Chapter 8: A Model for Cross-Classified Nested Repeated Measures Data

It is becoming increasingly common for educational researchers to measure students’ change in attributes, skills, or achievement over time while concurrently collecting data about those students’ teachers or classrooms. Multilevel modeling has been effectively utilized in this scenario to adequately summarize student growth and investigate determinants of this growth originating at both the individual and teacher levels. Embedded in this scenario is a methodological challenge that has yet to be addressed in the literature. Measuring student growth over time typically necessitates that students are changing classrooms, and likely but not always, changing teachers. An interesting nuance involves teachers themselves also being measured at multiple times each year, and these times may or may not coincide with when student measurements are gathered. Thus, a model that takes into account the multilevel cross-classification aspects of the longitudinal design and permits the modeling of nested change processes must be employed. This chapter is devoted to the modeling of this type of data. A real data example motivates the method’s development, and analyses will be used to demonstrate its utility in answering interesting substantive questions.