Quote:
Originally Posted by embee
Unfortunately it's when numbers get this low in data that the question of statistical significance becomes particularly relevant. Even with a test base of 40k people, when you get down to 1 or 2 or 3 cases, the effect of just another 1 or 2 is potentially doubling the rate, so levels of confidence become important. That's why in the Pfizer data a nominal efficacy of 52% gets a 95% credible interval of 29.5% to 68.4%, which is a pretty wide range in anyone's book. I doubt many politicians understand statistics to any detailed level.
|
Indeed, like I said any statistics around medicine are pretty dodgy......
I don't know if you've read the last BMJ? Interesting comments in there from some doctors involved in briefing the media on medical matters.........