mediation analysis with binary mediator and truncated DV

[michelle]

Thanks for reading my question. I want to assess if a binary variable is mediating the effects of a continued IV on a single DV, the DV is an truncated variable start from 0.
IV-continued
MV-binary
DV-truncated
I use stata12, what is the appropriate command to test the model?

[maljawadi]

Hi Michelle, to say a variable is a mediator is to say that it is in the causal pathway between the DV and IV. Therefore, if you adjust for the MV, you are blocking the relationship between the IV and the DV through this variable i.e. the relationship between IV and DV should diminish and the estimate will decrease. you can test if a variable is a mediator in the same way that you test for whether or not a variable is a confounder. The only difference is the theoretical argument.
hope this helps

[michelle]

Hi Michelle, to say a variable is a mediator is to say that it is in the causal pathway between the DV and IV. Therefore, if you adjust for the MV, you are blocking the relationship between the IV and the DV through this variable i.e. the relationship between IV and DV should diminish and the estimate will decrease. you can test if a variable is a mediator in the same way that you test for whether or not a variable is a confounder. The only difference is the theoretical argument.
hope this helps

Yes, there is a common procedure to test the mediation. I checked the stata manual, there are some special commands for binary DV and MV, e.g. binary_mediation, dv() mv() iv()
So, I wanna to know whether there are any commands for truncated variable as the DV