ABSTRACT
This chapter proposes to introduce a variant of the latent change score model that can be used as a possible alternative discrete dynamical model for analyses of longitudinal data. In John J. McArdle's dual change score model, latent change is determined by two components: an individualized time invariant latent constant score and a feedback effect of the time dependent preceding latent score. The chapter proposes an alternative latent change model to deal with repeated measures with such nonlinear characteristics. Bivariate latent quadratic curves are constructed simply by a system of two latent change scores which are determined by two components: an individualized linear constant, and an individualized quadratic constant with a time variant coefficient. In the prevailing and mainstream description of the latent change score model, the additive constant is treated as a time-invariant normally distributed latent variable and each individual is assigned a distinct unique quantity.
