Worobey et al.'s early-case clustering around the Huanan market persists under a social-media-mobility null instead of a population-density null
2 events · 1 assessment
Assessed Contested
verdict confidence 0.68
Whether the clustering of the earliest COVID-19 cases around Wuhan's Huanan Seafood Market holds up when the expected "center" of human activity is defined from human-mobility data rather than residential population density is genuinely disputed among statisticians, and the question is not settled. Worobey et al.'s 2022 analysis compared where the December 2019 cases lived against a residential-population reference and found them anomalously close to the market. Critics (notably Stoyan and Chiu, 2024) argue that residential population is the wrong benchmark: using social-media check-in mobility data, other high-traffic sites near the market — Wanda Plaza, Hankou railway station, the Wuhan CDC — are equally or more plausible "centers," so the test does not statistically single out the market. Débarre and Worobey's reply turns this around: because the market was a comparatively low-traffic destination (far less visited than Wanda Plaza and dozens of other markets), the concentration of early cases at that specific site is, if anything, more striking under a mobility view, not less. Both positions come from qualified statisticians and neither has been decisively refuted; the disagreement is partly about the validity of the mobility data itself (2013–2014 check-ins) and, more fundamentally, about whether a "center of the point cloud" test can identify an outbreak's origin at all. The matter would be resolved by an agreed, well-validated mobility-based reference distribution and a clustering test whose properties both camps accept.
Claim entered the graph