Testing whether marginal effects are statistically different from one another

[Tomas Milo]

Hello,

I am using the "margins" command following a Tobit model and I would love to know if there is a test statistic that I can use to compare the values of the marginal effects.

Below is the Tobit regression:

tobit io_for_us_t1 treat#post logfirmage yearlyreturn roa lnmv shareprice lnreturnvolatility lnturnover i.fyear, ll(0)

Below is the marginal effects code and the output of marginsplot:

margins, dydx(treat) at(post=(0,1)) vsquish

tobit graph.gph

To me it seems that going from post=0 to post=1 is not statistically different. How do I test this?

Thank you.

[Tomas Milo]

I think I can use margins contrast (https://www.stata.com/manuals13/rmar...rgins,contrast). I will leave this up for future users that are interested.