In academic publishing, a retraction is a mechanism by which a published paper in an academic journal is flagged for being seriously flawed to the extent that their results and conclusions can no longer be relied upon. Retracted articles are not removed from the published literature but marked as retracted. In some cases it may be necessary to remove an article from publication, such as when the article is clearly defamatory, violates personal privacy, is the subject of a court order, or might pose a serious health risk to the general public.[1]
Although the majority of retractions are linked to scientific misconduct,[2] they are often cited as evidence of the self-correcting nature of science.[3] However, some scholars argue this view is misleading, describing it as a myth.[4]
A retraction may be initiated by the editors of a journal, or by the author(s) of the papers (or their institution). Retractions are typically accompanied by a retraction notice written by the editors or authors explaining the reason for the retraction. Such notices may also include a note from the authors with apologies for the previous error and/or expressions of gratitude to persons who disclosed the error to the author.[5] Retractions must not be confused with small corrections in published articles.
There have been numerous examples of retracted scientific publications. Retraction Watch provides updates on new retractions, and discusses general issues in relation to retractions.[6][7]
An early example of a retraction in a scholarly, peer-reviewed publication can be traced to "A Retraction, by Mr. Benjamin Wilson, F.R.S. of his former Opinion, concerning the Explication of the Leyden Experiment," published in the Philosophical Transactions of the Royal Society on 24 June 1756. This is recognized as the earliest recorded retraction in scientific publishing.[8] In it, Benjamin Wilson, a British painter and scientist, formally withdrew his previous explanation of the Leyden jar experiment, a foundational study in the field of electricity. He acknowledged that subsequent discoveries, particularly those by Benjamin Franklin, had shown his original interpretation to be incorrect.[9]
A 2011 paper in the Journal of Medical Ethics attempted to quantify retraction rates in PubMed over time to determine if the rate was increasing, even while taking into account the increased number of overall publications occurring each year.[10] The author found that the rate of increase in retractions was greater than the rate of increase in publications. Moreover, the author notes the following:
"It is particularly striking that the number of papers retracted for fraud increased more than sevenfold in the 6 years between 2004 and 2009. During the same period, the number of papers retracted for a scientific mistake did not even double..." (p. 251).[10]
Although the author suggests that his findings may indeed indicate a recent increase in scientific fraud, he also acknowledges other possibilities. For example, increased rates of fraud in recent years may simply indicate that journals are doing a better job of policing the scientific literature than they have in the past. Furthermore, because retractions occur for a very small percentage of overall publications (fewer than 1 in 1,000 articles[11][12]), a few scientists who are willing to commit large amounts of fraud can highly impact retraction rates. For example, the author points out that Jan Hendrik Schön fabricated results in 15 retracted papers in the dataset he reviewed, all of which were retracted in 2002 and 2003, "so he alone was responsible for 56% of papers retracted for fraud in 2002—2003" (p 252).[10]
During the COVID-19 pandemic, academia had seen a quick increase in fast-track peer-review articles dealing with SARS-CoV-2 problems.[13] As a result, a number of papers have been retracted made "Retraction Tsunami"[14] due to quality and/or data issues, leading many experts to ponder not just the quality of peer review but also standards of retraction practices.[15]
Retracted studies may continue to be cited. This may happen in cases where scholars are unaware of the retraction, in particular when the retraction occurs long after the original publication.[16]
The number of journal articles being retracted had risen from about 1,600 in 2013 to 10,000 in 2023. Most of the retractions in 2023 were contributed by Hindawi journals.[17] The significant number of retractions involving Chinese co-authors—over 17,000 since 2021, including 8,000 from Hindawi journals—has led China to launch a nationwide audit addressing retractions and research misconduct.[18] Retractions are also measured among highly cited researchers.[19]
A low percentage of retracted papers can be due to unintentional error within the author(s) work. Rather than removing the entire article, retraction with replacement has been a new practice to help authors avoid being seen as dishonest for mistakes that were not purposefully done.[20] This method allows the author to fix their mistakes from the original paper, and submit an edited version to take the original paper’s place. The journal can decide to retract the original paper then upload the fixed version online, usually with a notice placed stating “Retraction and Replacement,” or “Correction,” on the article page. For example, JAMA will post the edited version with a retraction and replacement notice, along with a link to the original article, while Research Evaluation will use the term "correction" with a link posted on the updated article, referring to the old article.
Self-retraction is a request from an author and/or co-authors to retract its own work from being published. Self-retraction by an author is recommended because once it gets retracted from the journal, then it can affect the author(s) because investigations can begin which will have an effect the author's reputation. If one retracts their own work on their terms, it would show more integrity and honesty as they are owning up to their own mistakes,[21] just like the authors mentioned in The Wall Street Journal have done. Scientists at times have been asked to retract their work even though their work is exact and bold; the root cause of the problem should be looked into to avoid retractions.[21] A system to distinguish papers from "good" and "bad" would be beneficial to researchers. This system may save the reputation of scientists and researchers. Most researchers publish honest work and sometimes simple mistakes happen to be overlooked by the peer review process. Retraction should not be for simple spelling errors, but for inaccurate, skewed, and fraudulent data. For example, today new technologies are being developed in a culture of transparency to align the opportunity to record false claims.[21] Another solution is for researchers to use a term “self-citation” since citations look identical therefore they are classified in databases.[21] Recommending a same database to evaluate the researchers own work can help lessen retractions.