Registered reports

What are Registered Reports?

Registered Reports are a publication format in which researchers submit a study's question, methods and planned analyses for peer review before the main results are known. A journal can offer in-principle acceptance: a commitment to publish if the study meets the agreed requirements, regardless of whether its main result is statistically significant. The Center for Open Science describes the two-stage process.

What happens in the two stages?

At Stage 1, reviewers can challenge the question, measurements, comparison, sample-size rationale and analysis while the design can still be improved. Researchers then conduct the approved work. At Stage 2, the complete report adds results and interpretation, identifies departures from the plan, and distinguishes exploratory analyses. Acceptance remains conditional on the relevant protocol and quality requirements; it is not permission to submit an unusable study.

A result that deserves publication either way

Imagine a hypothetical study of whether a reminder helps students return library books on time. Reviewers may ask the researchers to define “on time,” account for repeated borrowing by the same student, and choose a sample that can detect a worthwhile difference. Those questions improve the planned comparison before anyone knows whether the reminder looks impressive.

If the completed study estimates almost no change with a narrow interval, that may be valuable information for libraries deciding whether to invest in the system. If it estimates an improvement, readers can compare the planned and reported analyses. In either case, the result's attraction is not supposed to determine whether the agreed study enters the literature.

How is this different from preregistration?

Pre-registration creates a dated record of a plan. Registered Reports add review and a publication decision tied to that plan before the main results. A preregistered study need not be a Registered Report, and posting a protocol does not itself mean a journal has accepted it.

The format targets incentives that contribute to publication bias and undisclosed analytical flexibility. It does not establish that an intervention is effective, that every measurement is valid, or that a result will transfer to a new setting. Judge the completed evidence as well as the process that produced it.