Assessment of Alternative Weighting Methods for the National Health Interview Survey

• Activity Number: 358
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): David Shaw and Joe Fred Gonzalez, Jr.*+ and Meena Khare
• Companies: University of Maryland and National Center for Health Statistics and National Center for Health Statistics
• Keywords: survey weighting ; raking ; nonresponse adjustment ; logistic regression ; recursive partitioning

Abstract

The National Health Interview Survey (NHIS) is a population-based survey which has collected health information from the U.S. civilian noninstitutionalized population since 1957. Major goals of the NHIS include the production of quality data as well as precise and reliable estimates of health conditions. This paper summarizes the results of alternative weighting techniques for NHIS data. The first approach applied different raking methods by age-sex-race-ethnicity, individually by education and income, as well as jointly by education and income with basic and full imputation. The second approach applied logistic regression to NHIS paradata using ten variables to predict the probability of response; the third approach applied recursive partitioning to the binary response variables. Raking was then applied to these nonresponse adjusted weights.