Does Place Matter?: Examining Community-Level Predictors of Intimate Partner Homicide Among a National Sample in Rural and Non-Rural Populations

Background and Purpose: Prior research has indicated that intimate partner homicide (IPH) occurs at a higher rate in rural areas than non-rural areas. However, studies have yet to indicate if the heightened association between rurality and IPH is due to the low density of people in rural areas or if there are community-level traits, e.g. the proportion of people living in poverty, specific to rural areas that drive the association. Few studies have explicitly looked at structural differences between rural and urban areas that drive IPH, particularly at the national level. Studies examining social disorganization and gender equity factors in specific localities or in urban areas found varying results in these community-level factors' influence on IPH. Building on these prior inquiries, this study examined which incident-level and community-level factors were associated with IPH, and whether or not these associations varied by rurality or victim sex. Methods: This study combined data from the FBI's Unified Crime Report and the American Community Survey to model homicides in 2014. Data were analyzed using a three-level multi-level model, organizing hierarchical data at the incident-, county-, and state- level. Results: The majority of the variation in the outcome across models was due to incident-level factors, but the hierarchical organization of the model remained important. In the full sample, women were more likely to be a victim or a perpetrator of an IPH than a non-IPH, and as age increased, there was a higher probability of a non-IPH. The association between sex and IPH was stronger in rural areas than in non-rural areas in the models that included community-level factors. In non-rural counties, higher unemployment was more predictive of non-IPH. In the models stratified by gender, among men, the use of a firearm and the presence of a firearm sanction law for protective orders in their state was associated with a decreased probability of IPH. In the female sample, the use of a gun in the incident and residency in a rural county increased the probability of IPH, while a higher unemployment rate decreased its probability. Conclusions and Implications: The full sample multi-level models suggest that overall community-level factors do not differentiate between non-IPH and IPH. If this finding holds in future studies, it would suggest that programs and policies aimed at preventing IPH should borrow from the well-established homicide prevention practices more generally. Less than 1% of the sample was rural, so the implications as to rurality are limited, however there was enough evidence to suggest that factors' influences on IPH vary by place, and attention to rurality should be included in future models. Women seemed particularly more at risk for IPH in rural settings, which highlights the importance of sex for studies of rural crime. This study underlined the importance of looking at community-level predictors of homicide separately among male and female victims, and underscores the need to consider gender in future studies as well as the design and deployment of prevention strategies.