Latent growth model ($Y_{it}=a_i+b_{i}t+{ϵ}_{it}$)

A longitudinal model in which unobserved intercepts, slopes or other trajectory parameters describe individual differences in change. The displayed linear example allows people to vary in both initial level and slope; actual trajectories need not be linear.

A model combining latent trajectory classes with a model for variation within classes. Class count, within-class variances, error assumptions and measurement timing affect the resulting patterns. A statistical class is not by itself an explanation of its members' adaptation.^[ref]