Comment on: the mean value of the W/sub u/ statistic

In a test to identify the uniform distribution, S.W. Cheng and F.A. Spiring (IEEE Trans. Reliability, vol. R-36, p.98, Apr. 1987) develop some theoretical results associated with the W/sub u/ test statistic. The central values obtained by their formula do not fit the quantiles of W/sub u/. The present authors derive three approximations for finding the mean value of W/sub u/ and compare the results with those of Cheng and Spiring. A table of simulated values of W/sub u/ is given.<
>