Imagine you have a simple one-factor measurement error model - just one latent variable u with say 5 indicators. How would you make the residual variance of the indicators a function of u? For example, a linear function of u. Is this possible using OpenMx?
Copyright © 2007-2024 The OpenMx Project