Document Type


Publication Date



Evidence-based policy is gaining attention, and legislation and agency regulation have been no exception to calls for greater uptake of research evidence. Indeed, current interest in “moneyball for government” is part of a long history of efforts to promote research-based decisions in government, from the U.S. Census to cost-benefit analysis. But although evidence-based policy-making (EBPM) is often both feasible and desirable, there are reasons to be skeptical of the capacity of EBPM in governmental decision-making. EBPM is itself bounded by limits on rationality, the capacity of science, the objectivity of science, and the authority we wish to give technocrats. Where values are highly contentious, efforts to produce and use evidence in legislative and regulatory decisions may go so far awry that they become “sham” versions of evidence-based choices. In this Article, I name several of these sham practices, including the distortion of evidence, the engagement in “terminal” experimentation that destabilizes governmental programs, and “ratcheting” actions that defund entire priorities rather than individual approaches. Broken experimentation is also common, with evaluations of government programming and policies neglecting or misusing opportunities to provide rigorous evidence. I argue that the stakes of these misuses are high, resulting in losses of welfare and efficiency, erosion of scientific legitimacy, and infringement on the dignity of human subjects. But where genuine engagement with empirical evidence is possible, the game is surely worth the candle. This Article proposes novel ways to promote responsible uses of empirical evidence in both legislation and agency regulation, including evaluation mandates, pre-registration of evaluation protocols and transparency of research reports, ex ante decision commitments, and more attentive uses of judicial review.


Administrative Law | Law | Public Policy