Linear Mixed Models II
In this chapter, we extend the methods of the previous chapter in several ways. First, we will relax the assumption of conditional independence of the responses given the random effects by introducing serial autocorrelations for longitudinal data. Then, we will discuss relaxing the assumption of constant variances of the random error terms. Finally, we consider models for multilevel data in which there are several levels of clustering. Examples are repeated measures on subjects in hospitals, patients in wards in hospitals, or patients of GPs in cities.
KeywordsHearing Loss Linear Mixed Model Peak Expiratory Flow Conditional Independence General Health Questionnaire
Unable to display preview. Download preview PDF.