In this illustrative section, the main objective is to examine whether the ML and REML estimators in linear mixed models yield different values of parameter estimates and approximates. The method for approximating the random effects for each subject is also displayed. Based on the prediction of the random effects, the model-based trajectories of the response variable are then displayed, both individually and group-averaged. The two datasets described in Chapter 3
continue to be used for illustrations: one concerning the effectiveness of acupuncture treatment on PTSD symptom severity and one analyzing the effect of marital status on an older person’s disability severity score.