Latent Transition Analysis Auxiliary Variables Methods and Misspecification

• In Event: Methods for Mixture Modeling and Classification

Abstract

Latent transition modeling offers the opportunity to explore the probability of transitioning between classes across time while factoring in the impact of auxiliary variables. Yet, there is no agreed upon standard for including auxiliary variables into these models or diagnosing if the variable influences the transition or only predicts classes at time points. This simulation study examines current practices implemented in other finite mixture models examining the incorporation and misspecification of time invariant auxiliary variables. Preliminary results are discussed in reference to the importance of correctly specifying auxiliary variables in latent transition analyses.

Authors

• Joseph E Miller, The University of Texas

• Tiffany Ann Whittaker, The University of Texas at Austin