That worked! Setting sigma2 = “sigma1” did the trick. Thanks!! As you can see below, the model recovered all of my simulated parameters and the pp check looks great.
Before your response, I was playing around with the priors and made them very informative to see if the model could converge and I could recover the parameters. That didn’t help. But, oddly enough, when I set sigma2 = “sigma1” and used the very informative priors, that didn’t work either. Only when I set sigma2 = “sigma1” and went back to the less informative priors I used in the simulation code above, did it work. Any idea why that would be?? I would have thought more informative priors would have helped…
Also, I am calling this a growth mixture model. That is the correct terminology for this type of mixture model on longitudinal data, correct? (I haven’t used these models before and am trying to learn about them).