No. A DOI proves that an identifier was registered and metadata deposited with an agency such as Crossref or DataCite. It is not evidence that experts reviewed the work: preprints, datasets, theses and conference abstracts carry DOIs too. To judge one paper, read what the DOI identifies, then look for evidence of review.
If you searched "does a DOI mean peer reviewed" because a paper you want to cite carries one, that is the short answer. The rest of this guide covers what doi.org does, which items carry DOIs without review, how to read the record behind a DOI, and how to check one paper. How a DOI is built and resolved is covered in what a DOI is and why every paper needs one.
Does a DOI mean peer reviewed? No — it records registration
A DOI is registered by an organisation, such as a publisher or a repository, through a registration agency. The registrant deposits the identifier, descriptive metadata and the web address the DOI should lead to. For journal articles the agency is usually Crossref; for datasets it is typically DataCite.
The agency processes that deposit. It does not read the manuscript, assess the research or confirm that any review took place. Crossref membership is open to organisations that produce professional and scholarly content and accept its membership terms, which is why our post on where we are indexed and where we are not calls DOI registration infrastructure, not endorsement.
A DOI still has a use: one that does not resolve is worth asking about. Treat a working DOI as a basic check that an article is registered, never as proof that it was reviewed.
Is doi.org a reliable source? It is a reliable signpost, not a source
doi.org is the resolver for the DOI system, which the DOI Foundation manages. When you open https://doi.org/ followed by a DOI, the resolver looks up the web address recorded for that identifier and redirects your browser there. It does not host the article, and it does not review, rate or rank the page it sends you to.
So doi.org is reliable at one job: delivering you to the address the registrant recorded. That is why a doi.org link is the right form for a citation. It is not a source in the sense a reference list means. The source is whatever the DOI lands on, and its reliability is that item's reliability. A preprint reached through doi.org is still a preprint.
Resolving has limits too. The address in the record is only as current as the registrant keeps it, so a DOI can redirect without error to a page that has since gone. The identifier and its metadata outlast a closed website, but the full text may not, as what happens to your article if a journal shuts down explains.
Items that carry DOIs with no journal peer review
Each item below can carry a DOI that resolves normally with no journal peer review behind it.
| Item | Who registers the DOI | What review, if any, stands behind it |
|---|---|---|
| Preprint | A preprint server, through Crossref or DataCite | No peer review before posting; a journal may review a later version |
| Dataset or software | A repository, typically through DataCite | Checks vary by repository; none is journal peer review of the findings |
| Self-deposited paper or file | A general repository such as Zenodo, through DataCite | No peer review by the repository; Zenodo's policies say information is provided "as-is" |
| Conference abstract or poster | The meeting organiser or its publisher | Selection by the meeting under its own rules; not review of a full paper |
| Thesis or dissertation | A university or its repository | Examination under the university's rules, not journal peer review |
| Report or working paper | The issuing organisation | Whatever the issuer applies |
| Editorial, letter or commentary | The journal, through Crossref | May be editorial review only, depending on the journal |
| Reviewer report | A journal or review platform that publishes its reviews, through Crossref or DataCite | None; the report is the review |
Preprints are shared before review by design, and a preprint and its later journal version carry separate DOIs. Preprints versus peer-reviewed articles sets out how the two differ and when to cite each.
Read what the DOI identifies: the agency and the record type
Three free lookups tell you what kind of item a DOI names before you rely on it. Replace [doi] with the identifier, starting at 10.
