09-24-2014 06:53 AM
I am new to structural equation modelling. The model I would like to test has several latent variables. Some of the indicator variables are categorical while others are continuous. Many of the continuous variables have ceiling and flooring effects. Missing data is present in the dataset. Can the Proc Calis command of SAS handle such a model or do I need to move to Mplus to estimate such a model? Should the WLS estimators be used to estimate the parameters? I do not see the weighted least square corrections to means and variances (WLSMV) option which is recommended for this type of data? Which estimator would be best to use?