Application of a Fay-Modified Balanced Repeated Replication Method to the National Health Interview Survey (NHIS)

• Activity Number: 191
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Van Parsons*+
• Companies: National Center for Health Statistics
• Keywords: variance estimation

Abstract

The NHIS is a complex survey targeting the health of the U.S. population. To allow for design-based standard errors via linearization, the public-use data provides limited design structures: strata, PSUs and a final survey weight. These structures are accommodated very well by popular complex-survey data analysis software. However, the linearization approach imposes limitations on statistical functional forms, and furthermore, this approach often does not account for statistical variability due to weighting adjustments. The Fay-modified balanced repeated replication (Fay-BRR) method provides a flexible alternative to variance estimation which can correct for some of the linearization method's weaknesses. In this paper the Fay-BRR method is implemented on the NHIS adult samples for the 1997-2005 NHIS design years, and the operating characteristics are discussed.