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

We study the effect of decreasing barriers to accessing non-police services on the demand for police services in cases of police-reported domestic violence. Variation comes from a large case-level randomised control trial designed to assist victims in accessing non-police services. Our data link information from local and national police administrative records, and a survey of victims. The intervention led to a robust 21% 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 the victims who do not provide a statement because of treatment had a relatively low statement effectiveness. Consistent with this result, we find treatment group statements are significantly less likely to be withdrawn than are control group statements.