A Note on Confidence Intervals for Proportions in Finite Populations