Study Sample
Individual-level data were drawn from the National Intimate Partner and Sexual Violence Survey (NISVS), a cross-sectional, nationally representative epidemiologic survey of the non-institutionalized English and Spanish-speaking U.S. population aged 18+.19 The survey was conducted in 50 states and the D.C. and administered from January 2010 to December 2010. NISVS is a random digit dial (RDD) telephone survey with dual-frame sampling that assessed multiple forms of interpersonal violence including IPV, sexual violence, and stalking among females and males. A total of 18,049 interviews were conducted using landline telephones (45.2%) and respondents’ cell phones (54.8%). The study protocol was approved by the Office of Management and Budget (OMB# 0920-0822) and the Institutional Review Board of the Research Triangle Institute, International. Further information on the study design, training, data collection, and study implementation are published elsewhere.19
The current study was exempted by [Institution masked for peer review] IRB because the focus was secondary data analysis of de-identified data. The current analyses were restricted to women and men who experienced lifetime physical, sexual, and/or psychological IPV (N=6,565). This data was analyzed in 2020.
Measures
State-level Data
State IPV Firearm Policy Climate. We examined 6 state-level firearm policies regarding IPV extracted from a compendium on IPV-related firearm policies created by Frattaroli and colleagues,20 which have been consistently used in previous research.7,21,22 The six policies are: 1) Prohibition of firearm possession by persons convicted of IPV-related misdemeanors; 2) Firearm relinquishment for IPV-related misdemeanor; 3) Prohibition of firearm possession by persons subject to IPV-related restraining orders; 4) Prohibition of firearm possession by persons subject to IPV-related protective orders; 5) Prohibition of firearm possession by persons convicted of stalking, and 6) Removal of firearms from the scene of an IPV incident. It is important to note that restraining orders are civil and initiated by the victim whereas protective orders are initiated in criminal procedures.23 For each policy, states were coded as a 1 (presence) or 0 (absence). The coding was based on 2010 policies and legislations given the timeframe of the data collection of the NISVS. An index variable was created by summing the responses across all 6 policy variables. These policy variables were then linked to the NISVS individual-level data.
State Covariates. We also controlled for the prevalence of firearm ownership and violent crimes. State firearm ownership was measured as the percentage of suicides committed with firearms for each state.24,25 Suicide data was collected from the Centers for Disease Control and Prevention’s Web-Based Injury Statistics Query and Reporting Systems database.26 State violent crimes was measured as the percentage of violent crimes (i.e., murder, rape, robbery and aggravated assault) for each state. Violent crime data was collected from the Uniform Crime Reporting Statistics database.
Individual-level Data
Nonfatal Injuries. NISVS assessed participant report of injuries as a result of an experience of IPV (e.g., “Were you ever injured when this/any of these things happened with any of these people?”). Participants were able to respond as Yes, No, and I don’t know. For the current analysis, this variable was dichotomized as Yes (1) vs. No (0).
Socio-demographics. Participants were asked to self-report socio-demographics: sex (female, male); age (<10, 11-17, 18-24, 25-34, 35-44, 45-54, and 55+); education (no schooling, 1-8th grade, some high school, high school graduate, technical or vocational school, some college, college graduate, postgraduate), ethnicity (Hispanic, not Hispanic); and race and ethnicity (Black, White, Hispanic, Asian, Native Hawaiian or Other Pacific Islander, American Indian or Alaska Native, and Other).
Analysis
Frequencies of socio-demographics were calculated for women and men who experience IPV. Weighted logistic regression models were conducted to examine associations between the IPV firearm policy climate index score and nonfatal injuries. Weights were based on the sampling weights. Additional models were conducted to examine associations between the six specific policies and nonfatal injuries independently. Moderation was also assessed by adding product terms between policy variables and sex variable (female vs. male). All models included age, race and ethnicity, and education. All analyses were conducted in 2020 using SAS to account for the complex sample design.