Address-Based Sampling for Subpopulation Surveys

Abstract

In this article, we describe the use of address-based sampling (ABS) for conducting a mail survey of a subgroup. Although ABS has been investigated as an alternative household sampling frame, its use for sampling subgroups is limited. Some researchers have used ABS to select households but then recruit households by telephone to the extent possible to screen for members of the subgroup. The approach we describe uses a two-phase methodology with mail as the primary mode of recruitment and data collection for all households. Facing declining response and coverage rates in its landline RDD surveys, the National Household Education Surveys (NHES) program embarked on a redesign effort. The two-phase ABS design using mail as the primary mode of data collection was chosen for the NHES:2009 Pilot Study. In the first phase, all sampled households were mailed a screener that identified members of the eligible subgroup (children) and provided the data needed for sampling among them. A second-phase substantive questionnaire was mailed to the sampled members of the subgroup. This use of the ABS and mail as the primary mode of data collection is new and largely untested. The findings from the NHES Pilot Study revealed that this approach is a feasible alternative for surveying targeted subgroups. Although the methods must undergo more robust and rigorous tests, it is an approach that may be able to fill some of the void created by the low response and coverage rates that afflict RDD surveys.