
John M. answered 12/09/16
Tutor
5.0
(87)
Statistics Master's with 17 years of Teaching Experience
Please note that without a standard deviation or the desired error, I can only give you a formula to use.
In confidence intervals, the error is always E = tcrit(σ/√n). If we solve this for n, we get n=(tcritσ/E)2.
For example, if the σ = 5 and the desired error E = 1.5, then the sample size would be
n = (2.704•5/1.5)2 = 81.24 or 82. Please note you always round up.