A dramatic increase in the volume of research literature referencing social determinants of health (SDH) since the report of the World Health Organization Commission on the topic in 2008 has not been matched by expansion of policies and interventions to reduce health inequalities by way of SDH. This article argues that familiar hierarchies of evidence that privilege clinical epidemiology as used in evidence-based medicine are inappropriate to address SDH. They misunderstand both the range of relevant evidence and the value-based nature of standards of proof. A richer conceptual armamentarium is available; it includes several applications of the concepts of epidemiological worlds and the lifecourse, which are explained in the article. A more appropriate evidentiary approach to SDH and health inequalities requires "downing the master's tools," to adapt Audre Lorde's phrase, and instead applying a multidisciplinary approach to assessing the evidence that adequately reflects the complexity of the relevant causal pathways. Doing so is made more difficult by the power structures that shape research priorities, yet it is essential.