Evidence Matters: Randomized Trials in Education Research

Opinions about education programs and practices are offered frequently—by children, parents, teachers, and policymakers. Credible studies of the impact of programs on the performance of children are far less frequent. Researchers use a variety of tools to determine their impact and efficacy, including sample surveys, narrative studies, and exploratory research. However, randomized field trials, which are commonly used in other disciplines, are rarely employed to measure the impact of education practice. Evidence Matters explores the history and current status of research in education and encourages the more frequent use of such trials. Judith Gueron (Manpower Demonstration Research Corporation), discusses the challenges involved in randomized trials and offers practical advice drawn experience. Robert Boruch (Wharton School, University of Pennsylvania), Dorothy de Moya (Campbell Collaboration Secretariat), and Brooke Snyder (University of Pennsylvania) explore the use of randomized field trials in education and other fields. David Cohen, Stephen Raudenbush, and Deborah Loewenberg Ball (all from the University of Michigan) review the history of progress in education over the past forty years and urge increased research on coherent instruction regimes. Maris Vinovskis (University of Michigan) examines the history and role of the U.S. Department of Education in developing rigorous evaluation of federal programs, and suggests a new National Center for Evaluation and Development. Thomas Cook and Monique Renee Payne (both from Northwestern University) take on the claim that randomized field trials are inappropriate in the U.S. education system. Gary Burtless (Brookings Institution) explores the political and professional factors that influence randomized field trials in economic programs, examining possible explanations for their lack of frequent use in education. Carol Weiss (Harvard University) provides a brief history of community studies in the United States and suggests a variety of alternatives to randomization. It is difficult to gauge the impact of various approaches in education. But the authors give a variety of concrete examples to illustrate the feasibility of randomized trials, and the circumstances under which they are appropriate. By offering a variety of suggestions to improve the methods used to evaluate education programs, the contributors to this volume seek to improve education in the United States. Frederick Mosteller is Roger I. Lee Professor in Mathematical Statistics, emeritus, in the department of statistics at Harvard University. Robert Boruch is the University Trustee Chair Professor the graduate school of education and statistics department at the Wharton School, University of Pennsylvania