedly has been criticized by Gary Kleck for flawed methodology.

The difficulties of social science studies on guns

Social science studies can be highly sensitive to small changes in their model and sample. See Steven N. Durlaf, et al., Model Uncertainty and the Effect of Shall-Issue Right-to-Carry Laws on Crime, 81 European Econ. Rev. 32 (2016). One technical analysis of gun law studies concluded that

most of the statistical methods commonly used to estimate the causal effects of gun laws at the state, county, or city level suffer from important limitations that make inferences from those models prone to error. Almost all models we investigated had SEs [standard errors] that were too small, leading to high false-positive rates. Common adjustments to the SE generally failed to fully correct this problem. All models were found to have insufficient power to detect causal effects that would correspond to an increase or decrease of 1,000 firearm deaths per year nationally, an effect size that we argue is a large enough change to be clearly important.

Terry L. Schell, et al., Evaluating Methods to Estimate the Effect of State Laws on Firearm Deaths: A Simulation Study (Rand Corp. 2018). Moreover, the mere showing of association or correlation in these studies does not prove causation. See Naomi Altman & Martin Krzywinski, Association, Correlation, and Causation, 12 Nature Methods 899 (2015). At best, it might suggest a hypothesis, not proof.

Other studies

The Donahue Brief cites several additional social science studies that link less restrictive public carry laws to higher violent crime. But the Law Enforcement and CPRC briefs identify many other studies that reach the opposite conclusion, namely, that less restrictive concealed-carry laws have little or no effect on homicide or other violent crime rates. The CPRC Brief cites additional studies showing that right-to-carry laws may actually reduce violent crime.

One study cited in the Law Enforcement Brief is Mark E. Hamill et al., State Level Firearm Con­cealed-Carry Legislation and Rates of Homicide and Other Violent Crime, 228 J. Amer. Col­lege of Surgeons 1 (2019). It uses state-level data from 1986 to 2015, and finds no statis­tically significant association between the liberalization of state carry laws and rates of violent crime, homicides, or firearm homicides. Another cited state level study found no evidence that RTC laws have increased crime. Carlisle E. Moody & Thomas B. Marvell, The Right-to-Carry Laws: A Critique of the 2014 Version of Aneja, Donohue, and Zhang, 15 Econ. J. Watch 51 (2018). A third cited state level study concludes that adoption of right-to-carry laws had no significant effects on the overall violent or property crime rates, and led to medium-term decreases in murder rates. Wei Shi & Lung-fei Lee, The Effects of Gun Control on Crimes: A Spatial Interactive Fixed Effects Approach, 55 Empirical Economics 233 (2018),

The Law Enforcement Brief also identifies two city level studies that reached similar conclusions. The first concluded that Arizona's 2010 repeal of its requirement for government-issued concealed carry permits had no appreciable effect on handgun-related violent crime in Tucson. Michael R. Smith & Matthew Petrocelli, The Effect of Concealed Handgun Carry Deregulation in Arizona on Crime in Tucson, 30 Crim. Justice Pol. Rev. 1186 (2019). Analyzing data from 56 cities from 1980-2010, the second study concluded that shall-issue laws reduced firearm homicides by 15 percent and total homicides by 13 percent. James M. La Valle, "Gun Control" vs. "Self-Protection": A Case against the Ideological Divide, 10 Justice Policy Journal 1 (2013).

Empirical evidence does not show that the adoption of RTC laws results in more violent crime. For defenders of the New York law, the social science evidence is at best a wash. The newer studies have not changed what Judge Posner observed in 2012: social science research has yet to establish a predictive relationship between less restrictive right-to-carry laws and violent crime.

Law-abiding concealed carry license holders

What threat of armed violence do concealed carry holders pose to others? The Law Enforcement Brief cites data showing that concealed carry permit holders, as a group, are vastly more law-abiding than the general public.

In Texas, data from 2020 show total convictions of a long list of crimes and the number of those crimes committed by carry license holders. Carry license holders were convicted of 114 out of a total of 26,304 such crimes, or 0.4334%. There were 1,626,242 active license holders in 2020. The population 18 and older is estimated by the Census Bureau to be 74.5% of the total Texas population of 29,145,505, or 21,713,401 adult individuals. Thus, carry license holders constituted 7.49% of the population 18 and older, but committed only 0.4334% of the crimes. The conviction rate of license holders is therefore about 1/17th of the rate for the adult population as a whole. (The minimum age for a permit in Texas is generally 21, but permits are issued to applicants 18-20 with present or past military service.)

