314 – Use of Vendor Data in Optimization of Address-Based Sampling Procedures

From Flagging a Sample to Framing It: Exploring Vendor Data That Can Be Appended to ABS Samples

at Marketing Systems Group

Address based sampling and random digit dialing designs are two of the most commonly used probability based sampling methods to obtain data aimed at answering questions in a broad spectrum of fields including social science, public health, and medicine and beyond. Enhancements to samples can be made in order to facilitate better frame construction including the appending of information on home ownership or presence of children or age of head of household to create strata that could be meaningful for a given study. These variables can also be used as part of a tailored design or in part for nonresponse and weighting adjustments after data collection is completed. The degree to which these appended variables can be used for various aspects of the survey process from sampling, to recruitment to weighting depends in part on their availability for all sampled or population units and on their accuracy at the time of appending. In this presentation we provide a broad overview of the types of variables that can be appended to address based samples with particular focus on vendor related, rather than public use, variables. We describe appending rates for a key set of socioeconomic and household variables, and discuss their variability across vendors. We discuss how survey researchers can properly evaluate the quality of these data and make recommendations about their potential uses for ABS sampling designs.