Dear list members,

I am sorry to ask about this again, as my other post was presumably phrased to technically to catch attention, and I can apparently not adjust the title or question of my earlier post aftward:

I have estimated IV Probit coefficients which are positive, but are claimed to yield negative marginal effects. I think this must be wrong, but do not know how I would need to use -ivprobit- or -margins, predict(pr)- differently to get it right.

The best solution I can think of is to estimate the IV Probit manually as 2SLS, but then I am not sure what I need to do to ensure my clustered standard errors are correct, given that the second stage contains a predicted regressor?

Thank you so much and best regards, Peter