Big Data Meets Survey Science

23Attitudes Toward Data Linkage: Privacy, Ethics, and the Potential for Harm

Aleia C. Fobia¹, Jennifer H. Childs¹, and Casey Eggleston²

US Census Bureau, Center for Behavioral Science Methods, Washington, DC, USA

23.1 Introduction: Big Data and the Federal Statistical System in the United States

Big Data offers both challenges and opportunities for national statistical systems (U.N. Economic and Social Council 2014). For the United States Federal Statistical System (USFSS), one of those opportunities is the use of existing administrative records. Administrative records are a subset of Big Data that include data collected as part of administering state, local, and federal programs (Connelly et al. 2016). Current uses of administrative records by federal agencies in the United States include construction of sampling frames, improving survey design, imputation for missing responses, and weighting to known population totals (NASEM 2017b). The 2020 Decennial Census will use administrative records to reduce the field workload for in‐person follow‐up for those who do not self‐respond to the census. Administrative records will help identify vacant units, units that do not meet the definition of a housing unit and will also help optimize the number and timing of contact attempts necessary to successfully enumerate a household. The data sources include both government and commercial sources (U.S. Census Bureau 2017).

Recent legislation proposes to expand the use of administrative records ...