Public Concern About Violence, Firearms, and the COVID-19 Pandemic in California

Key Points Questions Is the coronavirus disease 2019 (COVID-19) pandemic associated with changes in individuals’ worry about violence happening to themselves or others, the prevalence of and reasons for firearm and ammunition acquisition, and changes in firearm storage practices? Findings In this survey study of 2870 adults in California, worry about multiple types of violence for oneself increased during the pandemic. Individuals expressed concern that someone else might physically harm themselves because of pandemic-related losses; there was an increase in firearm acquisition and in unsecure storage practice of loaded firearms in response to the pandemic. Meaning The findings of this study suggest that the COVID-19 pandemic and efforts to lessen its spread have compounded the public health burden of violence.


Introduction
In the United States in 2018, there were nearly 68 000 violence-related deaths. 1 An additional 3.3 million people reported having been victims of nonfatal violent crime. 2 Most deaths (57%) and nearly 471 000 nonfatal violent victimizations involved a firearm. 1,2 Black, Indigenous, Latinx, and other communities of color endure a disproportionate share of this burden. [3][4][5] It is reasonable to expect that the emergence and progression of the coronavirus disease 2019 (COVID-19) pandemic-with more than 11 million confirmed cases and more than 270 000 deaths nationally as of December 2, 2020 6 -combined with the social, psychological, and economic fallout associated with efforts to lessen its spread, have intensified violence-related harms and inequities therein.
There are multiple possible mechanisms through which the pandemic may be associated with changes in the salience of, or concern about, violence and violence exposure. Pandemic-induced social isolation, hopelessness, and loss, particularly for people with existing mental health conditions, such as depression, may result in thoughts of suicide. Interpersonal violence in the home may increase in frequency and severity as household members, including intimate partners, children, and vulnerable elders, spend more time at home together under high-stress conditions. Having a firearm readily available in these situations creates additional risk. [7][8][9][10] The pandemic has also worsened long-standing injustices rooted in systemic racism and other oppressive systems of power that contribute to the underlying conditions (eg, poverty, unemployment, and lack of available resources) that elevate risk and compound the consequences of community violence. [11][12][13][14] Fear and scapegoating associated with COVID-19 may increase biasmotivated unfair treatment. 15,16 During the summer of 2020, largely peaceful 17 protests decrying structural inequities, which contribute to both police violence and the uneven burden of COVID -19, have been met, at times, by law enforcement use of crowd-control weapons 18 and heavily armed white supremacist and far-right vigilantes. 19 While most major news sources reported initial decreases in violent incidents, as measured by local police calls for service, following pandemic-related lockdowns and stay-at-home orders, the latest indications are that more serious acts of violence, particularly those involving firearms, have remained the same or increased. 20 In addition to a marked increase in shootings in several large cities across the country, 21 the pandemic appears to have fueled a surge in firearm background checks, an established proxy for firearm sales. Previous spikes in purchasing, such as those following mass shootings and political elections, have been associated with increased firearm violence. 22 Similarly, recent research suggests an excess of 2.1 million firearm purchases during the first 3 months of the pandemic, corresponding with an additional 216 to 1335 fatal and nonfatal firearm-related injuries nationwide. 23 However, the lack of data capturing individuals' self-reported experiences of, and behaviors associated with, violence in the context of the pandemic has limited our understanding of the intersection of these coinciding public health problems. One of the only past studies to survey individuals who purchased a firearm due to the pandemic relied on a nonrepresentative sample of respondents who did not reflect the sociodemographic profile of most firearm owners. 24 The current study provides what is, to our knowledge, the first population-representative estimates of individuals' worry about violence for themselves before and during the pandemic; concern someone they know might harm themselves or others due to a pandemic-related loss; experiences of unfair treatment related to the pandemic; firearm and ammunition purchasing; and changes in firearm storage practices due to the pandemic.

