Sample Designs for Measuring the Health of Small Racial/Ethnic Subgroups

Published in: Statistics In Medicine, v. 27, no. 20, Sep. 10, 2008, p. 4016-4029

Posted on RAND.org on January 01, 2008

by Marc N. Elliott, Brian Karl Finch, David J. Klein, Sai Ma, D. Phuong Do, Megan K. Beckett, Nathan Orr, Nicole Lurie

Read More

Access further information on this document at onlinelibrary.wiley.com

This article was published outside of RAND. The full text of the article can be found at the link above.

Most national health surveys do not permit precise measurement of the health of racial/ethnic subgroups that comprise <1 per cent of the U.S. population. We identify three potentially promising sample design strategies for increasing the accuracy of national health estimates for a small target subgroup when used to supplement a small probability sample of that group and apply these strategies to American Indians/Alaska Natives (AI/AN) and Chinese using National Health Interview Survey data. These sample design strategies include (1) complete sampling of targets within households, (2) oversampling selected macrogeographic units, and (3) oversampling from an incomplete list frame. Stage (1) is promising for Chinese and AI/AN; (2) works for both groups, but it would be more cost-effective for AI/AN because of their greater residential concentration; (3) is somewhat effective for groups like Chinese with viable surname lists, but not for AI/AN. Both (2) and (3) efficiently improve measurement precision when the supplement is the same size as the existing core sample, with diminishing additional returns as the supplement grows relative to the core sample, especially for (3). To avoid large design effects, the oversampled geographic areas or lists must have good coverage of the target population. To reduce costs, oversampled geographic tracts and lists must consist primarily of targets. These techniques can be used simultaneously to substantially increase effective sample sizes (ESSs). For example, (1) and (2) in combination can be used to multiply the nominal sample size of AI/AN or Chinese by 8 and the ESS by 4.

This report is part of the RAND Corporation External publication series. Many RAND studies are published in peer-reviewed scholarly journals, as chapters in commercial books, or as documents published by other organizations.

Our mission to help improve policy and decisionmaking through research and analysis is enabled through our core values of quality and objectivity and our unwavering commitment to the highest level of integrity and ethical behavior. To help ensure our research and analysis are rigorous, objective, and nonpartisan, we subject our research publications to a robust and exacting quality-assurance process; avoid both the appearance and reality of financial and other conflicts of interest through staff training, project screening, and a policy of mandatory disclosure; and pursue transparency in our research engagements through our commitment to the open publication of our research findings and recommendations, disclosure of the source of funding of published research, and policies to ensure intellectual independence. For more information, visit www.rand.org/about/principles.

The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.