Even carefully reviewed papers sometimes contain errors, and when one is found after publication, the correct response is a formal correction, not silence. A corrigendum or erratum is a short published notice that identifies the error, gives the correct information, and explains, briefly, how it affects the paper. Issued promptly and honestly, a correction protects readers, preserves the scientific record, and reflects well on its authors.
Authors often delay corrections because they fear embarrassment, but the record is worse without one: readers continue to cite wrong numbers, build on faulty details, and eventually discover the error from someone else. This guide explains the difference between a corrigendum and an erratum, when each applies, what a correction notice should contain, and how journals handle the process from notification to publication.
The terms corrigendum and erratum are often used interchangeably, but journals distinguish them by responsibility. An erratum is traditionally a correction for an error introduced by the publisher, such as a typesetting mistake, a misplaced figure, or a production error in an author's name. A corrigendum is a correction for an error made by the authors, such as a wrong value, a mislabelled table, or an omitted acknowledgement. Some journals use only one of the two terms for all corrections, so follow the terminology of the journal concerned.
Other post-publication notices serve different purposes and should not be confused with corrections. A retraction withdraws a paper because its findings are unreliable or its conduct breached ethics. An expression of concern alerts readers while an investigation is under way. A comment or reply debates interpretation without correcting the record. If you are unsure which notice fits your situation, describe the problem to the editor and ask; editors handle these distinctions routinely and will advise you.
A correction is needed when an error could mislead readers or affect the use of the paper: a wrong number in the results, an incorrect figure or table, a mistaken author name or affiliation, a missing funding or conflict statement, or an error in a method that would prevent reproduction. Such errors, however small they seem, belong in the published record, because the record is what other researchers rely on.
A correction is usually not needed for trivial typographical errors that do not change meaning, such as a misspelling in the discussion that no reader would misread, though journals differ and some will correct even these. When in doubt, tell the editor and let the journal decide: the cost of asking is a short email, while the cost of leaving a meaningful error uncorrected is borne by every reader who trusts the paper. Never quietly correct an error by changing a file on a personal website; only the journal's published correction amends the record.
Contact the journal's editor as soon as you discover an error, with a clear description of what is wrong, where it appears, and what the correct information is. Include the evidence: the original data, the correct calculation, or the source that shows the mistake. If co-authors are involved, inform them first and agree the correction together, since a correction, like the original paper, is the responsibility of all authors.
Be specific about the impact. State whether the error affects the paper's conclusions or only its details, and say so honestly: an error that changes a conclusion requires a more prominent correction and closer editorial scrutiny than one that does not. Editors appreciate candour here, because it lets them judge the right level of response. Concealing the impact of an error, only for reviewers or readers to discover it, converts a routine correction into a question of trust.
A correction notice is short, usually a few paragraphs, and it follows a standard pattern. It identifies the original article by title, authors, and publication details, states precisely what was wrong, gives the corrected information, and explains the effect on the paper's conclusions. It should be written so that a reader who has never seen the original error can understand exactly what changed and why.
Include the corrected text, table, or figure where needed, rather than describing the change vaguely. For example, state the incorrect value and the correct value, the incorrect label and the correct label, or the omitted acknowledgement in full. Close with an apology for the error where the journal's style requires one, and with a statement of whether the conclusions are affected. A correction that leaves the reader guessing about what changed has not done its job.
Who caused the error determines the notice's wording and, sometimes, who pays any associated cost, though corrections themselves are normally published without charge. If the error arose in production, for example a figure swapped during typesetting or an author name misspelled from a correct manuscript, provide the accepted manuscript or proof as evidence so the publisher can verify the source. The journal will then issue an erratum in its own name.
If the error was the authors', own it plainly. A corrigendum that attributes an author error to production, or that minimises it with vague language, will be read by the editor, the production team, and eventually the readers, none of whom will be reassured. Journals handle author corrections routinely and without judgement; what they judge is honesty about the source of the error and promptness in reporting it.
Most corrections do not change a paper's conclusions, and the notice should say so explicitly, because readers otherwise assume the worst. When a correction does affect conclusions, the notice must say how: which conclusion changes, what the revised conclusion is, and what evidence supports it. The editor may send the correction for peer review in such cases, and may ask for additional analysis or for the raw data to be checked.
In rare cases, an error is so serious that the findings can no longer be relied on, and the editor may propose retraction rather than correction. That decision belongs to the editor, informed by the authors' evidence and, where needed, by independent review. Authors who report such an error themselves, promptly and completely, are treated very differently from authors whose errors are exposed by others, and self-reporting is one of the strongest demonstrations of research integrity available.
Corrections to author lists, affiliations, funding statements, and acknowledgements are common and are handled formally, because these details affect credit, accountability, and compliance with funder requirements. An omitted author, a wrong affiliation, or a missing funder must be corrected through the journal, with written agreement from all authors, including anyone being added or removed. Journals require this agreement to prevent disputes being settled by altering the published record.
