Hi Linda, I am a new user of growth mixture modeling and have a question. I'm not sure I understand the c#1 ON cg#1 command in example 8.8 (pg. 166). I have run models (c=2 and cg=2) both with and without the command and get very different results. Could you explain what the command is doing. Could you also give me a reference that helps me deterimine the most parsimonious number of growth classes to include in a model. Thanks, Don
c#1 ON cg#1 specifies the regression of the categorical latent variable c on the KNOWNCLASS variable cg. This allows the class probabilities to vary across the observed groups in the sample. So you would get different results with or without the statement.
Much has been written about how to determine the optimal number of classes. A recent paper by Bengt in the edited Kaplan book might help. We also have a long reference list for mixture modeling that you can look at.