Pre-registration vs Registered Reports
Summary
Pre-registration and Registered Reports are two distinct mechanisms for committing to a plan before seeing results. A pre-registration time-stamps an analysis plan to a registry — no peer review required. A Registered Report adds results-blind peer review: the journal reviews the question and methodology before data are collected and grants in-principle acceptance, so publication does not depend on how the results turn out. Registered Reports therefore attack publication bias and selective reporting directly, while also improving study designs during review.
Overview
Both mechanisms enforce the Prediction vs Postdiction distinction, but they operate at different points in the workflow and solve different problems.
Pre-registration
Posting the research questions and analysis plan to an independent registry before observing outcomes. It establishes the prediction/postdiction bright line but says nothing about whether or where the work will be published.
Registered Report
A publishing model (offered by dozens of journals, originating with Chambers et al. 2014 and Nosek & Lakens 2014) in which the manuscript is reviewed and accepted before results exist.
Main Content
The two-stage Registered Report workflow
Stage 1 — results-blind review
Authors submit the research question and methodology for peer review before observing outcomes. If reviewers agree the question is sufficiently important and the methodology is of sufficiently high quality, the paper receives in-principle acceptance (IPA).
Stage 2 — execution review
Authors carry out the study and submit the final report. Reviewers do not evaluate the perceived importance of the outcomes. They evaluate quality of execution and adherence to the preregistered plan.
The decisive feature: because acceptance is decided at Stage 1, the result (positive, negative, or null) cannot affect whether the paper is published.
What each mechanism fixes
| Problem | Plain pre-registration | Registered Report |
|---|---|---|
| Postdiction recast as prediction | Fixed (bright line) | Fixed |
| Forking paths / undisclosed flexibility | Fixed for planned analyses | Fixed for planned analyses |
| Publication bias / file drawer | Not addressed | Fixed (IPA before results) |
| Selective reporting of outcomes | Detectable | Prevented by design |
| Weak study design caught too late | Not addressed | Improved during Stage 1 review |
| Peer-review prestige / acceptance | Independent of it | Built in |
Why results-blind review matters
In the ordinary system, reviewers and editors favor novel, positive, clean results, so authors are incentivized to produce (or selectively report) them. Reviewing before results exist removes the outcome from the acceptance decision and aligns the reward with rigor rather than with a favorable result. This is the “motive” lever in the means/motive/opportunity framing of Pre-registration and Open Science - Overview.
Related incentive mechanisms
- Pre-registration badges — journals award a visible badge for preregistered work; analogous open-data badges produced a >10-fold increase in data sharing in an initial test (Kidwell et al. 2016).
- The Preregistration Challenge — one thousand $1,000 awards for publishing the results of a preregistered study.
- TOP Guidelines — thousands of journals and many funders are signatories to the Transparency and Openness Promotion standards, which include pre-registration. (See Pre-analysis Plans and the Open Science Ecosystem.)
- Legal/editorial mandates — pre-registration of clinical trials is required by US law and by ICMJE-adhering journals (though clinical-trial standards require only identification of primary/secondary outcomes, not a comprehensive analysis plan).
Examples
Same study, two paths
A psychologist with a clear hypothesis can (a) pre-register the design and analysis on OSF, then submit the finished paper anywhere — publication still hinges on results; or (b) submit a Registered Report: get in-principle acceptance on the design, run the study, and publish regardless of whether the effect appears. Path (b) additionally protects against the file drawer.
Connections
- Pre-registration and Open Science - Overview — situates both mechanisms within the broader practice.
- Prediction vs Postdiction — the distinction both mechanisms enforce.
- Pre-analysis Plans and the Open Science Ecosystem — the registries and PAP contents that pre-registration relies on.
- Limits and Objections to Pre-registration — what neither mechanism fully solves.
See Also
- Garden of Forking Paths — the analytic-flexibility problem both mechanisms constrain
- Forking Paths and Bayesian Approaches — pre-registration recommended as a complementary safeguard