Study 1
Using a quasi-experimental design, Messing and colleagues (2014) evaluated the Lethality Assessment Program. A total of 414 women were included in the study (202 women were in the intervention group and 212 women were in the comparison group). The study was conducted in seven jurisdictions in Oklahoma. Police officers were responsible for participant recruitment and recruited women they suspected to be potential victims of partner abuse. All individuals who consented to participate had a 45-minute baseline interview and received the Lethality Screen. For those who could be contacted, a second interview was conducted 7 months later. Women in the intervention group who scored positive on the Lethality Screen, indicating a high homicide risk, were put into direct contact with a social service provider via telephone. Those women in the control group who scored positive for being at high homicide risk were not.
Women in the study ranged in age from 18–79 years. The race/ethnicity of the study participants was white (42.8 percent), African American (29.4 percent), Native American (10.0 percent), Latina (7.9 percent), multiracial (7.5 percent), and other (2.2. percent). The intervention and control groups were not significantly different on baseline measures, except for marital status and immigration status. Half of the sample (51.8 percent) completed high school or had a GED equivalent, and less than half (40.6 percent) had full-time or part-time employment. The overwhelming majority reported being victims of domestic violence by their intimate partners, and many reported being victims of severe violence.
The outcomes of interest were the frequency and severity of repeat intimate partner violence (IPV), and the rates of emergency safety planning and help seeking among women who experienced IPV and called the police in the participating jurisdictions during the time of the study. The first outcome was measured using the Conflict Tactics Scale (CTS-2), which measures severity and frequency of violence over time on a linear, 168-unit scale. Data was collected via self-report. Researchers followed up with participants through email or telephone. Scores from baseline were subtracted from scores at follow up. The second outcome was measured using a 16-item version of a safety-promoting behavior checklist, developed by McFarlane and colleagues (2004). These “yes” or “no” items were asked of women at baseline and then again at follow up.
Logistic and linear regression models controlled for differences between groups at baseline based on marital status, immigration status, Danger Assessment category or severe violence, and time between baseline and follow-up observations. Attrition is a notable limitation of the study, given that 40 percent of participants dropped out between baseline and follow-up observations.