Simulation-Based Power Analysis in Latent Transition Analysis

• In Event: 42.035 - Statistical Power and Missing Data

Abstract

Due to the complexity of latent transition analysis (LTA) model structure and qualitative decision making in LTA, power analysis has less been reported and studied. Baldwin (2015) and Gudicha, et al., (2016) recently reported results about power analysis in LTA, which included useful information. On the other hand, both studies are limited in the number of latent statuses from two-solution to four-solution models. In this paper, we conduct a Monte Carlo simulation study that is involved with larger number of latent statuses like five-solution and seven-solution models. The results did not only include an extension of Baldwin’s and Gudicha’s results but also provided applied researchers a framework on conducting power analysis in empirical study with larger number of latent statuses.

Authors

• Ji Hoon Ryoo, University of Virginia

• Robert H. Tai, University of Virginia