I'm doing a latent growth mixture model with two unobserved classes and two knownclasses that represent treatment group. When I save the cprobabilities file and do a crosstab of the treatment assignment variable with the knownclass variable I get a crossover, such that individuals who are assigned to one treatment group are incorrectly classified in the other treatment group under the knownclass variable. Based on my understanding of the knownclass option as a means for doing multiple groups analysis, I would expect the classification to be perfect.
Thanks for any advice!
bmuthen posted on Monday, March 06, 2006 - 8:15 am
Sounds like the Mplus team needs to see the details of what you are doing before answering - please send your input, output, data, and license number to firstname.lastname@example.org.