I didn't agree with early claims that large % were already seropositive, but did like that they were making a testable prediction: serology would eventually indicate whether the theory was correct. Think we need more of this in public discussions... 1/
If people think that pre-existing cross-reactive cellular or humoral immunity is reducing susceptibility, then what assay/samples would be required to conclusively show this? 2/
If people think that heterogeneity in contact network structure is dramatically reducing the herd immunity threshold, what contact data and seroprevalence would show this? 3/
If people think the IFR is actually higher/lower than 0.5-1%, what numerator and denominator data would give a conclusive lower/upper bound that would show this? 4/
If people think the proportion asymptomatic (not presymptomatic) is outside the 20-80% range, what community testing design or meta-analysis would be sufficient to show this? 5/
Vague claims about unseen infections can be shifted post-hoc, but this doesn't improve our knowledge. To tackle COVID, we need to define our questions clearly and define (and run!) the studies that will help us answer them. 6/6
