Preregistration and Registered Reports in Ageing Research

Key Takeaways

Preregistration creates a time-stamped record of research questions, methods, and planned analyses before the outcomes are known. [1] [2]
A Registered Report adds peer review before the study is carried out and, after in-principle acceptance, makes publication largely independent of whether the results are positive or statistically significant. [3] [4]
These approaches can make confirmatory and exploratory analyses easier to distinguish and can reduce opportunities for selective reporting, but they do not make a weak design strong. [1] [2]
In ageing research, prespecification is relevant to animal lifespan experiments, biomarker studies, longitudinal cohorts, and clinical trials, each of which presents different planning constraints. [5] [6] [7]

Who This Is Useful For

This page is useful for readers assessing whether the main questions and analyses in an ageing study were specified before researchers saw the results. It is especially relevant when a paper reports many biomarkers, subgroups, time points, or alternative statistical models.

Many datasets permit more than one defensible analysis. Researchers may have choices about exclusions, outcome definitions, covariates, transformations, stopping rules, and subgroup comparisons. When choices are influenced by the observed results and only the most favourable analysis is presented, the published evidence can appear more decisive than the underlying study warrants. [1] [2]

Preregistration and Registered Reports address this problem at different points in the research process. Neither replaces critical appraisal, replication, data sharing, or complete reporting. Their main value is to preserve information about which decisions were made before the outcomes were known. [2] [3] [8]

Preregistration and Registered Reports at a Glance

Feature	Preregistration	Registered Report
Core record	A time-stamped study and analysis plan	A detailed Stage 1 manuscript containing the rationale, methods, and analysis plan
Review before results	Not necessarily	Formal peer review before data collection or outcome inspection
Publication commitment	None by default	In-principle acceptance if the approved protocol is followed and the conclusions are justified
Main bias addressed	Undisclosed flexibility in design, analysis, and reporting	Undisclosed flexibility plus publication decisions based on the direction of results
Exploratory analysis	Allowed when identified separately	Allowed alongside the approved confirmatory analyses when identified separately

The distinction is therefore not simply that one record is more detailed. A Registered Report changes the publication sequence by placing the main review and provisional publication decision before the results are available. [3] [4]

What a Useful Preregistration Specifies

A useful preregistration normally identifies the research questions or hypotheses, primary and secondary outcomes, sampling or stopping rule, inclusion and exclusion criteria, data-processing steps, statistical models, and the criteria used to interpret the results. Specificity matters because a vague plan can leave most of the original analytic flexibility intact. [1] [8]

Preregistration does not prohibit changes. Equipment can fail, recruitment can be slower than expected, and an approved analysis may turn out to be unsuitable. The informative practice is to identify the deviation, explain why it occurred, and distinguish the revised or additional analysis from the original plan. Studies comparing registrations with publications show that undisclosed discrepancies remain common, so the presence of a registration badge is not enough by itself. [8] [9]

Why the Distinction Matters in Ageing Research

Ageing research spans short laboratory experiments, whole-lifespan animal studies, repeated biomarker measurements, secondary analyses of long-running cohorts, and human intervention trials. These designs create different opportunities for flexible decisions and require different forms of prespecification. [5] [6] [7]

Animal lifespan studies: strain, sex, intervention start age, randomisation, blinding, exclusions, and survival analyses can materially affect interpretation. A 2025 appraisal of 667 preclinical ageing-intervention studies found substantial gaps in the reporting of several such design features. [5]
Biomarker studies: preprocessing, feature selection, clock construction, covariate adjustment, and validation datasets can create many analytic paths. Prespecification can show which performance claims were planned and which arose during model development. [2] [6]
Longitudinal cohorts: the data may already exist and may have been examined by some members of the research team. Preregistration can still document an analysis plan, but readers need enough information to judge prior exposure to the data and whether the analysis is genuinely outcome-blind. [7]
Clinical trials: prospective trial registration and results reporting have regulatory and editorial roles, while a detailed preregistered analysis plan or Registered Report may specify substantially more than a basic registry entry. [2] [10]

What the Evidence About Registered Reports Shows

Meta-research provides encouraging but not definitive evidence. In one comparison from psychology, Registered Reports were much less likely than conventional articles to report support for their first hypothesis, a pattern consistent with reduced selection for positive findings. The comparison was not randomised, so it cannot by itself establish that the publication format caused the entire difference. [4]

