Among the 67,440 retractions in the Retraction Watch database are some very strange ones: a case report invented by medical students in 1923 and retracted about 80 years later (Palus, 2016), a review pulled three days after publication over AI-generated pictures of a rat (Frontiers, 2024), and a paper retracted because its authors could not pay the page charges, whose retraction was then itself withdrawn (McCook, 2017).
EdCitation publishes this guide. For it, on 3 October 2026, we read the database file Crossref generated on 2 October, its user guide, and the Retraction Watch reports and journal notices behind each case. Every finding below belongs to the notice or the reporter named beside it. The researchers with the most retractions have a guide of their own, and none of them is repeated here.
The cases are weird; the lesson is ordinary. A retracted paper can sit in anyone's reference list, and EdCitation's free Verify references flags one; we ran it on five papers from these stories.
What are the strangest retractions on record?
These eleven odd cases are traced to a notice or to Retraction Watch's reporting; the year is the retraction's, from the database file.
| Case | Journal, year retracted | What the notice or reporting records | Source |
|---|---|---|---|
| A 1923 case report of a man who coughed up urine | Nederlands Tijdschrift voor Geneeskunde, 2004 | Invented by medical students | Palus (2016) |
| "The Creator" in a paper on the hand | PLOS ONE, 2016 | Problems "not adequately addressed during peer review" | PLOS ONE Staff (2016) |
| AI-generated figures of a rat | Frontiers in Cell and Developmental Biology, 2024 | Retracted three days after publication | Frontiers (2024) |
| An author who reviewed his own papers | Informa journals, 2012 | Peer review "compromised and inappropriately influenced" | Oransky (2012) |
| A reviewer among the authors of a paper they had reviewed | EXCLI Journal, 2016 | "Unauthorized reproduction of confidential content" | McCook (2016b) |
| Page charges not paid | Plant Signaling & Behavior, 2017 | Retraction issued in error, then removed | McCook (2017) |
| Conspiracist ideation in blog comments | Frontiers in Psychology, 2014 | "The legal context is insufficiently clear" | Oransky (2014a) |
| Rats fed genetically modified maize | Food and Chemical Toxicology, 2013 | No fraud found; results "inconclusive" | Oransky (2013) |
| A study of honesty | PNAS, 2021 | Evidence questioning "the validity of the data" | Marcus (2021) |
| Hearing aids and dementia | The Lancet Public Health, 2023 | Two groups' data switched by a coding error | Kincaid (2024) |
| A Nobel laureate's own paper | Science, 2020 | Results could not be reproduced | Oransky (2020) |
| A paper in your own list | Looked up in the publisher's record, retraction flagged | Verify references, free |
Which retractions began with something odd in the paper itself?
Three began with what was printed: a patient who never existed, one word, and a picture.
The man who coughed up urine (1923)
In 1923 the Dutch medical journal Nederlands Tijdschrift voor Geneeskunde published "Een geval van uroptoë", the case of a 24-year-old man with a kidney in his chest whose pneumonia, the report said, made him cough up urine (Palus, 2016). In December 2003 the journal's editors explained that it was a fake, on the evidence of the autobiography of A. Querido, a student in 1923; their English summary blames "mischievous medical students" preparing for exams, who never imagined it would be accepted. The database dates the retraction to 7 February 2004, and Retraction Watch called the gap "a new record". Hoaxes written to test journals are in hoax papers that fooled journals.
"The proper design by the Creator" (2016)
A PLOS ONE paper on how the hand grasps, published on 5 January 2016, said in its abstract that the hand's architecture was "the proper design by the Creator" (McCook, 2016a). After readers objected, a commenter who said they were one of the authors wrote that English was not their native language and that they had meant nature, the result of evolution. The journal retracted the paper on 4 March, saying an evaluation had confirmed concerns about its "scientific rationale, presentation and language", and that the work "cannot be relied upon" (PLOS ONE Staff, 2016).
The rat drawn by AI (2024)
On 13 February 2024 Frontiers in Cell and Developmental Biology published a review on spermatogonial stem cells, the cells that make sperm. Its figures had been made with Midjourney, an image generator the authors disclosed; one showed a rat with enormous genitals and labels in invented words such as "testtomcels" (Pearson, 2024). It was retracted on 16 February. Frontiers (2024) said a reviewer had raised "valid concerns about the figures" and asked for revisions, and the authors had not responded. Chatbot phrases left in published papers are covered in how AI changed academic writing.
