Using Residential Mailing Lists as a Sampling Frame for In-Person Household Surveys
Iannacchione, V. G. (2006, May). Using Residential Mailing Lists as a Sampling Frame for In-Person Household Surveys. Presented at American Association for Public Opinion Research 61st Annual Conference, Montréal, Canada.
The benefits of using mailing addresses instead of traditional field enumeration as a sampling frame for household surveys include:
Cost Savings and Flexibility. A sampling frame based on mailing addresses can be developed in a matter of weeks instead of the months usually required for field enumeration. The time savings means that survey planners have more time to re-target the sample in reaction to findings from a pilot or other events that can happen during the development of the sampling design. The reduced costs associated with mailing addresses means that more resources to training interviewers and improving response rates.
Coverage. Our experience with the use of residential mailing addresses as a sampling frame for national household surveys has shown that mailing addresses, when supplemented with missed households discovered in the field, provide reasonably complete coverage of the household population. The only known sources of under-representation are the estimated 1.3 percent of households that are located in areas without home delivery of mail, and the less than 4 percent of households with ‘simplified’ rural addresses.
Geographic Diversity. Sampling frames that rely on field enumeration are constrained by costs to relatively small areas (or segments) such as Census Blocks. Small segments can reduce efficiency by introducing intra-cluster correlation into survey estimates. In contrast, the segments for a frame based on mailing addresses can be based on postal Carrier Routes, which are usually two or three times larger than Census Blocks. These larger segments produce a more geographically diverse sample of households which improves efficiency.
We describe our experiences with the use of mailing lists as a sampling frame for household surveys; how we evaluated the coverage of mailing lists; how the coverage of mailing lists can be supplemented; and, how demographic data can be combined with mailing lists to improve sampling efficiency.