In Florida, carry licenses are revoked if a licensee commits a disqualifying crime (all felonies plus others) or is disqualified because of mental health, substance abuse, domestic violence, or other reasons. 5,007,776 licenses were issued over the period October 1, 1987, through June 30, 2021. During that period, 17,134 licenses were revoked for all reasons, a rate of 0.34%. As of June 30, 2021, the number of valid licenses statewide was 2,321,146. In the year ending June 30, 2021, 1,050 licenses were revoked, an annual revocation rate of .045%. From 1987 through 2010, when the state stopped publishing this information, only 168 revocations were for a crime with a "Firearm Utilized."

The Law Enforcement Brief also points out that most homicides and non-fatal shootings are committed by persons with criminal records who are not eligible to possess firearms. Data from New York City, for example, shows that almost 90% of homicide suspects had at least one prior arrest, and seven out of ten homicide victims also had prior arrests. The largest part of violent crime problem is criminals shooting other criminals, not law-abiding citizens carrying concealed firearms.

Donohue's rebuttal

The Donohue Brief responds to the above data mostly with generalizations and anecdotes. It cites one study, Charles D. Phillips, et al., When Concealed Handgun Licensees Break Bad: Criminal Convictions of Concealed Handgun Licensees in Texas, 2001-2009, 103 Am. J. Pub. Health 86 (2013). The Phillips article states that "the proportion of deadly conduct offenses in convictions was five times higher for permit holders than for non-permit holders, indicating that these 'law-abiding citizens' are disproportionately involved in gun-related violent crime." In the "rare instances" when the licensees commit crimes, they are more likely to be convicted for serious weapons-related offenses than are non-license holders. The article also emphasizes that concealed handgun licensees "rarely 'break bad'" and are "almost universally a law-abiding population."

Stolen guns

The Donohue Brief also claims that "law-abiding citizens with firearms are major suppliers of weapons to criminals." The brief cites the Donohue Study for evidence that "a significant number of guns carried outside the home are lost and stolen—a figure estimated to be roughly 100,000 per year from permit holders carrying guns outside the home." (Donahue Brief at 22-23). The clear implication is that these guns are stolen from lawful permit holders while carrying outside the home. If the guns are stolen from the home, the statistic has no relevance to New York's public carry regime.

The theft claim reveals how social science evidence can be misused in legal briefs. The Donohue Brief cites the Donohue Study which, in turn, cites another study it claims "revealed that those who carried guns outside the home had their guns stolen at a rate over 1 percent per year." That study is David Hemenway et al., Whose Guns Are Stolen? The Epidemiology of Gun Theft Victims, 4 Injury Epidemiology 11 (2017). Hemenway's online survey revealed that 2.4% of gun owners reported having one or more guns stolen in the previous five years. Hemenway then estimates that about 380,000 guns are stolen each year. (Other studies have estimated far fewer.) Hemenway found that gun owners are more likely to have had guns stolen if they are non-white, are from the South, owned six or more guns, owned guns for protection, carried guns in the last month, did not store guns in the safest manner, or stored guns in the car. The Hemenway study nowhere identifies the number of guns stolen while carrying outside the home. Nor does Hemenway claim that respondents who carried outside the home had their guns stolen at a rate of over one percent a year. Instead, Hemenway and his coauthors frankly admit that "we know almost nothing about the actual event—the type of gun stolen, where the gun was stored (e.g., at home, in the garage), whether it was locked up, the time of day or day of week of the theft, who the thief was and whether he was known to the victim." The Donohue Brief's claim that law-abiding citizens who carry outside the home are "major suppliers" of weapons to criminals has no support in the social science literature.

The "weapons effect"

The Donohue and Education Fund Briefs raise another argument: Carrying a gun increases aggressive and confrontational behavior in both the carrier and those who observe the gun. They cite studies asserting that the mere presence of a firearm can produce aggressive behavior—the so-called "weapons effect." However, such research has already been rejected by the Supreme Court.

The briefs says the "weapons effect" was proven in C.A. Anderson, A.J. Benjamin,  & B.D. Bartholow, Does the gun pull the trigger? Automatic priming effects of weapon pictures and weapon names, 9 Psychological Science 308 (1998). In 2002, Kopel's Independence Institute colleagues Paul Gallant and Joann Eisen wrote an article describing Anderson's study:

Stimuli were presented to the subject on a computer screen in the form of "prime" words, and "target" words which were categorized as either "aggressive" or "non-aggressive." Two categories of prime words were used: weapon words (shotgun, machete, fist, bullet, dagger, and grenade), and animal words (rabbit, bug, dog, bird, butterfly, and fish).

