Modeling Multi-Time-Period State-Level Estimates for Wireless-Only, Wireless-Mostly, Landline-Mostly, and Landline-Only Households

• Activity Number: 159
• Type: Topic Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Nadarajasundaram Ganesh*+ and Michael Davern and Stephen Blumberg and Julian Luke and Michel Boudreaux
• Companies: NORC and NORC and National Center for Health Statistics and National Center for Health Statistics and University of Minnesota, SHADAC
• Keywords: NHIS ; Small Area Estimation ; State-level Wireless/Landline Estimates

Abstract

We use National Health Interview Survey (NHIS) data to derive small-area-level direct estimates for every six-month period from January 2007 to June 2009 for wireless-only, wireless-mostly, landline-mostly, and landline-only households for states and several metropolitan areas. Because the NHIS is not designed to produce unbiased state level or metropolitan level direct estimates, we use a small area model with covariates obtained from the American Community Survey (ACS). The model-based estimates are raked so they agree with the national-level direct unbiased estimates for the corresponding six-month periods from the NHIS, and the estimates are also raked so they agree with the complement of the "no phone" estimates obtained from the 2008 ACS. State-level and selected metropolitan-area-level estimates are obtained by appropriately aggregating the model-based small-area-level estimates.