My goal is to compare the AMEs between two subpopulations of my sample in a logit model. However, there are two different moments in my analysis I can think of on where I can divide my sample into those two subpopulations, leading to very different results.
In the example below y would be my binary dependent variable; a and b are covariates and subpop is a dummy, dividing my sample into those two subgroups.
Possibility 1:
forval i=0/1 {
logit y a b if subpop==`i'
margins, dydx(a b)
}
Possibility 2:
forval i=0/1 {
logit y a b i.subpop
margins if rich==`i', dydx(a b)
}
I know for the latter code, the over() and the subpop() command can be used as well, which does not change the results with my data. What does change the magnitude of the effects of a and b on y tremendously is whether I divide my population before the logit command is executed or after the logit command is executed (i.e. before the margins command is executed). So my question is, which option is the “correct” one or rather how do the two options differ in the interpretation of results?
Thanks a lot in advance!
Related Posts with When to divide in subpopulations, using the margins command (AME) in a logit model?
clustering error in gravity modelHi stata users, Currently, I'm researching for my thesis about the impact of non-tariff measures on…
Dummy trap(intercept) removal in stata. xtreg (AB9 Dne9* Dnr9* Dpe9* Dnr9*) note: Dnr9 omitted because of collinearity Random-effects GLS…
S-GMM interpretationHello everyone! I'm confused by my Stata-output. I'm doing a system-GMM estimation with bank risk a…
Rubustness Check in Panel Data Modelling: Sensitivity Analysis Modifying the Level of AggregationDear Statalist participants, I am running a panel data modeling in the municipality level of aggreg…
r(table) : when did this thing appear, why were we not told about it, and where is it documented?Good morning, Recently I (by chance) discovered that e-class commands such as -regress-, -probit-, …
Subscribe to:
Post Comments (Atom)
0 Response to When to divide in subpopulations, using the margins command (AME) in a logit model?
Post a Comment