Analyses of Longitudinal Models: Effects of Individually Varying Times of Observation

• In Event: 29.070 - Issues in Longitudinal Studies

Abstract

In latent growth modeling (LGM), time of observation is specified as fixed factor loadings, whereas in multilevel modeling (MLM), time is a metric variable. When times of observations are individually-varying in longitudinal studies, logically, fixed specified loadings lead to inaccurate estimation. Using piecewise growth modeling, this simulation study showed that (i) individually-varying times of observations with larger boundaries resulted in biased estimates and model fits when LGM was used; (ii) across all the simulation situations, estimating with PGM was robust within MLM, whereas LGM got identically good estimation with MLM only when time boundaries are ±1 months or shorter (in year as measurement unit); and (iii) larger changes of slope in piecewise modeling resulted in better estimation.

Authors

• Yuan Liu, Southwest University

• Hongyun Liu, Beijing Normal University

• Kit-Tai Hau, Chinese University of Hong Kong

• Xiaofang Wang, The Chinese University of Hong Kong