> When estimating the treatment on the treated, I want to do
> weighted least squares by weighting the treatment
> observations as 1, and control observations weighted by the
> number of times they are matched to a treatment observation
> (which is what Dehijia and Wahba (1999) calls "adjusted".)
>
> Can I do this using psmatch2?
first -psmatch2- is not an official stata command, so you might have more chance on a reply when you contact the authors
as to your question, -psmatch2- stores these weights in a variable called _weight
i suggest you give the help file a good read
edwin
<<winmail.dat>>
