The mixed model is an important class of models that is traditionally analyzed by procedures of analysis of variance. Three aspects are fundamental in the mixed models: estimation and tests of the hypotheses of the fixed effects, prediction of the random effects and estimation of the components of variance. In the analysis of unbalanced mixed linear models, the estimation of variance components is of fundamental importance and it depends on the covariance structure and methods of estimation used. In such a context, this article intends to present the main methods of estimation and analysis used in the study of mixed linear models with general covariance structures in the random effects, available in the MIXED procedure, of the SAS (Statistical Analysis System).
SAS; MIXED procedure; mixed models; variance components
