Stability of estimates of location and scale parameters in rank regression

Date

2001-05

Journal Title

Journal ISSN

Volume Title

Publisher

Texas Tech University

Abstract

We have several objectives or goals that we want to achieve or accomplish in the thesis. One goal is to determine which scale parameter estunate of the three being examined will be the best estimator of the actual value of r for each error term distribution. We wish to see if the choice of the scale estimate has any influence on the performance on the rank regression tests. We want to perform a hypothesis test on the analysis of covariance model to compare the aligned rank transform, rank regression, and least squares tests, as well as investigate the influence of the scale parameter estimates on the rank regression test. Another goal is to produce the empirical Type I error rates to investigate the robustness of stability of the tests. For the valid tests, the empirical power calculation is used to decide on the superiority of each test.

To implement the empirical results and the sunulation study of this thesis, we used SAS, SAS/IML and SAS/MACRO languages. The code and syntax can be found in the appendix.

Description

Keywords

Ranking and selection (Statistics), Linear models (Statistics), Regression analysis, Multivariate analysis

Citation