The Association Between State Laws Regulating Handgun Ownership and Statewide Suicide Rates

Abstract

Objectives. We examined the impact of 3 state laws (permit to purchase a handgun, registration of handguns, license to own a handgun) on suicide rates.

Methods. We used 2010 data from publicly available databases and state legislatures to assess the relationships between our predictors and outcomes.

Results. Results largely indicated that states with any of these laws in place exhibited lower overall suicide rates and suicide by firearms rates and that a smaller proportion of suicides in such states resulted from firearms. Furthermore, results indicated that laws requiring registration and license had significant indirect effects through the proportion of suicides resulting from firearms. The latter results imply that such laws are associated with fewer suicide attempts overall, a tendency for those who attempt to use less-lethal means, or both. Exploratory longitudinal analyses indicated a decrease in overall suicide rates immediately following implementation of laws requiring a license to own a handgun.

Conclusions. The results are thus supportive of the potential of handgun legislation to have an impact on suicide rates.

In 2010, suicide ranked as the 10th leading cause of death in the United States, and the use of firearms accounted for 50.5% of these deaths.¹ The fact that firearms account for such a high percentage of suicide deaths can largely be attributed to the high likelihood that self-inflicted gunshot wounds will result in death,^2–4 with reported lethality rates between 82.5% and 92%.^2,5,6

These findings point toward the potential importance of means restriction. Indeed, studies have demonstrated that restricting access to means has a robust impact on suicide rates whether one is looking at firearms^7–9 or other methods.^10–12 A concern with means restriction is the potential that an individual, thwarted in an effort to use one means, will substitute an alternative method, thereby reducing suicide rates by one method but not having an impact on rates as a whole. Although there is some evidence supportive of method substitution, that support is not universal.¹³ For instance, past data indicate that following the detoxification of gas in England, rates of suicide by other methods increased slightly (although overall rates dropped sharply)¹⁰; however, Lester and Abe¹⁴ did not observe the same increase in alternative methods in Japan. Furthermore, Lester¹⁵ only observed an increase in suicides by other means in men in the United States after the detoxification of gas.

Numerous potential methods for means restriction exist but one of the most controversial is legislation. Given the importance of firearms for many individuals within the United States^16,17 and different interpretations of individual rights listed in the Second Amendment,¹⁸ discussion of the potential impact of legislation on suicide is complicated. Nonetheless, empirical work examining this relationship exists and warrants discussion. Lester and Murrell¹⁹ reported that states with gun control laws enacted in 1968 had lower firearm-related suicide rates. Ludwig and Cook²⁰ noted a reduction in firearm-related suicides in individuals at least 55 years of age, with a substantially stronger effect within states that implemented both background checks and waiting periods (vs background checks only) as part of the Brady Act. More recently, Fleegler et al.²¹ found that states with stricter firearms legislation had lower firearm-related suicide rates.

Firearms legislation has also been associated with reductions in overall suicide rates.^19,22,23 Miller et al.²⁴ reported an association between the rate of handgun ownership and overall suicide rates at the state and national level. Boor and Bair²⁵ also found an association between handgun control laws and overall suicide rates in states with laws affecting both buyers and sellers of handguns. Furthermore, Yang and Lester²⁶ found that states with stricter gun control laws had lower overall and firearm-related suicide rates. In Washington, DC, the suicide rate abruptly decreased by 23% after individuals were required to obtain a license to buy a firearm.⁸

Despite these initial promising results, research specifically examining the impact of laws regarding handguns has been limited in scope. Past studies have found that handguns were used in 43% to 72% of firearms-related suicides in Iowa, Washington, and Tennessee between 1987 and 1991.^27,28 Furthermore, data from the National Violent Death Reporting System²⁹ indicate that, in 2011, handguns accounted for 65.4% of all firearm suicides across the 17 states involved within that reporting system. In addition, previous data indicate that, among individuals who purchased a handgun in California, suicide was the leading cause of death in the year the handgun was purchased. Moreover, there was an increase in the risk of suicide by firearms within a week after the handgun was purchased.³⁰ More recently, home access to handguns and not long guns has been found to be associated with an increased risk of a suicide attempt.³¹

