I'm new to mixture modeling and MPLUS, so apologies if this question is misguided. I am currently working through a dataset and learning off of Higher-Order Growth Curves and Mixture Modeling with MPlus.
If when you run an unconditional latent class growth model in MPLUS and the variance of the slope and intercept is insignificant but model fit is horrible, would there be reason to continue onto latent class growth analysis / growth mixture models to identify sub-trajectories?
TITLE: Unconditional LGCM DATA: FILE= Phys_traj.dat; Variable: NAMES ARE ID MX1 MX3 MX6 MX12; USEVAR = MX1 MX3 MX6 MX12; MISSING ARE ALL (99); ANALYSIS: type = missing H1; MODEL: I S | MX1@0 MX3@2 MX6@5 MX12@11; OUTPUT: SAMPSTAT STANDARDIZED TECH1;
How do you interpret the R-squared in output for MX1, MX3, MX6, and MX12?