An Analysis of Change Over Time: Latent Growth Models

Abstract

Latent Growth Models which are used in understanding how individuals change over time have been a topic of intense interest among the researchers during the past two decades. These models in the framework of Structural Equation Modeling have been recommended as an alternative to classical methods such as analysis of variance. In this study, Latent Growth Models were introduced by using a Monte Carlo simulation approach and the interpretation of the findings was discussed. In addition, the effect of different sample sizes (30, 50, 100, and 200) on power and parameter estimates were examined. For this purpose; (1) data generation was performed with Monte Carlo simulation, (2) the parameters of unconditional and conditional models were estimated and the findings were discussed and (3) the effect of sample size on parameter estimates, standard errors, coverage and power was studied. All of the analyses were performed by using Mplus 5.1 software. Results were discussed in the context of advantages and disadvantages of Latent Growth Models, and the effect of sample size.