I have a path model in which the dependent variable is nominal (with 6 non-ordered categories). Mplus error messages led me to believe that INTEGRATION=MONTECARLO was necessary for such a model. When I run the model with INTEGRATION=MONTECARLO, no fit statistics are calculated.
What is the best way to assess the fit of such a model?
There are no absolute fit statistics available when means, variances, and covariances are not sufficient statistics for model estimation. In these cases, nested models can be compared using -2 times the loglikelihood difference which is distributed as chi-square.