Contents
On this page
- Findings of a landmark analysis of retracted papers
- The most frequently recurring categories
- Trends in retraction frequency and their causes
- The contribution of the Retraction Watch database beyond journal notices
- The relationship between a formal finding of misconduct and a retraction
- Practical use of data on reasons for retraction
Academy · Research Publication · Retraction Causes & Trends
Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error
A widely cited analysis in PNAS found that misconduct (fabrication, falsification and plagiarism), rather than honest error, accounted for the majority of the retractions it examined, overturning the earlier assumption that most retracted papers resulted from inadvertent mistakes. This article examines those findings and their practical implications.
A retraction is a formal notice, issued by a journal, that withdraws a previously published paper from the citable scientific record because of a serious problem with its validity, integrity or ethical conduct. The Retraction Watch database, maintained since 2010, has become the principal public resource for tracking the reasons for retraction across disciplines, and its records, together with peer-reviewed analyses of retraction data, have overturned a previously common assumption that most retractions result from honest error rather than misconduct.
Learning objectives
- What a landmark analysis of retracted biomedical and life-science papers found about the balance between error and misconduct
- The specific categories of reportable misconduct and error that recur most frequently in retraction data
- How data on reasons for retraction can be used in evaluating a paper or planning research integrity practices
Common misconceptions
- “Most retracted papers were withdrawn because of an honest mistake identified by the authors themselves.” Fang, Steen and Casadevall’s analysis of 2,047 retracted biomedical and life-science articles found that misconduct (fraud or suspected fraud, duplicate publication and plagiarism) accounted for a majority of retractions with an identifiable cause, whereas error accounted for a minority, directly contradicting the assumption that most retractions reflect inadvertent mistakes.
Findings of a landmark analysis of retracted papers
The 2012 analysis by Fang, Steen and Casadevall in the Proceedings of the National Academy of Sciences (PNAS) examined 2,047 retracted biomedical and life-science articles indexed in PubMed, going beyond the often brief or vague wording of the retraction notices to classify the actual underlying reason with the aid of supplementary documentation where available1. Its central finding was that fraud or suspected fraud accounted for approximately 43% of retractions, duplicate publication for approximately 14% and plagiarism for approximately 10%, so that misconduct was responsible for approximately 67% of retractions with a clearly identifiable cause, whereas error accounted for only approximately 21%1. This finding substantially revised earlier, more limited analyses that had relied solely on the stated text of retraction notices, which the authors found often understated or omitted the true reason, probably reflecting institutional and legal caution in the wording of notices rather than the actual cause. This discrepancy between the stated wording of a notice and its documented underlying cause is itself a reason to treat vague retraction language as a prompt for further inquiry rather than its conclusion, a recommendation reinforced by the retraction guidelines of COPE (the Committee on Publication Ethics), which specify that a retraction notice should state its reason clearly rather than leave readers to infer it. Journals that apply this standard consistently make it measurably easier for subsequent researchers, and for resources such as the Retraction Watch database, to classify and track the reasons for which a given paper was withdrawn from the citable record.
The most frequently recurring categories
Beyond the broad distinction between error and misconduct, research on retractions identifies specific recurring categories that warrant distinction. Steen's separate 2011 analysis of 788 retracted papers in the Journal of Medical Ethics found that papers retracted for fraud took longer on average to be retracted after publication than papers retracted for error, and that fraudulent papers had been cited, and had therefore influenced subsequent research, more often before retraction than papers retracted for error2. Duplicate publication, that is, submission of substantially the same study to more than one journal without disclosure, is treated as a category distinct from both fraud and plagiarism because it does not necessarily involve fabricated data or copied text; it nevertheless distorts the literature by allowing the same findings to be counted twice in systematic reviews and meta-analyses that treat each publication as independent evidence. Image manipulation, the alteration or duplication of figures such as western blots or microscopy images, has become an increasingly recognized category in more recent analyses of retractions as dedicated image-screening tools have been more widely adopted by journals during peer review3.
Trends in retraction frequency and their causes
The absolute annual number of retractions has risen substantially since the early 2000s, but research investigating this trend finds that the increase is not primarily explained by a rise in actual misconduct. The analysis by Steen, Casadevall and Fang of the reasons for the increase found that it closely tracked the concurrent growth in the total number of published papers and the adoption by journals and the PubMed indexing system of more consistent and visible retraction-flagging practices over the same period, rather than reflecting a proportional increase in the underlying rate of fraud or error in the literature5. This distinction is important for the interpretation of the rising number of retractions: an increase may be consistent with a self-correcting system detecting a stable or even declining underlying rate of problems more effectively, rather than with a deterioration in research integrity, and the authors caution against treating the raw annual retraction count in isolation as a proxy for the overall trustworthiness of the literature.
