Article Text
PDF
Abstract
Objectives To describe firearm storage practices among US adults and examine the relationship between having received formal firearms training and firearm storage.
Methods In 2015 we asked a nationally representative online sample of 2072 gun owners how they stored household firearms, their reasons for owning guns, the number and types owned, had they carried a loaded handgun in the prior month and whether they had formal firearms training (and if so, whether training covered suicide prevention, accident prevention, firearm theft prevention, safe handling and safe storage). Unadjusted associations between gun owner characteristics and storage practices were estimated using Pearson’s χ2 tests; adjusted associations used multivariate logistic regressions. Final survey weights that combined presample and study-specific poststratification weights account for oversampling of firearm owners and survey non-response.
Results 29.7% (95% CI 27.4% to 32.1%) stored ≥1 firearm loaded and unlocked. Of the 61.4% (95% CI 58.9% to 63.9%) of gun owners with firearms training, 32.3% (95% CI 29.4% to 35.3%) stored ≥1 firearm loaded and unlocked, compared with 25.8% (95% CI 22.3% to 29.7%) of those without training. Storage did not differ by training component, age, sex or race. However, firearms were more likely stored loaded and unlocked when respondents owned for protection, owned >1 firearm, owned handguns or carried a loaded gun. After adjusting for firearm-related characteristics, firearms training was not associated with storing firearms loaded and unlocked (adjusted OR=1.11, 95% Cl 0.80 to 1.53).
Conclusion Firearms training, as currently provided, is unlikely to reduce unsafe firearm storage.
- firearm
- suicide/self-harm
- violence
- descriptive epidemiology
Statistics from Altmetric.com
Introduction
A large body of evidence has shown that the presence of firearms in a person’s home substantially increases the risk of suicide and unintentional firearm death.1–12 For children in homes with firearms, the risk of firearm injury is lower when firearms are unloaded (compared with loaded) and locked (compared with unlocked).13 14
Despite the well-established mortality risk associated with household firearms, especially with respect to suicide and unintentional firearm deaths, few gun owners appear to believe that the presence of guns in the home increases the risk of suicide,15 let alone that storage practices modify this risk or that of unintentional firearm injury. A recent paper that used the survey we report on here estimated that in 2015 approximately 20% of gun owners who lived with children stored at least one household firearm loaded and unlocked, more than twice the proportion estimated using data from the 1990s and early 2000s.16–20 Using data from 2016, others found that 45% of gun owners (those with and those without children) stored at least one gun unlocked and 54% stored at least one gun loaded.21 The proportion of gun owners who stored at least one firearm loaded and unlocked in 2016 was not reported.
Firearms training has long been looked to as a way to improve firearm storage practices. For example, over the last decade public health practitioners have developed and promoted online firearms training modules that incorporate suicide prevention messages that underscore the importance of routine safe storage, as well as temporary removal of firearms from homes with high-risk persons.22 To date, however, the only studies that have addressed whether firearms training is associated with safe storage use data from the 1990s,17–23 23 24 24–29 and, moreover, report either no relationship,18 19 23 or that people who have been trained are less likely to store their firearms safely.17 24
The current study uses data from the 2015 National Firearms Survey16 to examine whether the way gun owners store firearms in their homes is related to (A) having received formal firearms training, (B) specific training components (eg, suicide prevention vs safe handling), and (C) sociodemographic and firearm-related characteristics, including the number and type of guns owned, the reason(s) for owning guns and whether gun owners carried a loaded handgun on their person in the prior month.
