Managing Locating and Data Collection Interventions Through Adaptive Survey Design
Wan-Ying Chang
National Center for Science and Engineering Statistics
Zachary H. Seeskin
NORC at the University of Chicago
Chaoyi Zheng
NORC at the University of Chicago
Adaptive designs are increasingly being used for federal surveys to pursue survey goals in a cost-effective manner. These designs assign mid-data collection interventions to pursue such objectives as improving sample balance and increasing response within specific domains or overall. Particular challenges emerge for complex data collections that involve competing needs for locating cases and for obtaining responses from found cases. We describe the adaptive design strategies of the 2017 Survey of Doctorate Recipients, involving both differential locating and cooperation-gaining treatments at distinct phases over the field period. At each phase, high and low priority cases were separately identified for the locating and data collection activities where high priority cases would receive more intensive and costly treatment. We present the prioritization methods and describe the differential treatments for locating and data collection activities. Based on analysis of paradata and survey outcomes, we investigate the contribution of the adaptive design scheme toward improving the representativity of the sample and toward attaining targets numbers of completes for key analytic domains.