The CBS Replication Games addresses a blind spot in current research culture: how can research be independently verified when underlying data cannot be openly shared?

An increasing share of socially valuable research relies on restricted-access data that researchers can only access in secure environments. These restrictions are necessary to protect privacy and intellectual property, but they also make independent verification of research relying on such data uniquely difficult. The CBS Replication Games represent a new model for resolving this tension.

A first-of-its-kind initiative

For the first time, teams of independent researchers were given secure access to administrative microdata underlying published articles to assess those articles’ reproducibility and robustness. In doing so, the initiative extends the norms of transparency, accountability, and independent verification to research for which typical data sharing approaches are impossible.

In February 2026, ODISSEI and the Institute for Replication (I4R) brought together researchers to re-analyze published articles produced using restricted-access microdata from Centraal Bureau voor de Statistiek (CBS). The initiative centered on an intensive three-day hackathon in Utrecht, supplemented by a month of secure access to the CBS microdata. Each team produced a replication report assessing the reproducibility and robustness of their assigned study.

Why does this matter?

The significance of the CBS Replication Games extends far beyond merely repeating an established model of replication-based verification in CBS microdata. The dominant model of reproducibility assumes that research data can be openly shared. However, this assumption excludes an increasingly important class of socially valuable research. The CBS Replication Games develops a new framework through which the norms of reproducibility and independent verification can be extended to research involving restricted-access data. In so doing, this initiative promotes a culture of accountability for researchers privileged with the unique opportunity to analyze rich data to which the general public does not have access.

What comes next

This project will result in:

• Replication reports summarizing the reproducibility and robustness of selected papers produced using CBS microdata
• A meta-paper on the lessons learned from the initiative, including summary statistics on the reproducibility and robustness of results produced using this data
• Dialogue with stakeholders, infrastructure providers, and funders on how to best promote reproducibility and robustness in secure, restricted-access data environments