Hi,

I am using propensity score matching to reduce selection bias and then running a regression on the matched sample.

The basic stata command I am using is:

psmatch2 treatment x1 x2, out(wage)
reg wage treatment x1 x2 [fweight=_weight]

Is it possible to introduce new controls into the regression that were not used when selecting the matched sample?

E.g.
reg wage treatment x1 x2 z1 z2 [fweight=_weight]

(z1 and z2 are the new controls that I wish to introduce into the regression - they are not suitable to use when matching the sample)

Thanks!