- Find the agency. Open
https://doi.org/doiRA/[doi]. This address, described in the DOI Foundation's resolution documentation, returns a short answer naming the registration agency, such as Crossref or DataCite. - Read the type. For a Crossref DOI, open
https://api.crossref.org/works/[doi]and find the field "type". For a DataCite DOI, openhttps://api.datacite.org/dois/[doi]and find "resourceTypeGeneral". - Read the relations. In a Crossref record, the "relation" field can list has-review (peer review records, such as reviewer reports, registered and linked to the item), or has-preprint and is-preprint-of (the link between a preprint and its journal version).
| Type value | Agency | What it names |
|---|---|---|
| journal-article | Crossref | An item a publisher deposited as a journal article, which can be an editorial or a letter |
| posted-content | Crossref | Preprints, eprints and other content posted to a hosting platform |
| proceedings-article | Crossref | A paper in conference proceedings |
| peer-review | Crossref | A reviewer report, decision letter or author response |
| Preprint, Dataset, Software, Dissertation, JournalArticle | DataCite | DataCite's own vocabulary for these and other kinds of item |
Crossref publishes its full list of record types and its definition of posted content; DataCite publishes its metadata schema. Read every type as a label the depositor chose. Neither agency checks it against what happened to the manuscript. Crossref's list has no separate type for an editorial or a letter, so a record typed journal-article can be an item no external reviewer saw. DataCite's vocabulary includes JournalArticle too, so the label alone cannot tell a journal's version from a copy deposited elsewhere.
The label cuts the other way as well: some platforms link published reviews to preprint records, which Crossref's peer review records make possible. The type tells you what an item is; the relations and the checks below tell you what happened to it.
The reference below is invented to illustrate the lookups; the DOI is not real.
A draft cites 10.xxxx/abc.2025.0412 as a journal article. The agency lookup returns DataCite. The DataCite record gives resourceTypeGeneral "Preprint" and a repository as publisher. Conclusion: cite it as a preprint, and search for a journal version before relying on its findings.
A checklist for judging whether one paper was peer reviewed
Run these checks on the paper itself, not on the journal's home page. No single row settles the question; weigh the rows together.
| Check | Where to look | What it can show | What it cannot show |
|---|---|---|---|
| 1. The DOI resolves | https://doi.org/[doi] | The identifier is registered and leads somewhere | Anything about review |
| 2. Agency and type | The doiRA lookup, then the type field | Whether you hold a journal article, a preprint, a dataset or a thesis | Whether a journal article was reviewed |
| 3. The review policy for this article type | The journal's peer review and article types pages | Which article types the journal sends to external reviewers | That this paper received the review described |
| 4. Received, revised and accepted dates | The PDF or the article page | A history with time for reviewers to read and authors to revise | Proof on their own, though acceptance a few days after submission leaves little time for any review |
| 5. Published reviewer reports | The article page; has-review in the Crossref record | That review took place, in reports you can read | Anything when absent: a journal that keeps reports confidential publishes none |
| 6. The version you are reading | The PDF header, the landing page or the repository record | Whether you hold the preprint, the accepted manuscript or the version of record | How thorough the review of the final version was |
| 7. Correction and retraction notices | The landing page; Crossmark, where the publisher uses it | Whether the record changed after publication | Whether the paper was reviewed in the first place |
| 8. The journal in a curated index | The index's own records, such as the current indexing status shown in the National Library of Medicine catalogue, or the Directory of Open Access Journals | That an outside body assessed the journal | A verdict on one paper; absence does not show it went unreviewed |
Rows 3 to 5 bear on review directly. Rows 1, 2 and 6 establish what you are holding, which is still the first thing to settle.
What a DOI on our articles proves, and what it does not
The same rule applies to Directive Publications. We are a Crossref member, and every article we publish has a DOI registered under prefix 10.52338. That proves registration. It is not evidence of quality, and we do not present it as one.
What we publish about review is the process in our peer review policy: double-blind review, at least two independent expert reviewers sought per research manuscript, and a decision by the handling editor. Our reviewer guidelines describe that review as confidential, so row 5 returns nothing for our articles, and rows 3 and 4 cannot show that a given paper of ours received the review the policy describes. Row 8 does not help either: our journals are not in PubMed, Scopus, Web of Science or the Directory of Open Access Journals.
If one of our DOIs does not resolve, or lands on the wrong article, report the problem with that article so we can check the record.