Over-time, repeated measures, or longitudinal data are terms referring to repeated measurements of the same variables within the same unit (e.g., person, family, team, company). Longitudinal data come from many sources, including self-reports, behaviors, observations, and physiology. Researchers collect repeated measures for a variety of reasons, such as wanting to model change in a process over time or wanting to increase measurement reliability. Whatever the reason for data collection, longitudinal methods pose unique challenges and opportunities. This chapter has three main goals: (1) to help researchers consider design decisions when developing a longitudinal study, (2) to describe the different decisions researchers have to make when analyzing longitudinal data, and (3) to consider the unique properties of longitudinal designs that researchers should be aware of when designing and analyzing longitudinal studies. We aim to provide a comprehensive overview of the major issues that researchers should consider, and we also point to more extensive resources.

The two statistical approaches commonly used in the analysis of dyadic and group data, multilevel modeling and structural equation modeling, are reviewed. Next considered are three different models for dyadic data, focusing mostly on the very popular actor–partner interdependence model (APIM). We further consider power analyses for the APIM as well as the partition of nonindependence. We then present an overview of the analysis of over-time dyadic data, considering growth-curve models, the stability-and-influence model, and the over-time APIM. After that, we turn to group data and focus on considerations of the analysis of group data using multilevel modeling, including a discussion of the social relations model, which is a model of dyadic data from groups of persons. The final topic concerns measurement equivalence of constructs across members of different types in dyadic and group studies.