The federal government's premier education research agency has grown slow, disconnected from classrooms, and unable to deliver the timely findings that teachers and state policymakers desperately need, according to a blunt internal assessment submitted to U.S. Secretary of Education Linda McMahon. 
The 95-page report, authored by Dr. Amber M. Northern, a senior advisor at the Department of Education, recommends a fundamental restructuring of the Institute of Education Sciences (IES) — a $700-million-a-year agency that funds the bulk of rigorous education research in the United States. Despite two decades of scientific achievement, the report concludes that IES has too often prioritized academic precision over practical use, scattering its grants across hundreds of disconnected projects while producing data that can take years to publish and research that rarely makes it into a classroom.
The critique lands as the nation faces a deepening academic crisis. National test scores released last year showed that students across all grades and subjects were still performing below pre-pandemic levels, with no state achieving reading gains in 4th or 8th grade and a record one-third of 8th graders scoring below basic proficiency in reading. With the Trump administration pushing to return education authority to states and reduce Washington's footprint in schooling, the pressure on IES to demonstrate its value has never been greater.
The report, titled "Reimagining the Institute of Education Sciences: A Strategy for Relevance and Renewal," contains more than 50 recommendations built on interviews with nearly 400 people — researchers, state officials, district leaders, teachers, and technology experts — and more than 230 public comments. The findings describe an agency that, despite genuine scientific accomplishment, has drifted from its core mission of generating knowledge that improves what happens in schools.
Chief among the report's findings is that IES spreads its research dollars too thin. In recent years, the agency's main grant competition allowed applicants to submit work under roughly a dozen broad topic areas, producing what the report calls a "scattered research agenda" that rarely builds toward coherent solutions. Northern recommends that IES instead focus on three to five high-priority challenges — such as early literacy or algebra achievement — identified with input from state and district leaders, and that all four of the agency's centers work in parallel on those same problems.
The report saves some of its sharpest language for the National Center for Education Statistics, the agency's statistical arm. NCES manages dozens of surveys and data collections, many of which, Northern writes, have never been comprehensively reviewed since their creation. Some surveys overlap; others ask outdated questions or rely on inconsistent definitions. Critically, the various datasets are not designed to talk to each other, creating information silos that prevent the kind of holistic analysis policymakers need. The report also notes that two of NCES's flagship longitudinal studies — each costing more than $46 million — were quietly cancelled in early 2025.
Speed is another recurring complaint. IES products are "often out-of-date upon release," the report states, limiting their usefulness for practitioners making real-time decisions. Northern calls for automated data verification, standardized state reporting forms, and expanded use of application programming interfaces, the digital plumbing that would allow researchers, policymakers, and even AI tools to query federal education data directly, rather than waiting months or years for reports to be compiled, reviewed, and published.
Dr. Frank Fernandez, an associate professor of educational leadership and policy analysis at the University of Wisconsin, Madison, said that it is disingenuous to give IES a mandate to do work and "not give them the staff to do it, encourage them to contract the work out so that it will, ostensibly, be more efficient, and then critique them for using contractors."
The What Works Clearinghouse, the agency's well-known database of evidence-based interventions, would be dramatically scaled back. Northern proposes that the clearinghouse stop reviewing individual studies — a resource-intensive process that can take up to three years and that researchers have little incentive to supply — and instead focus exclusively on publishing practitioner guides, which stakeholders repeatedly described as its most useful product. AI tools, she argues, could take over the task of synthesizing research once enough evidence has accumulated on a given topic.
The report, said Fernandez, "raises some interesting points about IES research being so focused on prioritizing research that is rigorous and causal and that is not specific to local needs or groups of students." However, he cautioned that changing IES' approach to research still won't address the most local needs that impact whether and how well children can learn like poverty, hunger, underfunded schools, school segregation, or long commutes.
"If a child wakes up well before dawn to catch the bus, ride around in the cold for an hour, shows up at school hungry, and the school is under resourced and unsafe, an IES practitioner guide is not going to be the thing that is going to teach that child to read better or to do algebra," he said.
The report is also sharply critical of the Regional Educational Laboratories (REL), a network of ten geographically organized contractors that provide research and technical assistance to states.
Feedback from state chiefs was "generally more negative than positive," with leaders complaining that the labs are unresponsive, disconnected from local needs, and often staffed by researchers with little regional expertise. In an ideal world, Northern writes, many state leaders would simply prefer to receive their share of the REL funding directly and hire their own providers. Because that would require an act of Congress, the report instead recommends creating a coordinating hub to better align REL work with state priorities and giving states more flexibility in choosing who serves them.
In the report, NAEP — the National Assessment of Educational Progress, widely regarded as the gold standard of American education data — gets its own chapter, with Northern calling its technology "finicky" and "in desperate need of modernization." The report calls for AI-assisted test development, cross-platform compatibility, and a more transparent accounting of NAEP's contract costs. It notes that the current administration model, which sends trained contractors to schools nationwide to proctor the exam, accounts for nearly 29 percent of the program's budget.
"IES has struggled to stay fully relevant to on-the-ground realities," the report states, and has failed to "focus its efforts on the most pressing challenges, make its data truly usable to those who depend on them, and deliver timely insights."
Even the agency's strongest supporters, the report notes, say its research "sometimes strays into ideological waters and relies too heavily on contractors."
But Fernandez expressed concerns that the report might be focused on the wrong things. "If we're going to be serious about improving education in the United States, we should be candid about addressing the social problems that children and schools live with rather than argue about how IES should improve its contracting processes,” he said.
Dr. Matthew Soldner, IES's acting director, said that he looked forward to considering how Northern's recommendations "can further that work and help ensure educators and policymakers have the information they need to make a positive difference in the lives of our nation's learners."
IES accounts for just 0.3 percent of the Department of Education's annual spending — a figure the report uses to underscore the stakes of getting it right.
"This small statistics and research shop," Northern said, "remains committed to that mission, even as it needs to undergo significant reforms to carry out that mission much faster and better."
