I am testing a mediation model with two latent continuous predictors, one latent continuous mediator and one manifest dichotomous dependent variable. The default estimator for this type of analysis is a robust WLS.
The problem is that I do not get any of the usual Fit Indices but only the WRMW (Weighted Root Mean Square Residual) which is equal to 0.388.
As I was looking for some more information regarding this index I found the mplus discussion "Weighted Root Mean Square Residual" with the following posts:
bmuthen posted on Saturday, October 08, 2005 - 2:02 pm (...) A value less than 1 is generally good, (...) Linda K. Muthen posted on Friday, April 04, 2008 - 3:04 pm I would not worry about WMRMR. It is not a well-studied fit statistic and has not behaved as well as expected.
So that now I am even more confused. How can I evaluate the goodness of my model? Is there some way to get good fit indices with a binary dependent variable?