showed that the maximum likelihood estimator of population coefficient of variation is better than the sample coefficient of variation since the sample coefficient of variation may lie outside the bounds obtained for the population coefficient of variation.