Sample Redesign and Conversion to Windows-Based Survey Instrument: Evaluation of Impact on Part-Year Estimates from the Medical Expenditure Panel Survey (MEPS)

The MEPS is an ongoing longitudinal panel survey. A new panel is fielded annually and data are collected via five computer assisted personal interviewing (CAPI) interviews that cumulatively cover two consecutive years. Each annual MEPS panel is selected as a subsample of respondent households to the prior year’s National Health Interview Survey (NHIS). In 2006, a new sample design was implemented for the NHIS and the changes carried over to the 2007 MEPS (MEPS Panel 12). Also, in 2007 the MEPS survey instrument was converted from a DOSbased system to a Windows-based system. These two changes have the potential to affect point estimates and standard errors of MEPS estimates. In this paper, we conduct an initial assessment of the impact of the changes on 2007 estimates.