A Monte Carlo Power Study of k-Sample Tests for Exponentiality

Abstract

**Please note that the full text is embargoed** ABSTRACT: Suppose one has a collection of k independent samples where the ith sample size is [see pdf for notation]. Let [see pdf for notation] denote the ordered observations in the ith sample. A number of test procedures are available to jointly test for exponentiality of the collection of independent samples, that is [see pdf for notation] is the probability density function (pdf) of the ith population. These include the k-sample Durbin (1975) test, the k-sample Shapiro-Wilk (1972) W-exponential test, the k-sample Tiku (1974) test, and a test procedure derived by Dyer (1979) which is based on a characterization of the exponential distribution. The Pareto distribution can be jointly tested if the [see pdf for notation] are the natural logarithms of the original observations. It is the purpose of this paper to compare the power (i.e., the ability to detect non-exponentiality) of the aforementioned test procedures.