• Report

The CHUM: A frame supplementation procedure for address-based sampling

Citation

Shook-Sa, B., Harter, R., McMichael, J., Ridenhour, J., & Dever, J. (2016). The CHUM: A frame supplementation procedure for address-based sampling. (RTI Press Publication No. MR-0034-1602). Research Triangle Park, NC: RTI Press. DOI: 10.3768/rtipress.2016.mr.0034.1602

Abstract

RTI developed the check for housing units missed (CHUM) methodology to compensate for housing unit undercoverage of address-based sampling (ABS) frames for in-person, area probability surveys. The CHUM systematically identifies housing units missing from the ABS frame, giving each housing unit a chance of selection with known probability. The CHUM poses several advantages over alternative supplementation approaches. Because only a subset of housing units within selected areas must be evaluated, the CHUM is less costly than supplementation techniques that require the verification of all addresses within selected areas. Because it is conducted after housing units are selected instead of the frame-building stage, the CHUM provides timelier frame updates. This paper presents details for designing ABS studies that incorporate the CHUM, appropriately incorporating missed units into area probability samples, and training field personnel to implement the CHUM. It also compares the CHUM with other frame supplementation approaches and discusses the advantages and limitations of each approach.

Author Details

Rachel Harter

Rachel M. Harter, PhD, is a senior research statistician and program director at RTI. Areas of interest include household and establishment surveys, area probability survey designs, address-based sampling, imputation, and small area estimation.

Joseph McMichael

Joseph McMichael, BS, is a research statistician in RTI International’s Division for Statistical and Data Sciences.

Jamie Ridenhour

Jamie Ridenhour, MStat, is a research statistician at RTI. Her research interests are sample design, weighting, and methodological challenges associated with address-based sampling and dual-frame random-digit-dial surveys.

Jill Dever

Jill A. Dever, PhD, is a senior research statistician at RTI. Her current research interests are variance estimation with calibrated analysis weights for complex survey designs and statistical issues related to samples drawn without a defined probabilistic structure.