03 · What You Need to Know
What Can Happen to a Paper After Publication?
Publication Is Not Necessarily the Final Version of the Research Record
Errors and concerns can emerge after publication. Authors may discover mistakes. Readers may identify problems. Editors may receive allegations about data, images, analyses, authorship, ethics, or other aspects of the work. A journal may then take different actions depending on the nature and severity of the issue.
The terminology and procedures can vary among publishers and disciplines, but several forms of post-publication update are especially important when evaluating a paper.
| Publication Status |
What It Generally Signals |
What You Should Do |
| Correction or erratum |
An identified problem has been corrected or clarified, but the article is not necessarily invalid as a whole. |
Read the correction and determine whether it affects the information, analysis, or conclusion you intend to use. |
| Expression of concern |
The journal has formally alerted readers to potentially serious concerns while the matter remains unresolved or under investigation. |
Treat affected claims cautiously and read the notice to understand the nature and current status of the concern. |
| Retraction |
The journal or another authorized party has formally indicated that the article, or in some systems part of it, should no longer be treated as part of the reliable literature in its original form. |
Read the retraction notice and do not rely on the affected findings as valid evidence without a specific scholarly reason and explicit acknowledgment of the retraction. |
| Corrected and republished article |
A corrected version has been republished and linked to the earlier publication. |
Identify the corrected version and determine which version should be cited and interpreted. |
A Correction Is Not the Same as a Retraction
Not every error destroys a study. A correction may address an author affiliation, mislabeled figure, numerical error, omitted information, or a more substantive mistake while leaving the article's principal findings intact. In other cases, a correction may materially change a result or interpretation.
The label alone therefore does not tell you how consequential the change is. You need to read the correction.
The National Library of Medicine uses the category of errata broadly for corrections associated with previously published articles, including corrigenda, addenda, and some partial retractions. This is a useful reminder that terminology is not perfectly standardized across scholarly systems.
When assessing a corrected paper, ask the same broader question you would ask about any limitation that could affect confidence in a finding: what changed, and does that change matter for the particular claim you intend to use?
An Expression of Concern Is a Warning, Not a Retraction
An expression of concern alerts readers that significant questions have arisen about a publication but have not necessarily been resolved.
The National Library of Medicine links expressions of concern to the affected articles in PubMed. Such notices may later be followed by another editorial action, including a retraction, or the concern may eventually be resolved differently.
Do not treat an expression of concern as proof that every finding in the article is false. But do not treat the paper as unaffected either. Read the notice, determine what is under question, and consider whether your use of the paper depends on the disputed material.
Watch Out
An expression of concern represents an unresolved situation. If the paper is central to a consequential claim, check its current status again close to the time you submit, publish, or otherwise rely on your work because the editorial record may change.
A Retraction Is a Formal Change to the Publication Record
A retraction is much more consequential than an ordinary correction. It formally identifies a publication as having been withdrawn from the reliable research record in whole or, depending on the system and notice, in part.
Retractions can occur for different reasons. Problems may involve pervasive error, unreliable or irreproducible data, research misconduct, plagiarism, duplicate publication, ethical problems, or other circumstances that prevent the article from being relied upon as originally published.
Importantly, a retraction does not automatically tell you why the problem occurred. The National Library of Medicine, for example, does not distinguish retractions caused by honest error from those caused by scientific misconduct when assigning its retraction linkages.
That distinction matters ethically. “This article was retracted” and “the authors committed fraud” are not equivalent statements. If you need to discuss the reason, use what the formal notice and other authoritative records actually establish rather than inferring misconduct from the existence of a retraction.
The Retraction Notice Is More Informative Than the Word “Retracted”
If you discover that a paper has been retracted, find and read the actual retraction notice.
The notice may tell you which aspects of the work are unreliable, who initiated the retraction, and what reason was provided. The amount of detail varies, but the notice is the primary document for understanding the journal's formal action.
Do not stop after seeing a database label. A database can alert you to the existence of a retraction; the associated notice provides the context needed to interpret it.
Start With the Publisher's Current Article Page
For a specific paper, the publisher's current article record is an important place to check because the journal controls its formal publication status and should link relevant notices or updates.
Search using the DOI whenever possible. A DOI provides a persistent identifier for the work even if the publisher changes its website structure.
On the article page, look for visible notices such as:
- Correction;
- Erratum or corrigendum;
- Expression of concern;
- Retraction or withdrawal;
- Update;
- Corrected version;
- Version history.
The exact terminology and interface differ among publishers. Some place a prominent banner above the article; others link notices through publication history, related content, or Crossmark.
Crossmark Can Surface Post-Publication Updates
Crossmark is a Crossref service that publishers can use to communicate the current status of scholarly content. When implemented by the publisher, the Crossmark button can provide information about whether a document is current and can display updates such as corrections and retractions.
