Dear Statalisters,
I’m using Stata 15.1.
I am exploring the “sem” command to estimate some mediation-analysis models where both the mediators and the outcomes are estimated with error. In particular, I have an estimate of the specific standard error for each observation for each variable. I have found out that the "reliability" option allows me to introduce uncertainty in the mediators and outcomes measures. However, I am forced to assign a reliability index (for each variables) to all observations. Is there a way to allow individual variation in the reliability of measures (of outcomes and mediators)?
I’m using Stata 15.1.
I am exploring the “sem” command to estimate some mediation-analysis models where both the mediators and the outcomes are estimated with error. In particular, I have an estimate of the specific standard error for each observation for each variable. I have found out that the "reliability" option allows me to introduce uncertainty in the mediators and outcomes measures. However, I am forced to assign a reliability index (for each variables) to all observations. Is there a way to allow individual variation in the reliability of measures (of outcomes and mediators)?
