Correlated data are typically generated from studies where the outcomes under investigation are collected on clustered units. Specific examples include (1) longitudinal data where outcomes are collected on the same experimental unit (for instance, the same person) at two or more different points in time and (2) studies where outcomes are recorded at one single point in time on clustered units. Such studies have one major attraction, the ability to control for unobserved variables in making inferences. Sampled units serve as controls for other units in the same cluster. As an example, in a longitudinal study, each subject serves as his or her own control in the study of change across time. Therefore, these studies allow the researcher to eliminate a number of competing explanations for observed effects. The determination of causal ordering in making solid inferences constitutes another attraction for longitudinal studies.