Recently, the international collaborative research Reproducibility and Robustness of Economics and Political Science Research, co‑completed by Cui Jing, a postdoctoral researcher at BNUBS, was published in Nature on April 1. This study systematically assesses the reproducibility and robustness of research in economics and political science, providing important evidence for advancing transparency and reliability in social science research.

In recent years, the academic community has attached growing importance to the reproducibility of scientific research findings. The cumulative nature of scientific research relies on the robustness and reproducibility of existing research conclusions. Organized by the Institute for Replication (I4R), this study conducted systematic replication and robustness analyses on 110 research papers published in leading economics and political science journals between 2022 and 2023.

Figure 1 Comparison of effect sizes between original and replication studies

The research team first carried out computational reproduction of the analytical procedures in the original papers, and further conducted robustness tests using alternative model specifications and different estimation strategies. The results show that more than 85% of the original paper conclusions were successfully replicated. In further robustness checks, 72% of the originally statistically significant estimates remained significant with consistent sign, and the effect sizes obtained from replication closely matched those reported in the original papers (the median was approximately 99% of the original results). Overall, these findings indicate that social science research exhibits high reproducibility under conditions of open data.

Figure 2 Comparison of statistical significance between original and replication studies

This study offers systematic evidence for understanding the reproducibility of economics and political science research and carries important implications for promoting open science and enhancing research standards in the social sciences. It represents a key milestone outcome of the global social science replication research initiative.

The research was jointly conducted by scholars from universities and research institutions across multiple countries. As a researcher at the Institute for Replication, Cui Jing, postdoctoral fellow at BNUBS, was primarily responsible for designing parts of the paper replication studies, as well as data analysis and interpretation of results.

Edited by Sun Yue

Reviewed by Hu Conghui