Examples
Worked examples
- Is an instance
A journal publishes a correction notice titled 'Erratum to: [article title]' after discovering that Figure 3 and Figure 4 were swapped during typesetting, with the corrected figures reproduced in the notice and linked to the original DOI.
- Is an instance
A publisher issues an erratum to fix a co-author's name that was misspelled throughout the published PDF due to a production-system error, with no change to the study's data, methods, or conclusions.
- Is an instance
A journal corrects a transposed decimal point in a summary table that was introduced when the manuscript was converted to the journal's typesetting template, where the authors' original submitted manuscript had the correct value.
Counter-examples
Looks similar, but isn't
- Not an instance
An author discovers, after publication, that they used the wrong statistical test and the corrected analysis changes the paper's conclusions -- this is not an erratum (a production-side fix); depending on severity it is either an author-originated correction/corrigendum or grounds for retraction, not a typesetting or production error.
- Not an instance
A journal determines that a figure was manipulated to misrepresent underlying data -- this is a research-integrity matter handled through COPE's correction/retraction guidance and, if it invalidates the conclusions, retraction, not an erratum, since the issue originates in the research itself rather than in production.
Editorial commentary
An erratum (plural: errata) is a publisher-issued correction notice for an error that occurred somewhere between an accepted manuscript and its published form — typesetting, formatting, figure/table reproduction, or similar production-side mistakes — rather than an error in the underlying research. Conventionally, publishing and scholarly-communication guidance (including Council of Science Editors and journal-level style guides) draws the erratum/corrigendum split by origin: an erratum corrects a mistake the publisher introduced, while a corrigendum corrects a mistake the authors made in their own submitted work. In practice, many journals use the two terms loosely or interchangeably, and some style guides collapse both under a single ‘correction’ notice type — so the origin-based distinction is a useful default, not a universally enforced rule.
Erratum vs. corrigendum vs. retraction: who’s at fault, and how severe
The three notice types sit on two separate axes — who introduced the error and how serious its effect on the paper’s conclusions is — and it is the severity axis, not the fault axis, that determines whether a correction is enough or a retraction is required:
- Erratum — fault: publisher/production process. Severity: does not affect the validity of the results or conclusions by definition; the science was correct, only its published rendering was wrong.
- Corrigendum — fault: author(s). Severity: varies. A corrigendum can range from a minor, non-substantive fix (a wrong grant number, an affiliation update) to a substantive one that corrects a genuine error in the authors’ data or analysis, provided that error does not undermine the paper’s overall conclusions.
- Retraction — fault: either (author error, publisher/editorial error, or misconduct). Severity: the error is serious enough to invalidate the paper’s results and conclusions, regardless of who caused it or whether it was honest error or misconduct.
ICMJE’s Recommendations on corrections, retractions, republications and version control frame the correction/retraction line explicitly around impact on validity rather than around who was at fault: corrections are for errors of fact, including pervasive ones, that do not change the direction or implications of the results and conclusions; retraction (in some cases with republication of a corrected version) is for honest errors serious enough to invalidate or materially change those results, interpretations, or conclusions. The Committee on Publication Ethics (COPE) applies the same materiality test in its retraction guidelines: the operative question for an editor is whether a reader, informed of the correction, would draw materially different conclusions about the work. Neither ICMJE nor COPE treats fault (author vs. publisher, honest error vs. misconduct) as the determining factor for the correction-vs-retraction decision — it is one input into how the notice is worded and who is credited/blamed, not into which notice type applies.
Why the distinction matters in practice
For readers and citing researchers, the practical difference is trust calibration: an erratum signals ‘the record is now correctly rendered, the findings stand unchanged’; a corrigendum signals ‘the authors themselves had to fix something in their own work, but it doesn’t overturn the paper’; a retraction signals ‘do not rely on this paper’s conclusions.’ For metadata and discovery systems, all three are registered as distinct, linked update types — Crossref’s update-record schema enumerates erratum, corrigendum, and retraction (along with partial_retraction, expression_of_concern, and others) as separate values precisely so indexers, reference managers, and services like Crossref’s CrossMark can surface the correct status against the original DOI rather than lumping every post-publication notice together.
Related terms
- Correction (corrigendum) — the author-originated counterpart to an erratum.
- Retraction — the more severe outcome when an error invalidates a paper’s conclusions.
- Expression of concern — an interim notice used while an issue is investigated, before a correction or retraction is decided.
- CrossMark — the Crossref service that surfaces erratum/correction/retraction status on a published item.
References
- ICMJE, Recommendations: Corrections, Retractions, Republications and Version Control of Articles
- COPE Retraction Guidelines, version 2 (November 2019), DOI 10.24318/cope.2019.1.4
- Crossref, registering updates documentation (update-type schema)
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="Erratum"
vocab-term-identifier="https://casrai.org/dictionary/term/erratum" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/erratum",
"name": "Erratum",
"identifier": "https://casrai.org/dictionary/term/erratum",
"description": "An erratum is a formal, published correction notice that a journal or publisher issues to fix a factual error introduced during the publication process itself (e.g. a typesetting mistake, a mislabeled figure, an author name spelled wrong, a column transposed in a table) rather than an error in the authors' own research, analysis, or reporting. It is linked bidirectionally to the original article's metadata (Crossref lists 'erratum' as one of its enumerated update types) so that the correction is discoverable wherever the article is indexed or cited. An erratum does not withdraw or invalidate the article's findings -- by definition, the underlying science is not in question, only how it was rendered in print or online.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/research-integrity#set",
"url": "https://casrai.org/dictionary/term/erratum",
"sameAs": [],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"dateModified": "2026-07-18T04:50:17",
"inLanguage": "en"
}
