Hi all,

I am new to LCA and am running it on STATA with the gsem command.

My GOF statistics for a 3-class and 4-class model are identical. Class #2 and #3 for the 3-class model are more or less identifical to Class #3 and #4 for the 4-class model. The difference is in Classes #1 and #2 in the 4-class model vis-a-vis Class #1 in the 3-class model. In the 4-class model, none of the posterior probabilities of class #1 are >0.5, but lcprob gives me a 14% for it.

I'm not sure if this is sufficient information, but what might be the reason for this?

Thank you in advance.

Rayner