Network models for psychological constructs are becoming increasingly popular as an alternative to latent variable models for representing psychological constructs. We show how these models may be combined with crosslagged panel models to reveal longitudinal processes that occur within and across constructs over time. The proposed crosslagged network model uses regularized regression estimation to discover autoregressive and crosslagged pathways that characterize the effects of observed components of psychological constructs on each other over time. After describing the model we demonstrate its application to longitudinal data on students’ commitment to school and selfesteem.
|Journal||Multivariate Behavioral Research|
|Publication status||Accepted/In press - 2020|