Methods
Data for this survey study come from the 2020 wave of the California Safety and Well-being Survey, a recurrent statewide survey on firearm ownership and exposure to violence and its consequences in California. The 2020 California Safety and Well-being Survey was designed by the University of California Firearm Violence Research Center and the Violence Prevention Research Program, both at the University of California Davis, and administered online from July 14 to July 27, 2020, by the survey research firm Ipsos. 25 The 2020 California Safety and Well-being Survey was approved by the University of California Davis institutional review board. Participants read informed consent language, and study initiation constituted consent. This study followed the American Association for Public Opinion Research (AAPOR) reporting guideline.
Respondents were drawn from the Ipsos KnowledgePanel, an online research panel that has been widely used in population-based research, including other studies related to firearm ownership. 26,27 To establish a nationally representative panel, members are recruited on an ongoing basis through probability-based sampling with random digit dialing and address-based sampling from the US Postal Service's Delivery Sequence File. A probability proportional to size procedure was used to select a study-specific sample. All members who were aged 18 years and older and residents of California, except those currently serving in the US Armed Forces, were eligible to participate.
Invitations were sent by email; automatic reminders were emailed to nonresponders 3 days later. Of 5018 panel members invited to participate, 2870 completed the survey, yielding a 57% completion rate. The median survey completion time was 26 minutes.
A final survey weight variable provided by Ipsos adjusted for the initial probability of selection into KnowledgePanel and for survey-specific nonresponse and overcoverage or undercoverage using poststratification raking ratio adjustments based on cross-classifications of age, sex, race/ethnicity, education, household income, language proficiency, and geographic region within California.
Respondents tended to be older, men, and non-Latinx individuals and to have more years of education and higher income than nonresponders (eAppendix 1 in the Supplement). With weighting, the sample is designed to be statistically representative of the noninstitutionalized adult population of California as reflected in the 2018 American Community Survey.
Survey questions for this study covered 4 broad domains, as follows: (1) worry about violence happening to oneself, by type of violence (ie, homicide, suicide, mass shooting, assault, robbery, police violence, accidental shooting, and stray bullet shooting) and incident location, before and during the pandemic; (2) concern that someone else might physically harm another person or themselves in response to a pandemic-related loss; (3) experiences of unfair treatment attributed to the pandemic (eg, being threatened, harassed, or physically assaulted); and (4) firearm and ammunition acquisition and firearm storage practices (among current firearm owners) in response to the pandemic. Detailed survey items and response options are in eAppendix 2 in the Supplement.
Item nonresponse for key measures was small (<1.5%). Sociodemographic information was collected as part of ongoing panel membership and merged with individual respondents.

Statistical Analysis
To generate statewide prevalence estimates, we calculated weighted percentages and 95% CIs for each measure or cross-tabulation of measures using the survey and weighting commands in Stata version 15.1 (StataCorp). Using a retrospective pre-post design, respondents were asked to simultaneously report on their experiences during the pandemic and to reflect on the period before the pandemic. Differences in these measures, overall and by respondent characteristics, were examined using weighted repeated measures multinomial logistic regression models, including interactions between respondent characteristics and an indicator for the period of reference (during vs before the pandemic). The margins command was used to generate prevalence differences and 95% CIs. All statistical tests were 2-tailed, and statistical significance was set at P < .05.
reported that they themselves had tested positive, while 4.2% (95% CI, 3.2%-5.3%) reported that they had been sick with COVID-19 but had not been tested. Additional sociodemographic and firearm ownership-related characteristics of respondents appear in eAppendix 3 in the Supplement.

Worry About Violence
The percentage of respondents who reported that they were somewhat or very worried about violence happening to them was significantly higher during (vs before) the pandemic for all violence types except mass shootings, ranging from a 2. 8

Concern About Violence for Others
Of the 12.1% (95% CI, 10.4%-14.1%) of respondents who reported concern that someone they know might physically hurt another person on purpose, 1.8% (95% CI, 0.7%-4.4%) reported that their concern was at least in part because the person had experienced a major loss (eg, loss of someone they cared about, a job, or housing) related to the pandemic (Table 2). Likewise, of the 13.3% (95% CI, 11.5%-15.3%) of respondents who reported concern that someone they know might physically hurt themselves on purpose, 7.5% (95% CI, 4.5%-12.2%) reported that their concern was at least in part because the person had experienced a pandemic-related loss.
Among respondents whose concerns were due to a pandemic-related loss, most said they did concerned that someone they know might harm themselves due to a pandemic-related loss, 6.0% (95% CI, 1.2%-25.8%) said the person had access to a firearm ( Table 2).

