Solving the nonresponse problem with sample matching?

This paper explores the use of matched samples as an alternative for estimators based on surveys suffering from a substantial amount of nonresponse.