Abstract: In this paper, we aim to compare the robustness and the
power of Hotelling’s [3], with other recent type statistics such as the
statistics introduced by Tiku and Balakrishnan [5] and other corresponding
non-parametric test statistics, such as the statistic of Utts and Hettmansperger
[6]. A simulation study is used for small samples, in different multivariate
populations, through a simulation study – using the
Monte Carlo
method – to test the stability of type I error and the stability of the power
for the mentioned measures.
Keywords and phrases: robustness, power
Hotelling’s type statistics, simulation,
multivariate populations, Winsorizing proportion.
