Abstract
Computational reproducibility, or the ability to reproduce published results, tables, and other figures using the available data, code, and materials, through a process of reproduction, is necessary for ensuring that science is self-correcting. Reproducing published work can be used as a teaching tool to introduce students to scientific concepts, research methods, and fundamental scientific principles such as the Mertonian norms (Merton, 1973). In collaboration with Dr. Lars Vilhuber, the current American Economic Association's Data Editor, the Berkeley Initiative for Transparency in the Social Sciences (BITSS) has developed an adaptable curricular module to teach reproducible research through reproductions of published work. The module includes two complementary teaching resources: (1)The Guide for Accelerating Computational Reproducibility includes detailed steps, definitions of fundamental concepts, and criteria for assessing and improving reproducibility. (2) The Social Science Reproduction Platform (SSRP) is an open-source platform that crowdsources and catalogs attempts to assess and improve the reproducibility of published social science research. SSRP allows users to upload the results of their reproductions using a standardized form, receive feedback from peers through a discussion forum, and contribute citable evidence on the reproducibility of research. The presentation will describe our approach and illustrate how to use both tools in the classroom.
Speaker Bio
Fernando Hoces de la Guardia is a Project Scientist at the Berkeley Initiative for Transparency in the Social Sciences (BITSS) and an affiliate of the Berkeley Institute for Data Science (BIDS). Fernando works on bridging research-to-policy gaps in regards to transparency and reproducibility and improving the computational reproducibility of economics research. He has also led BITSS trainings in the US, South America, and Europe. Fernando received his PhD in Policy Analysis from the Pardee RAND Graduate School where his research focused on increasing the transparency and reproducibility of policy analysis as a way to strengthen the connection between policy and evidence. Before RAND, he studied economics and conducted impact evaluations and economic analyses of various social policies.