Psmatch2

[Lars Soendergaarden]

Hi all

First time posting, so please correct me if I'm not doing things the usual way.

I am estimating a propensity score model, which resembles this:

psmatch2 $treatment $xlist, outcome($ylist) common kernel

However, I want to use a different set of control variables, when estimating the treatment effects. So I include x1,x2 and x3 when estimating the propensity scores with the probit regression and x4 in the final regression.

How can I specify this in the PSMATCH2 command? Is it even possible or do I have to use a different approach.

Kind regards, Lars

[David Radwin]

By "in the final regression," do you mean a regression using the matched sample? If so, you can follow the example in an online appendix to DuGoff, E. H., Schuler, M., & Stuart, E. A. (2014). Generalizing observational study results: applying propensity score methods to complex surveys. Health services research, 49(1), 284-303. https://onlinelibrary.wiley.com/doi/...475-6773.12090

If you mean something else, would you please clarify?

Note: psmatch2 is a user-written command by Leuven and Sianesi, available from SSC. Statalist's FAQ 12.1 requests that posts include authors and locations of community-contributed (that is, not official) Stata programs.