Most researchers have heard about ‘retracted articles’ or ‘article retraction’, i.e., a formal removal of a published study from a journal because of major error or major misconduct. If you’ve wondered about why journals can first publish a study and then remove it (among other things that journals do as corrective measures), this newsletter is for you.
An article in The Hindu by R. Prasad
Why do retractions happen?
Most manuscripts go through several checks, at the level of journal administrator, editor, and peer reviewers, before it is considered for publication. Although at every stage thorough checking is done, feedback is given, and rejection of the manuscript can happen if there are fundamental errors, it is still a possibility that some error can remain unchecked. This error could be later caught by any reader of the paper, by the editor of the paper, or by the original authors themselves. In such a case, the concerned journal editor is notified and, if the error is significant enough, a retraction is issued to alert the readers of the journal. This error is usually something big- not a minor spelling mistake or missing reference. It is usually big enough that the main results and conclusion of the paper cannot be trusted.
Retraction notice for paper titled “On the same origin of quantum physics and general relativity from Riemannian geometry and Planck scale formalism”
While it can sometimes be a genuine methodological mistake, in other cases, the authors may have fabricated data or images using AI or copied images from other papers. This is a serious breach of trust in the scientific community and more and more reliable software are coming up to scan for this category of misconduct.
Many publishing houses have made their policies on the use of AI abundantly clear and are willing to retract papers if they discover that AI has been used to create or manipulate data or images.
What does a retraction mean for the authors?
A retraction does not necessarily prove that an author committed fraud. It can reflect an honest error, unreliable data, a publishing problem, etc. However a retraction is more serious than ordinary peer-review criticism.
It can damage authors’ image because other researchers, journals, funders, and potential collaborators may become more cautious about trusting their work. If the authors themselves openly reported the issue and helped correct the record, this can be seen as responsible scientific behavior rather than dishonesty. But if the retraction happened because of others reporting plagiarism, fabrication, or ethical violation, then it can seriously harm the author's credibility.
Retraction notice for article “Epidemiology of Sports Injuries Among High School Athletes in the United States: Data From 2015 to 2019”
Yes! Journals do their best to give the authors a chance to disprove the allegations against their paper. Editors send out emails highlighting the error or the allegation to the authors and seek justifications or corrections. The authors can respond to the editor by explaining why they disagree with the retraction and they can provide supporting materials like raw datasets, ethical approval, any correspondence, or analysis files. Usually journals give authors enough time to put their arguments together and respond before the journal issues any retraction notice.
However, what doesn’t work during appeal is stating that you, as one of the authors, were unaware that the other authors of the paper had committed fraud or manipulated/fabricated data.
We asked a Research Ethics Expert about how journals respond when one author says that she is unaware of data fraud committed by another author in the same paper. The expert said that journals make the duties of authors very clear in their research integrity and ethics policies. Authors are collectively responsible for the work they submit. One author cannot wash their hands off from the retraction by saying that they were only responsible for a part of the analysis or they were a offered co-authorship because they only contributed to the initial stages and brainstorming. That notice of retraction will appear on all authors' Google Scolar, Scopus, and Web of Science profiles. Journals hold all authors equally accountable, even if only one of the authors has committed fraud. Hence, it’s important to choose your research co-authors wisely. Don't collaborate with researchers that you don't know and whose work you cannot verify. Even if you're not a part of the data collection, cross-check the raw data yourself for possibilities of error. If someone is inviting you during revision rounds, make sure to preserve your records of collaboration. Don't delete emails or shared data folders after publication.
What are other corrective measures that journals take when they find a problem with an article?
When a problem is limited and the article’s main findings are still reliable, a journal will usually publish a correction rather than retract it. An erratum corrects a mistake introduced by the journal during editing or production, while a corrigendum corrects an author’s mistake, such as an inaccurate reference, a reporting error, or a calculation that does not change the main conclusion. The correction is linked to the original article so readers can see exactly what was changed. If the concern is serious but the evidence is not strong or conclusive, the editor can issue an expression of concern to warn readers. In rare situations, journals can replace an article with a clearly labelled corrected version. Article removal is generally a last resort; normally, journals preserve the original record and attach a clear notice instead.
This corrigendum was issued because “The lead author, LP, discovered that … in converting the raw data (downloaded from Qualtrics) to SPSS and prior to forming the variables, some participants in the calcium supplementation group had two survey items mis-labeled… This affected the final variables used in some of the analyses in the original article.”
Exploring AI Voice With SuperBloom
For Deel's "Feeling of Deeling" campaign, agency SuperBloom needed one consistent brand voice across markets—deployed fast, without sacrificing quality or consent. In this on-demand video session, SuperBloom and Voices break down the casting, production, and governance, plus what they'd do differently and where AI voice is headed next.
Stop making AI decisions in the dark.
Leadership is asking: where is AI delivering value for us and where is it creating risk? Right now, most teams have no idea.
With Harmonic Security’s Usage Explorer, you get a complete picture of how your organization actually uses AI, automatically categorized into custom use cases with complete tool-level granularity.
2 minutes. Your URL. A customer profile worth using.
Most founders can describe their product. They can't describe their customer. Not in a way that actually changes how they sell.
HubSpot for Startups built a free tool to fix that. Paste in your URL, answer a few quick questions, and it generates a structured profile of your best-fit customer. Firmographics, buying triggers, the works.
Takes 2 minutes. No spreadsheet required.