Distance and Shape: Factors Influencing Estimation of Piecewise Growth Mixture Model

• In Event: Marriott Roundtable Session 73
  In Roundtable Session: Precision, Power, and Small Sample Size Issues in Growth Curve Modeling Variations

Abstract

The piecewise growth mixture model (PGMM) is used in longitudinal studies to tackle the non-continuous trajectory and unobserved heterogeneity in a compound way. The present study aims at finding how PGMM is influenced by the potential factors such as latent distance and shape. Simulation results showed that the larger latent distance yielded better model fit and higher estimates precision. An unparallel shape slightly led to better overall model fit. The shape influenced the parameter estimates probably due to the parameter difference in the true value setting, where larger parameter difference yielded biased estimates.

Authors

• Yuan Liu, Southwest University

• Hongyun Liu, Beijing Normal University