On the statistical error of the half energy width

The half energy width (HEW) is the all powerful performance parameter used to characterize and compare X-ray optics. Values of HEW are always reported without an error bar or confidence interval, suggesting that the statistical error associated with such estimates are negligible. Is this true? And if it is, could one characterize the optics more quickly accepting a non-negligible but still acceptable statistical error in the estimate of the HEW? We try to answer these questions with the use of non-parametric statistical methods and X-ray pencile beam data. .