Gun violence is among the most consequential and most politically intractable public safety problems in the United States — a problem that kills approximately 45,000 Americans annually through homicide, suicide, and unintentional shooting, that injures many more, that disproportionately devastates specific communities, and that separates the United States from every other comparable wealthy democracy by margins that resist explanation through any factor other than the exceptional availability of firearms in American society. Criminology and Public Policy engages gun policy as a domain where the evidence base is substantial and growing, the policy implications are clearer than political paralysis suggests, and the specific policy interventions supported by the strongest evidence differ in important ways from both the most politically prominent firearms restrictions advocated by gun control advocates and the purely enforcement-focused approaches advocated by those who oppose regulatory intervention in firearms markets.
The relationship between firearms availability and violence is the empirical foundation on which gun policy debates must rest, and that relationship is real and substantial at the aggregate level across multiple methodological approaches. The United States has both the highest rate of private firearms ownership in the world — approximately 120 firearms per 100 civilian population — and the highest rate of firearms homicide among comparable high-income democracies, by margins that far exceed any plausible alternative explanation. Cross-national comparisons, within-country analyses exploiting variation in firearms availability over time and across states, and specific research on the consequences of particular policy interventions all contribute to an evidence base that supports more confident conclusions about the relationship between firearms access and violence than either the gun control or gun rights communities typically acknowledge.
Introduction
The political polarization surrounding gun policy in the United States has produced a policy debate frequently disconnected from the research evidence in both directions: gun control advocates sometimes claim more policy effectiveness than the evidence supports for specific regulatory approaches, while gun rights advocates sometimes deny the firearms availability-violence relationship that the evidence clearly establishes. Understanding what the evidence actually shows — and what it does not show, where it is contested, and where the strongest causal claims are most justified — requires engaging with the research literature in a way that neither advocacy community consistently does.
The evidence base for gun policy in criminology and public health has been complicated by the Dickey Amendment — a 1996 federal appropriations provision that effectively prohibited the Centers for Disease Control and Prevention from conducting research that might be used to “advocate or promote gun control,” which the CDC interpreted as prohibiting most firearms research and which led to a 20-year hiatus in federally funded firearms research at the nation’s premier public health research agency. The research gap created by the Dickey Amendment was partially filled by state-funded research, privately funded research at institutions including the Harvard T.H. Chan School of Public Health and the Johns Hopkins Bloomberg School of Public Health, and international comparative research, but the absence of federal research infrastructure produced a knowledge base less developed than the scale of the problem warrants. The partial reversal of the Dickey Amendment’s effect through the 2020 appropriations process, which explicitly authorized CDC and NIH gun violence research, began restoring federal research capacity, but the gap created by two decades of federal research absence has not yet been closed.
Three distinct categories of gun violence require separate analytical treatment because they have different demographic profiles, different causal mechanisms, and different policy implications. Firearms homicide — which accounts for approximately 55% of all homicides in the United States and is highly concentrated in specific communities, demographic groups, and conflict contexts — is most effectively addressed by policies that reduce the access of the highest-risk individuals to firearms in the specific social contexts where violence is most likely to occur. Firearms suicide — which accounts for approximately 54% of all firearms deaths, a proportion that the public debate underemphasizes relative to homicide — is most effectively addressed by reducing firearms access during acute mental health crises, particularly through secure storage requirements and crisis-oriented temporary removal mechanisms. Unintentional shootings — which kill approximately 500 Americans annually and injure thousands more — are most effectively addressed through safety education and storage requirements that reduce the unsafe handling and storage practices that produce them.
Firearms Availability and Violence: The Core Evidence
The International Comparison
The United States’ firearms homicide exceptionalism is not limited to comparisons with wealthy European democracies — it persists in comparisons with middle-income countries in Latin America and Asia that are otherwise less favorable comparators. While some Latin American countries have higher total homicide rates than the United States, they do not have higher rates after adjusting for economic development levels that predict overall violence — and the specific firearms homicide rate in the United States exceeds that of most Latin American countries when controlled for development level. Japan — with some of the world’s most restrictive firearms regulations and one of the world’s lowest firearms ownership rates — has a firearms homicide rate approximately 0.01 per 100,000, making the United States’ rate approximately 400–500 times higher. The scale of this difference is not explicable by any cultural or social variable other than firearms availability.