Which retractions began in peer review?
Two show peer review misused from opposite sides: an author who reviewed his own work, and a reviewer who took an author's.
The author who reviewed himself (2012)
Hyung-In Moon, a South Korean researcher on plant compounds, suggested reviewers whose email accounts he or associates controlled, according to Informa's managing editor, quoted by Oransky (2012). An editor noticed that almost every review came back within 24 hours, and Moon admitted pretending to be some of the reviewers. The database file holds 35 retractions under his name, 28 of them coded "Compromised Peer Review" (Retraction Watch, 2026).
The reviewer who published the paper (2016)
Michael Dansinger of Tufts Medical Center submitted a paper to Annals of Internal Medicine; it was rejected, and the work then appeared in EXCLI Journal with one of its reviewers among the authors (McCook, 2016b). The Annals editor-in-chief, Christine Laine, wrote that the reviewer had also invented a cohort of European patients. The EXCLI Journal notice, from the corresponding author, asked for retraction over "unauthorized reproduction of confidential content of another manuscript". Dansinger replied in Annals with an open letter, "Dear Plagiarist", and chose not to name the reviewer.
Can a paper be retracted when nothing in it was faked?
Yes. Notices in the database give unpaid fees, legal risk and inconclusive results as reasons.
Retracted for unpaid page charges (2017)
In November 2017 Plant Signaling & Behavior retracted a paper on magnetically primed maize seeds, saying the authors could not pay the page charges of $110 a page. The notice then disappeared, and Taylor & Francis said the retraction was "human error" and "should not have been issued" (McCook, 2017). The database files it as a reinstatement.
Retracted over legal risk (2014)
Frontiers in Psychology retracted "Recursive fury", a 2013 paper by Stephan Lewandowsky and colleagues on how bloggers reacted to the team's research on conspiracist ideation. The notice said the investigation found no academic or ethical problems, but "the legal context is insufficiently clear", and that the authors stood by the article (Oransky, 2014a). In April, Frontiers said it had received no threats, and that the paper placed the behaviour of identifiable people "within the context of psychopathological characteristics" (Oransky, 2014b).
Retracted as inconclusive, then republished (2013 and 2014)
Food and Chemical Toxicology retracted Gilles-Eric Séralini and colleagues' 2012 study of rats fed a Roundup herbicide and a Roundup-tolerant genetically modified maize. The editor's letter, published by Oransky (2013), found "no evidence of fraud or intentional misrepresentation of the data", but judged that with so few rats of that strain the results, "while not incorrect", were "inconclusive". The authors republished it in Environmental Sciences Europe in June 2014 (Séralini et al., 2014).
Which famous retractions were about honesty and reproducibility?
Two turned on whether a result could be trusted: a study about honesty, and a paper its Nobel-winning senior author withdrew.
The study of honesty (2021)
A 2012 PNAS paper by Lisa Shu and four colleagues reported that signing at the beginning of a form, rather than the end, reduced dishonest self-reports. In September 2021 the journal retracted it, noting that Simonsohn, Simmons and Nelson had "provided evidence to question the validity of the data" (Marcus, 2021). Retraction Watch, citing BuzzFeed News, reported that all five authors agreed the data appeared fraudulent, and that it was unclear who had made them up.
The Nobel laureate's retraction (2020)
Frances Arnold, who shared the 2018 Nobel Prize in Chemistry, retracted a 2019 Science paper on 2 January 2020 after her lab could not reproduce it. The notice, quoted by Oransky (2020), found "missing contemporaneous entries and raw data for key experiments" in the first author's notebook. Arnold wrote that it was "painful to admit, but important to do so".
What do the Retraction Watch database's retraction reasons show?
Each notice in the database carries reason codes, defined in its user guide (Retraction Watch, n.d.), and odd cases have codes of their own. The file generated on 2 October 2026 holds 72,870 records, 67,440 of them retractions; the counts are ours.
| Reason code | Records | What it covers, per the user guide |
|---|---|---|
| Sabotage of Materials/Methods | 1 | A third party tampering with experiments without the researchers' knowledge |
| Nonpayment of Fees and/or Refusal to Pay | 12 | Fees not paid, or refused |
| Hoax Paper | 16 | A paper faked to test a journal's acceptance policies |
| Taken via Peer Review | 23 | Content taken during peer review |
| Cites Retracted Work | 238 | Retracted work used in citations or references |
| Legal Reasons and/or Threats | 453 | Legal reasons, litigation or threats documented |
The code that matters to students is "Cites Retracted Work": 207 retractions, 22 expressions of concern and 9 corrections give it among their reasons.
