Improved hypothesis testing in a general multivariate elliptical model (original) (raw)

This paper presents improvements in hypothesis testing specifically within the context of a general multivariate elliptical model. Addressing limitations of traditional likelihood inference methods, it leverages Barndorff-Nielsen's and Skovgaard's adjustments to provide a more accurate test statistic with reduced type I error probabilities, particularly in small to moderate sample sizes. Through a comprehensive simulation study and real data applications, the modified tests demonstrate enhanced performance over conventional methods.