3 ms·
Yes, that's about right. The true FP rate is probably a fair bit lower than 2%. However, even FP rates that sound very low can still be a big problem when the
by mike_hearn 6y ago
Yes, that's about right.
The true FP rate is probably a fair bit lower than 2%. However, even FP rates that sound very low can still be a big problem when the signal itself is tiny. The signal ends up dominated by FPs, which is a problem given that a positive test is being used to trigger harsh actions, like being unable to leave your house for two weeks.
The bigger issue is the conceptual problem. What are tests for? The ultimate thing we care about is detecting ahead of time people who would get severe symptoms or die, or who would be spreading the infection (assuming we take for granted that quarantining and isolation is a reasonable policy to hold indefinitely for such a mild disease).
The test isn't actually detecting any of these things.