Comparison of the Performance of Growth Mixture Model, SEM Tree, and SEM Forest in Modeling Latent Heterogeneity of High School Academic Growth

• In Event: Data Science in Education

Abstract

The current study investigated the performance of the three methods (i.e., GMM, SEM Tree, and SEM Forest) in detecting latent heterogeneity in academic growth using an illustrative subsample of the Longitudinal Study of High School of 2009. The findings showed remarkable differences in detecting latent heterogeneity across the three methods. The scale of positive performance was tilted toward GMM, as indicated by a parsimonious number of classes with more unique growth trajectories. SEM Tree and Forest, in contrast, were better in tracking the influences of covariates in the model parameters’ heterogeneity by providing more accurate measures of covariates’ importance and a detailed description of the role of each covariate at each level of the tree or the forest.

Author

• Amal Alhadabi, Kent State University - Kent