When Unobserved Heterogeneity Presents in Cross-Lagged Panel Models

• In Event: 15.059 - Fairmont Royal York Roundtable Session One
  In Roundtable Session: 15.059-7 - Specialized Structural Equation Modeling Applications: Heterogeneous Cross-Lagged Data, Generalized Structured Component Analysis, and Item Wording Effects

Abstract

Recently Curran et al. (2014) and Hamaker et al. (2015) proposed a way to separate within and between effects in cross-lagged panel model (CLPM) framework. Although many empirical studies have begun to employ their methods, the reliability of within estimates under several conditions are not fully investigated. In this study, we thus investigate the within estimates of fixed effect CLPM (FE-CLPM), random intercept CLPM (RI-CLPM), and latent curve model with structural residuals (LCM-SR), and whether they provide stable and precise within estimates in the presence of unobserved heterogeneity. Our simulation result shows that the coefficients of within estimates of RI-CLPM and LCM-SR are sensitive to the inclusion of time-invariant covariates and the presence of unobserved heterogeneity.

Authors

• Baeksan Yu, University of Pittsburgh

• Richard James Correnti, University of Pittsburgh