[s1artibartfast]

I think the goal should be the most concrete data we have. A debate could be had on if self reporting or test data is more accurate. One factor to consider is if the self reporting data is being represented in a way commonly reserved for test data. Another factor to consider is that 33% of Americans think they may have had covid and 13% think they probably did [1]. This is grossly out of agreement with serology results.

https://www.businessinsider.com/small-percentage-americans-w...

[makomk]

In this case, the media was using the self reporting data to claim that Florida had Covid-19 cases as early as December (and it's this that Florida was supposedly firing her for refusing to cover up): https://www.miamiherald.com/news/state/florida/article242480... Needless to say, people self-reporting Covid-19 symptoms that early is the most unreliable self-reporting data of all due to the low or maybe even zero number of genuine cases and the much larger prevalence of things like the flu. If I remember rightly, researchers have tried serology tests on people who thought they had it that early and found that basically none of them had.