A A A

Could anyone explain the logic of propensity score matching in program evaluation?