Suppose you are given data from a simple comparative experiment in which nA observations have been made with treatment A and n8 observations with treatment B. lt is necessary that nA = nB? Suppose that nA = nB and the data are paired. Will the equation for the paired data always give a shorter interval? Why not?