Reproducibility Is Rising in Political Science—But Most Papers Still Don't Share Data

Political science shows clear improvement in sharing the data and code needed to reproduce quantitative results, but most articles remain unpublished with regard to reproduction archives.

📚 How Articles Were Sampled

A random sample of quantitative research articles published in political science from 1995 to 2022 was collected and inspected for references to reproduction archives.

🔎 What Counts as a Reproduction Archive

A reproduction archive is defined as the data and code supporting a quantitative research article that allows others to reproduce the computations described in the published paper.

📈 Key Findings

• Most quantitative research articles do not point to a reproduction archive, even in 2022.
• In 2014—around the time the DA-RT symposium was published in PS—about 12% of quantitative articles pointed to their data and code.
• By 2022 that figure had increased to 31%, demonstrating a substantial upward trend in sharing practices.
• The increase reflects a massive shift in norms, requirements, and infrastructure for reproducible research, but a minority of articles still share full supporting data and code.

⚖️ Why This Matters

• The limited availability of reproduction archives constrains the ability to verify, build on, and teach from published quantitative work.
• The 2014 observation by Lupia and Alter remains relevant: "Today, information on the data production and analytic decisions that underlie many published works in political science is unavailable." That statement could still be made today, underscoring that much work remains to be done to reach widespread reproducibility.