Correcting the scientific record: Errors in the scientific record explained

How serious is an erratum, notice or expression of concern, or a retraction? They should all be avoided as these errors do not reflect well on the authors, editors, editorial board, peer reviewers, journal, and publisher. An error introduced into the scientific record calls into question the integrity of the article and the journal, and erodes the public trust in scientific research.

Publishers, editors, authors, and the production team all share the responsibility for correcting the publication record when required, as it impacts the evolution of scientific knowledge. While there has been a general rise in the number of retractions, more of the articles have been retracted for unintentional errors.¹

How do errata, expressions of concern, and retraction notices arise? These are as a result of a request from the author, or observation from a reviewer, an editorial board member, another journal, a publisher, an employer, a research funder, a reader or “whistle-blower”, or a comparison of content using plagiarism software.

While misconduct or unintentional errors can be difficult to detect before publication, once an issue is raised, it is important to act as a matter of urgency. This document provides definitions, examples and policies on errata, expressions of concern, or retractions, and points to the resources of the Committee on Publication Ethics (COPE) for further reading.^2,3

Publication of any of these notices is not taken lightly, and investigation and verification is sought. Verification involves contacting the corresponding author and receiving a written response. Verification may be as simple as an author confirming there is a mistake in the paper or as involved as a full investigation conducted by the author’s institution or employer to explore whether the data and research is fraudulent. Correspondences on the concerns raised on a paper are dealt with as confidentially as possible. This involves only those people who can assist in the resolution of a claim, and can recommend whether a notice is warranted.

A notice is considered if the article is published and hence is publicly accessible. A published article can include a version that has appeared online, either in its final form or a prior version before a final version is placed in an issue. For articles that are not yet in a final version, an erratum is unlikely to be required as important changes could be incorporated before the final version is published. An expression of concern or retraction can still be applied to articles that have appeared online irrespective of whether it is a final version or not; as the article has been published the publisher is required to correct the publication record.

Each notice is published in the same way and appears in the table of contents of the journal (existing print and online). There will be a reference citation to the original publication, and in the electronic world it will be linked to the original article. The convenient electronic linkage of articles also occurs in databases such as PubMed. The person who is reading the article online has a distinct advantage over the print reader, who may be reading the original article without knowledge of notices published in later issues.

As a general principle the original article is not changed to accommodate the change in the erratum, or removed in the case of expression of concern or retraction. This is in line with the International Association of Scientific, Technical & Medical Publishers’ (STM) guidelines.⁴ This is because the publisher is alerting readers to a correction in the publishing record, and not changing publishing history, either in print or online. There are exceptions to this approach, such as when there is an inappropriate violation of privacy, errors that would pose a significant risk to health to the public, and defamatory comments, and then an explanation would be given clearly indicating that the electronic version of the article has been updated.⁴ Generally, the only change that would happen to the original publication is when there is a retraction, and this only happens to the electronic version—full text is removed, and the article remains in the form of a PDF which is watermarked to clearly indicate to the reader the article has been retracted.

So let us take these notices in turn, including special cases when republication of a whole article, or partial retraction and republication of an article need to be considered.