American education faces a curious mix of change and stasis. Public school closures during the pandemic fundamentally changed the relationship between schools and families, visible in historic declines in students’ test scores. K-12 education is also struggling with a broader stagnation: inflation-adjusted average per-pupil expenditures have more than doubled since 1970, yet long-term test scores have remained broadly flat. While several states are experimenting with various forms of universal parental choice programs, at the federal level, another idea is gaining steam: the establishment of an education research and development entity, modeled on the Defense Advanced Research Projects Agency (DARPA).

The Biden administration and some members of Congress have backed the creation of a new National Center for Advanced Development in Education (NCADE), which, according to Oregon representative Suzanne Bonamici and Pennsylvania representative Brian Fitzpatrick’s NEED Act, would be explicitly modeled on DARPA. Institute of Education Sciences director Mark Schneider summarized these efforts earlier this year, noting the “$30 million that Congress has appropriated for IES [the Institute of Education Sciences] to incorporate DARPA-like methods into the education R&D infrastructure.”

The Obama administration proposed an “ARPA” for education back in 2011, promising innovations such as “digital tutors as effective as personal tutors,” “courses that improve the more students use them,” and “educational software as compelling as the best videogame.” Congress never authorized or funded it. But with the Biden administration launching a new ARPA-like initiative, this time with funding from Congress, now is the time to think through what a new federal education R&D project should look like.

DARPA has its origins in a time in American history not so different from the present, when a geopolitical competitor, the Soviet Union, compelled policymakers to reassess the nation’s capabilities. While the 1957 Sputnik shock famously led to the creation of NASA the following year, it also motivated President Dwight Eisenhower to create the Advanced Research Projects Agency within the Department of Defense. The original priorities were defense and space technologies for winning the Cold War, enabled by a unique research approach and a degree of independence. As former DARPA leaders Regina E. Dugan and Kaigham J. Gabriel have explained, three essential features characterized DARPA projects: ambitious goals, which pushed teams to think big; temporary project teams, composed of experts assembled to solve a specific problem or pursue a specific outcome; and the autonomy to choose and run projects without outside interference.

DARPA’s breakthroughs have focused on defense applications, naturally—for example, stealth-capable aircraft and infrared sensors used for surveillance and reconnaissance. But some of its biggest accomplishments eventually found their way into broad commercial use. DARPA’s work on electronics helped make it possible to shrink GPS receivers, paving the way for Google Maps, and the agency’s development of the ARPANET communications network culminated in the creation of the Internet. It’s no wonder that policymakers are interested in applying this model to other realms like education.

The history of government support for education research dates back at least half a century, but it’s hard to find examples of education R&D projects that meaningfully improved student learning. Even when those projects have produced significant findings, implementation has often failed. The most notable example is Project Follow Through, which involved a national classroom evaluation that began in the late 1960s and aimed to identify promising instructional methods. While the project identified Direct Instruction as the best method for teaching children basic skills—a finding consistently replicated since—the education establishment broadly ignored those findings.

More recently, Emily Hanford and others have revealed the limited impact of education R&D findings on American classrooms. Few areas of education research are better understood than the science of reading, which shows that a focus on teaching children phonics is essential. Nevertheless, many states and school districts have instead adopted the controversial “reading recovery” program, which ignores the science and lowers students’ test scores.

Education R&D thus faces a core challenge: decisions about how children learn are made by people and bureaucracies with their own agendas, which may not involve “following the science.” Moreover, the Department of Education and researchers receiving federal funding already face questions about whether their work is being disseminated or used. A Government Accountability Office study cited concerns that the federally funded Regional Educational Laboratories and R&D Center were not publishing or promoting information in a useful manner. It will take great care in design and implementation to keep an education ARPA from becoming the next missed opportunity.

So what should ARPA-ED look like? First, it should strengthen the link between R&D findings and implementation. DARPA’s findings mattered only because they yielded practical applications and discrete products. The move from R&D to implementation often involved letting the private and public sectors take DARPA’s ideas and run with them beyond the confines of the Pentagon. Too often, the education sector’s interest groups and the dispersed nature of America’s federalist system, among other factors, have let important research in education go unnoticed or ignored.

Second, ARPA-ED should follow DARPA’s policy of having only temporary projects, with clear timelines and priorities instead of endless busy work. Far too much education R&D simply has nothing to show for it. Time-limited projects would push teams to focus on the most promising ideas and discourage the wastefulness of many existing R&D efforts. It would even be worth considering moving funds from some of those efforts into ARPA-ED. Strictly adhering to the “Heilmeier Catechism,” the set of commonsense questions asked by former DARPA director George H. Heilmeier, would also channel ARPA-ED’s R&D toward practicable work, rather than the jargon-ridden fad-following that constitutes much education research.

Finally, ARPA-ED should focus on developing new tools to help students, parents, and teachers improve the ways that children learn. The tools envisioned by the Obama administration, such as digital tutors and engaging educational software, are the kinds of applications that ARPA-ED could create. Schneider has explained how IES and the National Science Foundation are partnering to “develop a sophisticated AI-driven assessment and a tailored instruction program for students with speech and language disabilities.” These are the kinds of tools that a new ARPA-inspired model should aim to provide. Here, it would be important to copy the fruitful private-sector talent exchange that occurs within DARPA; developing the best education tools will require input not just from academics but also from industry innovators.

Institutional design is key. It would be all too easy to slap the “ARPA” label and a fat budget on a new project, while failing to give it the institutional structure needed to insulate it from political forces and the bureaucratic pressures of its parent organization. But a well-designed ARPA-ED could help address two of contemporary education’s greatest challenges.

The first is helping students recover from pandemic-era learning losses. High-dosage tutoring is an effective intervention, but many public schools have been unable to provide it. A December 2022 Department of Education survey revealed that only 37 percent of public schools offered high-dosage tutoring, with almost half reporting that “a lack of funding to hire staff limited their efforts.” Developing new technologies to help more schools offer such tutoring should help remediate learning losses.

A second, broader challenge comes from accelerating international competition in STEM fields, which has made education a genuine national security issue. As a recent report from the Aspen Strategy Group argues, the U.S. must not only strengthen its innovation strategy in sectors critical to national security but also invest in the human talent to support those sectors. In other words, just as important as the tech behind a moonshot is training for the students who will one day man the spacecraft.

Fifty years of federal education R&D spending has done little to improve student learning, and skepticism about follow-up efforts is entirely warranted. But designed correctly, ARPA-ED has the potential to turn that record around.

Photo by Kevin Sullivan/Digital First Media/Orange County Register via Getty Images


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next