The Replicability Crisis has shaken psychology over the past decade and, in response, Open Science has evolved from an ideal to a requirement. But beyond ethics, pre-registration is a strategic tool for any author. Essentially, it involves creating a public, time-stamped record where you detail your hypotheses, design, and analysis plan before collecting data. You publicly commit to your research plan.
One of the most frequent criticisms from reviewers today is the suspicion of p-hacking or of formulating hypotheses after seeing the results. When you include a link to your pre-registration in the manuscript, you are presenting proof of honesty. The reviewer can no longer suspect that you fished for results, which increases your credibility and disarms many potential methodological criticisms.
Null results: from problem to opportunity
Pre-registration is also the best defense against rejection due to null results. Without pre-registration, when you fail to find significant differences, the editor tends to think that your study failed or was poorly designed. But with a pre-registration, a null result becomes something scientifically valuable, because it demonstrates that an established theory might not be correct. And it does so with the guarantee that the analyses were not manipulated. Many journals regularly accept studies with null results if they are well pre-registered. Some even offer special formats such as Registered Reports, where the article is accepted in principle before you have the results.
The process on OSF
The process is straightforward on platforms like OSF (osf.io). You create a free account, start a project, and choose a template (the most popular is the OSF Prereg Template). In it, you fill in your specific hypotheses, the variables you will measure, the planned sample size, and the statistical tests you will use. Then you freeze the registration, which is assigned an immutable timestamp. You can make it public immediately or keep it under embargo until you publish.
A common mistake is being too vague when filling out the template. It is not enough to write "we will use an ANOVA." You need to specify the factors, the levels, the post-hoc comparisons, and the correction for multiple comparisons. The more specific your pre-registration is, the harder it will be for reviewers to question your analytical decisions.
Flexibility and exploratory analyses
A frequent concern is that pre-registration forces you to follow the plan to the letter. This is not the case. You can conduct additional exploratory analyses or change some aspect of the design if there are good reasons for doing so. The only rule is transparency: in your manuscript, you separate the confirmatory analyses (those you pre-registered) from the exploratory ones (those that emerged afterward). Reviewers respect this honesty, and in fact it is much worse not to pre-register and have them suspect that everything was exploratory disguised as confirmatory.
OSF, AsPredicted, ClinicalTrials.gov, and PROSPERO: choosing the right registry
OSF is the most flexible and widely used platform in psychology, but it is not the only option, and in some cases it is not even the most appropriate one. AsPredicted (aspredicted.org) was developed by researchers at Wharton and the University of Pennsylvania as a deliberately minimalist alternative. It consists of nine standardized questions that fit on a single page and force you to be concise. I find it useful for simple experimental studies where the analysis plan is essentially a single confirmatory test, but it falls short for designs involving multilevel models, multiple dependent variables, or complex contingencies. OSF, by contrast, lets you upload analysis code, simulation reports, materials, and even pre-analysis scripts that you can then run on the real data with minimal modification.
If you work with clinical trials in psychology (for example, a randomized controlled trial of a cognitive-behavioral intervention), ClinicalTrials.gov or the WHO ICTRP registries are not optional but a publication requirement in many journals that follow ICMJE guidelines. Registering on OSF alone will not be enough. For systematic reviews and meta-analyses, the equivalent is PROSPERO (the international prospective register hosted by the University of York). PROSPERO requires you to specify the eligibility criteria, the search strategy, the risk-of-bias instrument, and the planned synthesis method before extracting data, and editors of journals such as Psychological Bulletin or BMJ often ask for the PROSPERO ID in the cover letter.
Registered Reports: the format that changes the publication game
Registered Reports deserve a separate section because they invert the traditional publication workflow. In the classical model, you collect data, run analyses, write the manuscript, and then submit it. The editor evaluates whether the results are sufficiently interesting and clean to publish. This creates obvious incentives for selective reporting and HARKing (hypothesizing after results are known). In the Registered Reports format, peer review happens in two stages. In Stage 1, you submit only the introduction, hypotheses, methods, and detailed analysis plan, before collecting data. If the protocol is approved, you receive an in-principle acceptance: the journal commits to publishing the article regardless of whether the results turn out significant or not, as long as you faithfully follow the approved protocol. In Stage 2, you submit the complete manuscript with the results, and reviewers only check the protocol's faithful execution.
