What We Actually Need from a Reimagined Institute of Education Sciences
Note: This essay is the second in a three part series. You can read the first essay here.
TLDR: Not everything was working at IES, which Is exactly why we need to get the rebuild right.
There are often gaps between what researchers study and what practitioners need; these gaps vary in size and significance but they are real. The Institute of Education Sciences (IES) was not immune to this problem, as the recent report Reimagining the Institute of Education Sciences notes. Like any organization, public or private, government or not, IES had challenges. IES knew it. The field knew it. And a reform conversation about how IES should evolve was already underway well before the contracts were terminated. As just one example, when the Regional Educational Laboratories (REL) contracts were terminated by DOGE, all 10 RELs were developing professional learning toolkits to address the concern that teachers need additional support to implement What Works Clearinghouse (WWC) practice guide recommendations.
The Legitimate Critiques
IES could be slow. Within the programs I am most familiar with, the RELs and the WWC, as an example, all research went through rigorous but lengthy peer review processes. Each study typically involved multiple rounds of review at several points in the project, which tacked on many additional months over the course of the study. Rigor and quality in research are unarguably critical, but when review processes get so bogged down in the trees that the forest is forgotten, we have lost track of what is important. Meanwhile state and local education agency partners continue waiting on findings.
Similarly, IES’ prolonged annual grantmaking cycle meant districts that were ready to act had to wait a year or more for research to catch up; research grant proposals submitted in the summer did not receive a funding decision until late the following spring, an entire academic year later. Dissemination was strong in some corners but weak in others, and academic researchers were not necessarily incentivized to disseminate findings from IES research grants beyond journal articles and research conferences. Research agendas were sometimes driven more by what was fundable than what was needed. Some REL regions genuinely served their states, others felt distant and unresponsive. These critiques aren't new, and my sense was that many leaders at IES were working to address them.
What the Reimagining IES Report Gets Right, Almost
In the summer of 2025, Amber Northern, senior vice president for research at the Thomas B. Fordham Institute, was tapped by the US Department of Education to develop a set of recommendations for reimagining IES. To her credit, Northern took this task seriously, soliciting public comments from over 400 individuals and institutions. The report, released in late February 2026, included recommendations centered around six shifts. These shifts align with many of the recommendations in a 2022 report from the National Academies. They provide a solid framework, with some important caveats.
Focusing grantmaking on high-priority areas: This recommendation is hard to argue with; it was also recommended in the National Academies report, and IES seemed to already be taking steps in this direction. But if the broader goal is moving IES-funded research into practice, then states and districts need to genuinely co-define those priorities, not just receive them from Washington. That is a bigger lift than it may seem. Not only do states and districts vary in their priorities, but the co-definition process takes time and trust that can't be mandated.
Faster, more flexible grantmaking cycles: It is hard to overemphasize the importance of this recommendation. The annual cycle was genuinely broken in many ways, especially for research involving collaboration with states and districts (and was literally broken for research practice partnerships when this specific grant program was abruptly ended several years ago). By the time a grant was awarded, the practitioner partners had often moved on. Faster funding decisions and rolling submissions aren't just administrative improvements, they indicate a desire to better serve the needs and rhythms of schools.
WWC practice guides as the model: WWC practice guides distill evidence about the most effective practices in a specific area (e.g., reading instruction in the elementary grades) into a succinct and accessible summary for educators. The WWC was already focusing the bulk of its budget on practice guides, with just two percent of its funding allocated toward ad hoc study reviews. Arguably a greater challenge for the WWC was the so-called “last mile” problem: ensuring that practice guides reached the hands of educators and that educators had necessary supports to put them into practice. Professional learning toolkits being developed across each of the 10 RELs were focused on just this problem.
REL reform toward greater regional responsiveness: More than a decade ago, I was drawn to the REL program for two reasons: the opportunity to do research addressing on-the-ground state and local needs and IES’ ability to follow the evidence rather than political winds. I worked with just one of the 10 RELs and can only speak to that experience. But I saw firsthand the many months of work that partners in each jurisdiction and my REL colleagues put into defining each jurisdiction’s most pressing needs and co-creating research and technical support plans to serve these needs. Partners often expressed appreciation that REL program staff, because they were not tied to any specific local or regional entity, could deliver findings that local actors sometimes felt they could not. This is not to say that the REL program is in no need of improvement, but it is worth remembering that the trust and relationships that make REL partnerships productive take years to build – and, as the Reimagining IES report notes, some RELs were doing this well. Reform should involve learning from what those RELs were doing to achieve authentic regional responsiveness.
Building state and local research capacity: Yes, and. There is certainly value in continued efforts to support state and local partners in generating, interpreting and applying research evidence. However, capacity is not necessarily the only limiting factor when it comes to evidence-based policy and practice. Evidence competes with a range of other pressures and considerations, even when on-the-ground actors understand its importance. Too often overlooked is the need to build researcher capacity around state and local practice, preparing researchers to understand and attend to the competing priorities with evidence use.
What the Reform Conversation Underweights
There are at least two things that concern me about how the IES reform conversation could go without vigilance. First, the "last mile" problem is real yet chronically underfunded. The Reimagining IES report acknowledges it but treats dissemination largely as a communication and formatting challenge. It goes deeper than that; it is also an evidence use challenge, which makes it a relationship and trust challenge. No matter how compelling the evidence is, changing practice relies on local actors. Leaders and educators need to see that the evidence benefits their systems and their students, along with support to implement changes. That kind of work is expensive, slow, and hard to evaluate. It tends to get cut first.
Second, any rebuilt IES must keep equity at its center. The children and youth who benefit most from rigorous improvement-focused research are those who tend to get deprioritized when no one is watching. Students with disabilities, students in poverty, multilingual learners. Reform proposals that shift IES toward serving state leaders' priorities must always be filtered through an equity lens, as state leaders' priorities won’t always align with the needs of the most vulnerable students.
Where To From Here?
Reform requires infrastructure. And right now the infrastructure at IES is gone. So what do we actually do — right now, with what we have? This is the question we tackle in the upcoming final installment in this three-part series.