I am looking for published (including working papers) discussions of the use of propensity score matching and/or inverse propensity weighting in conjunction with sampling weights.
I am aware of the Zanutto (2006) Journal of Data Science paper as well as an informal discussion in my old report, with Juao Pedro Azevedo and Peter Dolton, on the evaluation of the New Deal for Lone Parents prepared for the UK Department for Work and Pensions, but not much else.
Here is a free idea for someone looking to get a zillion cites: update psmatch2 to incorporate weights, then write a paper for Stata Journal about it. Oh, and thank me for the idea.
It's really quite easy.
1 year ago