Thank you so much to anyone who can give any advice.
We have fit various models on testing data with a mixture of binary and polytomous items. There is a large amount of test form related structural missingness but this is not a worry as the sample is very large and the different test forms were randomly given to participants.
We used MLR as we wanted to get likelihood-based fit statistics (AIC, BIC, etc.), but we also wanted to report covariance matrix bases statistics too (RMSEA and CFI - specifically). So following the advice* here we also fitted the model using WLSMV to get the RMSEA and CFI. As such we decided to use WLSMV to extract the covariance matrix related fit statistics.
QUESTION: A review has told us that we can get RMSEA and CFI through fitting with MLR. Would it be possible to give us a steer on how to do that?
*"Maximum-likelihood estimation with censored variables does not have means, variances and covariances as sufficient statistics, but instead raw data, and therefore does not do the usual model test of fit. If you want a test of the fit to the covariance matrix [i.e., RMSEA and CFI] you can use WLSMV.”
I don't think it is possible to obtain RMSEA and CFI with MLR and categorical data (just with continuous data). Also not that if you want to compare the MLR model and the WLSMV model you will need to use analysis: link=probit; estimator=mlr; and analysis: parameterization=theta; estimator=wlsmv;