28. Expectation and variance of Wilcoxonstatistic. If the X 's and Y's are samples from continuous distributions F and G respectively, the expectation and variance of the Wilcoxon statistic V defined in the preceding problem are given by (57) E( ~) = P{ X < Y} = f FdG and (58) mnvar(~) = fFdG+(n-l)f(1-G)2 dF +(m - 1)f F2dG - (m + n - 1)(J FdGr Under the hypothesis G = F. these reduce to