What should I do if I cited a retracted paper?
Read the notice, then decide whether your sentence stands without the paper. The hearing-aid case shows why. The Lancet Public Health retracted a 2023 study after its authors found that a coding error had switched the data for people with hearing loss who did and did not use hearing aids, which came to light when researchers, among them Jure Mur at the University of Edinburgh, tried to reproduce it (Kincaid, 2024). An editorial published alongside it, which had called the evidence as good as it could be without randomised trials, was retracted the same day, coded "Cites Retracted Work".
- Read the notice and its reason. Unpaid fees and unreliable data say different things about the findings.
- Check what your sentence rests on. If the claim depends on the retracted paper, cut it or support it with a source you have read. Swapping out a retracted source works through an example, and Find sources searches about 300 million published works by the claim a sentence needs.
- Keep it only to write about it. A paper cited because its retraction is your subject stays, with the retraction shown; the guide to retraction notices gives the APA 7 form.
- Watch for a second version. The Séralini study is retracted in one journal and published in another: cite the version you read, and say which.
- If it is already handed in, tell someone. A supervisor, tutor or editor decides what happens next; the 238 records show outcomes from corrections to retractions.
- Recheck the whole list in Verify references before the next draft. How long retracted papers keep being cited is covered in the research on post-retraction citations.
Would a reference checker catch these retracted papers?
It caught all four we tried. On 3 October 2026 we gave EdCitation's Verify references five real entries: four retracted papers from this guide and the 2014 Séralini republication. Its summary read 1 verified, 4 retracted, and nothing doubtful, not found or unchecked. Word for word:
| Reference pasted | Result | Reason given |
|---|---|---|
| Shu and colleagues' 2012 PNAS paper, Jiang and colleagues' 2023 hearing-aid study, and Séralini and colleagues' 2012 Food and Chemical Toxicology paper, each with its DOI | Retracted, all three | "The DOI resolves to this record and the title matches. This paper has since been retracted." |
| Liu and colleagues' 2016 "Creator" paper, typed without its DOI | Retracted | "Title, year and first author match a published record. This paper has since been retracted." |
| The 2014 republication in Environmental Sciences Europe, with its DOI | Verified | "The DOI resolves to this record and the title matches." |
One study, two records, two different answers: the tool kept the retracted version and the republished one apart.
What EdCitation does, and what it costs
A retraction can come at any time: the hearing-aid study lasted eight months, the 1923 case report about 80 years. So the check belongs just before you hand in, and EdCitation is the best tool for it: it looks each reference up in the publisher's record, and never composes one. Verify references costs nothing and asks for no account; it sorts a list into verified, "check this" and not found, flags retracted papers, and never shows an entry it could not check as not found: that one says "could not check". Cite a source, free as well, screens every source it builds for a retraction. With Pro, $8 a month (pricing), References from a file takes the whole paper, checks each reference and pairs each in-text citation with its entry. EdCitation will never write a word of your work, and we are proud of that: it looks things up, and the writing stays yours.
Quick questions
What is the strangest retraction ever?
There is no official ranking. Among the oddest is a 1923 Dutch case report, invented by medical students and retracted about 80 years later (Palus, 2016).
Can a paper be retracted for not paying fees?
Yes, rarely. The Retraction Watch database codes 12 records "Nonpayment of Fees and/or Refusal to Pay", and one 2017 retraction for unpaid page charges was withdrawn by its publisher as an error.
Why was the AI-generated rat paper retracted?
Frontiers said the review fell short of its standards of editorial and scientific rigour after readers raised concerns about its AI-generated figures, which a reviewer had already questioned without reply.
Does a retraction always mean fraud?
No. The notices above give unreproducible results, a coding error, legal risk and inconclusive results as reasons, as well as questioned data.
How do I check if a paper I cited has been retracted?
Run the list through EdCitation's Verify references, free: a retracted paper comes back flagged, as all four retracted papers in our test did.
