Nate Silver @NateSilver538
Interesting thread here. Whenever there's no difference (in this case, between the rate of Delta vs Alpha hospital admissions) before you introduce statistical controls, but a large difference once you introduce those controls, there's a good chance you screwed up the controls. — PolitiTweet.org
Joël Mossong @joel_mossong
Presentation of results is puzzling. In appendix, main confounder is "week" and ethnicity, not age. How can calenda… https://t.co/xLq39bLOTw
