On Oct. 8, the U.S. National Science Foundation (NSF) and UK Research and Innovation (UKRI) agreed to jointly study how their agencies award research grants — an arrangement that applies the empirical lens of metascience to the funders' own grant-making workflows.

The agencies said they will "explore cooperation" through data analysis and coordinated experiments. UKRI has already piloted new approaches to grant review; the agreement positions NSF to evaluate those changes and run parallel studies inside the U.S. system.

What is metascience, and why are two funders running experiments on themselves?

Metascience applies the same empirical tools used in any research discipline — controlled studies, large datasets, peer review of methods — to the question of how science itself is funded, produced and rewarded.

The field has produced actionable findings on reproducibility, citation bias and grant-review reliability. Applying the same toolkit to the agencies' own processes is a logical extension, but one few large funders have attempted in public.

What does each agency bring?

NSF operates as the primary U.S. federal funder of non-medical basic research, distributing grants through disciplinary directorates that span fields from astronomy to the social sciences. UKRI consolidates the United Kingdom's research councils and innovation functions under a single roof.

Combining the two agencies' data — one populated largely by U.S. principal investigators, the other by UK-based labs — gives the joint project a cross-jurisdictional sample that has rarely existed in this kind of work.

What has UKRI already piloted?

The agencies' joint statement says UKRI has trialled "new ways of reviewing" grants. The announcement did not specify which mechanisms, which research councils or which funding lines were touched.

That omission shapes what the project can measure. The choice of pilot — alterations to reviewer panel composition, scoring rubrics, application length or asynchronous remote review — determines whether the joint analysis can isolate process effects cleanly or only broad stroke differences.

What does a "coordinated experiment" actually mean?

A retrospective comparison of historical award outcomes would be useful, but confounded by structural differences between the U.S. and UK systems. A coordinated experiment, by contrast, holds the research question constant and varies only the review mechanism.

That setup isolates the effect of process on outcome — measuring reliability, equity and downstream research productivity — rather than attributing differences to the underlying applicant pool.

What changes for applicants now?

Nothing operational. The collaboration sits at the design stage. Near-term work will likely centre on:

• building a shared, anonymised dataset of historical award records
• agreeing experimental protocols for prospective pilots
• defining shared evaluation metrics

If the joint work yields the effect sizes funders typically cite when revising process — review reliability, award-rate equity, citation trajectories of funded versus unfunded proposals — applicants could see changes to application forms, review timelines or eligibility criteria in future rounds.

What happens next?

The agencies did not announce a delivery date or first publication. The earliest concrete output is likely a shared dataset specification, with coordinated experiments following once both sides clear methodological review.

The Oct. 8 agreement places two of the world's largest basic-research funders on the same experimental footing; the first joint data products will set the terms under which each agency rewrites its review process through the rest of the decade.