For the life of me, I can't see why we haven't done random sampling of people for COVID. The self selection bias skews the data far away from any useful conclusions.