
David B. answered 04/15/21
Math and Statistics need not be scary
Have to make a LOT of assumptions.
I'll assume that the samples will be taken by simple random survey with a dichotomous choice.
Null hypothesis is that there is NO Preference between two drink preference choices.
Alternate hypothesis is that either one or the other drink choice is preferred.
Alpha is 2% (Confidence = 98%).
Error tolerance = margin of error (moe)
"determine the sample" is a typo error and should be "determine the minimum sample size"
The drink choices don't matter but for those who like to have that we will assume one choice is wine/beer and the other choice is "hard liquor"
Using cochran's formula. n0 = Z2pq/e2
(e = error tolerance [.05], Z = critical z for two sided 98% CI or 2.327)
p = .5 (non-assuming prior). q = 1-p = .5
minimum sample size = 542
note: if we were looking for a moe of 5% and an alpha of 5% the answer would be only 385