Routine screening for intimate partner violence (IPV) has been introduced in many health settings to improve identification and responsiveness to IPV. The debate about the level of evidence required to warrant routine screening continues. Three assumptions have impeded progress in measuring the impact of screening. The first is that routine screening is a test only which does not of itself have an impact on patients. The second is that it can be assessed by evaluating interventions provided to women after abuse is identified through screening. The third is that there can be an agreed appropriate intervention for IPV. Each of these assumptions is problematic. In addition, there are significant impediments to evaluating screening as an intervention through a randomized control trial. These include identification of the study group, isolating the control group from the intervention, ethics, lack of baseline data, and recall bias. A range of study designs is required and a rethink of assumptions is needed in researching this area.