Adaptive Designs for Managing Survey Nonresponse

I will present work that my colleagues and I are doing using adaptive survey design in the National Survey of Family Growth, a continuous survey with quarterly samples. Each quarter is divided into two phases of data collection, with subsampling of nonrespondents in the second phase. We track the survey paradata on a daily basis and execute planned interventions to address sample balance issues and potential bias. Some of these interventions are randomized experiments. In this talk, I will describe the process of adaptive survey design, present results of our work, and discuss the benefits of this approach both for controlling costs and for managing nonresponse error in surveys.