The ABS Frame: Quality and Considerations

Abstract

Address based sampling (ABS) is increasingly being used in the current survey environment of declining response rates and coverage concerns associated with landline random digit dial surveys. The goals of this research are to determine the implications of using ABS frame variables for data collection, to evaluate data quality of demographic variables provided on ABS frames, and to examine the possibility of using ABS frame variables to guide survey design. The implications of using ABS on data collection will be evaluated by examining the quality of frame data and by comparing these data to information reported by respondents. The ability to use ABS frame information for stratification and to guide operations will also be considered in light of the findings of these evaluations.

Address‐Based Sampling for Recruiting Rural Subpopulations: A 2‐Phase, Multimode Approach

“…Similar results were found in subsequent attempts using this approach. 15,16 Although Brick, Andrews, and Mathiowetz 17 have also had success with a single rather than 2-phase mail survey design for rare population groups, they agree with Montaquila et al 18 that a 2-phase design is ideal when within household sampling is needed, and when the survey is lengthy.…”

Variables appended to ABS frames: Has their data quality improved?

“…In addition to telephone numbers, vendors are able to append other characteristics of the address that can be leveraged for data collection (for example, number of household members or whether the household members rent or own the house). This appended data has been used or evaluated for a number of data collection purposes, including predicting household eligibility [ 3 ], stratification and sample allocation [ 4 – 6 ], and other purposes such as enhancing efficiency of survey operations [ 7 , 8 ]. [ 4 ] found some promise only in Hispanic origin or surname, but otherwise found availability and quality of appended variables needed improvement before considering them for sample design.…”

“…This appended data has been used or evaluated for a number of data collection purposes, including predicting household eligibility [ 3 ], stratification and sample allocation [ 4 – 6 ], and other purposes such as enhancing efficiency of survey operations [ 7 , 8 ]. [ 4 ] found some promise only in Hispanic origin or surname, but otherwise found availability and quality of appended variables needed improvement before considering them for sample design. [ 6 ] also found potential value in using Hispanic origin or surname, as well as presence of age 65+ in the household, in improving efficiency of locating members of those domains.…”

“…Thus, it is important to continue to periodically reassess the quality of the appended data and its viability for use in improving sample efficiency and other purposes. This research is in part a follow-up to [ 4 ] to determine whether the accuracy and completeness of appended frame data have changed over time. High quality frame information can be used in survey design to gain efficiencies in sampling and estimating outcomes in hard-to-reach populations.…”

Methodological considerations for a new household panel survey

“…In this situation, stratification of areas to achieve more precise estimates for subgroups is generally not very effective [81]. At the address level, data can be appended to the CDSF frame, but the data are missing for a large proportion of the subgroups reducing the ability to use this type of stratification to improve precision [60].…”