Latent Growth Modeling

Latent growth modeling (LGM) refers to a set of procedures for fitting statistical models to longitudinal or repeated-measures data. The term latent implies that these procedures are implemented using a [Page 790]latent-variable structural equation modeling (SEM) framework to specify and estimate models. In LGM, the latent variables are often referred to as growth factors, which represent the systematic features of an observed variable’s longitudinal trajectory across time. In many situations, a latent growth model specified and estimated using the SEM approach is equivalent to a multilevel model (i.e., mixed-effects model or mixed model) for the same longitudinal data; the means of the growth factors are analogous to fixed effects, and the variances of the growth factors (and their covariances) are analogous to random effects. This ...