Parametric Tests of Perfect Judgment Ranking Based on Ordered Ranked Set Samples

Abstract

We develop parametric and location-scale free tests of perfect judgment ranking based on ordered ranked set samples. The tests are based on the differences between the elements of the ordered ranked set samples and those of the original ranked set samples. We compare our proposed tests with the best existing tests of perfect judgment ranking in the literature by using Monte Carlo simulation. Our simulation results show that the proposed tests behave favorably in comparison with their leading competitors, especially under the fraction of neighbor rankings model. In comparison to the nonparametric competitors, the proposed tests have the advantage of not needing randomization to attain a specific size.