Ex post facto analysis

In correlational research, an attempt to reduce the third-variable problem by selecting subjects who are matched on characteristics that may be confounding.