ABSTRACT
Correlated data may arise as a result of repeated measurements in longitudinal studies. They may also arise from cluster sampling, such as studies of samples of families, within each there are multiple persons. The statistical issues are similar in these two situations, although the former has an additional issue of a time dimension. This chapter will mostly use the word panel to refer to a group of correlated observations, no matter if it is about repeated measurements or cluster sampling. However, in the terminology of the mixed models, it is typical to use the term “subject” to refer to a panel of observations because the context is often about repeated measurements of each person. So these words are used interchangeably sometimes.
