Report calls for rebuilding federal education data infrastructure with states at the center

The American Statistical Association is calling for a fundamental overhaul of the National Center for Education Statistics, arguing that the federal agency should become a more modern, data-centric partner to states rather than simply returning it to the structure it had before it was largely dismantled last year. 

The report, released in August, recommends shifting the Department of Education agency from a survey-centric model toward one focused on curating, linking and improving data collected by states and education organizations. The approach would require stronger data governance, improved interoperability and expanded technical capacity — both at the federal and state levels. 

Shelly Martinez, an independent statistical consultant who worked with ASA on the report, said the agency’s relationship with states is central to the proposed modernization. 

“NCES actually has a lot of responsibilities for bolstering, leading, and facilitating what we’re calling the national education data ecosystem,” Martinez said in an interview. “But states probably are the single biggest member of that ecosystem.”

NCES has historically provided grants, technical assistance and forums where states work together on voluntary data standards and common definitions, according to Martinez. Those efforts helped states develop comparable measures, such as how graduation rates are calculated, and produce longitudinal data.

But much of that work stalled after the agency was cut from 103 employees to three in early 2025 as part of the Trump administration’s efforts to reduce federal bureaucracy. Roughly three-quarters of its contracts were also cancelled.

“The Department of Government Efficiency came along through the Department of Education, and not unexpectedly —given what this administration’s views are about the Education Department — began to make some significant cuts,” Martinez explained. “NCES was hit harder than many other parts of the department.”

Martinez said the new report offers a guide to how NCES can better support state data infrastructure, improve the state-federal relationship and restore trust in federal education statistics.

“This idea of really partnering to create data once, to use multiple times in this sort of curation environment that we want to promote is really a key to future efficiency and effectiveness,” Martinez said. 

The idea builds on work already underway in some states. 

California has been developing its Cradle-to-Career Data Systems for years in order to link education, social service and workforce information that has historically been maintained in separate systems. The project’s first data installment included more than one billion data points across 200 different categories, including K-12, college admissions, and post-college employment opportunities. 

Experts, like Martinez, have described interoperability and data governance as foundational to making those datasets useful. 

“If this data system isn’t useful to students, to families, to the guidance counselor who has tried to help students in high school–if it’s not useful for all of them, we won’t have fully succeeded in our goal,” Mary Ann Bates, executive director of Cradle-to-Career, said in 2023 interview when the first dataset was released.

Martinez said states have also been leading the experimentation around linking education and workforce data. Regional groups of states have worked to connect education records with unemployment benefits and employment information;but, she said, without the ability to compare statistics at the national level, they face significant limitations. 

“What is lacking is national comparability,” Martinez said. “I think there is an interest and an understanding that there’s federal data that could really enhance those analyzes and that evidence is building.”

The report views NCES as a bridge between state education data systems and the federal statistical system. Martinez said federal statistical agencies are increasingly equipped to exchange data, while new privacy-enhancing technologies can allow accessibility without exposing sensitive information. 

 “It’s really more a question of what the laws permit and what the policies enable,” she said. 

State governments face similar governance challenges. While more than 30 states and the District of Columbia, have established chief data officer positions, those offices vary widely in their authority, structure and resources. A 2024 survey found that only 22% of states operated formal data quality programs, despite widespread recognition among state technology leaders that high quality data is necessary to implement artificial intelligence tools and other emergency technologies.

Recent state initiatives have shown that modernized governance structures can better support data sharing between agencies. 

Virginia’s Commonwealth Data Trust Portal replaced a manual data-sharing agreement process with a centralized system that allows agencies to request and share information more efficiently. Ohio also recently released a new statewide data-sharing compact intended to standardize agency agreements while addressing privacy, security and AI concerns. 

AI could ultimately become part of NCES’ modernization, the report notes. Martinez said the agency could use AI for data discovery, linkage and other efficiencies while also studying how schools and students are using the technology. 

The report also calls for NCES to take advantage of innovation efforts across the federal statistical system, including the National Secure Data Service, which Martinez described as a “potential innovation platform and sandbox” for AI tools.

But first, Martinez said, the agency needs to be rebuilt. 

“You can’t ask them to perform miracles every day,” she said, noting that the agency’s reduced staff leaves it without the capacity to take on any new major initiatives.