A combined test for differences in scale based on the interquantile range

摘要: A class of tests due to Shoemaker (Commun Stat Simul Comput 28: 189-205, 1999) for differences in scale which is valid a variety both skewed and symmetric distributions when location known or unknown considered. The based on the interquantile range requires that population variances are finite. In this paper, we firstly propose permutation version it does not require condition finite remarkably more powerful than original one.Secondlywesolvethequestionofwhatquantilechoosebyproposingacombined test our tests. showed extreme less ones, unless underlying very highly skewed. Since practice you may know if not, question arises. combined solves question, robust stand alone Thirdly conducted much detailed simulation study (1999) compared his F squared rank showing better. good scale, results suffer such drawback, reason instead considering consider, following suggestions several authors, Brown-Forsythe (J Am Assoc 69:364-367,1974), Pan 63:59-71,1999), O'Brien 74:877-880, 1979) Conover et al. (Technometrics 23:351-361, 1981).