How can you prove that the finite population correction factor when applied to the sample mean should be
I believe that you start off by considering the cases where the random variables $x_i$ in the sample are not independent, so that the additivity of variance must then include a covariance term. However, where do you go from there?