1/ The second part of my cautionary tale of bad #scicomm, continuing on from this thread here. https://twitter.com/Gurdur/status/1303182915501805568

2/ @SmritiMallapaty does a reasonable piece in Nature on IFR for age-groups. Not as detailed, nor as clear, as I would have liked, but reasonably informative. It mostly looks at #COVID19 IFR from Geneva, Spain and England. https://www.nature.com/articles/d41586-020-02483-2

3/ quick reminder:
IFR = number dying from the *total* number of infected. Obviously, often you don't know how many were infected, but often you can make a based estimate.

CFR = number dying out of all *confirmed* cases, usually the ones ill enough to be noticed (hospital etc.)

4/ Then @DKThomp writes in The Atlantic, ( https://www.theatlantic.com/ideas/archive/2020/09/what-young-healthy-people-have-fear-covid-19/616087/?utm_content=edit-promo&utm_source=twitter&utm_medium=social&utm_campaign=the-atlantic&utm_term=2020-09-07T10%3A00%3A24), the below.

And clearly, *something is wrong*. Because, as we saw in the previous thread, the risk for someone under 35 is (roughly) one in ten thousand, not one in one thousand.
Whoops.

5/ And the link he gives in his Atlantic article goes to the Nature article, the one first mentioned in this thread.

Now when I looked at that error, I tried to track how he might have made the error. At the same time as watching TV, tired out of my tiny mind. Bad idea.

6/ At first, I did some rough calculations (while watching TV at the same time, bad I know), & I thought @DKThomp had wrongly said "under the age of 35" when he meant "under the age of 50". I was challenged on that by @Mr_Steakhouse; I rechecked. And yes, I made an error myself

7/ So now, trying to recreate how @DKThomp might have made that error, I think it's much more likely that a zero simply got left off. One out of a 1,000 should have been one out of 10,000. Or the error might have come from somewhere else.
That's the second section concluded.