Data are from the FACTS Consortium’s cross-sectional web-based Gallop survey of 2,924 parents and primary caregivers of high-school teens (age 14–18). The survey was conducted from June to July 2020, with participants recruited from the Gallop Panel, a probability-based panel of non-institutionalized citizens constructed to represent the overall U.S. population. Gallop regularly recruits panel participants using address-based sampling and random-digit dialing methods encompassing landline and cellphone numbers. It intentionally oversamples young adults (age 18–34), those from lower educational backgrounds, and those from under-represented racial/ethnic groups. For this study, eligible panel participants included parents and primary caregivers for high-school age teens (age 14–18), defined to include both biological parents and other primary caregivers (e.g., grandparents) that reside with the teen at least some of the time. Known caregivers were randomly sampled from the panel, and if they were found to have children aged 14–18, they were emailed invitations to participate in the study. Following informed consent, panelists self-administered the web-based survey (approximately 15 minutes). Remuneration was $5 for survey completion. The University of Michigan (U.M.) and Gallop IRBs approved all study procedures. This analysis focused only on the adult Caregiver sample.
Measures
Parental Firearm Ownership: The dependent variable for this analysis was parental firearm ownership (yes/no). Ownership was measured using a single item (“Do you personally own a gun?”) from the 2015 National Firearm Survey (NFS)15,16.
Community Violence Exposure: Parent perceptions of community violence exposure, the primary independent variable, were assessed using a 9-item version of the "Things I have Seen/Heard" Survey17, measuring past 12-month exposure to a series of violent behaviors (e.g., hearing gunshots, seeing drug deals, seeing someone get shot) in their neighborhood. The standard response scale was modified to a dichotomous response (yes/no) for each item. A summary score (range 0–9) was created for analysis.
Control Variables:
The Gallup survey collected data on various sociodemographic characteristics, including age, biological sex, race/ethnicity, region, educational attainment, marital status, and military service. For analyses, race/ethnicity was categorized as non-Hispanic White or other, and educational attainment was grouped into those with or without post-secondary education. Additionally, respondents were queried about the presence of young children (under the age of 10) in their household, in addition to teenagers, as well as their receipt of public assistance, using questions derived from the 2016 American Community Survey. To assess the prevalence/frequency of partner (e.g., spouse, ex-spouse, dating partner) and non-partner (e.g., peers, friends, strangers, colleagues) victimization and aggression among various relationships, we used an adapted 9-item version of the Conflict Tactics Scale (CTS-2)18, which captured incidents of moderate (e.g., pushed, shoved) and severe (e.g., hit, punched, used a knife/gun) violent behaviors. The adapted scale measured both victimization (i.e., actions done to the respondent) and aggression (i.e., actions done by the respondent) on a response scale ranging from 0 (never) to 6 (> 20 times). These variables were then used to create a standard summary score by centering the variable using the mean and sum. Supplementary Table 1 illustrates the coding schema for all variables.
Statistical Analysis
We analyzed the data with SPSS software at the p < 0.05 significance level. We weighted all descriptive statistics and analyses to be nationally representative of parents with teenage children. Bivariate and multivariable analyses comparing the independent community violence exposure and the dependent variable parental firearm ownership were first examined among the entire sample. Eighteen independent variables based on prior research were selected and added to the analyses16–22. Stepwise regression modeling was used as it allows for clarity at multiple stages and simultaneous estimation effects at multiple level23. To elucidate how such factors may increase the likelihood of parental firearm ownership, we ran four models: Model 1 examined demographic variables and other control variables, marital status, and education; Model 2 introduced parent-partner and non-partner victimization; Model 3 introduced parent community violence exposure. To examine the relationship between the control variables of biological sex, age, and education level, interaction terms between community violence exposure and each of those variables were created, respectively. Model 4 introduced the interaction between biological sex and community violence exposure. These interactions enabled us to examine whether the relationship between community violence exposure and parental firearm ownership changes dependent on the sex of the parental respondant24.
Tables 1 and 2 approximately here.
Table 1
Descriptive Statistics for Exposure to Community Violence, Victimization, and Demographic Variables for Parent Firearm OwnershipN = 2,380
| | Chi-Squares |
| Parent Has a Firearm in the House | |
| No N (%) | Yes N (%) | |
Sociodemographic Variables |
Sex | | | 273.45*** |
Male | 503 (39.1) | 785 (60.9) | |
Race | | | 14.39*** |
Non-White and/or Hispanic | 449 (28.0) | 274 (21.8) | |
Marital Status | | | 12.20*** |
Married | 1,308 (54.5) | 1,093 (45.5) | |
Education | | | 46.11*** |
Some College or Trade School and Beyond | 1,209 (75.6) | 803 (63.9) | |
Violence Exposure and Victimization Variables and Age |
Scale and Continuous Variables | Mean | Std. Dev. | |
Age | 47.50 | 9.80 | |
Parent Community Violence Exposure | 0.13 | 0.85 | |
Parent Partner Victimization | 0.10 | 0.61 | |
Parent Non-Partner Victimization | 0.05 | 0.47 | |
*p < 0.05, **p < 0.01, ***p < 0.001 | |
Table 2
Descriptive findings regarding Community Violence Exposure among a Nationally Representative Sample of Parents of High-School Age Teens (age 14–18)N = 2,380
| % (N) |
Heard guns being shot | 30.1 (881) |
Have seen somebody arrested | 10.5 (307) |
Have seen drug dealing | 4.1 (121) |
Have seen somebody get beaten up | 1.6 (47) |
Have seen somebody get stabbed | 0.8 (22) |
Have seen someone pull a knife on another person | 0.8 (22) |
Have seen someone pull a gun on another person | 0.2 (7) |
Have seen someone get shot | 0.3 (10) |
My house has been broken into | 0.3 (8) |