Examining Ignition-Interlock Post-Arrest Treatments

Abstract The objective is to examine the effectiveness of the ignition-interlock device on DWI recidivism using a variety of counterfactual specifications. Matched samples are constructed that represent different counterfactual comparisons, and logistic regression is used to estimate the odds of DWI recidivism. A sample of four groups of DWI offenders who were all sentenced to probation was used. These groups include those that received no post-arrest ignition-interlock treatment (n = 2,810), pre-probation treatment only (n = 156), probation treatment only (n = 643), or combined pre-probation and probation treatment (n = 187) matched on the likelihood of receiving a certain treatment. Results indicate a lower odds of DWI recidivism for those who used the interlock device compared to those who did not use the device at all. When comparing treatment to an alternative treatment imposed, there is largely no meaningful difference in the odds of DWI recidivism. The estimated treatment effects indicate the device reduces DWI recidivism while installed but not after removal. Administering the treatment compared to not at all plays a meaningful role in reducing the odds of DWI recidivism, but the timing of that treatment exposure may not.