If I save the data from the GMM and export it to SPSS or Excel, I automatically get the latent classes based on their most likely latent class membership. As I read that you advise to use classes based on the estimated model, I was wondering if there is a possibility to directly export the classes based on the estimated model instead of their most likely class membership?
As a follow-up question: are the model results based on the estimated model? Or will the model results not be influenced by the choice of classes based on estimated model versus on most likely class membership?
In the estimated model, each subject has a certain probability of belonging to each of the classes. So exporting classes based on the estimated model is probably not what you want. I wonder what you are going to do with the exported classes - perhaps you should look into the "3-step" approaches where the classes using the estimated model are related to other variables not used in the latent class analysis. See Mplus Web Notes 15 and 21 on our web site.