Identifying Heterogeneous Growth Trajectories With Skew-T Growth Mixture Models

• In Event: 46.079 - Appropriate Modeling of Latent Growth and Relationships

Abstract

GMMs with skew t-distributions have the flexibility to accommodate non-normality in the data. However, up to date, there has been no simulation study to examine their performance characteristics, particularly, with respect to how well commonly used class enumeration measures can identify the correct number of latent growth trajectories with skew t-GMMs. This study intends to shed some light on this question by investigating the class enumeration capabilities of skew t GMMs under different experimental conditions. Preliminary results showed that skew t GMM can accurately identify the number of classes in the presence of high separation when sample size is 1500, but fails to do so when class separation is low and sample size is small.

Authors

• Ling Ning, University of California - Davis

• Wen Luo, Texas A&M University - College Station

• Shuqiong Lin, Texas A&M University - College Station