An evaluation of respondent selection methods for household mail surveys

Mail surveys are a staple of the survey industry; however, they are rarely used in surveys of the general population. The problem is twofold: (1) lack of a complete sampling frame of households and (2) difficulties with ensuring random selection of a respondent within the household. However, advances in electronic record keeping, such as the U.S. Postal Service Delivery Sequence File, now make it possible to sample from a frame of residential addresses. Unfortunately, less is known about the effectiveness of within-household selection techniques for household mail surveys. A six-state pilot study was conducted as part of the 2005 Behavioral Risk Factor Surveillance System using the Delivery Sequence File to sample addresses for a mail survey. The pilot study tested three respondent selection methods: any adult, adult with the next birthday, and all adults. The next-birthday and all-adults methods yielded household-level response rates that were comparable to the any-adult method, the method assumed to have the least respondent burden. At the respondent level, however, the response rate for the all-adults method was lower when we accounted for within-household nonresponse.