More than 300 journals now offer this format, including Cortex, Nature Human Behaviour, Psychological Science, and many specialized psychology journals. For doctoral students or early-career researchers, Registered Reports are a tremendous strategic asset: they reduce the risk of investing years in a project that ends up unpublishable due to null results, and they considerably strengthen the CV. The downside is that the Stage 1 review process can take three to six months before you even collect data, so they are not suitable if you have very tight deadlines.
How to specify your analysis plan to avoid problems later
The quality of a pre-registration is measured by its specificity. A vague pre-registration offers little protection because it allows enough decisional flexibility for any direction of result to be reframed as "what was planned." When I review pre-registrations, the seven elements I check most carefully are these: the exact operational definition of each variable (which scale, which scoring, which range of values), the inclusion and exclusion criteria for participants and observations (what counts as an outlier and what you do with it), the rule for stopping data collection (fixed N, sequential analysis, time period), the exact statistical model with all its terms (including covariates, interactions, and random effects), the method for handling missing data, the multiple-comparison correction, and the criterion for declaring a result confirmatory.
If the blank page is what stops you, paste the project into the Methodologist: it returns a draft preregistration with those sections already written, plus the design flaws worth fixing and a sample-size check. You still have to review the draft, but correcting one is much faster than starting from nothing.
A particularly useful technique is to write the analysis script in R or jamovi before collecting data, using simulated data with the same structure as the real data. This forces you to specify every decision and gives you a script ready to run when the real data arrive. The simstudy package in R or the simulation functions in faux are excellent for this. Once you have run the script on the simulated data and verified that it works, you upload it to OSF as part of the pre-registration. Then, when you collect the real data, you simply replace the dataset and run the script. The transparency is total.
Frequent reasons for rejection despite having pre-registered
Pre-registering does not guarantee publication, and there are recurring mistakes that lead reviewers to discount the pre-registration's value. The first is having a deviation from the protocol without justifying it transparently. If you changed the sample size, the analysis, or the operationalization of a variable, you must declare it explicitly in the manuscript, explain why, and discuss how it affects the interpretation. Hiding deviations is worse than not pre-registering. The second is pre-registering after collecting data (sometimes called "PARKing": pre-registering after results are known). The OSF timestamp helps detect this, but reviewers also look at the registration date relative to the IRB approval date and the recruitment dates. The third is pre-registering analyses that the data cannot support: for example, planning a structural equation model with a sample size that is clearly insufficient. The fourth is pre-registering hypotheses so vague that any result confirms them. If your hypothesis is "there will be a relationship between X and Y," any direction confirms it. Specify the expected direction and, when possible, the minimum effect size of interest.
That IRB approval date matters because registering before you have it is its own red flag: a design and analysis plan locked in before the committee has even seen the protocol is hard to defend if the committee later asks you to change something material. If your application is still being drafted, the ethics committee simulator anticipates the objections a committee is likely to raise and which paragraphs your application is missing, so the plan you register is the one that survives review, not one you end up revising afterward.
How to cite the pre-registration in the manuscript
APA 7 has clear guidelines for citing pre-registrations. In the method section, after describing the design, include a sentence such as: "The hypotheses, design, and analyses were pre-registered prior to data collection (Author, Year; https://osf.io/xxxxx)." If you have made the project public, include the persistent link. If it is still under embargo, indicate that it will be made available upon publication and provide a private anonymous link for reviewers (OSF generates one specifically for double-blind review). In the results section, when reporting each analysis, specify whether it is confirmatory (pre-registered) or exploratory (post-hoc). Some journals, such as those following the TOP guidelines, even require this distinction to be made visually in tables or with explicit subheadings. Consistency with the APA 7 reporting standards and a clear separation between confirmatory and exploratory analyses are signals of methodological maturity that reviewers value.
Before you submit: a pre-registration only protects you if the manuscript actually reflects it, and that is exactly where most authors slip. Once your draft is written, run it through the AI Paper Reviewer: a free Reviewer 2 style pre-review that flags undeclared deviations from the protocol, exploratory analyses presented as confirmatory, and results that do not match the plan you registered. Better to hear it now, while you can still fix it, than in the editor's letter three months from now.
Pre-registration does not take away your flexibility, but it does give you an armor of credibility that, in the competitive world of academic publishing, can be the difference between an acceptance and a rejection. If you need help designing your analysis plan before pre-registering, our journal publication support service can guide you so that your pre-registration is solid and your manuscript harder to reject.