The statistical methods for estimation and inference which are based on a function of sample observations, probability distribution of which does not rely upon a complete speci?cation of the probability distribution of the population from which the sample was drawn. Consequently the methods are valid under the relatively general suppositions about the underlying population. Number of times such methods include only the ranks of the observations rather than observations themselves. The example is Wilcoxon's signed rank test and Friedman's two directional analysis of the variance. In number of cases these tests are marginally less powerful than their analogues which suppose a particular population distribution even when the supposition is true. This is also generally called as the non- parametric method although the terms are not completely synonymous.