Evaluating the use of residential mailing addresses in a national household survey
Staab, J. M., & Iannacchione, V. G. (2003, August). Evaluating the use of residential mailing addresses in a national household survey. Presented at American Statistical Association Meeting, Section on Survey Research Methods, San Francisco, CA.
On-site enumeration is generally regarded as the gold standard for sampling frame construction for inperson household surveys. However, on-site enumeration is expensive and time consuming because of the amount of time needed to completely enumerate selected areas. A more expedient and cost-effective alternative method of householdenumeration is the use of residential mailing lists.In 2000, we used residential mailing lists, instead of a list derived from on-site enumeration, to develop a sampling frame for a probability-based in-person survey of 15,000 households in Dallas County, Texas(Iannacchione, Staab, and Redden 2003). We estimated that the expense of purchasing and processing the entire list of 800,000 residentialmailing addresses for Dallas County was less than one tenth of the cost of on-site enumeration. In addition, we found that the mailing lists, whensupplemented with missed households identified by a frame-linking procedure provided reasonably complete coverage of the metropolitan household population.Although these results demonstrated the utility and completeness of using residential mailing lists to develop a sampling frame, our evaluation in 2000was limited to one large metropolitan area. In this paper, we evaluate the completeness of mailing lists for a national household survey by comparing counts of residential mailing addresses to Census projectionsof households. We also demonstrate the utility of mailing lists by describing their use in a national household survey that we conducted in 2002.