Methods
Design and participants
We used data from a web-based nationally representative survey, designed by the investigators (DA, MM) and conducted by the survey firm Growth for Knowledge (GfK) in April 2015, to describe firearm ownership, storage and use in the USA. Respondents were drawn from GfK's Knowledge Panel (KP), a rotating panel which includes approximately 55 000 US adults sampled on an ongoing basis. Panel members are recruited by GfK randomly through probability-based sampling, and households are provided with access to the internet and hardware if needed. For recruiting, GfK uses address-based sampling methods (previously it relied on random-digit dialling methods). GfK assigns no more than one survey per week to individual members of the KP. Invitations to participate were sent by email; one reminder email was sent to non-responders 3 days later. Participants did not receive any specific incentive to complete this survey, although GfK has a point-based programme through which participants accrue points for completing surveys and can redeem them later for cash, merchandise or participation in sweepstakes. All panel members, except those serving in the US Armed Forces at the time of survey, were eligible to participate. To ensure reliable national estimates, firearm owners were oversampled from the KP. Respondents were more likely than non-respondents to be younger, female, unmarried, less educated and living in metropolitan areas. Respondents were about as likely as non-respondents to live in a firearm-owning house, but were more likely to personally own a firearm. Additional details about the survey design and participants are available elsewhere.16
Measures
Gun ownership status
Household firearm ownership was assessed by asking respondents, ‘Do you or does anyone else you live with currently own any type of gun?’ Among those who answered affirmatively, personal gun ownership status was assessed by asking respondents, ‘Do you personally own a gun?’
Only respondents who reported personally owning firearms were asked specific questions about household firearms including the number and types of guns in the home, storage practices and reasons for ownership. Specifically, respondents were asked: ‘Do you personally own any of the following types of guns (handguns, long guns, other guns)?’ and then, for each firearm type owned by the respondent, ‘How many (handguns, long guns, other guns) do you own?’
Gun storage
For each type of firearm (handgun, long gun, other), gun owners were asked to specify the number of guns they stored loaded and unlocked, loaded and locked, unloaded and unlocked, and unloaded and locked. Based on responses to these questions, gun owners were sorted into one of three hierarchical, mutually exclusive and collectively exhaustive categories: (1) those who stored at least one gun loaded and unlocked (the least safe storage method), (2) those who stored no guns loaded and unlocked but at least one gun loaded and locked, or unloaded and unlocked (the intermediate risk category), and (3) those who stored all guns unloaded and locked (the safest storage method). The two types of storage in the intermediate-risk category were combined because the RR of a loaded and locked gun, compared with an unloaded and unlocked gun, was determined to be too context specific to generalise about which is safer. Respondents (n=13) who refused to answer any questions about how their guns were stored were excluded. An additional 48 respondents did not furnish sufficient information to classify them into our storage categories.
Firearms training
To assess experience with firearms training respondents were asked, ‘Have you ever had any formal firearm training?’ Those who responded affirmatively were then asked to indicate if the training included information on (1) safe handling of firearms, (2) safe storage of firearms, (3) preventing firearm accidents, (4) preventing firearm theft, and/or (5) suicide prevention. Respondents could choose zero, some or all of the options.
Demographic, socioeconomic and firearm-related variables
Additional variables included were respondent’s age (18–29, 30–44, 45–59, ≥60), gender, race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, other), education (<high school, high school degree, some college, ≥bachelor’s degree), household income (<$25 000, $25 000–$59 999, $60 000–$99 999, ≥$100 000), political ideology (liberal/moderate/conservative), veteran status, residential region (North-East/Midwest/South/West) and rurality (urban/suburban/rural), and whether children under the age of 18 lived in their home. The types of guns owned by respondents were categorised as: at least one handgun versus own no handguns. Respondents were asked, for each type of gun owned (eg, handgun), their main reasons for ownership (protection against strangers vs people they knew vs animals, hunting, other sporting use or for a collection). For this analysis, we include a binary variable: owns any type of firearm for protection against people versus does not. We also examine additional characteristics plausibly related to storage and having received training, including having grown up in a home with firearms, and whether handgun owners carried a loaded handgun on their person in the last month.
