Creating Improved Survey Data Products Using Linked Administrative-Survey Data

Abstract

AbstractRecent research linking administrative to survey data has laid the groundwork for improvements in survey data products. However, the opportunities have not been fully realized yet. In this article, our main objective is to use administrative-survey linked microdata to demonstrate the potential of data linkage to reduce survey error through model-based blended imputation methods. We use parametric models based on the linked data to create imputed values of Medicaid enrollment and food stamp (SNAP) receipt. This approach to blending data from surveys and administrative data through models is less likely to compromise confidentiality or violate the terms of the data sharing agreements among the agencies than releasing the linked microdata, and we demonstrate that it can yield substantial improvements of estimate accuracy. Using the blended imputation approach reduces root mean squared error (RMSE) of estimates by 81 percent for state-level Medicaid enrollment and by 93 percent for substate area SNAP receipt compared with estimates based on the survey data alone. Given the high level of measurement error associated with these important programs in the United States, data producers should consider blended imputation methods like the ones we describe in this article to create improved estimates for policy research.