The headline is broadly defensible, but the qualifications matter. Effect sizes vary by population, the strongest claims rest on shorter trials, and credible voices push back on how it's typically framed.
the converse of that is what happens if you test negative well if you test negative you can take that to the bank you are as sure as God made little green apples free of coronavirus your negative predictive value is essentially a hundred percent
Every Sunday: the week’s new conflicts and verdict changes — and nothing else.
Native comments, Twitter mentions, and Reddit threads about this claim — surfaced together so the conversation isn't fragmented across platforms.
Bookmarking — the dossier-vs-overview split is the right call. Most of the time I want overview; sometimes I want receipts.
Would love a "what would change this verdict" RSS feed. Sign me up if it exists.
because in this situation I've described with a prevalence of only 1% there are so few cases out there and because I'm not allowing for us to just test exclusively symptomatic people in fact I'm specifically targeting asymptomatic people which is a very important question you're gonna have the same number of true positives as you do true negatives
well it turns out your positive predictive value would still stink it would be only 50% meaning the likelihood that if you have a positive test you have the coronavirus would only be 50%