Statistical analysis
Unadjusted associations between participant characteristics and outcome measures were tested using Pearson’s χ2 tests. We adjusted for additional characteristics in multivariate logistic regressions, presenting (A) adjusted ORs (AOR) for storing firearms loaded and unlocked after including an a priori set of demographic covariates (age, sex, presence of children in household, urbanicity, region of residence and veteran status) based on the prior literature, and (B) AORs after additionally adjusting for firearm-related characteristics. Final survey weights that combined presample and study-specific poststratification weights to account for oversampling and survey non-response were provided by GfK. To ensure results representative of the US adult (aged ≥18 years) population in 2015 we used these weights in all analyses and in reporting weighted percentages and 95% CIs. All analyses were conducted in Stata V.14 (StataCorp, College Station, Texas) using the ‘svy’ suite of commands.
Results
A total of 7318 panel members received the survey; 4165 began the survey and 3949 completed it, yielding a survey completion proportion of 55%.25 Of the 2072 gun owners in our sample, 2011 provided sufficient information to classify them into one of our three mutually exclusive and collectively exhaustive firearm storage categories.
Seventy-two per cent of our gun owners were men (72.1%, 95% CI 69.8% to 74.4%); most were non-Hispanic White (80.5%, 95% CI 78.0% to 82.7%), lived in either suburban (44.2%, 95% CI 41.7% to 46.7%) or rural (39.6, 95% CI 37.2 to 42.2) areas and did not have a bachelor’s degree or higher (70.9%, 95% CI 68.7% to 73.0%) (table 1). About a quarter (26.1%, 95% CI 23.9% to 28.4%) lived with at least one child, and had an annual income of over $100 000 (29.7, 95% CI 27.5 to 32.0). Approximately 20 (19.6%, 95% CI 17.8% to 21.5%) were veterans. Sixty-one per cent (61.4%, 95% CI 58.9% to 63.9%) had received formal firearms safety training. Nearly two-thirds (64.2%, 95% CI 61.7% to 66.5%) listed protection as a reason for ownership. Approximately three-quarters reported owning at least one handgun (75.7%, 95% CI 73.5% to 77.8%); 28.4% (95% CI 26.1% to 30.8%) owned one firearm, 20.6% (95% CI 18.6% to 22.8%) owned two, 25.9% (95% CI 23.8% to 28.2%) owned three to five and the remainder (25.0%, 95% CI 23.0% to 27.3%) owned six or more firearms (table 1). Seventy-seven per cent (77.1%, 95% CI 74.8% to 79.2%) grew up with a firearm in their home; 21.2% (95% CI 19.1% to 23.3%) carried a loaded handgun on their person in the past month.
Weighted characteristics of gun owning survey participants (n=2011): USA, 2015
Table 2 shows the distribution of storage practices among our gun owners. 29.7% (95% CI 27.4% to 32.1%) of gun owners reported storing at least one firearm loaded and unlocked (least safe) (table 2). One-quarter (24.6%, 95% CI 22.4% to 26.8%) reported storing all of their guns unloaded and locked (safest); 45.7% reported storing guns in our intermediate safety category, that is, no guns loaded and unlocked, but at least one of their guns either unloaded and unlocked, or loaded and locked (45.7%, 95% CI 43.2% to 48.3%).
Distribution of storage practices among a national sample of gun owners, by demographic and firearm-related characteristics (n=2011): USA, 2015
Storage varied by sociodemographic characteristics, including where people lived, income, the presence of children in the home, political affiliation and veteran status (table 2). Gun owners living in the South, for example, were more likely to store at least one of their guns loaded and unlocked (39.8%, 95% CI 35.9% to 43.8%) and less likely to store all guns locked and unloaded (17.3%, 95% CI 14.5% to 20.5%), compared with gun owners in any other region. The proportion of gun owners who stored any gun loaded and unlocked decreased as income increased. For example, 37.6% (95% CI 30.2% to 45.5%) of those who reported earning $25 000 or less also reported storing guns loaded and unlocked, compared with 26.3% (95% CI 22.6% to 30.3%) of those reporting earnings of $100 000 or more. Also, gun owners with children stored their firearms more safely. Of the 26% of gun owners who lived with children, 21.2% (95% CI 17.3% to 25.8%) stored at least one firearm loaded and unlocked, compared with 32.7% (95% CI 30.0% to 35.0%) of those who did not live with children.
