Aspects of SCHIP Oversampling and Subsampling in the 2000 Sample Redesign

• Activity Number: 255
• Type: Contributed
• Date/Time: Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods*
• Author(s): Alan Peterson*+
• Affiliation(s): U.S. Census Bureau
• Keywords: poverty stratum ; sample design ; State's Children Health Insurance Program ; subsampling/oversampling

Abstract

The goal of the State's Children Health Insurance Program (SCHIP) sample is to improve state estimates on the number of uninsured children. This estimate is derived from data collected in the Annual Demographic Supplement to the Current Population Survey (CPS). The SCHIP sample is combined with the supplement sample to obtain this improved measure. Currently, the sample designs for the SCHIP sample and the CPS are virtually identical. However, with a sample redesign slated to begin in April 2004, the sample designs for these two surveys will differ. To increase the likelihood of reaching the SCHIP target population, the uninsured, the SCHIP will employ a strategy to oversample census blocks classified as having a "high" proportion of low-income persons. This paper discusses the Census Bureau's proposed methodology for achieving the oversample, including research on determining a cutoff value on which to base the classification of census blocks in terms of minimizing variance and the impact on interviewer workloads.