The primary aim of this project was to investigate whether suicide-related outcomes are associated with laws requiring a permit to purchase handguns, a license to own handguns, and the registration of purchased handguns. We chose these laws because they (and not others) are tracked by the National Rifle Association Institute for Legislative Action. We selected this site in an effort to diminish concern that our selection of laws was biased toward the regulation of handguns. In this sense, we believed this approach would provide the most objective evaluation of the data and set the foundation for future work examining a broader range of laws. We hypothesized that states that had these laws in place in 2010 would have lower suicide rates, fewer suicides by firearms, and a lower proportion of death by suicide resulting from self-inflicted gunshot wounds during that same year. Furthermore, we hypothesized that the effect of these laws on overall suicide rates would be largely explained by the lower proportion of suicide deaths accounted for by firearms in states with these laws in place. Finally, we ran a series of exploratory analyses examining whether changes in these laws had an impact on suicide outcomes in the years immediately following the legislative changes. Should the data support our hypotheses, this would indicate that legislation regulating handguns has the potential to have a meaningful impact on suicide rates.

METHODS

To investigate the relationships between handgun laws and death by suicide across states, we used the Web-Based Injury Statistics Query and Reporting System, which is accessible on the Web site for the Centers for Disease Control and Prevention.¹ This system derives data from the National Center for Health Statistics and allows individuals to access annual state-level data regarding total number of suicide deaths, demographic characteristics of suicide decedents, and the frequency of use for specific methods. The data thus represented an optimal choice for this study as they provided reliable publically accessible records of our variables of interest across multiple years and within each state. Cross-sectional analyses used data from 2010, which was the most recent year available at the time of data extraction.

We originally accessed data on firearm laws via the National Rifle Association Institute for Legislative Action Web site.³² This Web site provided state-by-state information on requirements for permits, firearm registration, and owner licensing. To determine whether any states enacted or repealed relevant state gun laws during the time frame for which suicide data were available via the Web-Based Injury Statistics Query and Reporting System Web site, we consulted the Web site for Law Center to Prevent Violence.³³ From this Web site, researchers obtained information regarding changes in specific state legislation on gun laws. Three states (California, Maryland, and Michigan) underwent changes in firearm laws, so we consulted the appropriate state legislatures to determine the exact years in which these changes occurred.^34–36

In an effort to control for potential confounders, we also considered the percentage of each state under the poverty line as well as state-by-state population density. Both of these covariates have previously been shown to exhibit significant associations with statewide suicide rates.^37,38 In addition, these variables enable us to statistically control for at least a portion of the impact of burdensomeness, social isolation, access to mental health resources, and other robust risk factors for suicidal behavior. Although other potentially important variables were worth consideration as covariates (e.g., religiosity, psychopathology, cultural views toward suicide), such variables are difficult to quantify and 2010 state-by-state data for those variables were not available. We obtained information on population density from the 2010 US Census Web site.³⁹ We obtained data regarding poverty levels by state through the 2010 US Census as well.⁴⁰

To test the association between the presence of statewide handgun laws and suicide-related outcomes with control for statewide poverty rates and population density, we used analyses of covariance. We conducted these analyses with IBM SPSS Statistics for Windows, version 20.0 (IBM Corp, Armonk, NY). For each law, we tested for between-group differences on 3 outcomes: the mean statewide suicide rate for 2010, the mean statewide suicide by firearm rate for 2010, and the mean statewide proportion of suicide deaths by firearms for 2010. We examined partial η squared as an index of effect size. All rates derived from our between-group analyses will be presented as aggregate rather than average rates.

To test our hypothesis that the effect of 2010 firearm laws on 2010 suicide-related outcomes was indirect through (largely explained by) the percentage of statewide suicides resulting from self-inflicted gunshot wounds, we used bootstrapping methods (10 000 resamples, 95% bias-corrected confidence intervals), with the PROCESS macro for SPSS developed by Hayes.⁴¹ Predictors and outcomes mirrored those from the analyses of covariance. We examined κ² and the ratio of the indirect effect to the total effect as indices of effect size.

Given that 3 states had firearm laws change during the years for which state-by-state suicide rates were available (explained in the “Follow-Up Analyses” section), we lacked sufficient statistical power to conduct analyses of the effect these changes may have had on suicide rates. However, we conducted exploratory analyses to examine the extent to which changes in suicide rates corresponded with newly implemented firearm licensing laws.

RESULTS

Information regarding 2010 population and suicide-related outcomes by state can be found in Table 1.

