Learning how to write a data availability statement is now a normal part of preparing a research paper. Readers, reviewers, editors and funders expect to know what data support your findings, where those data live, and what limits apply to their use.
This guide explains the statement in plain language, with templates you can adapt. It covers journal and funder requirements, where the statement belongs, open, restricted and no-new-data wording, repository choice, dataset DOIs, licences, and sensitive human or animal data.
A data availability statement is a short paragraph, usually two to five sentences, stating where the data underlying your results can be found, or why they cannot be shared openly. It is not a second methods section or a summary of results. Its job is narrower: it gives a reader a working route to the data, usually a repository name plus a DOI or accession number, and any conditions on access.
A strong statement answers four questions: what data it covers, where they are, how a reader can get them, and what limits apply. Our guide to research data management: organising, storing and sharing explains how to prepare those files.
Requirements differ, so always read the target journal's author guidelines and any funder policy that applies to your grant before drafting. Many journals now require a statement for every research article, including articles that generated no new data. Some require deposit in a recognised repository; others accept supplementary material or a controlled-access archive. Fields such as genomics have discipline repositories of their own, and there the journal usually expects the recognised archive.
Funders often add a second layer: a data management plan before the project, open deposit after publication, or a statement explaining an embargo or restriction. Your statement should agree with that plan and with what you told participants or collaborators about sharing. For the wider process, see our guide to publishing a research paper.
Placement varies by journal, so follow the author guidelines for the journal you are submitting to. In many journals the statement appears near the end of the article, after the acknowledgements and funding statements and before the references. Others collect it in the submission system as a metadata field and display it on the article page automatically.
Avoid two placement mistakes. Do not bury the information only in the methods section, which describes how data were collected, not where they now live. Do not leave it only in the cover letter when the journal expects it in the manuscript, because a reader of the published article cannot see your cover letter. For double-blind review, keep the statement free of anything identifying you, such as a deposit under your own name; ask the editorial office if unsure. The same care applies to methods, as explained in our guide on how to write the methodology section of a research paper.
Decide, before drafting, which of four cases your paper falls into. Case one is open data: you generated data and can share them publicly, with no confidentiality or legal barrier. The statement should name the repository, give the DOI or accession number, and state the licence.
Case two is restricted data: the data exist but cannot be fully open because of privacy, consent limits, commercial agreements or safety concerns. Explain the restriction plainly, say what can be shared, and give a workable route for a request. Case three is no new data: your paper analyses existing datasets or published literature. Say so directly and cite the sources. Case four is third-party data: the data belong to someone else, and you used them under licence or permission. Credit the owner, cite the source, and explain whether readers can obtain the same data on the same terms.
Use the examples below as starting points, then replace every placeholder with your own details: repository name, DOI or accession number, licence, and contact or committee for requests. A template with placeholders left in it, or a DOI pointing to the wrong dataset, is worse than a short honest statement in your own words.
For open data, a typical wording is: "The datasets generated and analysed during this study are available in the [repository name] repository, [DOI or accession number]." For restricted data, explain the limit first and the route second: why the data cannot be public, what access is possible, who decides, and how long a request may take. For no new data, name the sources and cite them. For third-party data, name the owner and the terms. The table below summarises these patterns.
Use "available on request" only with a named restriction, a deciding body, criteria and a response time, and deposit anything shareable openly. Never promise guaranteed access to data you do not control.
| Situation | Example statement | What to include |
|---|---|---|
| Open repository data | The datasets generated during this study are available in [repository], [DOI]. | Repository name, DOI, licence, contents note |
| Restricted data | Data are available on reasonable request, subject to [restriction], from [committee or contact]. | Reason, request route, criteria, response time |
| No new data | No new data were generated; this study analyses the sources cited in the references. | Clear statement plus full citations |
| Third-party data | Data were obtained from [owner] under [terms]; readers may apply to [owner] on the same terms. | Owner, source citation, access terms |
Choose a repository before you finish the paper, not after acceptance, because depositing well takes time. Start with any discipline-specific repository your field or journal recommends, since the expected archive makes your data findable by the readers most likely to reuse them. If none fits, use a general repository that issues persistent identifiers and supports clear licences, such as Zenodo, Dryad or Figshare.
