Dear all,
I am trying to create a control group using Propensity score matching.
The idea is to select a control group based on covariates before treatment and compare several features after the treatment.
There are commands: pscore, psmatch, teffect:
My first question is that: for psmatch and teffect, what should be the outcome variables?
In my case, I would like to compare three (or more) dependent vars: let's say Y1, Y2, Y3.
Then which vars should I define as an outcome vars? If I change the dep vars, then, it would change the composition of control groups?
If I want to select control groups that are only similar pre-treatment stages: X-covariates, pscore and psmatch allow not to put outcome vars. In this case control group is correctly selected?
I guess there are many cases that researchers compare results using MANY DEP VARS between pre-treatment and post-treatment. What would be the solution?
Second question is that: if I get an error message that "Convergence not achieved". Could you please advise what to do in this case?
Thirdly, if I want to use panel data to create control group, how to use the command? I saw some previous post that advises to use "average" values before treatment,
Thanks a lot!
Related Posts with Propensity score matching and outcome variable
Creating variables with a sequence of numbers based on whether other variables stay the same (for case:control matching)Hello everyone. This is my first time posting here, and I do so because I haven't been able to find…
Problem with capture confirm and graph combine Hello. I have problems combining two graphs in a loop . I do not have all of the combination …
Drop if - variable not foundHi all, Just a very simple question and command yet I can't figure out why it is not working. I ju…
ivregress and country fixed effectsHi friends I am new to the use of STATA and Instrumental variables (IV) and using STATA 14.2 I am …
How to use Stata commands to calculate the total score of Likert Scale questions?In order to assess the attitude about antibiotic use, there are 6 questions with 5 levels of respons…
Subscribe to:
Post Comments (Atom)
0 Response to Propensity score matching and outcome variable
Post a Comment