Planning for a Modern U.S. Data Infrastructure Must Proceed
In ordinary times I would agree wholeheartedly with the key points made in the Science editorial, “Fixing the US statistical infrastructure” by Nancy Potok and Erica L. Groshen, published on May 29, 2025, by AAAS.
With federal statistical programs spread across 13 agencies and more than 100 programs, there is insufficient collaboration to ensure that new product development is effectively coordinated and prioritized.
Coordination and consolidation of statistical programs—long viewed as desirable by past administrations—have been stalled by territoriality, a lack of champions, and fears that such efforts would be poorly executed.
Many of the methods currently used to collect data were developed in previous centuries and are now difficult (and potentially expensive) to refine or update.
Unfortunately, I do not see the necessary work, coordination, or planning being effectively undertaken by the current administration, especially if a major goal is to simultaneously protect privacy and confidentiality. Based on personal experience, I know of at least one case where plans to significantly upgrade one federal agency’s regulatory data collection activities have been delayed—if not outright canceled. Were those efforts abandoned in light of a more strategic, long-term approach like the one recommended in the Science editorial? Unknown.
While we may need to place responsibility for developing a comprehensive data infrastructure strategy on the next administration, we should not delay laying the necessary groundwork. That work should include a serious review of existing planning documents, such as the National Academies of Sciences, Engineering, and Medicine's report on building a 21st-century national data infrastructure.
Planning documents like these must now incorporate documentation of the impacts caused by canceled or delayed data collection efforts, just as we need to document the impacts of research project cancellations in order to plan for how to repair the damage.
The AAAS editorial’s authors do compare the U.S. unfavorably with countries that have more centralized national data infrastructures. While such comparisons may overlook the scale and complexity of the U.S. government, they underscore a critical point: the importance of reliable, nationally collected data for shaping government policy and guiding private-sector decisions.
The challenge of developing a modern, sophisticated national data infrastructure should not be underestimated. Despite the enormity of the task, we must begin now—even if the current administration has different priorities.
Copyright © 2025 by Dennis D. McDonald
