Yuli Liang & Chengcheng Hao: Testing variance parameters in models with a Kronecker product covariance structure: using higher-order asymptotics
Time: Wed 2014-02-05 13.00 - 14.00
Location: Room B705, Department of statistics, Stockholm university
Inference which relies on the first order of approximation, such as Score test and the likelihood ratio test, may yield inaccurate results in some small sample applications. For small sample problems, more accurate testing procedures such as Barndorff-Nielsen (1991), DiCiccio et al. (2001) and Skovgaard (2001) have been developed by constructing pivot quantities to a higher order of approximation. In this work, we consider a random vector normally distributed of which the covariance matrix is a Kronecker product of two compound symmetry matrices. Testing the variance parameters based on n independent samples with the knowledge on higher order asymptotic is illustrated.
DiCiccio, T.J., Martin, M.A. and Stern, S.E. (2001). Simple and accurate one-sided inference from signed roots of likelihood ratios. Canadian Journal of Statistics, 29, 67-76.
Skovgaard, I.M. (2001). Likelihood asymptotics. Scandinavian Journal of Statistics, 28, 3-32.