Dear Statalisters,

To carry causal Mediation analysis, I am using the Mediation package (Medeff command). This command does not support the "mi" prefix used for imputed data in order to apply the Rubin rules and to calculate the estimates.

I thought that an older command "mim" could do the job but again unfortunately I was not able to find my way.

Do any one of you faced the same situation ? any help ?

It seems that a solution exist as mentionned in this paper, but I don'y know how to apply it.

https://www.ncbi.nlm.nih.gov/pmc/art...5216460/#ref43

Your support is really appreciated.

Regards.