Confounding Variable In Surveys