Qualifies· weakCOV-E002
The 2024 erratum: data files corrected, results unchanged
- what the source states
- The authors identified duplicated data files and a pseudoreplicate count of 10,000 "rather than the n = 1000 used for all related analyses," regenerated the files and re-ran every affected test: "All results remained the same as previously reported: p < 0.001 and p-Adjusted (BH) = 0.003." The notice also records that a correction was posted to the supplementary material on May 8, 2023, and that "Science should have indexed it as a formal Erratum," which occurred on March 15, 2024.
- what we infer
- Recorded because corrections must be as visible as findings, and because this one is routinely conflated with the separate and substantive correction to the companion lineage paper. This erratum changed no result. Its only substantive content is a ten-month indexing delay, which is a small process-integrity observation about the publication record and not evidence about origins.
- limitations
- Read via a dated archive snapshot because the publisher bot-walls automated clients and the repository deposit carries metadata only.