Use of Paradata in a Responsive Design Framework to Manage a Field Data Collection

Abstract

Every household survey confronts uncertainty in the performance of its recruitment protocol. Diverse modern populations present the data collecting organization with a myriad of practical problems of contacting household units, identifying the proper respondents, gaining their cooperation, and finding appropriate times to interview them. Responsive survey designs (Groves and Heeringa, 2006) use paradata to monitor costs and response rate features of a survey during data collection and actively intervene into the recruitment protocol to effect changes in the production process of interviews. When the paradata are informative about costs and potential error properties of the key estimates to be computed from the survey, responsive designs have the potential to improve the quality-per-unit-cost properties of a survey. This paper uses a simple production model articulating inputs to the data collection, the quality of the sample cases available, and the costs and data records produced. This model motivates a set of paradata that might be useful in managing a survey. The paper presents three examples of management interventions based on paradata collected throughout the Continuous National Survey of Family Growth (NSFG), and evaluates the effectiveness of these interventions. Use of Paradata in a Responsive Design Framework to Manage a Field Data Collection 3