CFE 2019: Start Registration
View Submission - CMStatistics
B0903
Title: Testing multiple means in two-level compound symmetry multivariate data Authors:  Ivan Zezula - P.J. Safarik University (Slovakia) [presenting]
Daniel Klein - P.J. Safarik University (Slovakia)
Abstract: A simple multivariate model $X=M+E$ is considered, where $M$ is a location or mean matrix, and $E$ is an error matrix. The variance matrix of $E$ is assumed to be doubly-exchangeable, i.e. block-wise compound symmetry. We develop several test procedures for multiple testing of possibly different mean matrices sharing the same variance matrix. Properties of different tests are compared via simulations.