Experiences of Unfair Treatment
The percentage of respondents who reported that they had experienced at least 1 form of unfair treatment in the past 12 months was 69.2% (95% CI, 66.6%-71.7%) (

Firearm Acquisition and Storage Practices
The percentage of respondents who reported that they or someone else in their household owned firearms was 23

Discussion
Violence is a significant public health problem that touches the lives of far more people than is typically recognized. Violence affects people not only through direct involvement but also through indirect and vicarious experiences that may increase the salience of, and concern about, violence across networks of individuals, families, and entire communities. This state-representative survey of California adults examined self-reported concerns about, experiences of, and behaviors associated with violence in the context of the COVID-19 pandemic. Our findings add support to a growing body of research suggesting that the pandemic, and efforts to lessen its spread, have compounded the burden of violence-related harms.
Our respondents expressed higher levels of worry about violence during (vs before) the pandemic, ranging from 2.8 to 5.6 percentage point increases in the estimated statewide prevalence of adults who reported that they were somewhat or very worried about multiple types of interpersonal violence (ie, robbery, assault, homicide, police violence), suicide, and unintentional firearm injury happening to them. As expected, 15,16 self-reported experiences of unfair treatment attributed to the pandemic were also reported and disproportionately common among Asian respondents. In addition to worry about violence for oneself, 13.3% of respondents reported concern that someone they know might physically harm themselves on purpose. Of those, 7.5% said this concern was at least in part because the other person had experienced a pandemic-related loss.
The processes underlying changes in worry about violence for oneself and others during the pandemic-as well as the absence of an observed change in worry about violence due to an intimate partner-warrant further investigation. Past research suggests that neighborhood social cohesion and collective efficacy inform both perceptions of safety and actual levels of crime and violence 28 ; however, pandemic-induced social distancing and stay-at-home orders, combined with more general fears regarding COVID-19 danger and contamination, 29 may make such social bonds and informal social controls difficult to establish and maintain. In prior work, worry about violence has been associated with higher perceived stress and higher depressive symptoms 30 as well as with health-risk behaviors, such as firearm carrying, that may increase the risk of future violence involvement. 31 Mounting concern about violence may also stem from the surge in firearm background checks, a proxy for firearm sales, in the months coinciding with the pandemic. 23 To our knowledge, however, this was the first population-representative study to estimate the prevalence of and motivations for ammunition and firearm acquisition in direct response to the pandemic. We found that 8.5% of firearm owners in California purchased ammunition in response to the pandemic, including an estimated 110 000 individuals who also acquired firearms (2.4% of owners in the state). Of those, an estimated 47 000 were new owners, who may have little past experience or training with firearms.
Consistent with the most common reasons for firearm ownership generally, 24,26,32 respondents who acquired firearms due to the pandemic usually did so for self-protection: three-quarters (75.9%) indicated worry about lawlessness and more than half (56.1%) endorsed worry about prisoner releases.
Although respondents' perceived need for self-protection continued to motivate firearm ownership amid the pandemic, an extensive body of evidence suggests that the presence of a firearm in the home elevates risk for firearm-related harm, particularly unintentional shootings (often involving children or teenagers), intimate homicide against women, and completed suicide. [7][8][9][10] Previous spikes in firearm purchasing have been associated with increased firearm violence, 22 and recent evidence suggests a similar association exists during the pandemic, particularly for suicide. 23,33 In addition, past research suggests that people who own firearms primarily for protection are more likely to store firearms in the home loaded and/or not locked up, 34 an independent risk factor for firearm injury and death. Our findings suggest the pandemic may be associated with increases in this risk: an estimated 55 000 people (1.2% of owners in the state) who stored at least 1 firearm loaded and not locked up reported adopting this unsecure storage practice in response to the pandemic.
Taken together, our findings add support to comprehensive public health-oriented prevention and intervention strategies designed to address the enduring psychological trauma associated with exposure to and worry about violence as well as the intermediary (eg, firearm ownership and storage) and upstream (eg, socioeconomic characteristics, education, and the environment) factors associated with violence risk. More immediately, given the impulsive nature of many types of violence and the multiple acute disruptions associated with the pandemic, short-term crisis interventions, such as options for temporary firearm storage outside the home, extreme risk protection orders, and efforts involving violence intervention workers in hospitals and communities, may be particularly critical for reducing the burden of violence.

Limitations
This study has some limitations. First, we relied on self-reported data, which is subject to social desirability, nonresponse, and recall biases. However, several administrative data sources provide an opportunity to broadly assess the validity of our estimates.