Near-exact Distributions for the Block Equicorrelation and Equivariance Likelihood Ratio Test Statistic

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

In this paper the authors combine the equicorrelation and equivariance test introduced by Wilks [13] with the likelihood ratio test (l.r.t.) for independence of groups of variables to obtain the l.r.t. of block equicorrelation and equivariance. This test or its single block version may find applications in many areas as in psychology, education, medicine, genetics and they are important “in many tests ofmultivariate analysis,e.g. in MANOVA, Profile Analysis,Growth Curve analysis,etc” [12, 9]. By decomposing the overall hypothesis into the hypotheses of independence of groups of variables and the hypothesis of equicorrelation and equivariance we are able to obtain the expressions for the overall l.r.t. statistic and its moments. From these we obtain a suitable factorization of the characteristic function (c.f.) of the logarithm of the l.r.t. statistic, which enables us to develop highly manageable and precise near-exact distributions for the test statistic.
Original languageUnknown
Title of host publicationAIP Conference Proceedings
Pages429-433
Volume1557
DOIs
Publication statusPublished - 1 Jan 2013
EventInternational Conference on Mathematical Sciences and Statistics (ICMSS) -
Duration: 1 Jan 2013 → …

Conference

ConferenceInternational Conference on Mathematical Sciences and Statistics (ICMSS)
Period1/01/13 → …

Cite this