Computational Reproducibility in Computational Social Science

David Schoch and Chung-hong Chan and Claudia Wagner and Arnim Bleier



In the last decade, replication and reproducibility crises have shaken the scientific landscape. As potential solutions, open science practices were heavily discussed and have been implemented with varying success in different disciplines. We argue, however, that the binary definition of reproducibility, specifically for computational-X disciplines such as computational social science, is insufficient since it is not explicit about the agents and conditions under which results can be reproduced. We expand the definition to avoid “open washing”, the practice of fabricating theoretical reproducibility but not supporting practical or verified reproducibility, and introduce a tier system of computational reproducibility based on the concept of verifiability. We identify common barriers to verifiable computational reproducibility, specifically in the field of computational social science, and provide suggestions on how to circumvent common access and computational barriers.

Arxiv 2307.01918