ECM APPROACHES TO HETEROSKEDASTIC MIXED MODELS WITH CONSTANT VARIANCERATIOS

This paper presents techniques of parameter estimation in heteroskedas tic mixed models having constant variance ratios and heterogeneous log residual variances that are described by a linear model. Estimation o f dispersion parameters is by standard (ML) and residual (REML) maximu m likelihood. Estimating equations are derived using the expectation-c onditional maximization (ECM) algorithm and simplified versions of it (gradient ECM). Direct and indirect approaches are proposed with the l atter allowing hypothesis testing about the variance ratios. The analy sis oi a small example is outlined to illustrate the theory.