Compare multiple regression coefficients across regression models

[Huong Quynh]

Hi guys,

I want to compare coefficients across two different regression models. I found a useful instruction using suest.

https://stats.idre.ucla.edu/stata/fa...r-more-groups/

However, this only has 1 independent variable. I have two independent variables and I want to see if their effects vary across two groups.
I don't know if that is possible and how to modify this to test that.


I am quite new to statistics so I hope you guys can help
Thank you so much.

[Jackson Monroe]

If you mean doing regressions within subgroups of your data then "regress" can do that with multiple independent variables just as well as it can with only one. You might find this https://www.stata.com/manuals13/dby.pdf useful. If you want to regress on multiple variables the syntax is in the documentation, but it is basically

regress DependentVariable IndependentVariable1 IndependentVariable2 IndependentVariable3

... etc.

You can also access documentation with

help by

help regress

If you want to use suest you can do

regress DependentVariable IndependentVariable1 IndependentVariable2 IndependentVariable3 if group==group1

estimates store group1est

and then you call

suest group1est ...

where the ... are other names similar to groupest1, the name Stata associates with your estimates from regressions on different subsets of your data.