Cross-national comparisons provide the most dramatic evidence for the relationship between firearms availability and violence. The United States’ firearms homicide rate — approximately 4–5 per 100,000 annually — exceeds the rate of comparable wealthy democracies by factors of 4–25: Canada’s rate is approximately 0.5, Australia’s 0.3, Germany’s 0.06, Japan’s 0.02. These differences in firearms homicide rates are not explained by differences in total violence: American non-firearms homicide rates are comparable to or only modestly higher than those of comparable countries, and Americans are not dramatically more violent than citizens of comparable democracies by most behavioral measures. The difference is specifically in firearms homicide, consistent with the interpretation that the United States’ dramatically higher firearms availability is the primary driver of its dramatically higher firearms homicide rate rather than some more diffuse cultural or social characteristic.
Research by Hemenway and Miller examining the relationship between state-level firearms ownership rates and firearms homicide rates found a robust positive association within the United States: states with higher rates of household firearms ownership have higher rates of firearms homicide even after controlling for other predictors of homicide, consistent with the interpretation that firearms availability elevates homicide risk at the state level (Hemenway & Miller, 2000). Research by Fleegler and colleagues found that states with more gun laws had lower rates of firearms fatalities, with the relationship holding after controlling for urbanization, poverty, and other potential confounders. These within-country analyses, while not as methodologically clean as genuine natural experiments, provide supporting evidence for the international comparison’s implication that firearms availability causally elevates homicide rates.
State-Level Variation Within the United States
Within-country variation in firearms ownership rates across American states provides a natural laboratory for examining the firearms-violence relationship that controls for the cultural and institutional differences that complicate cross-national comparisons. Research exploiting survey-based estimates of state-level household firearms ownership rates — developed because direct administrative data on ownership are unavailable in the United States — has found consistent positive associations between state firearms ownership rates and state firearms homicide rates after controlling for other predictors. The Harvard Injury Control Research Center has maintained a program of state-level firearms research that has documented these within-country associations across multiple data waves and analytical approaches, providing supporting evidence for the causal interpretation of the cross-national comparison from a data source whose methodological limitations differ in systematic ways that reduce the risk of correlated measurement error.
The gun ownership rate-crime rate relationship is substantially stronger for firearms violence than for total violence — that is, states with higher ownership rates have more firearms homicide but not proportionally more total homicide. This pattern is consistent with two complementary interpretations: that firearms substitute for other means of violence in states with higher ownership, maintaining total violence while shifting its weapon mix; or that firearms access amplifies the lethality of conflicts that would otherwise produce assault rather than homicide. Research on the relationship between gun ownership and non-firearms homicide rates has generally found null or slightly negative associations, more consistent with the substitution interpretation, but the lethality amplification interpretation also finds support in research documenting that firearms involvement dramatically increases the case fatality rate of violent assaults relative to comparable attacks with other weapons.
Natural Experiments and Causal Evidence
The most rigorous causal evidence on the firearms-violence relationship comes from studies exploiting natural experiments in firearms access — policy changes, regulatory shifts, and other events that produced quasi-random variation in firearms availability over time and across jurisdictions. The 1994 federal assault weapons ban — which prohibited the manufacture and sale of specific semi-automatic weapons and large-capacity magazines for ten years before its expiration in 2004 — provides one natural experiment, though the research on its effects is complicated by the law’s specific scope and the substitution of non-banned weapons. Research by DiMaggio and colleagues found that mass shooting fatalities increased following the ban’s expiration, consistent with the hypothesis that the assault weapons restriction meaningfully reduced the lethality of mass shootings during its operation (DiMaggio et al., 2019).
