Over the last two decades, state and federal laws and grant programs, such as state accountability polices, the No Child Left Behind Act (NCLB), Race to the Top, Title I School Improvement Grants, and State Longitudinal Data System Grants, have given state education agencies (SEAs) considerably more responsibilities for directing and guiding the improvement of low-performing schools. At the same time, they have pressed SEAs and school districts to incorporate research-based school improvement policies and practices in their statewide systems of support for low-performing schools, technical assistance for districts, professional development for teachers, and school improvement programs. Policymakers have urged SEAs to engage with organizations external to their own agencies to extend their strained capacity, and to help them collect and use research or other evidence. A variety of organizations involved in this enterprise have emerged over the last two decades. For example, the 2002 authorization of the Elementary and Secondary Education Act’s (ESEA) comprehensive assistance centers was specifically designed to provide and encourage SEA’s use of research. Although studies of districts’ and schools’ use of research exist, we know little about how SEAs search for, select, and use research and other kinds of evidence in their school improvement strategies. While one might assume similarities in research use behaviors, both the organizational structures of SEAs and the population of external organizations with which they interact are quite different than schools and districts, and the most recent in-depth study of SEAs was conducted nearly 20 years ago. The exploratory study on which this brief is based was designed to fill that gap by examining: 1) where SEA staff search for research, evidence-based, and practitioner knowledge related to school improvement; 2) whether and how SEA staff use research and these other types of knowledge to design, implement, and refine state school improvement policies, programs and practices; and, 3) how SEAs are organized to manage and use such knowledge.