A retraction is a formal notice that a published scientific paper should no longer be considered reliable. It is the primary mechanism journals use to correct the scientific record when a study’s findings, data, or conclusions turn out to be seriously flawed. Retractions can result from deliberate fraud, honest mistakes, plagiarism, or ethical violations. Their purpose is to protect the integrity of published research, not to punish authors.
Why Papers Get Retracted
The most common reason for retraction is errors in data or data analysis, which account for roughly 38% of all retractions in medical journals. These aren’t always cases of intentional deception. A miscalculation, a coding error in statistical software, or a mix-up in datasets can produce results that fall apart under scrutiny. When the errors are serious enough that the paper’s conclusions no longer hold up, a retraction is the appropriate response.
Plagiarism accounts for about 12% of retractions, and duplicate publication (submitting the same work to multiple journals) accounts for another 11%. Ethical concerns, such as conducting research without proper consent from participants, make up around 7%, with methodological flaws close behind at 7%. The remaining cases involve data fabrication, falsification, or a combination of issues.
The Committee on Publication Ethics (COPE), the body that sets international standards for scholarly publishing, outlines four broad situations where retraction is warranted: clear evidence that findings are unreliable (whether from misconduct or honest error), redundant publication without proper disclosure, plagiarism, and unethical research practices.
How the Retraction Process Works
Retractions don’t happen overnight. Most begin when someone raises concerns about a published paper. In about 73% of cases, the journal’s editor or editor-in-chief initiates the process. Authors themselves request retraction only about 9% of the time. The remaining cases are flagged by readers, peer reviewers, or institutional investigations.
Once concerns are raised, journals typically follow a graduated response. For minor issues that don’t undermine a paper’s conclusions, a correction (sometimes called an erratum) may be published. If the problems are more serious but still under investigation, the journal may issue an “editorial expression of concern,” a formal notice alerting readers that the paper’s reliability is in question. If the investigation concludes that the paper’s integrity is substantially compromised, a full retraction follows.
A retracted paper isn’t deleted. It remains in the journal’s archive, but a retraction notice is attached explaining why it was withdrawn, and the PDF is typically watermarked with the word “RETRACTED” across every page. This approach preserves the historical record while clearly signaling to anyone who encounters the paper that its findings are not trustworthy.
Not all authors accept these decisions gracefully. Of cases where editors raised concerns, authors agreed with the retraction decision only about 40% of the time. Roughly 7% actively disagreed.
Retractions Are Growing Fast
The number of retracted papers has surged in recent decades. In 2000, about 140 papers were retracted worldwide. By 2022, that figure exceeded 11,000. The compound annual growth rate for retractions stands at 22%, far outpacing the 6.25% growth rate for total published papers over the same period. An analysis of medical retractions alone identified over 16,000 retracted publications between 1975 and 2024.
This increase doesn’t necessarily mean scientists are becoming less honest. Better detection tools, including plagiarism software and image forensics, have made it easier to catch problems. Organizations like Retraction Watch, which maintains a public database of retracted papers, have also increased transparency and pressure on journals to act.
The Problem of “Zombie” Citations
One of the most troubling aspects of retractions is that retracted papers don’t stop influencing science. Studies show that the vast majority of citations to retracted articles are positive, meaning researchers cite the discredited findings as though they are still valid. This happens even when the retraction notice is clearly displayed on the publisher’s platform, and it happens regardless of whether the paper was retracted for honest error or deliberate fraud.
A striking example emerged during the COVID-19 pandemic. Two high-profile papers published in The Lancet and The New England Journal of Medicine in May 2020 relied on a database that turned out to be unverifiable. Both were retracted within weeks. The Lancet paper had claimed that hydroxychloroquine could harm COVID-19 patients, and its publication temporarily halted a major clinical trial. Yet of 200 papers examined after the retractions, more than half inappropriately cited one of the two discredited studies. In several cases, the retracted data served as a primary source for meta-analyses that combined multiple studies to draw broader conclusions.
This pattern has real consequences for public health. A retracted 1998 study in The Lancet that falsely linked vaccines to autism continues to be cited decades later. When retracted work with clinical implications is treated as valid evidence, it can directly influence treatment decisions and patient safety.
What Happens to Authors
For researchers found guilty of misconduct, the professional consequences are severe. A study examining scientists censured by the U.S. Office of Research Integrity (ORI) found that their publication output dropped by a median of 91.8%. More than half (55%) stopped publishing entirely within three years of being censured. Funding tells a similar story: NIH grants awarded to these researchers declined by 70.5% after the misconduct finding.
Formal censure by the ORI typically involves a voluntary agreement barring the scientist from receiving government research contracts for a set period, ranging from a few years to, in rare cases, a lifetime. The practical effect is often the end of a research career. Scientists who fabricated or falsified data experienced the steepest declines in productivity.
The consequences are less clear-cut for retractions caused by honest error. These cases don’t typically trigger misconduct investigations, and authors can often continue their careers. But a retraction on one’s record still carries a stigma in academia, which is one reason some researchers resist the process even when errors are clear.
Why Retractions Matter for Everyone
Retractions exist because science is self-correcting, but the correction only works if people pay attention. Medical guidelines, government policies, and drug approvals all rest on published research. When a flawed study goes uncorrected, or when a retracted study keeps getting cited as valid, the consequences ripple outward. Clinical trials get designed around faulty premises. Public health messaging gets distorted. Patients make decisions based on evidence that doesn’t hold up.
If you encounter a scientific paper online, checking whether it has been retracted takes seconds. Publisher websites display retraction notices prominently, and the Retraction Watch database is searchable and free. Given that thousands of papers are retracted every year, and many continue to circulate as though nothing happened, that quick check is worth the effort.