Adding treatment interactions to mtreatreg

[Wojciech Hardy]

Hi all,
I have a question about the use of mtreatreg. In particular, I would like to add interactions between some of the variables and the treatments to the outcome regression.

Is this doable? mtreatreg does not allow me to specify any interactions - I'd have to hardcode them by generating new variables. This approach provides me with estimation results, but I'm not sure if I should trust them.

In my code, this looks more or less like this:

gen x1Xtreatment1 = x1*treatment1 // where treatment1 and treatment2 are possible levels from the 'treatments' variable
gen x1Xtreatment2 = x1*treatment2

mtreatreg y x1 x1Xtreatment1 x1Xtreatment2, mtreatment(treatments = x1 x2) density(normal) vce(cluster country) simulationdraws(25)

So the question is: when I add user-generated interactions with treatments to the outcome equation in mtreatreg, are the acquired coefficients unbiased and consistent? If not, is there a different approach I could take?

Thank you and kind regards,
WH

[Phil Bromiley]

Please read the FAQ on asking questions.

mtreatreg appears to be user-written. That may be why it doesn't handle interactions. You didn't get a quick answer - you'll probably have to ask the author.