2 Comments

Can you do this test given you are sampling from a binomial distribution?

Expand full comment
author

I think technically, I should be integrating the tails of a binomial distribution, not the normal distribution. But, since the binomial distribution is well approximated by the normal distribution for large N, I think I'm okay. My error would have to be a factor of 1,000 to reach non- statistical significance.

Expand full comment