A separate blinded assessment compared 29 published Registered Reports with related conventional papers. Registered Reports received higher ratings for methodological and analytical rigour and for overall quality, while differences in novelty and creativity were uncertain. This evidence came mainly from psychology and neuroscience, so its size and generalisability to all areas of ageing research remain open questions. [11]

What Preregistration Does Not Guarantee

A sound question: a precisely registered study can still test an uninformative or poorly justified hypothesis. [2]
A strong design: preregistration alone does not correct low power, weak controls, unreliable measurements, or inappropriate models. [2] [5]
Complete transparency: a registration may be too vague to constrain decisions, and deviations may be difficult to detect if they are not reported clearly. [8] [9]
Confirmation: a preregistered result can still be affected by sampling error, bias, or limited external validity and may not replicate in another cohort, laboratory, species, or assay platform. [2] [5]
No discovery: exploratory analyses remain scientifically useful when they are labelled as exploratory and tested independently later. [1] [3]

How to Read a Preregistered Ageing Study

Check the timing: determine whether the registration was completed before data collection, before access to outcomes, or only after the researchers had seen the data. [2] [7]
Compare the records: match the paper's primary outcomes, sample size, exclusions, and main analyses to the registered plan. [8] [9]
Look for labelled deviations: a justified change is not automatically a flaw, but an undisclosed change prevents readers from separating planned tests from data-dependent decisions. [8]
Identify the publication format: ordinary preregistration does not imply that the journal reviewed the plan before results or committed to publish the outcome. [3]
Appraise the study itself: registration status should be considered alongside design quality, measurement validity, effect size, uncertainty, and independent replication. [2] [11]

Practical Interpretation Examples

If an animal study preregisters survival endpoints but adds an unplanned subgroup: the subgroup may generate a useful hypothesis, but it should not be presented as if it were the original confirmatory test. [1] [5]
If a cohort analysis is registered after investigators have explored the dataset: the record can still improve transparency, but it offers less protection against prior knowledge influencing the plan. [7]
If a Registered Report produces a null result: in-principle acceptance means the direction or significance of that result should not determine publication, provided the approved study was completed and interpreted appropriately. [3]
If a paper displays a preregistration badge: readers still need to inspect the plan's specificity and compare it with the published methods and analyses. [8] [9]

Summary

Preregistration records research decisions before outcomes are known; Registered Reports additionally subject the plan to peer review and make the initial publication decision before the results exist. In ageing research, these formats can clarify the status of analyses across complex animal, biomarker, cohort, and clinical designs. They reduce particular opportunities for bias, but the credibility of a finding still depends on the specificity of the plan, transparent reporting of deviations, study quality, and evidence beyond a single result. [2] [3] [5] [8]

References

Nosek, B. A., et al. (2018). Proceedings of the National Academy of Sciences. https://www.pnas.org/doi/10.1073/pnas.1708274114
Hardwicke, T. E., & Wagenmakers, E.-J. (2023). Nature Human Behaviour. https://www.nature.com/articles/s41562-022-01497-2
Chambers, C. D., & Tzavella, L. (2022). Nature Human Behaviour. https://www.nature.com/articles/s41562-021-01193-7
Scheel, A. M., et al. (2021). Advances in Methods and Practices in Psychological Science. https://journals.sagepub.com/doi/10.1177/25152459211007467
Parish, A., et al. (2025). npj Aging. https://www.nature.com/articles/s41514-025-00287-0
Higgins-Chen, A. T., et al. (2022). Nature Aging. https://www.nature.com/articles/s43587-022-00248-2
Weston, S. J., et al. (2019). Advances in Methods and Practices in Psychological Science. https://journals.sagepub.com/doi/10.1177/2515245919848684
Claesen, A., et al. (2021). Royal Society Open Science. https://royalsocietypublishing.org/doi/10.1098/rsos.211037
TARG Meta-Research Group & Collaborators. (2023). BMJ Open. https://bmjopen.bmj.com/content/13/10/e076264
DeVito, N. J., et al. (2020). The Lancet. https://pubmed.ncbi.nlm.nih.gov/31958402/
Soderberg, C. K., et al. (2021). Nature Human Behaviour. https://www.nature.com/articles/s41562-021-01142-4

Educational Disclaimer

This content is provided for educational purposes only and does not constitute medical advice.