Standards
Digital Object Identifier (DOI)
A persistent identifier for a registered digital object or other referent.
- Kind
- Identifiers
- Steward
- International DOI Foundation; ISO 26324
- Official source
- https://www.doi.org/
What it describes
A DOI is an identifier assigned through a DOI registration agency to a specific referent, often a scholarly article, dataset, or report. The identifier resolves through DOI infrastructure to metadata and a current landing page. This separates the durable identifier from a publisher URL that may change. DOI registration agencies maintain their own metadata rules and coverage, so the cited object must still be checked.
A DOI does not guarantee that the linked content remains available, open access, unchanged, or correct. Version relationships and corrections belong in metadata or the publisher record, not in an assumption about the string. For evidence work, retain the DOI, title, version or publication date, retrieval date, and a source copy where permitted. A DOI landing page is a locator, not the evidence excerpt itself.
Structure
Prefix beginning 10. followed by a registrant code, slash, and suffix; the suffix is assigned under the registrant's rules.
Useful for
- Link scholarly works across publishers.
- Resolve changing publication URLs.
- Reconcile citations to registered metadata.
Limits
- Does not guarantee access or preservation.
- Metadata can be corrected.
- A DOI can identify a version rather than all versions.
How it is used in practice
A DOI is the name printed in the reference list and resolved by a link. Registration agencies such as Crossref, which registers scholarly works from publishers, and DataCite, which registers research data, assign DOIs to their members' objects and deposit metadata about them. Following a DOI link through the doi.org resolver leads to the current landing page, and the agencies expose the metadata through programmatic services. That makes it possible to retrieve title, authors, dates, relations to other works and sometimes funding and licence information for a cited work.
The persistence is in the redirect, not in the content. When a publisher moves an article, it updates the target of the DOI, and the citation continues to resolve. When a publisher disappears or stops maintaining the target, the DOI may resolve to nothing useful, which is why the identifier is a locator for the work and not a guarantee of access.
A worked example
A fictional report states that a method reduces errors, citing a paper by DOI. A reviewer resolves the DOI and reaches a publisher page. The metadata retrieved from the registration agency gives the title, the publication date and a relation indicating that the article has a correction notice, which is registered under a different DOI.
The reviewer records the DOI, the retrieved metadata, the retrieval date and the relation to the correction. The cited sentence in the report is checked against the sentence in the paper, and then against the correction, which changes one figure. The finding is that the report cites the original figure, that the correction exists, and that the correction was published after the report's date or before it, whichever the dates show. The DOI made the correction discoverable. Reading it and deciding what it changes remained the reviewer's task.
Evaluating a claim that rests on a DOI
Resolution shows that the identifier is registered and points somewhere. It does not show that the target is the cited work, that it is the version the claim used, or that it is still accessible. Compare the title and authors from the metadata with the citation, and prefer the version of record when versions differ. Preprints, versions of record and datasets often have separate DOIs, with relations among them in the metadata.
For evidence work, a DOI should be accompanied by a retained copy where the licence allows it, the retrieval date and the passage cited. The landing page is a pointer. The text that supports a sentence is on the page behind it, and that text can change in place, for instance when a publisher adds an erratum notice. Capture it at the time of reading.
Confusions to avoid
A DOI is not a measure of quality or peer review. Predatory publishers, preprint servers and data repositories all assign DOIs, and the identifier says nothing about review. The reputation and review process of the publisher or repository are separate facts to be established from other sources.
It is also not an author identifier. An ORCID iD identifies a contributor, and the DOI identifies an object. The two are linked in metadata when authors supply their iD, and the link is a claim with a source. Finally, a DOI string is case-insensitive in resolution but systems differ in how they store it. Normalize in a separate field and keep the string as cited.
Questions a reviewer can ask
Does the DOI resolve, and to what? Do the title, authors and date in the registration metadata match the citation? Is the DOI for the version of record, a preprint or a dataset? Does the metadata list corrections, retractions or related versions? Is a copy of the cited text retained, with the passage located? When was the landing page read, and does it still show the same content today?
Related terms
Found an error on this page? See corrections for what counts as one, how to report it and the dated log.
Enquiry
Which of your sources carry Digital Object Identifier (DOI)?
Tell us which sources carry it, which key the same entity some other way and who may read which. We reply with whether and how a first record would use this identifier.
Or write to [email protected] in your own words.