Gun owners who reported having received formal firearms training were, on average, more likely to store a gun loaded and unlocked, compared with those who had not, 32.3% (95% CI 29.4% to 35.3%) vs 25.8% (95% CI 22.3% to 29.7%), respectively (table 2). No individual training component, or combination of components, was associated with safer storage. For example, among the 14% of gun owners whose training included all of the components listed, 32.0% (95% CI 25.8% to 39.0%) stored at least one gun loaded and unlocked, compared with 32.1% (95% CI 29.0% to 35.2%) among the 15% of gun owners who received training that involved accident prevention.
Firearm-related characteristics predicted storage practices (table 2). Owners who listed protection as a reason for ownership, compared with those who did not, were more likely to store at least one gun loaded and unlocked (40.3, 95% CI 37.3 to 43.4) vs (8.0, 95% CI 5.8 to 10.9). Handgun owners were also more likely than those who owned no handguns to store at least one gun loaded and unlocked (35.9%, 95% CI 33.2% to 38.8% vs 9.9%, 95% CI 6.9% to 13.9%), as were 57% of handgun owners who reported having carried a loaded handgun in the last 30 days (95% CI 51.4 to 63.1).
Table 3 shows crude OR and AOR for storing at least one firearm loaded and unlocked. ORs are adjusted (A) for sociodemographic characteristics, and (B) after additionally adjusting for firearm-related characteristics. Crude associations between unsafe storage (ie, storing firearms loaded and unlocked) and sociodemographic characteristics remained largely unchanged after adjustment, with the exception of veteran status, with differences in how veterans versus non-veterans store their firearms largely disappearing after adjustment for sex and other sociodemographic factors, partially AOR=0.90 (95% CI 0.68 to 1.20) (table 3).
Factors associated with storing at least one household firearm loaded and unlocked among a national sample of gun owners, crude and adjusted ORs
After adjustment for sociodemographic characteristics, gun owners who had received versus had not received firearms training remained more likely to store at least one gun loaded and unlocked: AOR=1.66, 95% CI 1.27 to 2.16 (table 3). Additional adjustment for firearm-related characteristics, however, rendered the association between training and storage null: gun owners who had received versus had not received firearms training were similarly likely to store at least one gun loaded and unlocked: AOR=1.11, 95% CI 0.80 to 1.53.
In fully adjusted models (table 3), storing guns loaded and unlocked remained more likely if gun owners owned for protection, AOR 4.70 (95% CI 3.17 to 6.99); owned handguns, AOR 1.63 (95% CI 1.04 to 2.57), owned several (compared with few) guns (eg, AOR=4.49 [95% CI 3.23 to 6.25] for owning 6+ guns compared with one gun), or carried a loaded handgun on their person in the past month, AOR=2.78 (95% CI 1.95 to 3.97).
Discussion
Household gun ownership rates in the USA have remained relatively stable over the past two decades, with most studies estimating that about one-third of all US households contain firearms.26 Based on our findings, nearly 30% of these homes (ie, 10% of all households) contain at least one loaded and unlocked firearm (ie, least safe). In addition, only one-quarter of current gun owners store all of their guns unloaded and locked (safest); half store guns in some other manner (our intermediate safety category). Consistent with prior work, we found that firearms are stored more safely when gun owners live with children, are of higher socioeconomic status, own long guns only and do not own firearms for protection.16 21 23 As in prior work,16–19 21 23 we also found that the strongest predictor of unsafe storage was owning firearms for protection.
Approximately 60% of our gun owners report having received formal firearms training (a proportion that has remained relatively stable over the past two decades).27 In unadjusted analyses, gun owners who had received training were more likely than those who had not received training to store a gun loaded and unlocked: one-third of those with training stored at least one household firearm loaded and unlocked, versus one-quarter of those without training. This was the case regardless of whether the training included safety components like suicide prevention. Consistent with our crude findings, Hemenway et al found that gun owners in 1994 who reported having ever received firearms training were less likely to store firearms safely.17 The association of less safe storage with a history of firearms training persisted in Hemenway’s adjusted analyses (which included firearm-related characteristics),17 whereas in our study, as in other studies from the 1990s,18 19 23 after adjustment for firearm-related characteristics training no longer predicted storage practices.