Australia’s mandatory buyback program — implemented in 1996 following the Port Arthur massacre, which killed 35 people with a semi-automatic rifle, and which removed approximately 650,000 firearms from private ownership through a federally funded mandatory buyback combined with registration requirements and licensing standards — provides the most dramatic natural experiment in comprehensive firearms regulation in a comparable democracy. Research by Leigh and Neill found that the buyback was associated with substantial reductions in firearms homicide and a cessation of mass shootings, with no compensating increases in non-firearms homicide that would suggest displacement, and that the pre-buyback trend in firearms homicide would have predicted far higher rates than actually occurred in subsequent years (Leigh & Neill, 2010). The Australian evidence is frequently invoked in American gun policy debates, where its direct applicability is limited by the constitutional and political differences between Australian and American contexts, but its evidentiary significance — as the most rigorous available test of comprehensive firearms reduction — is not diminished by those differences.
Table 1. Gun Policy Interventions: Evidence Summary
| Policy | Mechanism | Study Design | Effect on Violence | Evidence Strength | Key Qualification |
|---|---|---|---|---|---|
| Background check expansion (universal) | Prevent prohibited persons from purchasing firearms | Quasi-experimental (state variation) | Negative (reducing) | Moderate | Only effective with comprehensive enforcement; private sales loophole limits impact |
| Red Flag / ERPO laws | Temporary removal during crisis | Quasi-experimental (state variation) | Suicide reduction; some homicide reduction | Moderate-strong | Most evidence on suicide; implementation and compliance vary |
| Permit-to-purchase requirements | Licensing and background check at point of sale | Quasi-experimental (state adoption/repeal) | Firearms homicide reduction (16–40% in some studies) | Moderate-strong | Connecticut study (adoption) and Missouri study (repeal) both find effects |
| Child Access Prevention / safe storage laws | Reduce youth access to household firearms | Quasi-experimental (state variation) | Youth firearm suicide and homicide reduction | Moderate | Larger effects in studies with verification mechanisms |
| Assault weapons and high-capacity magazine limits | Limit lethality of mass shootings | Federal ban natural experiment | Mass shooting fatality reduction during ban | Moderate; contested | Substitution to non-banned weapons partial; effects concentrated in mass shooting context |
| Focused deterrence / violence interruption | Reduce gang violence specifically | Quasi-experimental; comparison group | Firearms violence reduction in targeted communities | Moderate-strong | Requires credible deterrence and social service investment |
| Australian-style mandatory buyback | Substantial reduction in civilian firearms stock | Natural experiment (Port Arthur response) | Substantial firearms homicide reduction; mass shooting cessation | Strong in context | Constitutional and political barriers to US adoption; different baseline context |
Specific Policy Interventions
Background Checks and Point-of-Sale Regulations
The federal National Instant Criminal Background Check System (NICS), established by the Brady Handgun Violence Prevention Act of 1993, requires background checks for firearms purchases from licensed dealers but contains the “gun show loophole” — the absence of background check requirements for private party sales, which the Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) estimates account for approximately 20–40% of all firearms transfers. Research on the crime consequences of background check expansion — exploiting variation across states that have or have not extended background check requirements to private sales — has found consistent reductions in firearms crime in states with universal background check requirements, with the effect concentrated in firearms trafficking pathways that the private sale loophole enables (Webster et al., 2014).
Permit-to-purchase laws — which require individuals to obtain a license or permit before purchasing any firearm, typically involving an in-person application, a background check, and a waiting period — have generated some of the most compelling quasi-experimental evidence in the gun policy literature. Research by Webster and colleagues examining Connecticut’s adoption of a permit-to-purchase requirement in 1995 found a 40% reduction in firearms homicide rates following adoption, controlling for pre-adoption trends and using synthetic control comparisons with other states (Webster et al., 2014). Research on Missouri’s repeal of its permit-to-purchase requirement in 2007 found a subsequent 25% increase in firearms homicide rates — the mirror image of the Connecticut adoption finding — providing convergent evidence from adoption and repeal studies that permit-to-purchase requirements causally reduce firearms homicide.