TABLE 1—

State-by-State Population, Suicide Rate, and Handgun Law Status Information: United States, 2010

State or District	Permita	Registrationb	Licensec	2010 Population	2010 Suicide Rated	Suicide Rate Ranke	Suicide Firearm,f %
Alabama	No	No	No	4 779 736	14.21	24	67
Alaska	No	No	No	710 231	23.09	2	65
Arizona	No	No	No	6 398 017	17.08	11	57
Arkansas	No	No	No	2 915 918	15.33	15	60
California	Yes	No	Yes	37 253 956	10.50	44	38
Colorado	No	No	No	5 029 196	17.20	8	49
Connecticut	Yes	No	Yes	3 574 097	9.88	45	31
Delaware	No	No	No	897 934	11.80	39	41
District of Columbia	Yes	Yes	Yes	601 723	6.81	51	32
Florida	No	No	Yes	18 801 310	14.83	20	52
Georgia	No	No	Yes	9 687 653	11.70	40	63
Hawaii	Yes	Yes	No	1 360 301	15.22	16	18
Idaho	No	No	Yes	1 567 582	18.50	6	63
Illinois	Yes	No	Yes	12 830 632	9.18	46	38
Indiana	No	No	Yes	6 483 802	13.33	29	53
Iowa	Yes	No	No	3 046 355	12.21	37	48
Kansas	No	Yes	Yes	2 853 118	14.05	25	52
Kentucky	No	No	Yes	4 339 367	14.54	21	64
Louisiana	No	No	No	4 533 372	12.29	35	69
Maine	No	No	No	1 328 361	14.00	26	51
Maryland	Yes	No	Yes	5 773 552	8.69	48	44
Massachusetts	Yes	No	Yes	6 547 629	9.13	47	23
Michigan	Yes	Yes	No	9 883 640	12.78	31	48
Minnesota	Yes	No	No	5 303 925	11.43	42	46
Mississippi	No	No	No	2 967 297	13.08	30	66
Missouri	No	No	No	5 988 927	14.29	22	57
Montana	No	No	No	989 415	22.94	3	62
Nebraska	Yes	No	No	1 826 341	10.57	43	55
Nevada	No	No	No	2 700 551	20.26	4	53
New Hampshire	No	No	No	1 316 470	14.89	18	52
New Jersey	Yes	Yes	Yes	8 791 894	8.18	49	26
New Mexico	No	No	No	2 059 179	20.06	5	49
New York	Yes	Yes	Yes	19 378 102	7.98	50	30
North Carolina	Yes	Yes	No	9 535 483	12.31	34	60
North Dakota	No	No	No	672 591	15.76	14	53
Ohio	No	No	No	11 536 504	12.47	32	50
Oklahoma	No	No	No	3 751 351	16.47	13	61
Oregon	No	No	No	3 831 074	17.88	7	55
Pennsylvania	No	No	No	12 702 379	12.41	33	48
Rhode Island	No	No	No	1 052 567	12.26	36	23
South Carolina	No	No	No	4 625 364	13.77	28	62
South Dakota	No	No	No	814 180	17.20	9	46
Tennessee	No	No	No	6 346 105	14.86	19	62
Texas	No	No	No	25 145 561	11.50	41	59
Utah	No	No	No	2 763 885	17.11	10	58
Vermont	No	No	No	625 741	16.94	12	62
Virginia	No	No	No	8 001 024	12.04	38	60
Washington	No	No	No	6 724 540	14.23	23	48
West Virginia	No	No	No	1 852 994	15.06	17	75
Wisconsin	No	No	No	5 686 986	13.94	27	48
Wyoming	No	No	No	563 626	23.24	1	63

Note. Each law scored such that 0 = no such law in place in a specific state and 1 = a law is in place in a specific state.

^aPermit = law in place requiring an individual to have a permit to purchase a handgun.

^bRegister = law in place requiring an individual to register any handguns that he or she owns.

^cLicense = law in place requiring a license to own a handgun.

^dSuicide rate = number of individuals per 100 000 who died by suicide.

^eSuicide rate rank = ranking relative to all US states and District of Columbia in 2010 suicide rate per 100 000. Lower numerical rank indicates higher suicide rate.

^f2010 firearms = proportion of 2010 deaths by suicide that resulted from self-inflicted gunshot wounds.

A total of 14 states (out of 50 states and the District of Columbia) had a law in place requiring a permit to purchase handguns in 2010, 7 states had a law in place requiring registration of handguns in 2010, and 14 states had a law in place requiring a license to own a handgun in 2010. A summary of which states did and did not have each of these laws in place can be found in Table 1.