Provide the evidence the journal asks for: the original submission records, the funding award details, or the contribution records that justify the change. Changes to authorship after publication receive particular scrutiny, since they affect careers and responsibilities, and the editor may consult the journal's ethics guidance before approving them. Raise such corrections early, with all authors informed, rather than as a surprise to co-authors who learn of them from the published notice.
Once approved, a correction is published as a separate, citable item, linked to the original article from the article page and, in most systems, from the correction itself. The original article is usually updated with a notice pointing to the correction, while the original version remains available so that the history of the record is preserved. Indexing databases and reference managers pick up the correction through the publisher's metadata, which is why the formal route matters more than any informal announcement.
Check the published correction when it appears, and confirm that the corrected values, the linked article, and the metadata are all correct. If the correction itself contains an error, tell the editor at once so that it can be corrected in turn. Also update any repositories, institutional profiles, or personal pages that carry the paper, so that readers who find the work through those routes see the corrected version rather than the original error.
Errors are sometimes found by readers, reviewers of later work, or commentators, and reported to the journal or to the authors directly. Respond promptly, courteously, and on the evidence: verify the claim against your data, acknowledge it if it is correct, and work with the editor on a correction. A gracious, fast response to an external report reflects as well on authors as self-reporting does, and it often prevents a public dispute.
If the claim is incorrect, explain why with evidence, calmly and specifically, and let the editor adjudicate. Avoid public arguments about the error before the facts are established, and avoid attributing motives to the person who raised it. Most people who report errors do so in good faith, and the scientific record is better for their attention, whether the particular claim proves right or wrong.
Corrections are a normal part of publishing, but most are preventable with a disciplined final check. Before submission, verify every number in the abstract, tables, figures, and text against the source data, and have a co-author check them independently. Confirm author names, order, affiliations, funding, and acknowledgements with every author, not only the corresponding author. At proof stage, check the same items again against the accepted manuscript, character by character for names and numbers.
Keep the data, analysis code, and records that support each paper, organised and accessible, so that a query months or years later can be answered from evidence rather than memory. When an error is found despite these checks, the same records make the correction fast and convincing. A laboratory or group that treats corrections as routine maintenance, rather than as embarrassment, produces a more reliable record over time.
Use this checklist when an error is found: inform all co-authors and agree the facts; contact the editor promptly with the error, the correction, and its effect on conclusions; provide the evidence the journal asks for; check the published correction and its links when it appears; and update repositories and profiles that carry the paper. If the error was found by someone else, respond courteously and on the evidence, and let the editor adjudicate any disagreement.
Finally, remember that correcting the record is a professional strength, not a confession of failure. Readers, editors, and funders trust authors who correct errors openly far more than those who leave them standing. Act promptly, state the facts plainly, and the correction will protect both your work and the readers who rely on it. For related guidance, see our guide on how to publish a research paper, our explainer on double-blind peer review explained, and our guide on how to avoid predatory journals.
| Notice | Used when | Who issues it |
|---|---|---|
| Erratum | Publisher introduced the error | Journal or publisher |
| Corrigendum | Authors made the error | Authors, via the journal |
| Retraction | Findings are unreliable | Editor, with authors |
| Expression of concern | Investigation is under way | Editor |
| Comment or reply | Interpretation is debated | Reader or author |
These authoritative external resources explain the standards and tools mentioned in this guide:
What is the difference between a corrigendum and an erratum?
Traditionally, an erratum corrects an error introduced by the publisher, such as a typesetting mistake, while a corrigendum corrects an error made by the authors, such as a wrong value. Some journals use only one term for all corrections, so follow the terminology of the journal concerned.
When should I request a correction?
Request one when an error could mislead readers or affect the use of the paper: a wrong result, figure, table, author name, affiliation, or funding statement. Trivial typographical errors that do not change meaning usually do not need a correction, but ask the editor if unsure.
How do I notify a journal of an error?
Contact the editor promptly with what is wrong, where it appears, what the correct information is, and whether the conclusions are affected. Inform all co-authors first, agree the facts together, and provide the evidence, such as the original data or calculation.
Do corrections change the original article?
The original article usually remains available, with a notice linking to the correction, which is published as a separate citable item. This preserves the history of the record while ensuring readers who find the paper also find the correction.
What if a correction changes the conclusions?
State exactly which conclusion changes and what the revised conclusion is, with evidence. The editor may send the correction for peer review, and in rare cases where findings can no longer be relied on, may propose retraction instead of correction.
Can authorship be corrected after publication?
Yes, but formally and with written agreement from all authors, including anyone added or removed. Journals scrutinise authorship changes closely, since they affect credit and responsibility, so raise them early and with all authors informed.