Mediation analysis on multiply imputed data with binary dependent outcomes

[Kunle Oginni]

I am new to STATA and I am trying to carry out mediation analyses. The data was imputed using multiple imputation by chained equations, the outcome and mediator variables are binary while the covariates are both continuous and binary. I was going to use paramed but could not combine it with the mi estimate command, I also wanted to use sureg but found that it was designed for continuous dependent outcomes.

Is there any way of carrying or mediation (or path analysis) on multiply imputed data with binary outcomes in STATA?

Thank you.
Kunle.

[Kjell Weyde]

Hi, have you found any solutions to your problem?

Kjell