Address frames were developed from marketing lists as lower cost alternatives to field enumeration for in-person area probability designs. In the United States, many of the available address frames are based on U.S. Postal Service (USPS) databases, so mail surveys have enjoyed a resurgence with the availability of the address frames. Mail surveys, mail and web combination surveys, and other multi-mode surveys have achieved response rates higher than random digit dial surveys. Address-based sampling (ABS) has become an important tool in the survey designer’s toolkit.
The American Association of Public Opinion Research has produced a detailed introductory report that explains the current state of ABS for the survey community (AAPOR 2016). Other useful overviews include Iannacchione (2011) and Link (2010).
To do ABS well requires familiarity with ABS frames and samples. Many of the details about ABS frames and samples are described by AAPOR (2016). Very few survey organizations, however, have direct access to an ABS frame for exploration and research. RTI International is fortunate to be one of those few organizations. More details may be found in RTI’s ABS flyer (PDF).
Our ABS Atlas provides interactive maps and expert information on ABS, allowing users to see how they could leverage this sampling method. The information is intended to aid survey designers in anticipating their sample needs and discussing their plans with sample providers.
The ABS Atlas provides users information about residential addresses in a specific geographical area (total U.S., states, or counties), including:
- Total residential addresses
- Address types
- Addresses flagged as vacant or seasonal
- Drop points
- Unlocatable addresses
- Available telephone appends
The best way to learn the Atlas’s capabilities is to explore for yourself.
Our primary address file from Compact Information Systems (CIS) is essentially the USPS’s Computer Delivery Sequence (CDS) file (U.S. Postal Service, 2013). We supplemented it with 8.5 million records from the USPS’s No-Stat file, also obtained from CIS. Samples from the CDS and No-Stat files were submitted to Marketing Systems Group for information about appended telephone numbers.
The full address file was used to summarize the types of address records available for the ABS Atlas. After summarizing the full address file, we subset the file to addresses most likely to be included in a sampling frame. Most of the summaries on this site are now based on our frame from CIS’ January 2017 file.
One of the strengths of ABS is the ability to append auxiliary variables. For our clients’ studies, we have appended the frame with geocodes and other data from federal and commercial sources including the decennial census, the American Community Survey, the U.S. Department of Agriculture, and Acxiom® InfoBase™ marketing data.
- American Association for Public Opinion Research. (2016). Address-based sampling. Report prepared for AAPOR Council by the Task Force on Address-based Sampling; R. Harter, Chair. Oakbrook Terrace, IL: AAPOR. https://www.aapor.org/Education-Resources/Reports/Address-based-Sampling.aspx
- Iannacchione, V. G. (2011). The Changing Role of Address-Based Sampling in Survey Research. Public Opinion Quarterly, 75(3), 556–575. https://doi.org/10.1093/poq/nfr017
- Link, M. W. (2010). Address-based sampling: What do we know so far? http://ww2.amstat.org/sections/srms/AddressBasedSampling11-29-2010.pdf
- U.S. Postal Service. (2013). CDS user guide. http://ribbs.usps.gov/cds/documents/tech_guides/CDS_USER_GUIDE.PDF