For the experimental procedure, a prime word was presented to each subject for 1.25 seconds, followed by a blank screen of 0.5 seconds duration. Then, a target word was presented. The subject's task was to recite the target word as quickly as possible. The computer was equipped with a microphone to measure the time between the presentation of the target word and the first sound made by the subject.

In this part of the study, the researchers found that, on animal-primed trials, subjects were 0.005 seconds slower at naming aggressive target words than at naming non-aggressive words. For weapon-primed trials, however, subjects named aggressive target words 0.009 seconds faster than they named non-aggressive words. The authors claimed that these results provided "clear support for the priming interpretation of the weapons effect," i.e., that "the mere cognitive identification of a weapon increases the accessibility of aggression-related concepts in semantic memory."

In the second experiment. . . the prime stimuli consisted of black-and-white line drawings of weapons (guns, swords, and clubs—3 different pictures for each category, for a total of 9 weapons) and of plants (fruits, trees, and flowers, also 3 different pictures for each category). The prime stimulus was presented as in the previous experiment, and the subject was instructed to call out the category as quickly as possible. Again, a blank screen appeared for 0.5 seconds. Then the target word was presented and remained visible on the screen until the subject called it out.

The researchers found that after exposure to plant pictures subjects were 0.005 seconds faster at naming aggressive target words compared to non-aggressive words. However, after exposure to weapon pictures, subject reaction time decreased, and subjects were 0.011 seconds faster at naming aggressive target words compared to non-aggressive words. . . .

The authors concluded: "These two experiments demonstrate that simply identifying weapons increases the accessibility of aggressive thoughts . . . that thinking about weapons increases accessibility of aggressive concepts in general….Does the gun pull the trigger? Extant research suggests that it does. Our research demonstrates one way that exposure to weapons might increase aggressive behavior—by increasing the accessibility of aggressive thoughts."

But did the authors really demonstrate what they claimed?

Insomuch as "gun" might well be associated with "shoot" or "murder," when it comes to the non-weapon primes they selected, there is no such logical link. For example, while butterfly was used as a prime word, the words "flutter," "fly," and "cocoon" were nowhere to be found. If the idea was to explore whether a certain word would trigger a class of words, such as "gun" triggering the entire class of aggressive words, why did not the authors compare this effect with similar effects for animal primes? The word "rabbit" is likely to trigger "carrot," "ears," "chew," and "hop," but that was not tested. In addition, potentially threatening primes like "lion," "shark," or "rattlesnake" should have been used to determine whether these would have elicited the same aggressive tendencies.

Paul Gallant & Joanne D. Eisen, Trigger-Happy: Re-thinking the "Weapons Effect", 14 Journal on Firearms & Public Policy 89 (2002). (The Journal was published from 1989 to 2016 by the Second Amendment Foundation. At the time, Kopel was the editor.)

The U.S. Supreme Court said what it thought of Dr. Anderson's research in Brown v. Entertainment Merchants Association, 564 U.S. 786 (2011). The case was a successful First Amendment challenge to a California statute requiring minor to obtain parental permission to purchase violent video games. California relied heavily on Dr. Anderson to show that violent video games made children more violent. Justice Scalia's opinion for the Court found Anderson's claims and methodology laughable:

The State's evidence is not compelling. California relies primarily on the research of Dr. Craig Anderson and a few other research psychologists whose studies purport to show a connection between exposure to violent video games and harmful effects on children. These studies have been rejected by every court to consider them,6 [FN 6 lists 3 Circuit and 4 District Court decisions] and with good reason: They do not prove that violent video games cause minors to act aggressively (which would at least be a beginning). Instead, "[n]early all of the research is based on correlation, not evidence of causation, and most of the studies suffer from significant, admitted flaws in methodology." Video Software Dealers Assn. 556 F. 3d, at 964. They show at best some correlation between exposure to violent entertainment and minuscule real-world effects, such as children's feeling more aggressive or making louder noises in the few minutes after playing a violent game than after playing a nonviolent game.7

[Note 7.] One study, for example, found that children who had just finished playing violent video games were more likely to fill in the blank letter in "explo_e" with a "d" (so that it reads "explode") than with an "r" ("explore"). App. 496, 506 (internal quotation marks omitted). The prevention of this phenomenon, which might have been anticipated with common sense, is not a compelling state interest.

Entertainment Merchants Association v. Brown rejects the Anderson and other "violent video game effect" studies for failing to demonstrate a compelling state interest. Indeed, Entertainment Merchants Association suggests that the studies do not even rise to the level of a trivial state interest.

The Anderson "weapons effect" research finds milliseconds of differences seconds in response rates to target words, based on whether subjects had been primed with weapons-related cues. It never seems to have occurred to Anderson et al. that the tiny difference might have been because the weapons cues were followed by relevant target words and the non-weapons cues were not.

