Background and significance: Lethal violence, including homicide and suicide, takes the lives of 45,000 Americans per year. Intimate partner (IP) homicide is a type of lethal violence that is classified as a form of homicide by policymakers, law enforcement personnel, healthcare providers, and others. It accounts for only 7% of all homicides, but nearly one-third of all homicides of women. The rates of all types of lethal violence have decreased since 1985, but the distribution of IP homicide has gone from nearly half male and half female victims to one-quarter male and three-quarters female victims during that time. Because of the relationship shared between the perpetrator and the victim, the nearly unlimited access the perpetrator has to the victim, and the sociodemographic distinctions between IP homicide and non-intimate homicide, the author hypothesizes that it is not a type of homicide per se, but instead a separate entity deserving distinctively different approaches to prevent it.
Design and Methods: The primary aim of this study is to determine if IP homicide is a type of homicide similar to non-intimate homicide in seasonal, regional, economic, and sociodemographic distributions, if it is more like suicide in those distributions, or if it a separate form of lethal violence with its own set of predictive criteria. Using a combination of statistical techniques including Student’s t-test, Mann-Whitney’s U-test, ordinary least squares regression, and negative binomial Poisson regressions, the author of this study examines a comprehensive database consisting of homicide data from the Supplementary Homicide Reports, suicide data from the Mortality Multiple Cause of Death Reports, as well as Census, firearm, alcohol, and religiosity data from a variety of resources covering the entire United States from 1985-2004. To create a panel for evaluation, the author aggregated these data at the county level to create a dataset with 54,037 county months of observations.
Results: Intimate Partner Homicide does not correlate to regional, seasonal, economic, and sociodemographic variables in the same way as non-intimate homicide or suicide. Instead, IP Homicide does not vary with season or region of county. It correlates with increases in population density, but is lower in counties designated as rural. Its only economic response is to negatively correlate with unemployment. Further, it is higher in counties with higher numbers of churchgoers. It also higher in counties with higher proportions of single-parent households and with broader age gaps between male and female intimates. The finding most important to policymakers is that since the passage of the Violence Against Women Act, IP homicide incidence has significantly decreased, but that decrease is disproportionately in favor of male victims.
Discussion and Policy Implications: The results of this study support the post that Intimate Partner Homicide is a separate “Current of Lethal Violence”, which many say is the most preventable type of lethal violence. Since IP homicide is not like non-intimate (NI) homicide, it might not be amenable to the same prevention strategies applied to NI homicide. Examples of these strategies include an increased police presence in areas with higher IP homicide rates. Enhancements to the Violence Against Women Act that focuses on women not capable of taking risks to protect themselves, such as leaving the abuser, could prove very beneficial. These enhancements could include interventions focusing on education of mental health and emergency department healthcare providers, as well as improvements in law enforcement methods used to stop intimate violence. An approach that recognizes that IP homicide is unlike NI homicide; therefore, prevention should begin with recognition of the intimate dyad and its effect on the persons within it. Some couples might not want to be separated, but instead they might just want the violence to end.