ABSTRACT
Modern applications in the field of sociology and economics, and in many other fields, are frequently based on longitudinal (or panel) data. These data arise from the repeated observation of the same subjects, or more generally sample units, at a certain number of time occasions. From many points of view, longitudinal data are similar to time-series data, even if the contexts of application are usually very different. The main difference is that time series usually refer to only one unit observed at many time occasions, whereas longitudinal data refer to several units observed at few time occasions.
