Dear Statalist,
I use margins, dydx(*) predict(p(0, .)) after ivtobit and get no result from Stata - it has been loading for more than an hour already. However, following Stata instructions on ivtobit postestimation and an example provided with Fictional data on hours spent on social media, I get the results almost immediately.
Therefore, I have two questions:
1) Is that normal that it takes Stata so long to get marginal effects after my ivtobit estimation?
2) Most importantly, is there a way I could decompose ivtobit into two stages and use tobit with dtobit afterwards (as it is performed immediately)? Could you please kindly share your vision of the code for such a case?
Thank you for taking your time!
UPD: I got my ivtobit marginal effects, 1.5 hours later. Out of curiosity, still would like to know is there is a manual way.
I use margins, dydx(*) predict(p(0, .)) after ivtobit and get no result from Stata - it has been loading for more than an hour already. However, following Stata instructions on ivtobit postestimation and an example provided with Fictional data on hours spent on social media, I get the results almost immediately.
Therefore, I have two questions:
1) Is that normal that it takes Stata so long to get marginal effects after my ivtobit estimation?
2) Most importantly, is there a way I could decompose ivtobit into two stages and use tobit with dtobit afterwards (as it is performed immediately)? Could you please kindly share your vision of the code for such a case?
Thank you for taking your time!
UPD: I got my ivtobit marginal effects, 1.5 hours later. Out of curiosity, still would like to know is there is a manual way.