The "weapons effect" is a deeply held article of faith in the gun prohibition community, but it has no basis in science.

Violence Policy Center

Donohue and Education Fund Briefs do present anecdotes involving carry permit holders who have killed after becoming angry over insignificant issues ranging from cutting someone off on a highway to texting in a movie theater to playing loud music at a gas station. Another amicus brief, by the Violence Policy Center (VPC), presents more anecdotes. The VPC also cites statistics from its database Concealed Carry Killers.

We haven't fact-checked the VPC database. But Clayton Cramer (coauthor of several law review articles with Kopel) did. Clayton E. Cramer, Violence Policy Center's Concealed Carry Killers: Less Than It Appears (2012). Cramer found numerous instances of VPC incorrectly claiming that a perpetrator had a concealed carry permit, or VPC counting events that had nothing to do with the carry permit, such as a permit-holder committing suicide at home.

Or consider this example from the VPC brief:

In May 2014, Michael Bowman, who also possessed a valid concealed carry handgun permit, shot and killed police officer Kevin Jordan in Griffon, Georgia. Officer Jordan was working an off-duty security job in uniform at a Waffle House restaurant. Bowman was drunk when he and his girlfriend Officer Jordan attempted to arrest Bowman's girlfriend. Officer Jordan—a father of seven—was on the ground attempting to restrain Bowman's girlfriend when Bowman shot him multiple times in the back. Bowman was convicted of felony murder and sentenced to life in prison without parole.42

Footnote 42 cites to the VPC's database. The database in turn cites: Former soldier gets life without parole for murdering police officer, Atlanta Journal-Constitution, Feb. 17, 2017 (article does not mention a permit); Griffin Tragedy; Drunk and armed at 2 a.m., Atlanta Journal-Constitution, June 8, 2014 (we couldn't find this article in the newspaper's archive, or in Westlaw News, although both databases contain many articles on the crime); Funeral next Monday for slain Griffin police officer, Atlanta Journal-Constitution, June 2, 2014 (found at 2014 WLNR 14878704). This last article explains who had the concealed carry permit: officer Kevin Jordan's brother, Raymond Jordan.

Raymond Jordan was in the parking lot when the gunfire erupted and grabbed his own gun, [police officer Mike] Richardson said. A civilian, he had a permit to carry a gun and was at the restaurant to visit his brother, who often worked there on weekends, police said. Raymond Jordan was not charged Saturday in Bowman's shooting, and police offered no explanation.

So in this case, the man with the concealed carry permit was not the criminal; the man with the carry permit was the person who shot the criminal who was attacking the officer.

There definitely are cases of persons with concealed carry permits perpetrating crimes in public places, including homicide. However, the VPC database is not necessarily a reliable guide. The better approach is to look at comprehensive datasets, such a statewide revocation figures, described above.

Right to carry laws and self-defense

The Donohue and Educational Fund Briefs argue that carrying guns outside the home does not promote self-defense, but rather increases the risk of becoming the victim of violent crime.

The Educational Fund argues that public carry presents a high risk of gun violence, citing one study it says shows that carrying a firearm increases the risks of both gun violence exposure and gun victimization. But this study examines illegal gun carrying among young people in two large cities; it has nothing to do with lawful public carry.

The Donohue Brief begins with the claim that firearms are rarely used in self-defense. It cites David Hemenway & Sarah J. Solnick, The Epidemiology of Self-Defense Gun Use: Evidence from the National Crime Victimization Surveys 2007-2011, 79 Preventive Med. 22, 23 (2015). That article uses National Crime Victimization Survey (NCVS) data from 2007-11 to find just 127 victims present at a crime incident (about 1 percent) who used a gun defensively. This study is an outlier when estimating defensive gun uses (DGUs). The NCVS provides much lower estimates of DGUs than other research. The reason for the low number is that the NCVS survey does not ask respondents whether they have used a gun defensively. Instead, some respondents disclose DGUs while being interviewed. Because the NCVS never asks about DGUs, it records fewer instances of such use. Jacob Sullum, A Survey Not Designed to Measure Defensive Gun Use Finds Little of It, Reason, Sept. 7, 2015.

There is much disagreement among researchers over the number of DGUs, which include not only firing the weapon but also merely displaying it to deter an attacker. The large majority of DGUs are displays, without a shot being fired. The National Research Council (NRC) noted in 2013 that "[a]lmost all national sur­vey estimates indicate that defensive gun uses by victims are at least as common as offensive uses by criminals, with estimates of annual uses ranging from about 500,000 to more than 3 million, in the context of about 300,000 violent crimes involving firearms in 2008." National Research Council, Priorities for Research to Reduce the Threat of Firearm-Related Violence 15 (National Academies Press 2013) (citations omitted).