Primary Analyses

Test of whether states in which a higher proportion of suicide deaths result from self-inflicted gunshot wounds account for a disproportionate amount of suicidal behavior.

The 9 states with the lowest percentage of suicide deaths attributable to firearms (18%–38%), accounted for 91 090 901 individuals (29.5% of the total US population), 8685 deaths by suicide (22.6% of the national total), and 2906 deaths by suicide resulting from firearms (15.0% of the national total). If each of these states exhibited suicide rates equivalent to the national average (12.43 per 100 000), a total of 11 381.52 deaths by suicide would have been expected. As such, a total of 2633.52 fewer deaths occurred than would have been anticipated.

The 9 states with the highest percentage of suicide deaths attributable to firearms (63%–75%), accounted for 31 001 858 individuals (10.0% of the total US population), 4252 deaths by suicide (11.1% of the national total), and 2772 deaths by suicide resulting from firearms (14.3% of the national total). Despite having 60 089 043 fewer residents, there were only 134 fewer deaths by suicide attributable to firearms in these states relative to the 9 states with the lowest percentage of death by suicide attributable to firearms. If each of these states exhibited suicide rates equivalent to the national average (12.43 per 100 000), a total of 3852.14 deaths by suicide would have been expected. As such, a total of 399.86 more deaths occurred than would have been anticipated (Table 2).

TABLE 2—

Suicide Data and Handgun Law Status in the States With the 9 Highest and 9 Lowest Proportion of Suicide Deaths Accounted for by Firearms: United States, 2010

State	2010 Population	Suicide Ranka	Firearm, %b	Total Suicidesc	Expected Suicidesd	Differencee	Firearm Suicidesf
9 lowest percentages of suicides attributable to firearms
Hawaii	1 360 301	16	18	207	169.02	38.98	37
Massachusetts	6 547 629	47	23	598	813.58	−215.58	136
Rhode Island	1 052 567	36	23	129	130.79	−1.79	30
New Jersey	8 791 102	49	26	719	1 092.44	−373.44	187
New York	19 378 102	50	30	1 547	2 407.83	−860.83	459
Connecticut	3 574 097	45	31	353	444.10	−91.10	110
District of Columbia	301 723	51	32	41	37.49	3.51	13
Illinois	12 830 632	46	38	1 178	1 594.27	−416.27	442
California	37 253 956	44	38	3 913	4 629.00	−716.00	1 492
Totals	91 090 109			8 685	11 318.52	−2 632.52	2 906
% of nation	29.5			22.6			29.5
9 highest percentages of suicides attributable to firearms
West Virginia	1 852 994	17	75	279	230.24	48.74	210
Louisiana	4 533 372	35	69	557	563.60	−6.30	358
Alabama	4 779 736	24	67	679	593.91	85.09	454
Mississippi	2 967 297	30	66	388	368.70	19.30	256
Alaska	710 231	2	65	164	88.25	75.75	107
Kentucky	4 339 367	21	64	631	539.19	91.81	404
Georgia	9 687 653	40	63	1 133	1 203.74	−70.74	718
Wyoming	563 626	1	63	131	73.03	60.97	83
Idaho	1 567 582	6	63	290	194.78	95.22	182
Totals	31 001 858			4 252	3 855.44	399.84	2 772
% of nation	10.0			11.1	10.0		14.3

^aSuicide rank = ranking relative to all US states and District of Columbia in 2010 suicide rate per 100 000. Lower rank indicates higher suicide rate relative to other states.

^bFirearm % = percentage of 2010 suicides attributable to self-inflicted gunshot wound.

^cTotal suicides = total number of deaths by suicide in 2010.

^dExpected suicides = number of deaths by suicide expected if state exhibited national suicide rate (12.43 per 100 000).

^eDifference = total 2010 deaths by suicide minus expected suicides.

^fFirearm suicides = total number of deaths by suicide attributable to self-inflicted gunshot wound in 2010.

Test of whether the presence of a law requiring a permit to purchase handguns is associated with suicide-related outcomes.