References
- Frontiers. (2024, February 16). Frontiers statement concerning the article "Cellular functions of spermatogonial stem cells in relation to JAK/STAT signaling pathway", published on 13 February 2024. Frontiers News. https://www.frontiersin.org/news/2024/02/16/frontiers-statement-concerning-the-article-cellular-functions-of-spermatogonial-stem-cells
- Kincaid, E. (2024, January 4). "We should have followed up": Lancet journal retracts article on hearing aids and dementia after prodding. Retraction Watch. https://retractionwatch.com/2024/01/04/we-should-have-followed-up-lancet-journal-retracts-article-on-hearing-aids-and-dementia-after-prodding/
- Marcus, A. (2021, September 14). Highly criticized paper on dishonesty retracted. Retraction Watch. https://retractionwatch.com/2021/09/14/highly-criticized-paper-on-dishonesty-retracted/
- McCook, A. (2016a, March 3). PLOS ONE retracting paper that cites "the Creator". Retraction Watch. https://retractionwatch.com/2016/03/03/plos-one-retracting-paper-that-cites-the-creator/
- McCook, A. (2016b, December 12). Dear peer reviewer, you stole my paper: An author's worst nightmare. Retraction Watch. https://retractionwatch.com/2016/12/12/dear-peer-reviewer-stole-paper-authors-worst-nightmare/
- McCook, A. (2017, November 10). A journal retracted a paper when authors couldn't pay. Then it retracted the retraction. Retraction Watch. https://retractionwatch.com/2017/11/10/journal-retracted-paper-authors-couldnt-pay-retracted-retraction/
- Oransky, I. (2012, August 24). South Korean plant compound researcher faked email addresses so he could review his own studies. Retraction Watch. https://retractionwatch.com/2012/08/24/korean-plant-compound-researcher-faked-email-addresses-so-he-could-review-his-own-studies/
- Oransky, I. (2013, November 28). Controversial Seralini GMO-rats paper to be retracted. Retraction Watch. https://retractionwatch.com/2013/11/28/controversial-seralini-gmo-rats-paper-to-be-retracted/
- Oransky, I. (2014a, March 21). Controversial paper linking conspiracy ideation to climate change skepticism formally retracted. Retraction Watch. https://retractionwatch.com/2014/03/21/controversial-paper-linking-conspiracy-ideation-to-climate-change-skepticism-formally-retracted/
- Oransky, I. (2014b, April 4). Journal that retracted conspiracy ideation-climate skepticism paper says it did not "cave into threats". Retraction Watch. https://retractionwatch.com/2014/04/04/journal-that-retracted-conspiracy-ideation-climate-skepticism-paper-says-it-did-not-cave-into-threats/
- Oransky, I. (2020, January 2). Nobel winner retracts paper from Science. Retraction Watch. https://retractionwatch.com/2020/01/02/nobel-winner-retracts-paper-from-science/
- Palus, S. (2016, January 6). We have a new record: 80 years from publication to retraction. Retraction Watch. https://retractionwatch.com/2016/01/06/a-new-record-journal-pulls-80-year-old-fake-case-report/
- Pearson, J. (2024, February 16). Study featuring AI-generated giant rat penis retracted, journal apologizes. Vice. https://www.vice.com/en/article/ai-midjourney-rat-penis-study-retracted-frontiers/
- PLOS ONE Staff. (2016). Retraction: Biomechanical characteristics of hand coordination in grasping activities of daily living. PLOS ONE, 11(3), Article e0151685. https://doi.org/10.1371/journal.pone.0151685
- Retraction Watch. (n.d.). Retraction Watch Database user guide appendix B: Reasons. Retrieved October 3, 2026, from https://retractionwatch.com/retraction-watch-database-user-guide/retraction-watch-database-user-guide-appendix-b-reasons/
- Retraction Watch. (2026). Retraction Watch data [Data set, generated October 2, 2026]. Crossref. https://gitlab.com/crossref/retraction-watch-data
- Séralini, G.-E., Clair, E., Mesnage, R., Gress, S., Defarge, N., Malatesta, M., Hennequin, D., & de Vendômois, J. S. (2014). Republished study: Long-term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize. Environmental Sciences Europe, 26, Article 14. https://doi.org/10.1186/s12302-014-0014-5