How to Use the OSF Preregistration Writer Prompt to Lock Your Analysis Plan Before Data Collection

A preregistration is only as useful as its least specific sentence. Lines like we may control for baseline differences or outliers will be handled appropriately leave choices open, and those open choices are exactly what reviewers and readers worry about. The OSF Preregistration Writer: Hypotheses, Sampling Plan, Exclusion Rules, and Confirmatory Analysis Plan Before Data Collection prompt turns a lab's study plan into a complete OSF Preregistration in the template's own section order, and it audits your notes for vague lines before you submit.

What the prompt produces

1. Study information with a title and numbered, directional hypotheses that name the variables.
2. A design plan covering study type, blinding, study design, and randomization.
3. A sampling plan with an existing data statement, procedure, sample size, the power calculation spelled out, and a stopping rule.
4. Variables including manipulated and measured variables and the exact formula for each index.
5. An analysis plan with one confirmatory model per hypothesis, inference criteria, exclusion rules, missing data handling, and a separate list of exploratory analyses.
6. A vagueness audit that quotes each open ended line from your notes and replaces it with a fixed rule.
7. A one page AsPredicted style summary for co-authors.

How to fill the inputs

ResearchQuestion is a few sentences on what you are testing and why. DesignPlan gives the conditions, whether they are between or within subjects, how randomization works, and who is blind to what. Sample covers the population, recruitment platform, and eligibility rules.

PowerInputs should include the effect size and where it came from, alpha, power, and the test. The prompt will not invent an effect size or a citation, so if you do not have a source it marks NEED SOURCE. Measures lists items, scales, and scoring, including reverse scored items. AnalysisNotes is the plan as the principal investigator described it, vague parts and all, because that is what the audit works on. DataStatus tells the prompt whether any data exist yet.

Reading the example output

The example is a two condition experiment testing whether a short self affirmation task makes regular sugary drink consumers more accepting of a health message:

• The hypotheses are directional and specific. H1 predicts higher message acceptance in the affirmation condition, and H2 predicts higher intention to cut down.
• The sample size is explained. A pilot effect of d = 0.30, alpha .05 two tailed, and power .90 give 235 per group, and the plan notes that a pilot estimate is uncertain.
• The stopping rule is concrete. Recruit 520, stop when 470 pass exclusions or recruitment reaches 600, with no interim analyses.
• Exclusions are decided now. Attention check failures, very fast completions, and very short writing responses are excluded, and no outcome based outlier removal is allowed.
• The audit closes three loopholes. The maybe covariate becomes a single prespecified robustness model, outliers will be handled becomes no outlier exclusion, and the gender question becomes clearly exploratory.

Tips for better results

• Paste your AnalysisNotes honestly, including the parts you are unsure of. The audit is most useful on rough notes.
• Run your own power analysis to confirm the numbers before submitting.
• Share the AsPredicted style summary with co-authors first, since it is easier to agree on one page.
• Keep the exploratory list short and specific so readers can see what you planned to explore.

Mistakes to avoid

• Do not mix confirmatory and exploratory analyses in the same section.
• Do not set exclusion rules after looking at the data.
• Do not cite an effect size from a paper you have not read.

Who it is for

Graduate students preparing their first preregistration, lab managers who review registrations before submission, and researchers in psychology, behavioral economics, and health communication.

Related PromptDig links

Open the OSF Preregistration Writer: Hypotheses, Sampling Plan, Exclusion Rules, and Confirmatory Analysis Plan Before Data Collection prompt and paste your study plan to begin. For more prompts built for researchers, Browse more prompts. If you have a methods or open science prompt your lab uses, Share a prompt.