Check four points before depositing. Persistence: will the link still work in years to come, and does the repository assign a DOI or stable accession number? Metadata: can you record title, authors, description, methods, variables, units and licence so a stranger can understand the data without emailing you? Versioning: if you correct the dataset, can you publish a new version while keeping the old one reachable? Access control: if part of the data must be restricted, does the repository support controlled access or embargoes? A dataset DOI is worth obtaining whenever offered. Like an article DOI, it permanently identifies one version of one dataset and lets you cite it in your references.
A licence tells a reader what they may do with your data, and a dataset without one leaves reuse in legal uncertainty, however open the repository page looks. Common open options are a public-domain dedication or an attribution licence similar to CC BY, under which others may reuse the data provided they credit the source. Restrictive licences forbidding commercial use or adaptation may suit some datasets, but they limit reuse and should be chosen deliberately, not by default.
The licence on your data and the licence on your article are separate choices. An open access article under CC BY 4.0, as published by Global Academic Publishers, does not automatically place its dataset under the same licence; the dataset's own licence governs reuse of the data. If third-party data are in your deposit, check you have the right to redistribute them; often the correct approach is to deposit only your derived data and point readers to the third-party source.
Openness has genuine limits, and privacy is the most important. Data about identifiable people, including health information, interview transcripts, location data and small-sample responses that could identify an individual, must not be published in identifiable form. Removing names is rarely enough: combinations of age, place, occupation and a rare condition can identify a person in a small community. The same risk applies to animal owners, farms or businesses identifiable from detailed location or production data.
There are honest ways to share without breaching confidentiality. Deposit anonymised or aggregated data that cannot reasonably identify individuals, with a description of the anonymisation steps. Use a controlled-access repository, where approved researchers apply and agree to conditions. Share analysis code and summary results openly while keeping individual-level data restricted. The data statement should also agree with the ethics statements described in our guide to research ethics, consent, approval and animal welfare.
Run through this checklist on the final manuscript, with the repository page open beside you. The statement names the correct case, or describes a mixed case part by part. Every repository name, DOI and accession number matches the live deposit, and every link has been tested. The licence in the repository agrees with the statement and with your funder and institution rules. Sensitive data are anonymised or under controlled access, and the statement explains the restriction and the request route.
The most common mistakes are easy to avoid: copying a template without replacing the placeholders; citing a DOI pointing to a draft, a private deposit, or a different version from the one analysed; writing "available on request" with no restriction, contact route or timeframe; promising open data that a collaborator, owner or ethics committee has not agreed to share; and forgetting the statement for a review paper, on the mistaken idea that no new data means no statement is needed. This article makes no claim about indexing in any database, offers no promise of acceptance, and reports no statistics about sharing, because a data availability statement is a record of what you have done, not a marketing claim. Keep it factual, specific and checkable.
What is a data availability statement?
A short, formal paragraph in a research paper that tells readers where the data behind the findings can be found, or why they cannot be shared openly. It names the repository, accession number or DOI where applicable, and any access conditions.
Do I need a data availability statement if my paper uses no new data?
Yes, in most journals that require a statement. State clearly that no new data were generated and cite the existing sources you analysed, so readers know the position rather than guessing.
Can I write that data are available on request?
Only when a genuine restriction prevents open sharing, and many journals now discourage or reject this wording on its own. Explain the restriction, name who to contact, and state the conditions and timeframe for a response.
Where does the data availability statement go in a manuscript?
Usually after the acknowledgements and funding statements and before the references, or in the dedicated field of the submission system. Follow the target journal's author guidelines, because placement varies.
What if my data contain personal or sensitive information?
Do not publish identifiable personal data. Share anonymised or aggregated data, use a controlled-access repository, or explain the legal and ethical restriction and describe what can be shared and how access may be requested.
Should my dataset have its own DOI?
Yes, whenever the repository offers one. A dataset DOI gives the data a stable, citable link, lets readers find exactly the version you used, and lets you cite the dataset in your references.
These authoritative external resources explain the repositories and standards mentioned in this guide: