Retraction
A journal's formal withdrawal of a published paper because its findings are unreliable, usually through error or misconduct.
A retraction does not delete the paper. The article stays online, stamped RETRACTED, with a notice explaining why — that permanence is deliberate, so the record of what was claimed and then withdrawn survives.
Never cite a retracted paper as evidence. Check before you rely on anything: the publisher's page carries the notice, and Retraction Watch's database and PubMed's retraction flags catch most cases. Retracted papers keep accruing citations for years, which is exactly the failure to avoid.
See also
Erratum and corrigendumPublished corrections to an article: an erratum for the publisher's error, a corrigendum for the author's.Peer reviewThe process by which independent researchers in a field evaluate a manuscript before a journal agrees to publish it.Version of recordThe publisher's final, typeset, citable version of an article — the one page numbers and DOIs point at.Citation countHow many later publications cite a given work, according to whichever index you ask.