I have clustered data on ADHD symptoms among twins and used EFA to identify the factors. However the results were not approved and I was told that not accounting for the cluster effect (family ID) is an unacceptable error, on the ground that the observations are not uniquely identifiable (whatever that means). A potential solution is to use choose robust SE method, but I do not see any such option in Stata 14. Does anyone know about any user written command or other tricks to address this?