How to Select the Best Fit Model among Bayesian Latent Growth Models for Complex Data


  • Laura Lu University of Georgia Author
  • Zhiyong Zhang Author



Model Selection Criterion, Bayesian Estimation, Latent Growth Models, Missing Data, Robust Method


Bayesian approach is becoming increasingly important as it provides many advantages in dealing with complex data. However, there is no well-defined model selection criterion or index in a Bayesian context. To address the challenges, new indices are needed. The goal of this study is to propose new model selection indices and to investigate their performances in the framework of latent growth mixture models with missing data and outliers in a Bayesian context. We consider latent growth models because they are very flexible in modeling complex data and becoming increasingly popular in statistical, psychological, behavioral, and educational areas. Specifically, this study conducted five simulation studies to cover different cases, including latent growth curve models with missing data, latent growth curve models with missing data and outliers, growth mixture models with missing data and outliers, extended growth mixture models with missing data and outliers, and latent growth models with different classes. Simulation results show that almost all proposed indices can effectively identify the true model. This study also illustrated the application of these model selection indices in real data analysis.

Author Biography

  • Laura Lu, University of Georgia

    Dr. Lu has been rigorously trained as a quantitative methodologist. Before she joined the Quantitative Methodologies (QM) program at the University of Georgia (UGA), she obtained a PhD on Quantitative Psychology at the University of Notre Dame (ND) and a MA on Mathematics at Temple University. The training she received has provided her a solid grounding in mathematics, statistics, programming, and quantitative methodologies. In general, she has expertise in structural equation modeling (SEM), longitudinal data analysis, hierarchical linear modeling (HLM), and computational statistics such as Bayesian. During her ten years' professional career at UGA, her research has focused on developing innovative statistical approaches to address perennial challenges in statistical modeling such as mixture structure, reliabilities, model selection, missing data, outliers, topic modeling, and also promoting statistical models to applied research areas through collaborating, mentoring, and classroom teaching.






Theory and Methods