Errors in the sampling frames that underlie critical national health surveys threaten the quality of the statistical estimates used to track the nation's health. Yet coverage error is largely unmeasured in area probability surveys. Given the importance of the data collected by these surveys and their expense to the federal government, it is critical that we gain a better understanding of the error in these housing unit listings and the impact of the errors on health estimates.
The first aim of this research is to identify influences on undercoverage and overcoverage errors in housing unit listing. Any attempt to reduce the effects of frame errors on survey estimates must be informed by a thorough understanding of housing unit and segment characteristics that make accurate listing difficult.
The second aim i s to estimate coverage bias and variance due to listing errors in a national health survey. I will conduct research towards these aims in conjunction with the National Survey of Family Growth (NSFG). This research will begin to fill in our knowledge of coverage error in housing unit listing and thus will benefit many other surveys conducted by the CDC and other government agencies.
Errors in the sampling frames that underlie critical national health surveys threaten the quality of the statistical estimates used to track the nation's health. Yet coverage error is largely unmeasured in area probability surveys. Given the importance of the data collected by these surveys and their expense to the federal government and the Centers for Disease Control and Prevention in particular, it is critical that we gain a better understanding of the error in these housing unit listings and the impact of the errors on health estimates.
