I am trying to predict a categorical distal outcome in the context of a 3-class model. If I am understanding Example 8.6 correctly, you simply treat the distal outcome as another indicator of the latent classes? Is this true? If so, do you then have to constrain the conditional response probabilities to equivalence to determine if they statistically differ? If not, what syntax would I use to predict this binary outcome with the categorical latent variable? Thanks!
You add the distal to the USEVARIABLES list. It is another latent class indicator. The effect is the varying of the threshoolds of the distal across classes. You can use MODEL TEST to test if these differences are significant.