Provenance

Retractions and corrections in research

Last checked 7 November 2026 · 5 sources

Short answer

A correction fixes an error in a paper that still stands. A retraction tells readers a paper's findings cannot be relied upon. An expression of concern flags doubt when facts are unclear. COPE, whose retraction guidelines reached version 3 on 4 September 2025, says the purpose of retraction is to correct the literature, not to punish authors ¹. ICMJE says honest errors require a published correction ². A US research misconduct finding requires proof by a preponderance of the evidence ³.

COPE Retraction Guidelines: Version 1 in 2009, version 2 on 11 December 2019, version 3 on 4 September 2025 ¹
Who decides to retract: The editor, per COPE ¹
Correction trigger: Errors of fact that should have been recognised at publication, per ICMJE ²
US misconduct standard of proof: Preponderance of the evidence, 42 CFR 93.103 ³

Three notices compared

Notice	When it is used	Source
Correction	Honest errors of fact that should have been recognised at publication. Matters of debate and evolving science are not errors ²	ICMJE ²
Retraction	Findings and conclusions cannot be relied upon, from honest error, naive mistakes or research or publication misconduct ¹	COPE ¹
Retraction with republication	Honest error leads to a major change in results, the science appears valid and the changed paper survives further review ²	ICMJE ²
Expression of concern	The editor is uncertain about reliability owing to insufficient information, delays or inability to obtain information ¹	COPE ¹

What COPE's guidelines say

Editors can retract if they no longer have confidence in the results and conclusions ¹.
Retraction might be needed for unethical research practices, compromised peer review or undisclosed conflicts of interest ¹.
Retraction notices should link to the retracted article, identify it by title and authors, be published promptly and be freely accessible ¹.
The editor should make the final decision, except where the editor is compromised or has a conflict of interest. Publishers can support investigations ¹.
Batch retractions might be needed where there is evidence of systematic manipulation of publication, such as paper mills ¹.
If there is a delay in deciding, editors should publish an expression of concern ¹.

Timing and version history

COPE says retraction notices should be published as soon as the editor is confident that a publication is seriously flawed, misleading or falls into one of the listed categories ¹. The aim is to limit harm such as citation of erroneous work, acting on its findings or drawing incorrect conclusions ¹. Where the decision is delayed, COPE says editors should publish an expression of concern ¹.

COPE's guidelines were first published in 2009 ¹. Version 2 was published on 11 December 2019 ¹. Version 3, published on 4 September 2025, added sections on batch retractions and on communicating with authors, and aligned with new 2025 guidelines on expressions of concern ¹.

What ICMJE says about errors

ICMJE says pervasive errors from a coding problem or miscalculation that do not change the direction of the results should be handled by a correction ². Errors serious enough to invalidate the results may require retraction ². For papers published long ago, for example more than 10 years, it says an editor may reasonably decide a correction is not warranted ².

On duplicates, ICMJE says an undisclosed duplicate found after publication might warrant retraction, with or without the author's explanation or approval ⁴.

US research misconduct findings

Under 42 CFR 93.103, a finding of research misconduct requires a significant departure from accepted practices of the relevant research community, that the misconduct be committed intentionally, knowingly or recklessly, and that the allegation be proven by a preponderance of the evidence ³. A finding is a formal outcome. A retraction is a journal decision, and the two can happen apart.

Checking a paper's status

Look for linked notices on the article page. COPE says notices should link to the retracted article ¹.
Check errata and retraction statements for every study you rely on. The Cochrane Handbook tells review authors to examine them, since errata can reveal fatal flaws ⁵.
Search for later comments and letters. Cochrane says searching for comments, letters or retractions may uncover more information ⁵.
Record the version you read. ICMJE's recommendations cover corrections, retractions and version control together ².

Common questions

Is a retraction a finding of misconduct?

No. COPE says unreliable content can result from honest error, naive mistakes or misconduct, and retraction aims to correct the literature, not punish authors ¹.

Who decides to retract a paper?

COPE says the editor should make the final decision, with the publisher assisting ¹.

What is an expression of concern?

A notice an editor can publish when unsure about a paper's reliability, where information is insufficient or delayed ¹.

Can an old paper escape correction?

ICMJE says for papers published long ago, for example more than 10 years, an editor may reasonably decide a correction is not warranted ².

Rules and charges change. Check the official source linked on this page before you pay or travel. Spotted an error? Email hello@evidencefamily.com.

Sources

Duplicate publication rules in research
What the ICMJE Recommendations say about duplicate and secondary publication, preprints and registry results, and why it matters for evidence.
Multiple reports of one study: how reviews link them
How the Cochrane Handbook, PRISMA 2020 and ICMJE rules stop one study from being counted as several, and the criteria used to link reports.
Clinical trial registration explained
How ICMJE, WHO, US and EU rules make trials record their plans and results in public registries, with the deadlines, data items and registries that count.
Sourcing and provenance glossary
Official definitions of sourcing, corroboration and provenance terms from intelligence directives, EU law, C2PA, IPTC, ICMJE, COPE and Cochrane.