IES and Accelerating Learning Recovery

Mark Schneider, Director of IES | July 5, 2023

America's students are struggling.

That's according to the recent release of NAEP's long-term trend report on the math and reading scores of 13-year-old students in the United States, which adds the finishing touches to NAEP's current cycle of reports showing just how harmful the COVID-19 pandemic was to student achievement. As reported widely, scores fell dramatically across the entire population of 13-year-olds and across just about every subgroup of students NAEP reports on. The depth of the learning losses that NAEP and other assessments have documented shows that learning recovery is the key task facing educators and policymakers today.

Here are several avenues IES is pursuing to help speed recovery.

Making better use of NAEP data

NAEP calls itself "The Nation's Report Card"—and like almost all report cards, it serves as an invaluable marker for student performance at a particular point in time. NAEP is not currently as useful as it could be identifying strategies for how to improve student performance and speed recovery. We could be doing more, but this would require us to broaden our view of NAEP beyond its essential role as a report card.

The unfolding AI revolution gives NAEP the opportunity to occupy center stage. As is now well known, the large language models that underlie generative AI require vast quantities of high-quality data to produce results. But there aren't many large, high quality educational data sets available and far too much of the data that is now used in existing generative AI models is questionable. As John Whitmer and I observed in a recent Ed Week article, large-language models often do not have access to enough student academic work to make useful observations about the kinds of mistakes students make and how we can help them improve. This is where NAEP can come into play, with its huge nationally representative samples of student work in reading, writing, math, science, and civics.

IES is encouraging the best researchers throughout the country to access and use NAEP data to employ AI to identify and improve interventions that could improve student learning. Part of this encouragement has come in the form of IES-sponsored competitions using NAEP data to improve student performance in the subjects NAEP assesses.

In short, NAEP data can and should be used to help identify ways of speeding student recovery—but we must do more to ensure researchers and developers know that this resource is available to them.

Supporting Research on Learning Recovery

IES received $100 million in American Rescue Plan funding to help the nation's students recover from the learning losses of the pandemic. Unfortunately, in the recent debt ceiling deal, all unobligated ARP money was returned to the Treasury. Because IES funds grants annually, around $44 million of the $100 million was being held to fund continuation of ARP-funded research programs. IES had to return that money. Four grant programs were most affected by this recission:

NCER Improving Pandemic Recovery Efforts in Education Agencies (2 Research Networks)
NCER Using Longitudinal Data to Support State Education Policymaking
NCER Learn Network
NCSER Research to Accelerate Pandemic Recovery in Special Education

Consistent with the terms of the ARP money, all these grant programs specifically focused on learning recovery. Given the importance and quality of the work these four grant programs support, we have decided to continue funding them using our "regular" appropriations. We realize that learning about how to improve and accelerate recovery after the most significant education challenge of our lifetimes must take precedence—but this means that less money will be available in the next two years to fund new work.

NCADE as a Path Forward

The COVID crisis has exposed many of the weaknesses of the education R&D infrastructure — in particular, the need to get evidence-based interventions rapidly developed and deployed in an instructional context that is radically changing. One potential fix is to employ DARPA-like approaches to education research and product development. Discussions of establishing a unit within the Department of Education that could emulate DARPA's rapid cycle, high-risk/high-reward transformative work have been going on for decades. It is probably not coincidental that discussions of establishing an ARPA-ED gained traction post-pandemic. This is evident in legislative action: ARPA-ED is included in ongoing discussions about the reauthorization of the Education Science Reform Act and sits center stage in in the proposed NEED act, which would create an ARPA-ED unit named the National Center for Advanced Development in Education (NCADE) within IES.

We believe that NCADE is essential to the modernization of education R&D in the long run, and, in the short run, it's needed to identify ways that we can employ education technology and the learning sciences to improve student performance and help students recover from the pandemic.

On June 8, we released an RFA asking for proposals that support innovative and transformative research. This is the first program tapping the $30 million that Congress dedicated to ARPA-like research.

Common among established ARPAs is the use of "seedlings." IES is planning a seedling program that will invest a limited amount of money in a cohort of programs for about one year. At the end of this period, the programs are evaluated, and the successful ones become eligible for a second round of funding for another year or so. Those that meet their second-round objectives will be eligible for a third round of greater funding focused on scaling up the proven intervention or product to affect tens of thousands of students. A common theme or set of themes will unite our seedlings (for example, neurodiversity) so that the seedlings can grow into a force that might affect a whole ecosystem. In the meantime, as the first cohort matures, IES will continue to fund new cohorts.

Watch for an official request for information (RFI) and other outreach activities to engage our stakeholders in this exciting new approach. And watch for other announcements as IES builds out more APRA-like activities.

The tighter fiscal environment that seems inevitable for the next two years or so will lead IES to make difficult choices. But the pandemic has increased awareness that IES has a central role to play in leading the nation out of the education crisis that was forming even before COVID-19. The paths outlined above should lay the foundation for a more modern, more agile, and more effective IES in the years to come.

As always feel free to contact me: mark.schneider@ed.gov