What a retracted paper means for your reference list

A paper can be retracted after you read it, after you cite it, and after you hand in your draft. The article usually stays online, and nothing tells you when the status changes. The only way to know is to look.

This post explains what a retraction is, why retracted papers keep turning up in reference lists, how to check your own sources, and what to write when you need to cite one.

What a retraction is

A retraction is a journal’s formal statement that a published paper should not be relied on. The Committee on Publication Ethics (COPE), which sets guidelines that many journals follow, describes the purpose as correcting the literature and alerting readers to seriously flawed content. It says the purpose is not to punish the author.

COPE updated its retraction guidelines in September 2025. The 2019 version named reasons such as major error, plagiarism and redundant publication. The update adds image irregularities, fictitious data, undisclosed use of artificial intelligence, and systematic manipulation of the publication process, such as paper mills. It also describes a “retraction with replacement” for articles whose serious errors change the findings but do not invalidate the underlying work, and a “retraction with removal” that it reserves for extremely rare cases, such as legal or public health risks.

Two things follow for a student:

• A retraction does not always mean fraud. Honest mistakes lead to retractions as well. The notice normally states the reason, and the reason is worth reading.
• The paper is usually still there. Removal is the exception. In most cases the article stays online with a notice attached, so a link to it keeps working and the text still looks like any other paper.

Retractions, corrections and expressions of concern

Journals have more than one way to flag a problem, and the labels matter when you read a notice.

Crossref’s data on this lists four kinds of update: retractions, corrections, expressions of concern and reinstatements. A correction fixes an error in a paper that otherwise stands. An expression of concern is a notice that raises doubts about a paper without retracting it. A reinstatement means a paper has been restored after an earlier notice. Only a retraction tells you the paper should not be relied on. Treat an expression of concern as a reason to look closer before you cite the paper, and treat a correction as a reason to cite the corrected version.

Why retracted papers end up in reference lists

A retracted paper is easy to cite by accident. It still appears in search results and library databases. It is still in the reference lists of papers published before the retraction. A reference copied from another paper’s bibliography carries no warning with it, and a PDF you saved last spring does not update itself.

Studies of citation patterns show this happening. A 2020 study in BMJ Global Health followed 304 retracted articles indexed in MEDLINE. It found that the papers kept being cited after retraction, and the authors found no long-term link between a retraction and the number of citations a paper received. The authors recommended stronger safeguards against citing retracted papers, including reference-manager tools that flag them.

None of this is a reason to assume your own sources are affected. Most are not. It is a reason to check, because the check is quick and the paper will not announce itself.

How to check your sources

Check each reference for a retraction notice. Do it again shortly before you submit, since a paper can be retracted at any point.

1. Open the article’s page on the publisher’s site. A retracted article normally carries a notice on that page, often near the title.
2. Look the paper up in PubMed if it is biomedical. PubMed marks a retracted article with the publication type “Retracted Publication,” and its citation includes a “Retraction in:” line that links to the notice. The notice itself is marked “Retraction Notice” and begins with “Retraction of:”. PubMed does not separate retractions for honest error from retractions for misconduct, so read the notice to see the reason.
3. Search the Retraction Watch database. Crossref acquired the database from the Center for Scientific Integrity in September 2023 and makes it publicly available. It is updated every working day. It collects retractions from publisher websites, database searches, web searches and reports from the community.
4. Run the list through a checker. For a long reference list, doing steps 1 to 3 by hand takes a while. PaperCitation matches each reference to its published record and shows a warning on top of the verdict when the record has been retracted or carries an expression of concern. It reads this from retraction information published through Crossref, which includes Retraction Watch data.

No tool sees every retraction. The Retraction Watch database is mainly a collection of retractions, and Crossref describes the other notice types in it as less comprehensive. A new notice takes time to reach the databases. If a checker shows no warning, the paper is probably fine, but it is not proof, and step 1 is the check that does not depend on any database.

If one of your sources was retracted

What to do depends on how you used the paper.

You used it as evidence for a claim. Remove it. A retracted paper cannot support a claim, so the claim needs another source or has to go. If the retracted paper is the only support for a point, that point probably does not belong in your paper.

A replacement exists. In a “retraction with replacement,” the journal publishes a corrected version. If you find one, read it, because the findings may have changed. Cite the version that stands.

You are writing about the retraction itself. Citing a retracted paper is acceptable when the paper is your subject, for example in an essay on research misconduct or on how a field responded to a withdrawn study. APA has two formats for this.

To cite the retracted article itself, which is the format for discussing its contents, give the publication details of the original article, then add the retraction information in parentheses at the end, with its year, journal, volume, issue and page:

Author, A. A., & Author, B. B. (Year). Title of the article. Journal Name, volume(issue), pages. https://doi.org/xxxxx (Retraction published year, Journal Name, volume[issue], page)

To cite the retraction notice, which is the format for explaining why a paper was retracted, treat the notice as its own article. Its title is reproduced as shown and usually includes the word “retraction.” The author may be an editor, the editorial board, or some or all of the original authors, so check the notice to see who it credits. APA’s page on journal article references shows both formats with real examples.

In either case, say in your text that the paper was retracted. A reader who follows the reference should not have to find that out for themselves.

What a clean result does and does not tell you

A source that has not been retracted is still not automatically a good source. A retraction check answers one question: has the journal formally withdrawn the paper? It does not say whether the study was well designed, whether you read it correctly, or whether it supports your sentence. The post on why AI tools invent citations covers a different failure, references that do not exist at all, and the DOI guide covers confirming that a reference points to the paper you meant. A reference list that survives all three checks has a much better chance of holding up.