Our findings also suggest that in 2015, compared with the 1990s, fewer gun owners store firearms safely. In 1994, 21% of gun owners reported storing ≥1 firearm loaded and unlocked (27% of those with vs 14% of those without firearms training), compared with 30% of our respondents (33% with vs 26% without training).17 Changes in patterns of and reasons for gun ownership may help explain why US adults today store firearms less safely. Over this period, for example, millions of handguns have been added to the US gun stock.28 Handguns are more often owned for personal or household protection, compared with long guns,27 and historically are more likely to be stored loaded and unlocked.17 18 21 Consistent with the shift in the gun stock towards handguns and towards less safe household firearm storage, public opinion regarding the risks and benefits of having a firearm in the home also appears to have changed. According to polling conducted by Gallup, in 2000 approximately 35% of US adults believed that ‘a gun in the home makes it a safer place to be’; by 2014, that figure had increased to 63%.29 This perception is at odds with the reality that, on average, guns in the home are more likely to lethally imperil, rather than protect, members of that household.30
Online panel surveys, such as used here, have been shown to reduce social desirability bias and yield more accurate estimates of respondent characteristics than telephone surveys.31 32 In addition, prior research has validated survey responses to firearm questions on random-digit dial surveys, with false denials of gun ownership limited to approximately 10%.31 33 Nevertheless, as with findings from all self-reported surveys, our study’s results should be interpreted in light of potential inaccuracies due to social desirability, recall and other biases. We also do not know how recently respondents were trained, by whom or if storage practices changed after training. Notwithstanding these considerations, our finding that formal firearms training is not associated with safer storage practices should not come as a surprise given that historical data generally find either a null relationship or that people who report firearms training store their firearms less safely. Consistent with our findings, a recent audit of basic firearms training in the North-East found that only half of the instructors recommended storing guns unloaded when not in use, 15% recommended hiding guns, 10% actively promoted always keeping handguns loaded, unlocked and ready and 10% covered suicide prevention.34
The factors that influence how a firearm is stored are likely complex, governed by societal norms as well as perception of the risk trade-offs involved, neither of which have been trending in a way that would augur safer storage. It may be unreasonable at this point to expect that an intervention relying solely on firearms training as historically taught would in and of itself motivate gun owners to store firearms locked and unloaded. However, continued efforts to improve firearms training curricula can only help, especially if trainings serve as a vehicle to provide better information about the injury risks of gun ownership under different storage regimens, and information about the RRs that a household firearm will be used to kill an intruder versus that it will be used by a household member to kill himself (or herself) or another family member unintentionally or in a domestic violence homicide. Although the success of training efforts with respect to safer storage has remained elusive, future success is more likely if training efforts are undertaken in conjunction with broader campaigns that reinforce societal norms about responsible firearm ownership and storage.
What is already known on the subject
The presence of firearms in a person’s home substantially increases the risk of suicide and unintentional firearm death.
For children, the risk of firearm injury is lower when household firearms are unloaded (compared with loaded) and locked (compared with unlocked).
Studies from the 1990s suggest that people who report that they have received formal firearms training are no more likely to store household firearms safely.
What this study adds
Contemporary data about the relationship between formal firearms training and household firearm storage practices.
The first assessment of whether firearm storage practices are associated with discrete training components (eg, suicide prevention vs safe handling).
References
Footnotes
Contributors JB, DA and MM conceptualised the study. JB conducted analyses and managed the data. JB wrote the first draft and DA and MM critically reviewed and edited the manuscript throughout. All authors contributed to interpretation of data, further critical revision of the article and writing of final drafts. All reviewed the final drafts. DA and MM supervised the study.
Funding This study was funded by Fund for a Safer Future and Joyce Foundation.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval The Northeastern University Institutional Review Board approved the study.
Provenance and peer review Not commissioned; externally peer reviewed.