Hi,

I hope there is some simple reason for this, but I just could not wrap my head around this. Perhaps someone here could help me out

I am using panel data and differences-in-differences method with time and id -fixed effects.
These fixed effects are collinear with treat and post variables, so Stata should omit treat and post variables, like in this example:
Code:
webuse nlswork, clear

keep if inrange(idcode, 1, 10)
gen byte treat = (idcode > 5)
gen byte post = (year >= 80)

xtset idcode year

xtreg ln_wage i.year treat##post , fe  vce(cluster idcode)
However, when I add weights, like birth_yr in this sample data (only using this because it's the same for each id over time and this is just an example.), with code looking like this
Code:
xtreg ln_wage i.year treat##post  [aweight=birth_yr], fe  vce(cluster idcode)
Stata no longer omits the treat variable. Why is this? And how should I "correct" this if this is somehow wrong?