Accession Number : ADA135759

Title :   Testing Hypotheses in Linear Models with Weighted Rank Statistics.

Descriptive Note : Technical rept.,

Corporate Author : WESTERN MICHIGAN UNIV KALAMAZOO DEPT OF MATHEMATICS

Personal Author(s) : Sievers,G L

PDF Url : ADA135759

Report Date : Nov 1983

Pagination or Media Count : 32

Abstract : Tests of hypotheses for the parameters in a general linear model are considered based on weighted rank statistics. Results are presented for tests based on a rank estimate, tests based on drop in dispersion and aligned rank tests. Weights can be used to focus the analysis on simple effects and provide an additional degree of robustness to rank tests. Several analysis of variance applications are discussed. (Author)

Descriptors :   *Statistical tests, *Hypotheses, *Mathematical models, Weighting functions, Rank order statistics, Analysis of variance, Parameters, Linearity, Dispersions, Estimates

Subject Categories : Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE