An Examination of Parameter Recovery in Latent Transition Models With Distal Outcomes Using the Three-Step Approach

• In Event: Poster Session 15
  In Poster Session: Important Issues in Model Selection, Measurement Invariance, Reliability, and Mixture Models

Abstract

Latent transition analysis (LTA) is an increasingly popular research method
used to categorize subsets of individuals within a population. given their prevalence in applied research The current study sought to investigate parameter recovery of a distal outcome effect in LTA models for the first time using the three-step approach with Monte Carlo methods. The outcome of interest was the difference between the estimated and true distal outcome effects expressed as a percentage of bias. The manipulated design factors were sample size, number of indicators, transition effect, class prevalence at time 1, and true distal outcome effect size. The findings from our study suggests that, on average, distal outcome estimation is accurate, but the precision is another story. Caution is advised.

Authors

• Courtney Moore

• Grant B. Morgan, Baylor University