ABSTRACT
In biomedical studies, interests are often focused on evaluating the effects of treatments, medication dosage, risk factors or other biological and environmental covariates on certain outcome variables, such as disease progression and health status, over time. Because the changes of outcomes and covariates and their temporal patterns within each subject usually provide important information of scientific relevance, longitudinal samples that contain repeated measurements within each subject over time are often more informative than the classical cross-sectional samples, which contain the measurements of each subject at one time point only. Since longitudinal samples combine the characteristics of cross-sectional sampling and time series observations, their usefulness goes far beyond biomedical studies and is often found in economics, psychology, sociology and many other scientific areas.
