I have a potentially elementary question about covariates in LCGA. We are running a LCGA for four time points. Our basic goal is to investigate trajectories of deterioration in functioning in our sample of patients, and then relate these trajectories to distal variables. Per a paper by Shaunna Clark (2009) posted on your website, we want to take into account uncertainty in class membership when testing for equality of means on these distal variables; thus, we have included these variables using the AUXILIARY command and followed them with the (e) option.
However, we have another set of covariate variables (age at time of testing, sex, race, and family SES) that we want to control for when relating trajectory classes to distal outcome variables. However, we don't want to use them as x variables using the ON command in the analysis, because we don't want class membership, or slope and intercept variables, to be informed by these covariates; we simply want to use them as covariates when relating the classes to distal outcome variables.
I'm not sure if there is any way to do this within the framework of Mplus, but I would greatly appreciate if you might have any insight on this.
Thank you so much, and I hope you have a great weekend.