Detecting Treatment Effects With Latent Growth Models Among Groups With Linear and Nonlinear Growth Trajectories

• In Event: Saturday Roundtable Session 11:25 am
  In Roundtable Session: Issues and Developments in Structural Equation Modeling (Table 10)

Abstract

The current study examined the performance of latent growth models (LGMs) in detecting treatment effects when the control group demonstrates linear growth and the treatment group demonstrates nonlinear growth. Three between-groups LGM approaches were evaluated, including the multiple-group LGM, the dummy-indicator LGM, and the added-growth LGM, under simulated conditions that varied sample size, the number of measurement occasions, nonlinear growth trajectories, and treatment effect size. Type I error and power rates of detecting the treatment effect and correct model selection by information criteria were assessed. The findings suggest that the multiple-group and added-growth LGMs outperform the dummy-indicator LGM. Additionally, model selection accuracy is most optimal when the control group has linear growth and the treatment group has quadratic growth.

Authors

• Natashia Bibriescas, University of Texas at Austin

• Tiffany A. Whittaker, University of Texas at Austin