Do referees give dark-skinned footballers more red cards?
✓ Numbers check out
Counts as holding up
The code and data produced the same result. That doesn't prove the study's method is right.
Recomputing the median of the answers that 29 research teams gave about dark-skinned footballers and red cards gives 1.31, matching the number published in the paper.
Silberzahn and colleagues asked 29 research teams to answer the same question with the same data: are footballers with dark skin more likely to be shown red cards by referees? The teams' answers varied widely, from 0.89 to 2.93 times the odds, with a typical answer of 1.31. This shows how much researchers' choices can affect the result even when the data and question are identical.
A small independent program was run on the project's published table listing each team's answer, and it recalculated the median, giving 1.31 — the same number the paper reports. This confirms only that the same published numbers give the same median; it does not redo the 29 analyses, check whether the data were sound, or say whether the original findings hold, so a mismatch would only have ended as inconclusive.
The text was written by a private AI model on NEAR AI (z-ai/glm-5.3-flash). The verdict and numbers come straight from the blockchain, and the AI can't change them.
- What the paper said
- effect sizes ranged from 0.89 to 2.93 ( Mdn = 1.31) in odds-ratio units. Abstract of the paper, which also says 20 teams found a significant effect and 9 did not
- What we got
- 1.31
- Whose code ran
- Independent rebuild
- How well known
- Cited by 773 studies
- What forecasters thought
- Nobody forecast this one. We posted it with the result already known, to show how a famous paper looks.
- Proof
- Read the receipt or see claim 28
What we didn't check
- Whether the sample was fair or large enough.
- Whether test data leaked into training or tuning.
- Whether the result holds on new data.
- Whether the study's method was the right one for its question.
- The 29 teams' own analyses. We only recompute the summary from their published answers.
- Which team's method is the best one.
Community notes
Notes are people's own words, kept on the blockchain. They never change a verdict, and the Lab's AI never reads them. Posting costs a little NEAR for storage, and what isn't used comes back.