Results indicated that states that had a law in place in 2010 requiring a permit to purchase a handgun exhibited a lower rate of death by suicide (10.3 per 100 000 vs 15.4 per 100 000; F = 23.8; P < .001; _pµ² = 0.34), a lower rate of suicide by firearms (4.0 per 100 000 vs 8.8 per 100 000; F = 27.20; P < .001; R² = 0.50; _pµ² = 0.37), and a lower percentage of deaths by suicide caused by firearms relative to states without such a law (38% vs 56%; F = 16.43; P < .001; R² = 0.52; _pµ² = 0.26; Tables 3 and 4).

TABLE 3—

Suicide-Related Outcomes in States With Versus Without Specific Laws Regulating Handgun Ownership: United States, 2010

Laws	No.	Statewide Suicide Rate for 2010,a Mean (SD)	Statewide Suicide by Firearm Rate for 2010,a Mean (SD)	Statewide Percentage of Suicide Deaths by Firearms for 2010, Mean (SD)
Permit to purchase handguns
Yes	14	10.3 (2.27)	4.0 (1.76)	38 (12)
No	37	15.5 (3.20)	8.8 (2.51)	56 (9)
Registration of handguns
Yes	7	11.0 (3.33)	4.3 (2.51)	38 (16)
No	44	14.6 (3.63)	8.0 (2.97)	53 (11)
License to own handguns
Yes	14	11.2 (3.35)	5.3 (2.78)	43 (14)
No	37	15.2 (3.34)	8.3 (3.09)	54 (11)

^aStatewide suicide rates (total and by firearms) are presented as number of individuals per 100 000.

TABLE 4—

Descriptive Data and Between-Group Differences on Suicide-Related Outcomes in States With vs Without Specific Laws Regulating Handgun Ownership: United States, 2010

	Statewide Suicide Rate for 2010			Statewide Suicide by Firearm Rate for 2010			Statewide Proportion of Suicide Deaths by Firearms for 2010
	R2	b (95% CI)	pµ2	R2	b (95% CI)	pµ2	R2	b (95% CI)	pµ2
Permit to purchase handguns	0.44	4.968*** (2.919, 7.017)	0.34	0.50	4.204*** (2.582, 5.826)	0.37	0.52	0.130*** (0.066, 0.195)	0.26
% of state below poverty line		–0.131 (–0.413, 0.151)	0.02		0.085 (–0.138, 0.308)	0.01		0.014** (0.005, 0.023)	0.18
Population density		0.000 (–0.001, 0.000)	0.05		0.000 (–0.001, 0.000)	0.06		0.000* (0.000, 0.000)	0.10
Registration of handguns	0.19	2.240 (–0.884, 5.364)	0.04	0.28	2.638* (0.170, 5.106)	0.09	0.43	0.112* (0.022, 0.201)	0.12
% of state below poverty line		0.081 (–0.239, 0.401)	0.01		0.259* (0.006, 0.512)	0.08		0.019*** (0.010, 0.029)	0.28
Population density		–0.001* (–0.002, 0.000)	0.09		–0.001* (–0.001, 0.000)	0.08		0.000* (0.000, 0.000)	0.11
License to own handguns	0.29	3.259** (1.073, 5.445)	0.16	0.31	2.352* (0.546, 4.158)	0.13	0.41	0.072 (0.004, 0.140)	0.09
% of state below poverty line		0.053 (–0.247, 0.354)	0.00		0.246 (–0.002, 0.495)	0.08		0.019*** (0.010, 0.029)	0.27
Population density		–0.001 (–0.001, 0.000)	0.08		–0.001* (–0.001, 0.000)	0.10		0.000** (0.000, 0.000)	0.14

Note. CI = confidence interval. Reference to laws, % of state below poverty line, and population density refer to status and numbers from 2010.

*P < .05; **P < .01; ***P < .001.

Test of whether the presence of a law requiring registration of handguns is associated with suicide-related outcomes.

Results indicated that states that had a law in place in 2010 requiring registration of handguns exhibited a lower rate of death by suicide relative to states without any such law (11.0 per 100 000 vs 14.6 per 100 000), but this effect became nonsignificant after we controlled for the percentage of the state below the poverty line and state population density (F = 2.08; P = .16; R² = 0.19; _pµ² = 0.04). Even after we accounted for the effects of poverty and population density, states that had a law in place in 2010 requiring registration of handguns exhibited a lower rate of suicide by firearms (4.3 per 100 000 vs 8.0 per 100 000; F = 4.62; P = .037; R² = 0.28; _pµ² = 0.09) and a lower percentage of deaths by suicide caused by firearms (38% vs 53%; F = 6.28; P = .016; R² = 0.43; _pµ² = 0.12; Tables 3 and 4).

