How Much Do Additional Mailings and Telephone Calls Contribute to Response Rates in a Survey of Medicare Beneficiaries?

Surveys often spend substantial money on multiple mailings and telephone calls to ensure high overall response rates and adequate representation of hard-to-reach demographic subgroups. We examine the extent to which an additional mailing and additional sets of telephone calls are effective in attaining these goals across a variety of subgroups in a large, national multimode survey of Medicare beneficiaries. We also examine the relative data quality of the responses that come with each level of extra effort. We find that additional mailings appear more effective in some groups, while additional telephone calls appear more effective in others. Tailoring the fielding strategy differently by subgroup may improve response rates at a lower cost per complete than using the same fielding protocol for all potential respondents, although data quality is likely to decline with additional efforts in either mode.