June 5, 2024 – A distinguished panel of experts convened to discuss the evolving role of blended data in the United States’ national data infrastructure. The session, titled The Role of Blended Data in a New National Data Infrastructure, was held virtually and featured leading statisticians, economists, and policymakers who provided insight into the opportunities and challenges presented by blended data.
Moderated by Melissa Chiu from the National Academies of Sciences, Engineering, and Medicine, the panel included Erica Groshen (Senior Economics Advisor, Cornell University), Sharon L. Lohr (Emeritus Distinguished Professor, Arizona State University & Vice President and Senior Statistician, Westat, retired), and Jerry Reiter (Department Chair and Professor of Statistical Science, Duke University). Together, they explored the advancements and hurdles in leveraging blended data for economic and social research.
The Role of Blended Data in Modern Analytics
Blended data—data compiled from multiple previously collected sources—has emerged as a crucial tool for enhancing statistical analyses while reducing the burden and costs of traditional data collection methods. As recent federal legislation and guidance emphasize the role of such data in policymaking, experts are working to address the technical and ethical considerations associated with its use.
The discussion was framed around three reports produced by committees appointed by the National Academies of Sciences, Engineering, and Medicine, each focusing on different aspects of modernizing the national data infrastructure. Panelists outlined the vision for a comprehensive data system, highlighted the role of surveys in supplementing blended datasets, and examined practical applications in areas such as income, health, crime, and agriculture statistics.
Key Insights and Challenges
Dr. Erica Groshen emphasized the potential for blended data to improve labor market analytics and enhance economic decision-making. Dr. Sharon Lohr elaborated on the statistical methodologies required for effective data integration while ensuring accuracy and reliability. Dr. Jerry Reiter provided insights into privacy challenges and discussed strategies for balancing data accessibility with confidentiality protections.
A key theme of the session was data equity—ensuring that advancements in blended data infrastructure do not disproportionately disadvantage certain populations. Speakers stressed the importance of implementing robust policy controls and technical safeguards to address privacy risks and biases in data integration.
Another major discussion point revolved around technical controls and data governance frameworks necessary for securing sensitive information while still enabling effective research. The panelists emphasized the importance of interdisciplinary collaboration in addressing these challenges, with statisticians, policymakers, and technology experts working together to implement best practices for data stewardship.
Case Studies and Real-World Applications
The session also highlighted real-world applications of blended data across different sectors. Panelists presented case studies in labor market trends, health disparities, and criminal justice statistics, demonstrating how data integration has led to more precise and actionable insights. These examples underscored the value of blended data in informing policy decisions and addressing pressing societal challenges.
One compelling case study involved the use of blended data in measuring income inequality. By integrating survey data with administrative tax records, researchers were able to paint a more comprehensive picture of wage distribution and economic mobility. Another example highlighted the role of blended health data in tracking disease outbreaks and improving public health interventions. These case studies showcased the power of blended data to drive evidence-based policymaking and improve social outcomes.
The Future of National Data Systems
The session concluded with a discussion on the future of national data systems and the need for continued collaboration among government agencies, researchers, and policymakers to build a resilient, transparent, and effective data infrastructure. Panelists also discussed how public trust in government data collection efforts could be strengthened through increased transparency and accountability in data handling.
One of the main takeaways was the necessity of ongoing investment in data infrastructure to support the evolving needs of researchers and policymakers. Attendees were encouraged to consider how their own work in data science, economics, and statistics could contribute to a broader national effort to modernize data collection and analysis.
Acknowledgments
NISS would like to extend a heartfelt thank you to the esteemed speakers and moderator for sharing their expertise and valuable insights. Their contributions played a crucial role in advancing the conversation on blended data and its implications for national statistics. Special appreciation goes to Melissa Chiu for expertly moderating the discussion and facilitating an engaging and thought-provoking exchange among panelists. We also thank the attendees for their participation and thoughtful questions, which enriched the dialogue and provided further perspectives on the subject.
We are especially grateful to Erica Groshen, Sharon Lohr, and Jerry Reiter for their time and dedication in sharing their knowledge on this important topic. Their insights provided a deeper understanding of the potential and challenges of blended data, inspiring further research and collaboration. Their efforts are instrumental in shaping the future of national data infrastructure and ensuring that blended data is used responsibly and effectively to benefit society.
The insights shared by the panelists underscored the transformative potential of blended data and reinforced the ongoing efforts needed to achieve a modern, equitable, and efficient national data ecosystem. With continued collaboration and innovation, the future of national data infrastructure looks promising.