Test of whether the presence of a law requiring a license to own a handgun is associated with suicide-related outcomes.

Results indicated that states that had a law in place in 2010 requiring a license to own a handgun exhibited a lower rate of death by suicide (11.2 per 100 000 vs 15.2 per 100 000; F = 8.99; P = .004; R² = 0.29; _pµ² = 0.16), a lower rate of suicide by firearms (5.3 per 100 000 vs 8.3 per 100 000; F = 6.86; P = .012; R² = 0.31; _pµ² = 0.13), and a lower percentage of deaths by suicide caused by firearms relative to states without such a law (43% vs 54%; F = 4.52; P = .039; R² = 0.41; _pµ² = 0.09; Tables 3 and 4).

Test of whether the association between the presence of these 3 gun laws and rates of death by suicide is explained by the proportion of deaths by suicide resulting from self-inflicted gunshot wounds in each state.

Results from bootstrapping analyses indicated a significant indirect effect of the presence of a law requiring the registration of handguns on 2010 statewide suicide rate through the 2010 percentage of statewide suicides resulting from firearms. Similarly, results from bootstrapping analyses indicated a significant indirect effect of the presence of a law requiring a license to own handguns on 2010 statewide suicide rate through the 2010 percentage of statewide suicides resulting from firearms. These results indicate that the impact of these 2 laws on overall suicide rates is largely explained by the lower percentage of suicide deaths accounted for by firearms in states with such laws in place. Bootstrapping analyses did not support an indirect effect of the presence of a law requiring a permit to purchase handguns and 2010 statewide suicide rates through 2010 percentage of statewide suicides resulting from firearms. Results did, however, indicate that the presence of a law requiring a permit to purchase handguns maintained a significant, direct relationship with statewide suicide deaths (b = −2.06; SE = 0.59; P = .001; Table 5).

TABLE 5—

Indirect Effect of Handgun Laws on Statewide Suicide Rates Through the Proportion of Suicide Deaths Accounted for by Firearms: United States, 2010

Independent Variable	Dependent Variable	Adjusted R2	Path	b ±SE	P	Path	Bootstrap Coefficient ±SE (95% CI)	Ratio I:Ta	κ2
Purchase permitb	Statewide suicide	0.41	Total	−2.59 ±0.47	< .001	Indirect	−0.53 ±0.37 (−1.31, 0.18)	0.20	0.13
	Rate		Direct	−2.06 ±0.59	.001
Register handgunc	Statewide suicide	0.22	Total	−1.97 ±0.53	< .001	Indirect	−0.61 ±0.33 (−1.48, −0.14)	0.31	0.15
	Rate		Direct	−1.37 ±0.52	.012
License to ownd	Statewide suicide	0.11	Total	−1.77 ±0.73	.019	Indirect	−1.03 ±0.48 (−2.29, −0.28)	0.58	0.19
	Rate		Direct	−0.75 ±0.73	.311

Note. CI = confidence interval. In each analysis, the indirect effect is through the statewide proportion of suicide deaths accounted for by self-inflicted gunshot wounds. Each law is scored such that 0 = no such law in place in a specific state and 1 = a law is in place in a specific state.

^aI:T = ratio of the indirect effect to the total effect.

^bPurchase permit = law in place requiring an individual to have a permit to purchase a handgun.

^cRegister handgun = law in place requiring an individual to register any handguns that he or she owns.

^dLicense to own = law in place requiring a license to own a handgun.

Follow-Up Analyses

Only 3 states enacted or repealed any of the 3 laws during the time period during which we were able to gather data 2 years before and 2 years after enactment of the change. Specifically, California enacted laws requiring a permit to purchase handguns and a license to own handguns in 2003, Maryland enacted a law requiring a license to own a handgun in 2003, and Michigan enacted a law requiring registration of handguns in 2004. (In our interpretation, the new laws in all 3 states [Maryland, California, and Michigan] represented a change from previous legislation, specifically stating information only implied in earlier laws or clarifying ambiguous text. On the basis of the need to clarify the earlier laws, we would assume that enforcement had previously been suboptimal and thus required more specificity in wording.) Because 2 states enacted laws requiring a license to own a handgun in 2003, we opted to focus on that law.

