Email Record: Reliability of estimates with alternative cluster sizes in the Health Interview Survey :