Extreme Risk Protection Orders
Extreme Risk Protection Orders (ERPOs) — known colloquially as Red Flag laws — provide courts with the authority to temporarily remove firearms from individuals who present an imminent danger to themselves or others, based on petition by law enforcement, family members, or in some states school personnel. As of 2024, 21 states and the District of Columbia have enacted ERPO laws with varying procedural requirements. Research on the effects of ERPO laws has found consistent evidence of suicide reduction — particularly firearm suicide among older men, the demographic group accounting for the largest share of firearms suicides — with evidence on homicide effects less consistent but generally positive in direction. Research by Swanson and colleagues estimated that one suicide was prevented for approximately every 10–20 ERPOs issued, a substantial preventive effect given the scale of firearm suicide (Swanson et al., 2017).
The crisis intervention logic of ERPOs — temporarily reducing access to the means of self-harm during acute crisis, when the individual’s own long-term preferences are most likely to be violated by impulsive action — is well-supported by the broader literature on means restriction as suicide prevention. Research documenting that suicidal crises are often acute, time-limited, and method-substitution-resistant — that individuals who survive serious suicide attempts rarely go on to die by suicide, and that individuals denied access to highly lethal methods do not reliably substitute other methods — supports the specific crime prevention logic of ERPOs and of safe storage requirements more broadly.
Community Violence Intervention
Community violence intervention programs — which deploy credible messengers with lived experience of violence to interrupt conflicts, mediate disputes, and connect high-risk individuals to services — represent the most direct policy response to the specific social dynamics of gun homicide in American cities, where most firearms violence is concentrated in the gang and social network conflicts of highly disadvantaged urban communities. The evidence for community violence intervention’s effectiveness in reducing gun violence has grown substantially, with evaluations of programs including Cure Violence, the Group Violence Intervention, and hospital-based violence intervention programs showing meaningful reductions in violent incidents and gunshot victims in targeted communities.
The concentration of firearms homicide in specific social networks and geographic locations within American cities — documented in research by Andrew Papachristos and colleagues using social network analysis to trace violence transmission along network ties — has fundamental implications for intervention design (Papachristos et al., 2015). If gun homicide is network-concentrated, not random, then interventions that disrupt the network transmission of violence — by changing norms, interrupting retaliatory cycles, and providing alternative paths for conflict resolution — are more efficient than interventions that treat the full population as equally at risk. Papachristos’s research in Chicago found that approximately 70% of all gunshot victimizations occurred in a social network comprising less than 6% of the city’s population, a concentration that makes targeted network-based intervention far more cost-effective than diffuse population-level approaches.
The RAND Corporation’s Gun Policy in America project — a systematic review of the evidence on 18 specific gun policies across multiple outcomes — provides the most comprehensive summary of the gun policy evidence base available, rating each policy-outcome combination by the quality and consistency of available evidence. RAND’s findings are consistent with the conclusions of this article: permit-to-purchase and background check expansion have moderate to supportive evidence for reducing firearms homicide; child access prevention laws have supportive evidence for reducing youth firearms injuries; stand-your-ground laws have some evidence of increasing homicide; and the evidence on many other commonly discussed policies is inconclusive due to insufficient research rather than evidence of ineffectiveness.
Stand-Your-Ground Laws and Permissive Carrying
Stand-your-ground laws — which remove the duty to retreat before using lethal force in self-defense, expanding the legal conditions under which firearms use is legally justifiable — have been evaluated in multiple studies exploiting variation in the timing of state adoption. Research by Cheng and Hoekstra using state panel data found that stand-your-ground adoption was associated with approximately 8% increases in homicide rates in adopting states, with the effect concentrated in conflicts between acquaintances and strangers rather than in criminal victimization contexts — consistent with the interpretation that expanded legal justification for lethal force escalates conflicts that would otherwise be resolved non-lethally (Cheng & Hoekstra, 2013). The RAND Corporation’s systematic review rated the evidence on stand-your-ground laws as supportive of their increasing homicide, making it one of the few gun policies for which RAND found evidence of crime increase rather than crime reduction.