It is useful when available, but absence of a Crossmark indicator does not prove that no correction or retraction exists. Not every publisher participates or supplies the same metadata.
The broader principle is more important than the particular interface: follow the publication's persistent identifier back to current authoritative metadata rather than assuming an old local copy is current.
PubMed Explicitly Links Many Post-Publication Notices
For biomedical and life-sciences literature indexed in PubMed, the database provides particularly useful linkages between original articles and several kinds of post-publication notice.
The National Library of Medicine links citations for original articles with citations for retraction notices, errata, expressions of concern, and corrected and republished articles, among other related records.
A retracted article can be indexed with the publication type Retracted Publication, while the corresponding notice is indexed as a Retraction Notice. PubMed records can also display relationships such as “Retraction in” and “Retraction of.”
For corrections, NLM links the erratum notice with the affected article. Corrected and republished articles receive their own linked designation.
If the paper is indexed in PubMed, therefore, open the full PubMed record rather than relying solely on a citation exported years earlier.
Crossref Metadata Can Help You Check Updates Across Disciplines
Crossref provides metadata for scholarly works registered by its members and supports relationships between publications and later updates. Its REST API includes metadata indicating when a work has been updated by another DOI, such as through a correction or retraction, where such metadata are available.
Crossref also integrates Retraction Watch data into its production services. Retractions identified through publisher metadata and Retraction Watch can be surfaced through Crossref's REST API, and the Retraction Watch dataset is openly available.
This makes Crossref useful beyond disciplines covered by PubMed. It also illustrates why persistent identifiers matter: a DOI can connect a paper to metadata about changes in its publication status.
Still, metadata depend on information being supplied, identified, and linked correctly. Treat database checks as powerful safeguards, not infallible proof that no update exists.
The Retraction Watch Database Provides an Additional Check
The Retraction Watch Database was developed specifically to track retracted scholarly literature and related information. Crossref acquired the database in 2023 and subsequently integrated Retraction Watch retraction information into its REST API while continuing to make the dataset openly available.
For researchers checking an important publication, this provides another route for identifying a retraction that might otherwise be missed in ordinary literature searching.
Searching by DOI is preferable when available because article titles, author names, and journal details can vary in formatting and may produce ambiguous matches.
Google Scholar Is Useful for Discovery but Should Not Be Your Only Status Check
Researchers often encounter papers through Google Scholar, institutional repositories, author websites, reference managers, academic social networks, or ordinary web searches. These are useful discovery routes, but discovery and verification are different tasks.
A copy found outside the publisher's current record may not visibly carry subsequent editorial notices. Even if a search platform displays some status information, you should verify an important correction or retraction against the publisher's notice or another authoritative scholarly record.
The same principle applies to PDFs already stored in your reference manager. Your local file does not automatically update when the publication record changes.
Do Not Assume the PDF Will Warn You
This is one of the easiest ways to miss a retraction.
A researcher may have downloaded the original PDF before a correction or retraction occurred. Copies may remain in institutional repositories, personal websites, email attachments, shared drives, reference managers, and other archives. Those copies can continue circulating long after the publisher has updated the official record.
When a paper matters substantially to your argument, checking the DOI or current bibliographic record is safer than relying on what appears inside the PDF.
A Paper Can Be Corrected After You First Read It
Publication status is time-sensitive. A paper that had no warning when you added it to your literature review may acquire a correction, expression of concern, or retraction months or years later.
This is particularly relevant to long research projects. A doctoral dissertation, systematic review, guideline, or multi-year research program may cite papers collected considerably earlier.
You do not necessarily need to manually recheck every citation repeatedly. But papers that are central to your argument, provide key evidence, support consequential decisions, or have known integrity concerns deserve renewed verification before final submission or publication.
Checking Status Is Not the Same as Critically Appraising the Study
A paper with no correction or retraction is not thereby proven reliable. Most methodological weaknesses never produce formal editorial notices.
Likewise, peer review does not guarantee that a study will never require later correction. A paper may remain formally published while containing limitations that should substantially reduce your confidence in its conclusions.
Checking publication status should therefore complement, not replace, critical evaluation of the research itself. It is one additional question: has the formal scholarly record changed since this article was published?
What You Do Next Depends on What the Notice Says
If you find a minor correction that does not affect the claim you are using, the paper may remain entirely appropriate to cite, although you should work from the corrected record.
If the correction changes a result central to your argument, revise your interpretation accordingly.
If you find an expression of concern, determine whether your argument depends on the disputed material and consider whether independent evidence can support the same claim while the matter remains unresolved.
If the article has been retracted, the situation requires more careful treatment. Whether and how a retracted article can appropriately appear in scholarly work is a separate question, because there are legitimate reasons to cite a retracted paper when the retraction itself is the subject of discussion. The key point is that it should not unknowingly continue functioning as ordinary valid evidence.
