If I examine the average treatment effect on the treated using a matching estimator, only the treated variables should be matched.
However, when I examine the matching variable (generated by "gen(match) "), I find that all observations (treated or not treated) are matched.
I am confused about why this is the case. Example below.
use http://www.stata-press.com/data/r15/cattaneo3.dta
teffects nnmatch (bweight fage) (mbsmoke), nn(1) gen(match) atet
tab mbsmoke if match1 ~= .
Related Posts with teffects nnmatch, atet generates wrong matching variable?
Recovering p-values from multiple bootstrapped iterations of a logit or glm regression- Crosstabs automation
- changing numeric string to numeric variable where e5 next to a number means 10^5
- An indicator variable
- How to find nonnumeric characters within a string variable and lists the observations that have this issue
Subscribe to:
Post Comments (Atom)
0 Response to teffects nnmatch, atet generates wrong matching variable?
Post a Comment