Shall-issue concealed carry laws — which require law enforcement to issue carry permits to all applicants meeting statutory criteria rather than maintaining discretion to deny permits on public safety grounds — have been examined in a contested research literature where multiple studies using the same general approach have reached substantially different conclusions depending on methodological choices. John Lott’s influential “more guns, less crime” thesis, based on county-level panel data analyses, found that shall-issue laws reduced violent crime; subsequent research by Duggan and others using similar designs found null or positive crime effects, and the National Academies of Sciences‘ review of the evidence found the existing research too methodologically heterogeneous to support confident conclusions in either direction (Duggan, 2001; National Academies, 2004). The RAND Gun Policy in America project’s review rated the evidence on concealed carry laws as inconclusive for violent crime outcomes — meaning insufficient evidence to draw confident conclusions rather than evidence of no effect — a more cautious assessment than the confident claims on both sides of this specific policy debate.
Mass Shootings: Frequency, Pattern, and Policy
Mass shootings — events in which four or more people are shot, a definition used by Gun Violence Archive — have become a defining feature of American public life in the 21st century, generating intense policy debates about the specific regulatory and institutional changes that might reduce their frequency and lethality. Research on mass shooting patterns has documented that they are increasing in frequency, that they disproportionately occur in spaces — schools, workplaces, houses of worship, entertainment venues — where potential targets are concentrated and often unarmed, and that they are predominantly perpetrated by individuals with documented histories of domestic violence, mental health crisis, or prior threatening behavior that background check and red flag mechanisms might have identified before the event.
The debate about the crime prevention effectiveness of armed civilians in public spaces — the proposition that more widespread carrying of firearms by law-abiding citizens would reduce mass shooting casualties by enabling bystander intervention — has been evaluated in research by Everytown for Gun Safety and by criminological researchers who have examined the actual frequency and effectiveness of armed civilian intervention in mass shooting events. The research consistently finds that armed civilian intervention in mass shootings is extremely rare, that most mass shootings end before police arrival through the shooter’s suicide or departure rather than through armed civilian resistance, and that the presence of armed civilians in complex, high-stress situations creates genuine risks of additional casualties through friendly fire, misidentification, and escalating dynamics that make the armed civilian defense scenario substantially less viable in practice than in theory. The evidence does not support the proposition that expanding civilian carrying would meaningfully reduce mass shooting casualties, and the FBI’s active shooter research documents that the specific dynamics of active shooter events are poorly suited to effective civilian armed intervention in the vast majority of cases.
Firearms Suicide and the Neglected Majority
Firearms suicide accounts for more annual deaths than firearms homicide in the United States, yet receives substantially less policy attention and research investment — a disproportion that reflects both the lower political salience of suicide relative to homicide and the different demographics of each problem. While firearms homicide disproportionately affects young Black men in urban environments, firearms suicide disproportionately affects older white men in rural environments — a demographic reversal that tracks the political geography of gun policy advocacy in ways that are not coincidental. Addressing firearms suicide requires policies calibrated to the specific population at highest risk: safe storage requirements that encourage the secure storage of firearms in homes where at-risk individuals live, ERPO laws that enable temporary removal during acute mental health crises, and lethal means counseling programs that work with mental health providers to address firearms access as a component of suicide risk management.
Research on the effectiveness of safe storage promotion — counseling programs that encourage firearms owners to store firearms securely when not in use — has found positive effects on storage behaviors in randomized trials, with the largest effects among programs that provide low-cost storage devices alongside the counseling message. The Brady Campaign’s End Family Fire initiative and the American Foundation for Suicide Prevention‘s lethal means counseling training for healthcare providers represent two implementation pathways for safe storage promotion that the research evidence supports, and that address the firearms suicide problem through means compatible with a wide range of positions on firearms regulation more broadly.
The Political Economy of Gun Policy
The persistent gap between what criminological and public health evidence supports for gun policy and what American politics produces has been extensively analyzed as a political economy problem that evidence quality alone cannot resolve. The National Rifle Association and allied organizations have built one of the most effective political advocacy infrastructures in American history — through campaign contributions, voter mobilization, legislative lobbying, and the cultivation of firearms ownership as a powerful political identity — that has made firearms regulation extraordinarily difficult to achieve even when public opinion polls show majority support for specific measures. The Dickey Amendment’s two-decade suppression of federal firearms research was itself a product of this political advocacy, representing the most direct intervention of firearms industry political power into the scientific research process in contemporary American history.
