Reproducibility in Social Sciences and Statistics: Context, Concerns, and Concrete Measures When Data Are Restricted-Access

Replicability is at the core of the scientific enterprise. In the past 30 years, recurring concerns about the extent of replicability (or lack thereof) of the research in various disciplines have surfaced, including in economics. In this talk, I describe the context in which the current discussion in the social science is occurring: what are the definitions of replicability and reproducibility, what is failing, and to what extent. In particular, I discuss the concerns in economics: to what extent is this a problem in economics, what are the approaches that are being considered, and what are the possible broader implications of those approaches. Finally, in this version, I specifically discuss the implications for restricted-access environments such as the Federal Statistical Research Data Centers. The solutions to these problems will change the way research will be taught and conducted, in economics in particular, and in the social sciences more broadly. The implications affect undergraduate and graduate teaching, research infrastructure, and habits. Presented at the 2019 Federal Statistical Research Data Conference at the University of Wisconsin-Madison on 2019-09-06.The opinions expressed in this talk are solely the authors, and do not represent the views of the U.S. Census Bureau, the American Economic Association, or any of the funding agencies