Association for Public Policy Analysis & Management

Major US federal regulatory decisions are developed and justified using regulatory impact analyses (RIAs) mandated by executive order. While regulators may be motivated by several political and practical reasons to rely on scientific evidence when developing RIAs, they are not required to do so, and previous research has shown that science-use significantly varies across agency and the surrounding political context. Nevertheless, in the “Golden Age of evidence-based policy,” agencies are expected now more than ever to use the best evidence available to craft regulatory policies. There are few systematic analyses regarding the type and quality of evidence used. We examine the research cited in RIAs to understand what characteristics of research, scientific or otherwise, are most valued by policymakers. Specifically, we employ a novel dataset of citations in all 108 economically significant RIAs produced by US federal regulatory agencies between 2008 and 2012 to examine knowledge utilization in regulatory policymaking. Building off previous work that focused on peer-reviewed scientific citation patterns in RIAs, which is approximately 20 percent of total RIA citations, we summarize the use of evidence from all sources, including reports, other regulations/laws, law journals, etc.. Our analysis offers important contributions to scholarly research on public policy as well as a practical assessment of the types of research that are useful in the policymaking process. Understanding patterns of evidence use in RIAs advances the collective aim to develop the most effective policies and improve evidence-based policy practices.