Intensive longitudinal research designs are increasingly used to study personality processes. The resulting data can be highly informative in ways that other data cannot, but these data also pose statistical challenges. Most often a multilevel or mixed effects modeling approach is adopted which is appropriate but may not be optimal. Surprisingly little attention is given to reliability of measurement, and the models often lack adequate complexity to test theoretical questions of interest. These limitations can be addressed with multilevel structural equation modeling (MSEM), which weds the ability to deal with nested data structures with the strengths of structural equation modeling (e.g., latent variable models, multiple outcomes and mediators). This article provides a gentle introduction to MSEM for personality researchers. Following an initial review of the relevant challenges facing researchers interested in studying personality using intensive longitudinal data, basic issues in MSEM are summarized, and a series of example models are presented. The online supplementary material provides Mplus syntax for the models presented.
Using structural equation modeling to study traits and states in intensive longitudinal data.