The contribution of the Retraction Watch database beyond journal notices
Journal retraction notices are often brief and, as Fang, Steen and Casadevall documented, frequently vague about the underlying cause, which is precisely the gap that the Retraction Watch database was established to address. A detailed investigation in Science by Brainard and You, drawing on the records of the database, found that a comparatively small number of individual researchers with multiple retractions accounted for a disproportionate share of all retracted papers in the dataset, and that specific institutions and countries showed meaningfully different retraction rates and reasons when the more detailed, independently researched categorization of the database was applied instead of the journal notices alone6. Such patterns, which become visible only when the reasons for retraction are systematically tracked and categorized rather than read notice by notice, constitute the practical value that the database offers to researchers, journalists and institutions seeking to understand where integrity problems are concentrated rather than treating each retraction as an isolated event. The comprehensive survey of retracted articles by Grieneisen and Zhang similarly found that retractions are disproportionately concentrated among a comparatively small number of repeat authors and specific journals, confirming that systematically tracked, searchable retraction data reveal patterns of concentration that a case-by-case reading of individual notices would not10.
Practical use of data on reasons for retraction
This evidence base has practical implications for researchers, editors and institutions. Because retractions for fraud take longer on average to emerge and accumulate more citations before detection, the citation history of a retracted paper provides a meaningful, verifiable indicator of the extent to which a retraction may already have influenced subsequent research, which the Retraction Watch database and its associated citation-tracking tools are designed to help readers trace. For journals and institutions, the finding by Fang, Steen and Casadevall that misconduct, rather than error, accounts for the majority of retractions has informed a shift in emphasis toward stronger pre-publication integrity checks, image-screening software, statistical consistency checks and clearer institutional investigation procedures, rather than treating retraction primarily as a mechanism for correcting honest statistical or methodological errors after publication1. The COPE retraction guidelines, which specify what a retraction notice should state and how clearly it should identify the reason, reflect a direct response to the problem identified in the PNAS analysis: vague or incomplete retraction notices make it more difficult for the scientific community to learn from, and guard against, the specific failure that caused the retraction4.
The relationship between a formal finding of misconduct and a retraction
Not every confirmed finding of research misconduct results in a retraction, and not every retraction follows a formal misconduct investigation, a gap documented by Resnik and Dinse in their analysis of retractions and corrections linked to confirmed misconduct findings of the U.S. Office of Research Integrity (ORI). Their study found that an appreciable proportion of papers connected to a confirmed ORI misconduct finding were never formally retracted or corrected, remaining in the citable literature without any notice of the institutional finding; a retraction search alone, however useful, is therefore not a complete substitute for checking the known misconduct history of an author or paper through other available channels where this is relevant to a specific decision9. This gap forms part of the continuing case for standardizing and strengthening retraction practices across journals and institutions, since a confirmed finding that never produces a visible correction leaves precisely the misleading, uncorrected record that the retraction system exists to prevent. For institutions and funders, it also supports the direct cross-checking of ORI findings against journal retraction notices, rather than the assumption that either record alone provides a complete account of the integrity history of a researcher.
Key points
Key points
Retraction data contradict the assumption that most retracted papers reflect honest error. A landmark analysis found that misconduct accounts for the majority of retractions with an identifiable cause, and that retractions for fraud take longer to emerge and accumulate more citations beforehand; this evidence has contributed both to stronger pre-publication integrity checks and to clearer standards for retraction notices.
Practical application
Practical application
- The retraction status of a paper, and the stated reason, should be checked before it is cited, including in the case of older papers. The Retraction Watch database permits direct lookup; because retractions for fraud may take years to emerge, an older and heavily cited paper is not exempt from this check.
- A vague retraction notice should be treated as a prompt for further inquiry. COPE’s retraction guidelines call for clear and specific reasons; where a notice is vague, institutional statements or the reporting of Retraction Watch may clarify the actual cause.
- Checks for image consistency and statistical consistency should be incorporated into the pre-submission process. Given the frequency with which image manipulation and data inconsistency recur in analyses of retractions, screening figures and statistics before submission directly addresses two of the most common identifiable causes.
Frequently asked questions
No. The causes of retraction range from deliberate fabrication and falsification to honest error, duplicate publication and plagiarism; the analysis by Fang, Steen and Casadevall found that misconduct accounts for a majority of retractions with an identifiable cause, but an appreciable minority are retracted for genuine, unintentional error.