The Centers for Disease Control (CDC) cites the NRC report in estimating DGUs to be in the range of 60,000 to 2.5 million. Centers for Disease Control and Prevention, Firearm Violence Prevention. The Rand Corporation's 2018 metastudy concludes that estimates at the high end (e.g., 2.5 million DGUs annually) "are not plausible given other information that is more trustworthy, such as the total number of U.S. residents who are injured or killed by guns each year." Likewise, at the low end, "the NCVS estimate of 116,000 DGU incidents per year almost certainly underestimates the true number." Rand Corporation, The Science of Gun Policy: A Critical Synthesis of Research Evi­dence on the Critical Effects of Gun Policies in the United States 280 (2018). For an extended discussion, see the third edition of our textbook Firearms Law and the Second Amendment: Regulation, Rights, and Policy at pages 20-25. The Law Enforcement Brief provides a summary of the various estimates. The brief also cites English's 2021 National Firearms Survey (July 14, 2021), which indicates that 75% of defensive gun uses take place outside the home.

Are firearms harmful for self-defense?

The Donohue Brief makes a further claim: firearms do not reduce the defender's risk of injury. It says the best evidence is from the analysis of NCVS data in the Hemenway and Solnick study cited above: "4.2% of victims were injured after using a gun in self-defense, and 4.1% of victims were injured without using a gun." Hemenway and Solnick caution against overinterpreting their data: "the sample of those injured after using a gun (5/127) is really too small to warrant strong conclusions." As the Law Enforcement Brief points out, when the injury rate is calculated for the entire incident from beginning to end, the Hemenway and Solnick study reveals that "there is a lower likelihood of injury at any time during an incident when a gun was used in self-defense." (original emphasis).

Another study cited by both the Donohue and Education Fund Briefs is Charles C. Branas et al., Investigating the Link Between Gun Possession and Gun Assault, 99 Am. J. Pub. Health 2034 (2009). This study shows that when victims have a chance to resist, those who possess a gun are about four-to-five times more likely to be shot during the assault. The Branas article provides no information to indicate that any of the homicide victims had a carry permit. Only six percent of the homicide victims were carrying a gun when they were shot. Fifty-three percent of the victims had criminal records. The Branas study may just show that persons, such as drug dealers, are more likely to possess guns.

As the National Research Council wrote in 2013:

Studies that directly assessed the effect of actual defensive gun uses (i.e., incidents in which a gun was 'used' by the crime victim in the sense of attacking or threatening an offender) have found consistently lower injury rates among gun-using crime victims com­pared with victims who used other self-protective strategies.

Right to carry laws and risks to law enforcement officers

Both the Donohue and PAGV Briefs claim that adoption of right-to-carry laws pose significant risks of violence to law enforcement officers. The studies they cite, however, only show correlation between homicide rates of law enforcement officers in states with low and high firearm ownership. They do not attempt to show any correlation with right to carry laws. Although the PAGV Brief asserts that "a higher incidence of armed individuals in public would increase the likelihood of violent encounters between civilian and law enforcement officials," it cites in support only affidavits from public officials and police administrators filed in support of laws challenged in Kachalsky v. County of Westchester, 701 F.3d 81 (2d Cir. 2012), and Woollard v. Gallagher, 712 F.3d 865 (4th Cir. 2013), and Gould v. Morgan, 907 F.3d 659 (1st Cir. 2018), and general data indicating the number of officers who died from criminals using firearms. Quoting an amicus brief filed in Peruta v. County of San Diego, 824 F.3d 919 (9th Cir. 2016), the PAGV Brief states that concealed weapons licensees have killed at least 11 law enforcement officers since 2007.

The Law Enforcement Brief, filed on behalf of eight law enforcement associations, presents the results of multiple surveys showing that law enforcement overwhelmingly supports the right of law-abiding citizens to carry concealed firearms.

It is unlikely that the decision in Bruen will turn upon social science evidence. The constitutional right to keep and bear arms does not depend on statistics. In Heller, Justice Breyer's dissent summarized the pro/con data and social science studies from the amicus briefs. He argued that because the evidence was mixed, the Court should defer to the judgment of the D.C. Council. The Heller majority did not discuss the modern social science. Rather, said the majority, the interest-balancing had already been done by the American people, when they ratified the Second Amendment.

Even so, as the English Brief concludes, "social science provides valuable empirical confirmation that exercise of this right does not exact the social costs claimed by Donohue and others who share his policy preferences."