Replies to post #231271 on Biotech Values

[semi_infinite]

This was the most cogent explanation of criticism from Columbia stat prof. He is now in contact with Stanford authors about the math.

https://statmodeling.stat.columbia.edu/2020/04/19/fatal-flaws-in-stanford-study-of-coronavirus-prevalence/

[GrthzGd]

Thanks but I'm no statistician and I'm hoping someone can break down Luis Coelho's mathematical analysis of the Santa Clara study on Twitter in the following terms:

* The study's authors detected a total of 50 antibody positives out of 3300 sampled, or one out of 66.
* The Premier Biotech test the authors used has a false positive rate of one in 185.5 true negatives.

If all 3300 test subjects were negative, you'd get 18 positives, so 32 of the 50 positives were statistically significant. Couldn't one generate a range from those two figures, as the study authors did?