Policy research is valuable as a source of information for decisionmakers. The value of research is the expected social gain from policy decisions influenced by the information generated. The gain from a decision depends on choosing the best policy given the state of the world, which is uncertain. The output of policy research is a conclusion about that state. Taking a Bayesian approach, the ex ante value of a research program is estimated from information about decisionmakers’ prior probabilities of the state of the world and the likelihood of correct and incorrect research findings. Case studies of U.S. agricultural policy research cover consequences of trade liberalization, farm commodity program analysis, effects of publicly supported commodity promotion, and the value of publicly funded crop forecasting and research on agricultural technology. The quantification of the value of these research programs is highly conjectural, but consideration of their likely magnitude, as compared to the costs of the research, suggests that there are substantial net social gains.