Random Digit Dialing versus Address-Based Sampling using Telephone Data Collection
Dual-frame random digit dialing (RDD) and address-based sampling (ABS) are typically considered for household survey frames. Inaccurate geographic assignments using RDD can lead to undercoverage of the target population and challenges in oversampling geographically-clustered subpopulations. Geographic targeting is straightforward with an ABS design, but ABS poses challenges related to contact for telephone data collection. The Aligning Forces for Quality: Assessment of Consumer Engagement (AF4Q) survey was originally based on an RDD design and was migrated to an ABS design due to concerns with coverage, sampling efficiencies, and data collection costs. We discuss the challenges associated with each sampling method and compare key measures between the designs. We find that the ABS design offered higher frame coverage, lower costs, and improved sampling efficiency, while the RDD design had higher response rates. ABS should be considered as a viable option for telephone-only data collection efforts that feature geographic targeting or oversampling of geographically-clustered populations
Shook-Sa, B., Warren, L., Roe, D., & Currivan, D. (2016). Random Digit Dialing versus Address-Based Sampling using Telephone Data Collection. Survey Practice, 9(3), e1-e9.