Katie Hoeberling
Program Manager, Berkeley Initiative for Transparency in the Social Sciences (BITSS)
University of California, Berkeley
Fernando Hoces de la Guardia
Project Scientist (BITSS)
University of California, Berkeley
Rigorous replication, robustness checks, and extensions of research are possible only to the extent that published findings are first computationally reproducible—i.e., that tables and figures can be reproduced within a reasonable margin of error using available data, code, and materials. In the social sciences, reproductions are routinely conducted by students as part of graduate curriculum. Their work however, as well as that of non-student researchers, is seldom published in journals or shared outside of their classrooms, presenting lost opportunities for reproducers to receive credit, or for the wider community to learn from their valuable work. In collaboration with the Data Editor of the American Economic Association, the Berkeley Initiative for Transparency in the Social Sciences (BITSS) is developing the Accelerating Computational Reproducibility in Economics (ACRE)* platform to enable researchers to systematically and transparently assess and improve the computational reproducibility of published social science research. This platform will house these reproductions; assign DOIs and attribute appropriate credit; display distributions of reproducibility by claim, paper, journal, and field; and provide a forum for discussing them. Importantly, an accompanying guide, which serves as a teaching resource for instructors who include reproductions in their courses, also provides guidance on how to have constructive conversations with original authors. We hope to source reproductions from several courses in 2021, refining the platform and guide accordingly, with the goal of making this type of publication normative across social science disciplines. *While ACRE began with a focus on Economics, its platform and principles can be applied to computational work in other social science disciplines, such as Political Science, Psychology, and Sociology.
ACRE Project information: https://www.bitss.org/ecosystem/acre/
ACRE Guide: https://bitss.github.io/ACRE/intro.html
Beta platform: URL forthcoming in November