Steen's analysis found that fraud is often more deliberately concealed, and more difficult to detect through ordinary peer review or post-publication scrutiny, than an honest statistical or methodological error, which tends to emerge more quickly once other researchers attempt to build on or replicate the work.
Research-integrity bodies generally regard it as a less severe but still significant breach; it does not involve fabricated data, but it distorts the literature by allowing the same findings to be counted more than once in evidence synthesis, which is why COPE addresses it as a distinct category of violation with its own guidance.
The Retraction Watch database permits direct searching by title, author or journal, and most major databases, including PubMed, display retraction status on the article record; checking before citing, especially for older papers, is therefore a quick and concrete step.
Not necessarily. The analysis by Steen, Casadevall and Fang found that the increase closely tracked growth in total publication volume and the adoption by journals of more consistent retraction-flagging practices, which is consistent with a self-correcting system detecting problems more effectively rather than with a proportional increase in misconduct.
References
- Fang FC, Steen RG, Casadevall A. Misconduct accounts for the majority of retracted scientific publications. Proceedings of the National Academy of Sciences. 2012;109(42):17028-17033. read 10 October 2026. View on PNAS
- Steen RG. Retractions in the scientific literature: do authors deliberately commit research fraud? Journal of Medical Ethics. 2011;37(2):113-117. read 10 October 2026. View on Journal of Medical Ethics
- Bik EM, Casadevall A, Fang FC. The prevalence of inappropriate image duplication in biomedical research publications. mBio. 2016;7(3):e00809-16. read 10 October 2026. View on mBio
- Committee on Publication Ethics (COPE). COPE Retraction Guidelines. read 10 October 2026. View COPE Retraction Guidelines
- Retraction Watch. The Retraction Watch Database. read 10 October 2026. View the Retraction Watch Database
- Steen RG, Casadevall A, Fang FC. Why has the number of scientific retractions increased? PLOS ONE. 2013;8(7):e68397. read 10 October 2026. View on PLOS ONE
- Brainard J, You J. What a massive database of retracted papers reveals about science publishing's 'death penalty'. Science. 2018. read 10 October 2026. View on Science
- International Committee of Medical Journal Editors. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals: Corrections, Retractions, Republications and Version Control. read 10 October 2026. View ICMJE Retraction Recommendations
- Resnik DB, Dinse GE. Scientific retractions and corrections related to misconduct findings. Journal of Medical Ethics. 2013;39(1):46-50. read 10 October 2026. View on Journal of Medical Ethics
- Grieneisen ML, Zhang M. A comprehensive survey of retracted articles from the scholarly literature. PLOS ONE. 2012;7(10):e44118. read 10 October 2026. View on PLOS ONE
Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
Pubrica Academy. (2026, October 10). Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error. Pubrica. https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/
Pubrica Academy. “Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error.” Pubrica, 10 Oct. 2026, pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/.
Pubrica Academy (2026) Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error. Pubrica. Available at: https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/ (Accessed: 15 October 2026).
Pubrica Academy. “Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error.” Pubrica. October 10, 2026. https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/
Pubrica Academy, ‘Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error’ (Pubrica, 10 October 2026) <https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/> accessed 15 October 2026.
Pubrica Academy. Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error. [Internet]. Pubrica; 2026 Oct 10 [cited 2026 Oct 15]. Available from: https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/
Pubrica Academy, “Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error,” Pubrica, Oct. 10, 2026. [Online]. Available: https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/
@misc{/la_retraction_watch_common_reasons_2026,
author = {{Pubrica Academy}},
title = {Retraction Watch: The Most Common Reasons Papers Are Retracted and the Balance of Misconduct and Error},
howpublished = {Pubrica},
year = {2026},
month = {oct},
url = {https://pubrica.com/academy/study-guide/research-writing/retraction-watch-common-reasons-scientific-publishing/},
note = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
Pre-Submission Retraction-Risk Checklist
For use in addressing the most common identifiable causes of retraction before submission.
[ ] Checked all figures (blots, micrographs, gels) for unintended duplication or manipulation [ ] Re-verified all reported statistics for internal consistency [ ] Confirmed the manuscript has not already been substantially published elsewhere [ ] Checked all cited sources for retraction status before finalizing the reference list [ ] Confirmed all data underlying the manuscript can be produced and defended if questioned
Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.
How Pubrica can help
Pubrica’s publication ethics team screens manuscripts before submission for the most common identifiable causes of retraction, namely image inconsistency, statistical irregularities and risk of duplicate publication, and checks reference lists for retracted sources.