Study
Kaufman and Wiebe (2016) conducted a quasi-experimental study and 50-state difference-in-differences analysis to evaluate the effect of statewide (or universal) ignition interlock requirements on reducing alcohol-involved crash fatalities in the United States, during a 15-year intervention period (1999 to 2013). For the evaluation, the study authors grouped states into treatment conditions (i.e., the intervention group or the control group) based on state policies collected from the Trust for America’s Health (a health policy organization); the years of implementation of ignition interlock polices were corroborated with state legislative records and state highway offices. Between 2004 (when the first state adopted the universal ignition interlock requirements) and 2013, 18 states had adopted universal ignition interlock requirements, which mandated that all individuals convicted of driving under the influence of alcohol have an ignition interlock device installed into their vehicle. These 18 states were included in the intervention group. States that had not adopted the policy initiative by 2013 were included in the control group (n = 32).
The 18 intervention states were Alaska, Arizona, Arkansas, Colorado, Connecticut, Hawaii, Illinois, Kansas, Louisiana, Maine, Nebraska, New Mexico, New York, Oregon, Tennessee, Utah, Virginia, and Washington. Of these states, 44.4 percent were located in the western region of the United States, 22.2 percent in the Southeast, 16.7 percent in the Northeast, and 16.7 percent in the central region of the United States. More than half of the population in these intervention states were White (69.3 percent), with 22.9 percent of the population living in rural areas. Of the 32 states in the control group, 28.1 percent of states were located in the central United States, 18.8 percent were in the Northeast, 15.6 percent were in the Southeast, and 15.6 percent were in the West. The region for the other 21.9 percent of states was mistakenly omitted due to a typographical error; nevertheless, all 50 states were included in the study analysis. (This information was provided in correspondence with the study authors).
Approximately 73 percent of the population in control states were White, with 28.4 percent of the population living in rural areas. There were no statistically significant differences between intervention and control states’ basic demographic characteristics (such as age, region, political orientation, and total vehicle miles driven) during the study period; however, intervention states were statistically significantly more often located in the West and were less rural, and people in intervention states drove fewer total miles on average, compared with the control states.
The outcome of interest was alcohol-involved crash death rates. Alcohol-involved crash deaths were defined as a fatal crash in which at least one of the drivers had a blood–alcohol concentration level greater than zero (zero was selected because interlock devices were typically set to a low blood–alcohol content threshold of 0.02 percent). Data for fatal death rates were collected from the National Highway Traffic Safety Administration’s Fatality Analysis Reporting System. Population denominators were pulled from the U.S. Census. The effect of universal ignition interlock devices on alcohol-involved crash deaths was estimated using difference-in-differences analysis, comparing state rates of alcohol-involved crash deaths before and after policy implementation with the change over the same period in states without the policy. An ordinary least squares (OLS) regression was used to determine differences in alcohol-involved crash deaths between intervention states and control states, with standards errors adjusted for clustering at the state level across years.
In addition, the study authors conducted subgroup analyses, using regression analysis to evaluate differences in the total vehicle crash rates between alcohol-involved and non–alcohol-involved crashes.
