Why is denominator sometimes “n” and sometimes “n-1”?
if population n, if sample (n-1)
So how can you determine if a set of data is a population or a sample? In my mind, a sample would be a subset of data from which a larger population exists… population would be a set of data using all available points. If that is the case, problems in the CFA text do a poor job of offering sufficient evidence in determining between population and sample. For anyone using CFA texts, please see reading 53 problem 1C. The answer uses n (as opposed to n-1) in computing the covarience of a stock’s return over a 6 month period. How am I supposed to know to use n in the denominator?
It’s not really the case - the issue is whether or not the means for the populations are known. If they are, use n. If not, use n-1. In the real world, you would probably never know the means for the populations so use n-1. Of course, omniscient test question writers know population means.
you are actually right about that Joey. Since I work with real data I automatically use (n-1) to get an unbiases estimate but the real reason is because mean is estimated from the sample.
I am “actually right about that”? Good thing that Statistics Ph.D. isn’t totally wasted.
Sorry about how it came across, I wasn’t trying to put you down.