In some of the online Mplus trainings, I found what I think might be the solution to what I am trying to model. I have four time points for two measures of reading: one letter word and one vocabulary. I have two time points of comprehension. I would like to extract classes of growth based on the constellation of these three measures. From what I saw presented as a semi-continuous growth model, this looks like it might work. Theoretically, could I extract clusters of growth using lw and pv and then use the passage comprehension as a covariate to the classes in order to include its influence.
Would I need to extract classes of growth on each measure separately and then extract classes of growth based on the class memberships? Then on top of that add comprehension as a covariate to the latent class of the probabilities of growth from lw and pv? Or can I extract classes from lw and pv simultaneously, as well as applying a covariate. To me this would look similar to example 6.16, only with classes extracted, a covariate, and both outcomes as continuous.
I can send the path diagram I created, if that would help. As always, thanks so much.