There does seem to be some conflicting information. The report I read said the effective rating was based on blood work, not who did/didn’t get sick. Even the Pfizer press release is vague.
Who knows, maybe the report I read was wrong. But it’s exceeding difficult to demonstrate 100% with attribute sampling. The attribute sampling table I’m looking at says you need a sample size of over 9000 to have 95% confidence of the actual rate with an error of 1%.
So I’ve had what the first report stuck in my head as correct (which I can’t find, dammit).
If anyone is wondering what I’m going on about, try this quick experiment: Flip a coin 5 times. Calculate the percentage heads vs tails. Congrats! Your data now proves your coin is not 50/50. That percentage calculation is precisely the same math as saying the vaccine is 100% effective if no kids got sick. Turns out a different math is needed: statistics.
