Sign-up links for the Santa Clara covid19 study were shared through social media and used multiple times by people who had symptoms and wanted a test. The authors think they handed out one-time use links randomly via Facebook ads! Nope! The selection bias is through the roof!

You could fix the selection bias by comparing the number of participants who reported symptoms in the last two months and comparing to the general population. The study authors asked about those symptoms, but haven't reported the responses. They are sitting on critical data!

You could also estimate the selection bias by checking back in the survey logs (if they have them) to see how many times each link was used.

Every critical analysis I've seen so far assumes that participants were selected the way the authors say they were, and that they don't have any data about symptoms. Neither of those are true!

Now you can see the results of this study being used to support speculation that there are different strains of the virus with hugely different infectiousness and fatality rates. The study can't support that, because it's even worse than people suspect!

This study has sown immense confusion and is hugely damaging to the public understanding of the severity of this disease.

How do I know all this? I was in the study.

And I took pictures.