Dear Stata user,
I'm using stata 14 to estimate the effect of public aids on the recipients. In order to do that, I'm using psmatch2 and teffects psmatch.
My issue is that I have a very small number of recipients (only 40) compared to the number of controls (700 000). I'm afraid that the probit/logit used by teffects will be biased by this low number of recipients. My first idea was to the command on a subsample of the control group, but I think that it is a pity to not use all the observations to find the match. Do you know if there is a way to implement the probit/logit made by teffects on a subsample, to use it to predict the pscore for all observations and to redo teffects on that?
I hope my presentation of this issue was clear enough...Thanks for your future comments!
Best,
Juliette R.
I'm using stata 14 to estimate the effect of public aids on the recipients. In order to do that, I'm using psmatch2 and teffects psmatch.
My issue is that I have a very small number of recipients (only 40) compared to the number of controls (700 000). I'm afraid that the probit/logit used by teffects will be biased by this low number of recipients. My first idea was to the command on a subsample of the control group, but I think that it is a pity to not use all the observations to find the match. Do you know if there is a way to implement the probit/logit made by teffects on a subsample, to use it to predict the pscore for all observations and to redo teffects on that?
I hope my presentation of this issue was clear enough...Thanks for your future comments!
Best,
Juliette R.