> Sample size is not everything "Literary Digest poll was also one of the largest and most expensive polls ever conducted, with a sample size of around 2.4 million people" The large size by itself does not guarantee correctness https://www.math.upenn.edu/~deturck/m170/wk4/lecture/case1.h...
But isn't that really only the case when you're trying to use statistical inference to generalize from a sample? It seems like the right way to think about this poll is as a direct measurement of a floor of support for the challenger (others have said that is >2x the her official vote count), and the right statistical question is to ask is what's the probability that the official results are true given that floor.
People who can vote officially and telegram accounts are not the same thing.
I'm sure we even teenagers could have found multiple phone numbers to vote in the telegram poll.
We can't exclude the possibility that If I were a foreign power with a military budget measured in trillions then the telegram poll would say exactly what I wanted it to say.
But isn't that really only the case when you're trying to use statistical inference to generalize from a sample? It seems like the right way to think about this poll is as a direct measurement of a floor of support for the challenger (others have said that is >2x the her official vote count), and the right statistical question is to ask is what's the probability that the official results are true given that floor.