Oversampling small low income communities in California using telephone surveys
*Ismael Flores Cervantes, Westat
Keywords: RDD, rare population, geographic stratification, list samples, design effect
This paper describes the challenges in oversampling small low-income communities in a random-digit-dial (RDD) telephone survey. The California Health Interview Survey (CHIS) is a biannual telephone survey conducted since 2001, to explore issues in public health and health care and to monitor changes over time in California. As part of CHIS 2009, additional samples were included to increase the representation of young families in 12 low income California communities. As with any survey of rare populations, the size of the communities and other eligibility requirements presented challenges for sample design, data collection, and weighting. Different alternatives considered to oversample these communities such as list samples, disproportionate stratified sampling, and screening are described. We also discuss data collection strategies for identifying the eligible population in these communities. Finally, we describe the impact these additional samples would have on CHIS estimates.