Political Science, Psychology, and economics have each recently taken their turn in the spotlight with instances of influential research that falls apart under scrutiny. There is growing evidence that much social science research features false positives or is not reproducible. I review the source of these problems, then discuss new techniques to increase reproducibility, including changes in journal publication processes and the use of public registries for preanalysis plans.
Garret Christensen is a Financial Economist in the Division of Insurance and Research at the FDIC. His research interests include housing and consumer finance, poverty programs, and meta-science and reproducibility. Before joining the FDIC, Garret was an Economist with the US Census Bureau, a BIDS Data Science Fellow, and a research fellow with the Berkeley Initiative for Transparency in the Social Sciences (BITSS), a program of the Center for Effective Global Action. He also taught economics at Swarthmore College and conducted water, sanitation, and hygiene research in western Kenya. He received his PhD in economics from the University of California, Berkeley, and has since conducted research for the WASH Benefits public health randomized trial for Innovations for Poverty Action and Emory University and has taught economics at Swarthmore College.