Big Data Meets Survey Science

3Linking Survey Data with Commercial or Administrative Data for Data Quality Assessment

A. Rupa Datta, Gabriel Ugarte, and Dean Resnick

Center for Excellence in Survey Research, NORC at the University of Chicago, Chicago, IL, USA

3.1 Introduction

As response rate challenges and data collection costs continue to climb, survey sponsors and data collection organizations face constant pressures to find less costly means of answering high‐priority research questions. As surveys have grown more challenging to conduct, administrative and commercial data sources have become increasingly ubiquitous and accessible. A hybrid strategy is to link these data sources with survey data at the record level to create expanded datafiles. Administrative or commercial data can potentially enhance survey data through linkage in a variety of ways, including

1. (1) expanding the set of feasible analyses by supplementing survey responses with additional variables, possibly on new topics, at new time points, or at a level of precision or accuracy that cannot be obtained in the survey data;
2. (2) introducing efficiencies or quality improvements into survey processes such as constructing more efficient sampling frames, assessing quality of survey responses, or informing data collection processes; and
3. (3) achieving cost or schedule efficiencies by using already available administrative or commercial data to reduce the scope of survey data collection, for example, shortening questionnaires, reducing sample sizes, ...