Regressions for each subsample of my dataset

[omarbensaid]

Hi,

I'm new to STATA so my problem is probably pretty simple but I'm stuck:

my data st consist of 3 variables: bloomberg, return and flow and I want to perform a mlogit on return and flow. Bloomberg serves as a group identifier.

bysort bloomberg: mlogit return flow works perfectly fine but how can I add another command, in my case mfx to estimate marginal effects?

bysort bloomberg: mfx... won't work since it has to be estimated after every regression (I think).

So, is there a way to define a subsample based on my bloomberg variable and run a couple of similar commands for each group?

Thanks in advance!

[Sergiy Radyakin]

Yes, there are a few ways to do that. Most appealing imho is write your subsample analysis as a byable program. See help for details and example. Best, Sergiy Radyakin

[Elan Cohen]

levelsof bloomberg
foreach lev in `r(levels)' {
  mlogit ... if bloomberg==`lev'
  mfx ... if bloomberg==`lev'
}

[omarbensaid]

Sergiy, Elan, your hints where very helpful! Thanks a lot! I managed to do what I wanted.