Second Order Latent Growth Modeling in Ordinal Responses : Focusing on Model Specification and Evaluation

• In Event: AER Roundtable Session Friday 1:45 pm - Gold 4
  In Roundtable Session: Innovations in Item Response Theory for Modeling Student Performance and Growth

Abstract

Second-order latent growth modeling (SOLGM) extends conventional LGM by incorporating individual indicators to capture indicator-specific effects, such as common item effects. Despite its strengths, applications to ordinal data remain limited. This study applies SOLGM to ordinal responses under a graded response model (SOLGM-GRM), evaluating model fit and parameter accuracy under conditions of common item effects and violations of longitudinal measurement invariance (MI) on item slopes. Results show that including common item factors improves model fit and sustains stable and accurate model parameter estimates, and assuming MI on item slopes is generally robust unless slope drift is substantial. A stepwise approach is suggested: first test common item effects, then evaluate MI when applying SOLGM-GRM.

Authors

• Sook Hyun Park, University of Texas at Austin

• Tiffany A. Whittaker, University of Texas at Austin