Understanding access barriers to public services : lessons from a randomized domestic violence intervention

We study the effect of reducing barriers to accessing non-police services on the demand for police services in cases of police-reported domestic violence. Variation comes from a large randomized controlled trial designed to assist victims in accessing non-police services and we link information from local and national police administrative records and a survey of victims to form a unique dataset for the evaluation. The intervention led to a 18% decrease in the demand for police services, as measured by the provision of a statement by victims. Despite a strong correlation between statements and criminal sanctions against perpetrators, we do not find a corresponding effect of the intervention on perpetrator arrest, charges, or sentencing. This suggests that treated victims who do not provide a statement do so because their potential statement was relatively less effective for pursuing criminal sanctions. Consistent with this result, we find treatment group statements are significantly less likely to be withdrawn than are control group statements.