Does anyone have a link to the actual study? https://twitter.com/skarlamangla/status/1252327455609634817

Apparently it was 863 people? I’m sorry. AYFKM?!?

I think to really declare numbers like this for a serology study you need to test...what, 10,000 people in a nice random sample? What do you think #medtwitter

I will add that I have ZERO doubt that cases are grossly underreported, but the whole 55x thing is a bit too much, and a lot of extrapolation from tiny numbers.

It completely fascinates me, by the way, that there are whole disciplines of science and math that have not encountered Bayes in the wild before.