The Effect of Between-Level Residual Structure in Multilevel Latent Growth Model

• In Event: Poster Session for Quantitative Methods and Statistical Theory

Abstract

The purpose of this study is to evaluate the effect of restricting and freely estimating between-level residual structure in multilevel latent growth model. Under various conditions, parameter estimation, bias, and coverage were examined. The results show that the parameter estimate for the variances for the intercept and growth factor in between-level was estimated larger in restricted between-level residual structure model. However, in the random effect parameter estimates, the between-level residual variances of the intercept and slope growth factors were underestimated across conditions. The result indicates in different parameter estimates in variances for intercept and growth in between-level. Thus, it is recommended to start modeling without restricting the between-level residual variances to zero.

Author

• Suyoung Kim, The University of Texas at Austin