We examined rates in the United States as a whole, states that had no law requiring a license to own a handgun throughout that time period, states that did have a law requiring a license to own a handgun throughout that period, and both California and Maryland. As can be seen in Table 6, all 5 groups saw an overall increase in suicide rates from 2000 to 2006; however, whereas the other 3 groups saw a continued increase from 2003 to 2006, California (–3.9%) and Maryland (–4.8%) saw deceases in their suicide rates during that time period, which immediately followed the enactment of a law requiring a license to own a handgun.

TABLE 6—

Suicide Rates in the Years Preceding and Following Enactment of Firearms Laws in California and Maryland in 2003

Year	United States	Yes Licensea	No Licenseb	California	Maryland
2001
Suicide rate per 100 000	10.75	8.39	12.35	8.21	8.45
% increase in suicide rate from previous year	3.1	6.3	3.5	−6.3	−5.6
% suicide attributable to firearms	55.1	43.5	58.9	51.2	48.5
2002
Suicide rate per 100 000	11.01	8.25	12.59	9.26	8.77
% increase in suicide rate from previous year	2.4	−1.6	1.9	12.7	3.8
% suicide attributable to firearms	54.0	41.6	58.9	44.3	48.7
2003c
Suicide rate per 100 000	10.85	8.00	12.39	9.64	8.93
% increase in suicide rate from previous year	−1.5	−3.0	−1.6	4.1	1.9
% suicide attributable to firearms	53.7	42.3	58.5	44.3	48.7
2004
Suicide rate per 100 000	11.08	8.14	12.70	9.47	9.01
% increase in suicide rate from previous year	2.1	1.7	2.5	−1.7	0.9
% suicide attributable to firearms	51.6	40.5	56.5	40.6	50.0
2005
Suicide rate per 100 000	11.04	8.20	12.72	8.95	8.44
% increase in suicide rate from previous year	−0.4	0.8	0.2	−5.5	−6.4
% suicide attributable to firearms	52.1	41.8	56.6	41.5	46.6
2006
Suicide rate per 100 000	11.16	8.46	12.71	9.26	8.50
% increase in suicide rate from previous year	1.1	3.1	−0.1	3.4	4.2
% suicide attributable to firearms	50.7	38.7	55.6	39.6	48.1
% change in suicide rate from 2001 to 2006	3.8	0.8	2.9	12.8	0.6
% change in suicide rate from 2001 to 2003	0.9	−4.6	0.3	17.4	5.7
% change in suicide rate from 2003 to 2006	2.9	5.8	2.6	−3.9	−4.8

^aYes license = law requiring license to own a handgun in place throughout 2001–2006.

^bNo license = no law in place requiring license to own handgun throughout 2001–2006.

^cLaw requiring license enacted in California and Maryland.

DISCUSSION

The primary aim of this study was to examine the association between 3 state laws regulating handgun ownership (requirement of a permit to purchase a handgun, requirement to register purchased handguns, and requirement of a license to own a handgun) and 3 suicide-related outcomes (statewide suicide rate, statewide suicide by firearms rate, statewide proportion of suicide deaths accounted for by firearms). Results were largely consistent with our hypotheses.

With respect to laws requiring a permit to purchase handguns, results indicated that states with such laws in place in 2010 had lower overall suicide rates, lower suicide by firearms rates, and a smaller proportion of suicide deaths accounted for by firearms. In each case, effect sizes were large (_pη²; small = 0.01; medium = 0.06; large = 0.14).⁴² With respect to laws requiring the registration of purchased permits, results indicated that the effect on the overall suicide rate became nonsignificant after we controlled for poverty and population density, but that states with such laws in place had significantly lower rates of suicide by firearms and a lower proportion of suicide deaths accounted for by firearms. Each significant effect exhibited a medium effect size. With respect to laws requiring a license to own a handgun, results indicated that states with such laws in place had lower overall suicide rates, lower suicide by firearms rates, and a smaller proportion of suicide deaths accounted for by firearms. The effect on the overall suicide rates met criteria for large effect size and the other 2 effects met criteria for a medium effect size.

