Programs for domestic violence (DV) victims and their families have grown exponentially over the last four decades. The evidence demonstrating the extent of their effectiveness, however, often has been criticized as stemming from studies lacking scientific rigor. A core reason for this critique is the widespread belief that credible evidence can derive only from research grounded in randomized control trials (RCTs). Although the RCT method has its strengths, we argue that it is rarely an optimal--or even a possible--approach for evaluating multifaceted DV programs. This article reviews the reasons that RCT is a poor fit for such programs and argues that a more inclusive conceptualization of credible evidence is critical to expanding our knowledge base about how DV programs affect survivors' safety and well-being.