Residential Mobility and Big Data: Assessing the Validity of Consumer Reference Datasets

Alex Ramiller
Taesoo Song
Madeleine Parker
University of California, Berkeley

Karen Chapple
University of California, Berkeley
University of Toronto

The increasing availability of privately produced longitudinal Consumer Reference Datasets (CRDs) presents substantial opportunities for housing and urban studies research, permitting the analysis of processes, including residential mobility, migration, and neighborhood change. Despite their growing popularity in academic and policy research, these datasets—which are produced by private companies for sale primarily to commercial interests—are not explicitly designed for research purposes and have not been comprehensively assessed in terms of data quality or representativeness. This article carries out a comparative analysis of the CRDs that two of the most prominent sources of consumer reference data—Data Axle and Infutor Data Solutions—produce for King County, Washington. Comparing these datasets with estimates from the American Community Survey at the county and census tract scales, this article identifies substantial limitations associated with each dataset in terms of population counts, demographic characteristics, distribution across census tracts, and residential mobility rates. It concludes that despite notable advantages, including the ability to provide valuable and novel insights into heretofore unobserved patterns of residential mobility at a range of spatial scales, these datasets contain systematic biases. These biases may lead researchers to underestimate population counts and mobility rates for low-income households, renters, young adults, and people of color, and should, therefore, be used with caution in social, demographic, and policy research.