Such results provide support for the notion that specific laws regulating the ownership of handguns are associated with a range of suicide-related outcomes beyond the effects of poverty and population density. These results speak to the presence of an effect; however, they do not directly test proposed mechanisms. That being said, the nature of the laws, which either slow down the process of and present barriers toward the acquisition of firearms or require individuals to alert the state that they own personal firearms, certainly speak to likely mechanisms. Indeed, the significant between-group differences for license and permit (but not registration) laws is consistent with results from the means restriction literature in that these laws render access to lethal means more difficult, even if only temporarily. The effect size for laws requiring a permit to purchase handguns was more than twice that of any other law, further supporting the notion that creating impediments to the initial acquisition of a handgun has a particularly robust impact on suicide-related outcomes.

Our second set of analyses more directly addressed the issue of mechanisms. These analyses indicated that, for laws requiring registration of or a license to own handguns, there was a significant indirect effect through the proportion of statewide suicide deaths resulting from self-inflicted gunshot wounds. These results indicate that the relationship between those laws and overall statewide suicide rates was largely explained by the lower proportion of suicides attributable to firearms in states with such laws in place. In each case, the effect size for the indirect effect met criteria for a medium-sized effect (κ²; small = 0.01; medium = 0.09; large = 0.25).⁴³ These results indicate that the impact of legislation on overall suicide rates is largely accounted for by a decrease in the percentage of statewide suicide deaths that result from self-inflicted gunshot wounds. This point is further highlighted by the finding that the states with the 9 lowest proportions of suicide deaths accounted for by firearms had more than 60 000 000 more people than the states with the 9 highest proportions, but only 134 additional fatal self-inflicted gunshot wounds (Table 2). By diminishing access to handguns, it appears possible that states with these laws ultimately decrease the number of suicide attempts, lead individuals who would have attempted using a handgun to instead use less-lethal means, or both.

Our tests of indirect effects more directly addressed the issue of mechanisms; however, the cross-sectional nature of the data still precluded strong conclusions. Such data are not capable of truly countering the potential that states with suicide rates already lower for other reasons implemented laws regulating handgun ownership and that those laws did not account for the differences. It is unclear what other variable would have accounted for those differences and our decision to control for poverty and population density partially addressed that limitation; however, consideration of longitudinal trends would offer a clearer picture of causality. Although conclusions drawn from exploratory analyses should be considered with caution, it should be noted that, in the years immediately after implementing a law requiring a license to own handguns, both California and Maryland saw a decrease in their suicide rates whereas the United States as a whole and states that either did or did not have such laws in place throughout that time period saw a continued increase.

Our results provide support for the possibility that legislation limiting access to handguns may have an impact on suicide rates as a whole. It is unclear to what extent additional laws regarding handguns would further influence suicide rates or whether other laws related to handguns may have a more potent impact. Indeed, we chose to use the National Rifle Association Institute for Legislative Action Web site in an effort to diminish concerns that this study was introducing selection bias by selecting laws that spuriously favored the regulation of handguns. In doing so, however, we were unable to track other laws that are likely relevant to suicide-related outcomes (e.g., waiting periods). Future work that considers a broader range of laws would enhance confidence in legislation as an effective tool overall rather than as confined to these 3 specific laws. However, the relatively consistent effects on suicide rates across methods indicate that legislation could be a key aspect of prevention efforts.

Our results should be considered within the context of their limitations. As noted previously, many of our analyses were cross-sectional and those that were longitudinal were underpowered. In addition, it is unclear to what extent our findings extend beyond the United States. As such, implications for global suicide prevention efforts are limited. Third, statewide data for nonlethal suicide attempts were not available, so we were unable to test the possibility that fewer individuals who attempt suicide in states with laws in place ultimately die as a result of their attempt. It was also not possible to rule out other plausible but more difficult to measure explanations for our findings. For instance, local cultural norms related to suicide, firearms, and violence may have had an impact on both the propensity toward specific behaviors and the likelihood that a particular state would pursue and enact gun ownership legislation. Furthermore, although relatively easy to assess on an individual level, statewide levels of other suicide risk factors (e.g., hopelessness, access to adequate mental health care) are difficult to measure and could conceivably have influenced our findings.

Despite these limitations, we believe that our findings exhibit important scientific, clinical, and legal implications. Our data provide support for the potential utility of handgun legislation in efforts to reduce suicide rates. They also speak to the possibility that limiting access to highly lethal means can have a broad effect. These results are not intended as a political statement or a criticism of components of US culture that value firearms. Instead, these are meant as an objective assessment of the impact of legislation on suicide-related outcomes. The cost–benefit analysis of such results with respect to valued cultural norms is a question to be considered by the various individuals involved in the legislative process.