Objective. Our objective is to explain how scientists interpret less‐than‐certain scientific findings to inform policymakers' choices on controversial science policy issues. We focus on two particularly difficult policy cases concerning global climate change and low‐dose radiation protection.Methods. Our method is to analyze data from a unique multination survey of scientists to analyze the ways their views about what is scientifically correct are translated into judgments about appropriate policy. The surveys asked scientists, randomly drawn from U.S. and E.U. subscribers to the journal Science, to indicate the “most likely” relationships between greenhouse gas emissions and average global temperatures and between radiation dose and incidence of cancer in humans. Follow‐up questions asked for their judgments about appropriate policy targets for reductions in greenhouse gas emission and safety standards for radiation exposure. The data permit analysis of the relationships between scientific certainty and policy judgments in these two cases.Results. Our results shed light on when and how scientists reach precautionary policy conclusions, demonstrating that scientists' application of precaution is dependent on context. In the case of radiation protection, greater certainty is associated with less precaution. But with respect climate change, we found the opposite relationship.Conclusions. We conclude with a discussion of the implications for the role of scientists, and scientific advice, in the policy process.