The political economy of gun policy has begun shifting in the 2010s and 2020s, driven by the proliferation of mass shootings at schools, entertainment venues, and houses of worship that has mobilized a counter-advocacy movement. Organizations including Everytown for Gun Safety, Giffords, and the Brady Campaign have developed political infrastructure increasingly competitive with the NRA’s, contributing to the passage of state-level gun legislation in several states and to the first federal gun legislation in nearly 30 years through the Bipartisan Safer Communities Act of 2022. The BSCA’s provisions — enhanced background check procedures for under-21 purchasers, incentives for states to adopt red flag laws, closure of the boyfriend loophole in domestic violence firearms restrictions, and funding for crisis intervention programs — represent a partial evidence-based policy response that the political balance of forces made achievable in 2022, even if substantially less than the evidence base would support in a less politically constrained environment.
Conclusion
Gun policy and crime in America requires engagement with an evidence base that is both clearer than political polarization suggests and more complex than the simple “more guns, more crime / fewer guns, less crime” framing that advocates on both sides deploy. The evidence clearly establishes that firearms availability causally elevates homicide rates — at the individual level, the state level, and the cross-national level — and that specific regulatory interventions, particularly permit-to-purchase requirements, background check expansion, and extreme risk protection orders, reduce firearms violence with effect sizes comparable to or larger than many conventional criminal justice interventions.
The evidence also establishes that the specific policy interventions with the strongest evidence are not necessarily those most prominently featured in American gun policy debates. Permit-to-purchase requirements — which have more evidence of effectiveness than assault weapons restrictions for reducing firearms homicide — receive less political attention because they are more demanding of firearms purchasers. Extreme risk protection orders — which address the specific mechanism of acute crisis access that drives the large majority of firearms suicides — are more politically feasible than comprehensive registration requirements but less well-known among the general public. And community violence intervention — which addresses the specific social dynamics of urban gun homicide most directly — requires investment in social services and community organizations rather than in enforcement or regulation, placing it outside the conventional gun policy debate while representing the most direct evidence-based response to the gun violence problem that most Americans identify as most urgent.
References
- DiMaggio, C., Avraham, J., Berry, C., Bukur, M., Feldman, J., Klein, M., Shah, N., Tandon, M., & Frangos, S. (2019). Changes in US mass shooting deaths associated with the 1994–2004 federal assault weapons ban. Journal of Trauma and Acute Care Surgery, 86(1), 11–19. https://doi.org/10.1097/TA.0000000000002060
- Hemenway, D., & Miller, M. (2000). Firearm availability and homicide rates across 26 high-income countries. Journal of Trauma, 49(6), 985–988. https://doi.org/10.1097/00005373-200012000-00001
- Leigh, A., & Neill, C. (2010). Do gun buybacks save lives? Evidence from panel data. American Law and Economics Review, 12(2), 509–557. https://doi.org/10.1093/aler/ahq013
- RAND Corporation. (2023). The science of gun policy: A critical synthesis of research evidence on the effects of gun policies in the United States (3rd ed.). https://www.rand.org/pubs/research_reports/RR2088-1.html
- Swanson, J. W., Easter, M. M., Alanis-Hirsch, K., Belden, C. M., Norko, M. A., Robertson, A. G., Frisman, L. K., Bhatt, M., Parker, G., Lin, Z., & Swartz, M. S. (2017). Criminal justice and suicide outcomes with Indiana’s risk-based gun seizure law. Journal of the American Academy of Psychiatry and the Law, 47(2), 188–197.
- Webster, D. W., Crifasi, C. K., & Vernick, J. S. (2014). Effects of Missouri’s repeal of its handgun purchaser licensing law on homicides. Journal of Urban Health, 91(2), 293–302. https://doi.org/10.1007/s11524-014-9865-8