Thread by @Ayjchan

If you were taken in by the beautiful figures in the new Proximal Origin preprints, please get in touch with @sciencecohen

https://twitter.com/sciencecohen/status/1497957724017074178

@sciencecohen The authors' own words, not mine.

https://twitter.com/Ayjchan/status/1497943297762992134?s=20&t=I3TmUmUHLY-UwZ9rTFz57A

Imagine if the same approach was applied to clinical trials of covid-19 vaccines. "We don't know if the data is authentic. We don't have the full data. But we're sure our analysis is robust."

Should science journalists be seeking out scientists who were convinced by the paper?

At the very least, when you're interviewing these scientists for quotes, please make sure they are aware of the source of the data (this Epoch Times leak; citation 41 in the preprint) and that the data is incomplete and likely prone to sampling bias.
epochtimes.com/gb/20/5/31/n12…

If you're aware that @ScienceMagazine is peer reviewing these preprints, please help to point the editor to this glaring problem with the scientific analysis in one of the preprints.

https://twitter.com/Ayjchan/status/1497925079845462021

@ScienceMagazine I've searched all 3 preprints and none of them have the data on how many samples were collected at each stall.

@ScienceMagazine Although I will note that the only preprint with access to full data wrote that the market SARS2-positive surface samples are likely derived from human beings at the market. "No animal host of SARS-CoV-2 can be deduced."

They also wrote that they did not observe any association of cases with different products, e.g., wildlife. And their results suggest the virus might have been circulating in the market for a period to time in Dec 2019, leading to extensive distribution facilitated by crowds.

Naturally, I'd like to see the data in full first. Which is why I added my name to @jbloom_lab's open letter to request the data.

https://twitter.com/jbloom_lab/status/1497743165226303489