Network Estimation from Time Series and Panel Data

This chapter discusses how to estimate graphical vector auto-regression (GVAR) network models from time series and panel data. The GVAR model can be used to estimate temporal networks (within-person relationships over time), contemporaneous networks (within-person relationships in the same window of measurement), and between-person networks (relationships between the means of persons in the data). The chapter explains how such network structures can be estimated using the R-packages graphicalVAR, psychonetrics, and mlVAR. The chapter concludes with a discussion of current practical and methodological challenges, including the power of N = 1 networks, heterogeneity, missing data, model assumptions, and the importance of identifying appropriate time scales.