ABSTRACT
Longitudinal data, or repeated measures data, can be viewed as multilevel data with repeated measurements nested within individuals. In its simplest form, this leads to a two-level model, with the series of repeated measures at the lowest level, and the individual subjects at the highest level. Longitudinal measures can be taken at fixed or at varying occasions. Multilevel analysis for longitudinal data can handle both situations. Since multilevel modeling does not require balanced data, it is not a problem if the number of available measurements is not the same for all individuals. This is an important benefit if there is (non-informative) panel dropout, or other forms of missing measurements within individuals. Since longitudinal data collected at fixed occasions is the simplest situation, this chapter starts with fixed occasions, and discusses varying occasions later. The chapter ends with a discussion of statistical issues specific to longitudinal modeling.
