But couldn't they just tell people about the uncertainty? (like "90-98% effective"?) Or underestimate the efficacy (in some objective way, like "at least 90% effective"). Or something else, as long as they're ultimately measuring what matters and what people actually care about. It's not like that 100% with their sample sizes would realistically ever suddenly turn into 50%, would it? It seems just inherently wrong to measure the wrong thing and convey it with certainty than to measure the right thing and convey it with the associated uncertainty, does it not?