This page assumes you already know what a data availability statement (DAS) is required to contain — for that, see How to Write a Data Availability Statement for Reproducibility. What follows is a set of complete, ready-to-adapt sample statements across the three situations authors most often face: data openly deposited, data under restricted or controlled access, and no new data generated.
These samples are illustrative composites. The study details, repository accession numbers, and institutional names below are invented for demonstration and follow realistic formatting conventions (a Zenodo-style DOI suffix, a dbGaP-style phs accession number) — they are not real datasets, real accession numbers, or attributable to any specific real institution, in keeping with the same illustrative-composite framing used elsewhere on this site. Do not cite these accession numbers or DOIs as if they resolve to real records; replace every bracketed and invented value with your own before submission.
Sample 1: Open data deposited in a public repository
This is the simplest case: the underlying data has no legal, ethical, or commercial restriction and can be deposited somewhere openly accessible before or at publication.
“The datasets generated and analysed during the current study, comprising the behavioral response times and eye-tracking fixation data reported in Figures 2–4, are available in the Zenodo repository at https://doi.org/10.5281/zenodo.[XXXXXXX]. The analysis code used to produce all reported statistics is available in the same deposit and is separately archived on GitHub at github.com/[org]/[repo], tagged at the version used for this manuscript (v1.2.0).”
Why this works: it names a specific repository rather than a category (“Zenodo,” not “a public repository”), gives a persistent identifier rather than a URL that can rot, states plainly that there is no restriction, and — per the Center for Open Science’s TOP Guidelines Level 3 standard covered in the linked reproducibility guide — addresses code as well as data, since the raw data alone would not let a reader reproduce the reported statistics.
A shorter version for data with no associated custom code:
“All data generated during this study are available in the Dryad Digital Repository at https://doi.org/10.5061/dryad.[xxxxxxx].”
Sample 2: Restricted or controlled-access data
Used when data cannot be openly shared — typically because it contains identifiable human-subjects information, is governed by a data use agreement, or is otherwise subject to a documented legal or ethical constraint. The key requirement, per PLOS’s data-availability policy and equivalent publisher policies, is that the restriction is named and backed by a real basis, with an institutional (not individual) point of contact.
“Individual-level genomic and clinical data from this study cannot be made openly available because participants did not consent to unrestricted public sharing and re-identification risk cannot be fully excluded. De-identified data are deposited in dbGaP under accession phs[XXXXXX].v1.p1 and are available to qualified researchers under controlled access, subject to review and approval by the [Institution Name] Data Access Committee. Requests should be submitted through the dbGaP request process; the Data Access Committee, not the study authors individually, determines approval.”
A version for data governed by a bilateral data sharing agreement rather than a centralized controlled-access repository:
“Data are available from the corresponding author on reasonable request, subject to execution of a data sharing agreement with [Institution] and approval by the [Institution] Research Ethics Board, because the dataset contains potentially identifying clinical information from a small patient cohort that precludes public deposit.”
Why this works, and where it differs from Sample 1: notice the second example still uses “available on reasonable request” — language the parent guide flags as insufficient on its own — but pairs it with a stated legal/ethical basis and a named institutional mechanism (the Ethics Board, not just “the author”), which is the specific condition major publishers require before that phrasing is acceptable. A bare “available upon request” with no such basis does not meet this bar.
Sample 3: No new data generated
Applies to purely theoretical, computational-review, or secondary-analysis work where nothing new was produced, or where all data used already has a public source.
“No new data were generated or analysed in the course of this study.”
Where the manuscript is a secondary analysis of already-published or third-party data, name the original source rather than using the blanket sentence above:
“This study is a secondary analysis of publicly available data from the [Dataset Name], accessed via [repository/portal] at [DOI or URL]. No new data were collected for this manuscript. Analysis code is available at https://doi.org/10.5281/zenodo.[XXXXXXX].”
Why this works: a reader who wants to check the analysis needs to know where the source data actually came from, not just that the current authors didn’t generate anything new — “no new data” and “no data available” are not the same claim, and conflating them is a common way this scenario gets under-specified.
A fourth situation worth naming: mixed availability
Many real studies combine an open primary dataset with a restricted supplementary one (for example, open survey data alongside restricted linked administrative records). Don’t force these into a single sentence — state each component’s availability separately:
“Survey response data are available in the OSF repository at https://doi.org/10.17605/OSF.IO/[XXXXX]. Linked administrative health records cannot be shared publicly due to state data-sharing restrictions; access requires a data use agreement with the [State] Department of Health, arranged through the corresponding author’s institution.”
Checklist before you submit any of these
- Replace every bracketed placeholder. An unresolved [XXXXXXX] or [Institution] is worse than a slightly plain statement — it signals the statement was copied rather than written for this manuscript.
- Get the real identifier before you write the final version. Deposit first, then cite the DOI or accession number the repository actually issued — don’t submit a placeholder expecting to fill it in at proof stage.
- Match your journal’s own required template if it has one. Springer Nature, PLOS, and others publish their own standard wording; use theirs where it’s mandated, and treat the samples above as the underlying structure to adapt, not a fixed script to override a journal’s own form.
- If this is a clinical trial, check whether the ICMJE data-sharing requirement applies. Manuscripts reporting clinical trial results submitted to ICMJE-following journals need a data-sharing statement addressing individual participant data specifically, on top of the general DAS — covered in the linked reproducibility guide.
- Keep it in sync with the actual deposit. If an accession number changes between submission and publication, update the statement in the final version.
Frequently asked questions
Can I copy one of these statements word for word?
Only after replacing every invented detail — the study description, repository name (if you deposit somewhere other than the example), accession number, and institution name all need to reflect your actual work. The structure and phrasing pattern are what’s reusable, not the specific content.
My journal has its own required DAS template — should I still use these?
Use the journal’s required wording where one is mandated; these samples are for understanding the underlying structure (identifier, access condition, restriction basis, contact) so you can fill in a journal’s template correctly, or write one from scratch where no template is imposed.
Is “data available on reasonable request” ever acceptable on its own?
Most major publishers now treat it as insufficient without a stated legal or ethical basis and a named institutional contact — see Sample 2 above and the fuller discussion in the linked reproducibility guide, which cites the specific journal-policy language this comes from.
Do I need a separate statement for data and for code?
Not necessarily separate statements, but both need to be addressed if the analysis needs code to reproduce it — Sample 1 above shows both covered in one statement. A data-only statement that omits necessary code leaves a reader unable to actually check the result.
For the underlying data management plan that a well-executed DAS often draws from, see the Data Management Plan worked examples guide, which follows the same fully-worked-example format for two complete DMPs.
