Skip to content
Caravel Press

Corrections and retractions

How the published record is put right when something in it is wrong.

The principle

A published paper is a permanent record. It is never quietly edited, replaced or removed. Where something in it is wrong, the correction is itself published, is linked to the paper from both directions, and carries its own DOI, so that anyone who has the original also finds the correction.

Correcting the record is a sign that a journal is working, not that it has failed. Web of Science states that journals are not penalised for issuing retractions. We would rather publish a retraction than leave a wrong result standing.

The four instruments

  1. Correction

    For an error that affects a part of the paper but not its conclusions. A typographical error in an equation, a mislabelled figure, a missing funder. The paper is amended, the amendment is described, and the date is recorded.

  2. Expression of concern

    For a serious concern that has been raised but not yet resolved, where readers should be told while an investigation proceeds. It is withdrawn or replaced by a correction or retraction once the matter is settled.

  3. Retraction

    For findings that are unreliable, whether through honest error or misconduct, and for plagiarism, redundant publication, unethical research, or unauthorised use of material. The paper remains available, clearly watermarked as retracted, with a retraction notice stating who retracted it and why.

  4. Removal

    Used only where leaving the paper available would itself cause harm or break the law, for example defamation or a serious breach of privacy. The metadata and a note remain so the citation still resolves.

Which instrument applies is decided following the COPE retraction guidelines.

How a correction reaches a reader

  • The notice is published as its own item with its own the DOI system, and is free to read.
  • The original paper links to the notice, and the notice links to the original paper.
  • The change is deposited with Crossref, so every index and reference manager that reads Crossref learns of it.

Asking for a correction

Anyone may ask for a correction, including a reader who is not an author of the paper. Write to [email protected] with the paper DOI and what you believe is wrong. You will be told what has been decided and why.

This policy applies to every journal we publish. Last revised 19 October 2026.