Good morning, friends, I'm performing a second-order confirmatory factor analysis. I'm using the WLSMV estimator. For the linear regression of the general factor on each of the other factors, which estimator does the system use: ML, robust ML, or does it decide based on homoscedasticity?