Using Dynamic Structural Equation Models for Analysing Intensive Longitudinal Data from Online Measures in Education

This chapter presents structural equation models (SEM) to analyse the dynamics of intensive longitudinal data nested in individuals obtained in online measures in educational research. We introduce multi-level SEM under the Bayesian statistical framework, including testing for stationarity, model and prior specification, model evaluation, and presentation of the findings. We then illustrate the approach with an example and provide computer code in Mplus and R for the individual analytical steps and demonstrate how observation and heart rate data obtained via electrocardiography (ECG, an indicator of cognitive stress) from a deep-reading study of N=1 can be meaningfully matched and analysed. The methods we explore have implications for future research seeking to appropriately model intensive longitudinal educational data, and in turn, improve substantive and practical conclusions that can be drawn from educational research.