Second Order Latent Growth Models and Measurement Equivalence

Research problems related to individuals' behaviors and attitudes requires examining inevitable changes over time. Because learning by nature implies change, analysis of longitudinal data becomes an important topic especially in the field of education. In this article, linear changes of a particular attribute over time was studied in the framework of the second order latent growth models by using data generated from Monte Carlo simulation. All analyses were performed by using Mplus 5.1 software. Related Mplus syntaxes were introduced and the interpretation of the model parameters was discussed. Additionally, it was explained how to study measurement equivalence in these models. Analyses were performed in three steps: (1) basic latent growth model, (2) latent growth model with weak measurement equivalence, and (3) strong measurement equivalence.