An Application of Alternative Weighting Matrix Collapsing Approaches for Improving Sample Estimates

• Activity Number: 153
• Type: Contributed
• Date/Time: Monday, July 30, 2007 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Linda Tompkins*+ and Jay J. Kim
• Companies: Centers for Disease Control and Prevention and National Center for Health Statistics
• Keywords: Coverage ratio ; Sample weighting ; Cell collapsing

Abstract

In sample weighting, most surveys combine small race groups such as American Indians, Asians, and other races with Whites. One problem with this approach is that Whites have a much better coverage ratio than the smaller race groups. This results in an overestimation of Whites and an underestimation of smaller race groups. For example, in weighting the 2003 National Health Interview Survey (NHIS) sample, the number of American Indians was underestimated by 29.70 percent and the number of Whites children under age year one were overestimated by 7 percent. The overall coverage ratio for Whites was 84.8 percent, while that for American Indians was only 50 percent. In this paper, alternative approaches for cell collapsing applied to the 2003 NHIS and the resulting estimates will be compared.