Preregistration and Registered Reports: Planning Research Transparently

Understand the problem these practices address

Research teams often make many decisions about outcomes, exclusions, models and interpretation. When decisions are made after examining results, it can become difficult for readers to distinguish an original test from a later explanation. Transparent planning helps preserve that distinction.

Preregistration records a research plan at an appropriate point before the relevant analysis or outcome knowledge. The Center for Open Science describes it as a way to distinguish planned confirmatory work from exploration. ^[1] It does not eliminate the need for good design, ethical review or careful analysis. A vague plan can be registered and still leave important decisions unclear.

Distinguish preregistration from Registered Reports

A preregistration can be created independently of a journal submission. It provides a dated account of questions and methods, with access arrangements depending on the registry and project. It is not automatically peer reviewed or an assurance that the study will be published.

Registered Reports are a publication format in which a journal evaluates the question and methods before results are known, with later review against the approved plan and relevant conditions. ^[2] Requirements vary by journal. Researchers should inspect the particular process, eligibility and timing rather than assuming that all journals offer the same arrangement.

Write a plan that guides real decisions

Specify the research question, hypotheses where appropriate, design, sampling, outcomes and analysis. Explain inclusion and exclusion rules, missing-data handling and any planned subgroup comparisons. Define when collection will stop and how multiple analyses will be interpreted.

The appropriate detail depends on the design. If parts of the work are intentionally exploratory or adaptive, describe the decision process rather than pretending every detail is fixed. A useful plan makes clear what was known when decisions were made. Avoid registering a polished explanation after examining the same results that it is supposed to predict.

Preserve flexibility through transparent amendments

Unexpected events occur: an instrument fails, a partner withdraws or an assumption proves untenable. The responsible response is to document the change, explain the reason and identify when it occurred relative to data access and analysis. Follow the registry, journal and institutional procedures that apply.

Separate analyses that follow the original plan from those added later. Exploratory work can be valuable, especially when it generates new questions or reveals an unanticipated pattern. The problem is not exploration itself but presenting it as if it were a pre-specified test. Transparency allows readers to judge the different evidential roles.

Illustrative example: a study of feedback timing

Suppose a team plans to compare two feedback schedules using a defined assessment outcome. It records the allocation process, sample-size assumptions, primary outcome and analysis before collecting results. During the study, an assessment date changes for one group because of an institutional disruption.

The team documents the disruption, consults relevant reviewers and explains the resulting amendment. It reports the planned analysis where appropriate and adds a sensitivity analysis addressing the timing difference. The final paper clearly labels the additional work. Readers can then assess the impact of the change instead of discovering an unexplained mismatch between the original plan and report.

Consider fit, timing and responsibilities

Preregistration is especially useful when a study tests specified claims, but its form must suit the discipline and design. For existing datasets, describe what information the team has already seen and what analysis remains unexamined. For qualitative or iterative work, consider whether a transparent protocol can clarify intent without imposing inappropriate rigidity.

Assign responsibility for maintaining the plan and checking the final report against it. Include all relevant collaborators in the planning discussion. A registration completed by one person at the last minute may not reflect the actual project. Protect confidential or sensitive information when choosing what to make publicly accessible.

Use a planning and reporting checklist

• The question and purpose of registration are explicit.
• Relevant prior data access is disclosed.
• Outcomes, sampling and analysis decisions are sufficiently clear.
• Legitimate adaptive elements have a documented process.
• Amendments are dated and explained.
• Exploratory analyses are labelled accurately.
• The final report links to the correct plan and version.

These practices improve the visibility of research decisions, but they are not substitutes for judgement. Readers still need to evaluate the question, methods, execution and interpretation. The strongest use of transparent planning is to make that evaluation easier, including when findings are unexpected or inconclusive.