ARPA for Education FAQ
The Alliance for Learning Innovation has been advocating to get an Advanced Research Projects Agency, or ARPA, for education off the ground. Since the creation of the Defense Advanced Research Projects Agency (DARPA) in 1958, the U.S. government has invested in informed-risk, high-reward research and development (R&D) that’s led to groundbreaking innovations like the internet, GPS, and mRNA vaccines. Given the success of DARPA’s innovative approach, nearly every sector now has its version of DARPA – from energy to intelligence to health.
At a moment when American students are still struggling to recover from the pandemic’s learning disruptions and the education status quo is failing to adequately equip young people for the jobs of the future, the call for an ARPA for education is growing. A bill soon to be reintroduced in Congress, the New Essential Educational Discoveries (NEED) Act, would create an ARPA for education called the National Center for Advanced Development in Education (NCADE), housed within the Institute for Education Sciences (IES).
NCADE would have a nimble management structure that prioritizes solutions aligned with the science of learning and development, and that have the potential to improve student achievement dramatically and close longstanding gaps between advantaged and disadvantaged students. In particular, NCADE would invest in breakthrough technologies; new pedagogical approaches; innovative learning models; and more efficient, reliable, and valid forms of measurement of student learning, experiences, and opportunities.
As an increasing number of education leaders and community members advocate for NCADE, some common questions are cropping up. We aim to address them with this FAQ.
What makes a National Center for Advanced Development in Education (NCADE) different from other R&D programs at the U.S. Department of Education?
NCADE would invest in high-potential, high-impact ideas to drive innovative solutions to education’s most challenging problems. These would be “moonshot” ideas, similar to those developed by other ARPA-type models.
What would set NCADE apart from other Centers at IES is its focus on the development side of R&D. All of the existing Centers at IES are dedicated primarily to education research but NCADE would ensure much-needed attention to the development of new approaches to teaching and learning. Building on existing basic research, NCADE would allow for the rapid testing of solutions and provide space to take risks and learn from failures.
Other existing R&D programs within the U.S. Department of Education, like Education Innovation and Research (EIR) grants and IES’ Small Business Innovation Research (SBIR) grants do not support early-stage development of bold innovations. Rather than duplicating NCADE, EIR and SBIR would complement it by scaling its successful innovations, or preparing them for commercialization. Examples of projects that are ripe for early-stage development but not yet ready for EIR or SBIR include:
The development of a virtual or augmented (VR or AR) reality tool to teach geometry through enhanced visualization and student engagement.
The creation of a robust dataset that maps effective instructional behaviors (e.g., prompting a diversity of students to participate in classroom discussion) to positive student outcomes.
Leverage machine learning to develop a Netflix-inspired auto-recommender for personalized learning exercises that match a student’s readiness and interest.
Specifically, NCADE differs from EIR in that it would not require its projects to fit into one of the four Tiers of Evidence described in the Every Student Succeeds Act (ESSA). NCADE could therefore invest in informed-risk, high-reward projects based on scientific and technical merit, like other ARPAs. EIR uses three grant programs to generate, validate, and scale educational innovations. The grant programs – early phase, mid-phase, and expansion – are awarded depending on whether the innovation is under development or has already proven to be effective. Even for EIR’s early phase grants, prior evidence is a requirement – so EIR could not support the moonshot ideas that NCADE or an ARPA-like process could.
NCADE would be distinct from SBIR in that it would focus on early-stage development, rather than later-stage commercialization, and it would not solely fund edtech. SBIR provides funding to promising edtech for rapid prototyping, development, and evaluation whereas NCADE would fund a more diverse set of highly innovative projects, which could include bold new approaches to pedagogy or learning models.
Why should the federal government rather than industry fund these types of high-potential, high-impact projects?
Most innovation in education happens in the private sector. As a result, the innovations are proprietary, which dramatically limits the amount of new research and development that can come from them. Furthermore, the education market is highly fragmented and does not lend itself to transformative R&D. There are nearly 14,000 school districts in the U.S., making it extremely difficult to achieve scalable impact for new products. For many private companies, it is a better business decision to build safe tools that are guaranteed to sell as opposed to taking the risks associated with the development of breakthrough solutions.
What about philanthropy? Isn’t this the sort of thing big foundations could fund?
Many smaller nonprofits and social entrepreneurs already rely on philanthropic capital to support their research and development. However, government funding would provide a more sustained and reliable source of support for high-impact, high-potential.
I’ve heard that some ARPAs with small budgets haven’t been successful. Is NCADE worthwhile if the federal government can only make a small investment in it?
There is evidence that smaller investments can lead to groundbreaking innovations. The Experiment Foundation is testing this out for programs with budgets totaling $50,000 to $250,000 – at a cost of less than $10,000 per project. Early results are promising, such as CalWave, which started as a $9,000 project and is now one of the longest-duration wave energy projects in the world. While other ARPAs tend to be funded at significantly higher levels, there is value in establishing a smaller version of NCADE, which would simply fund fewer projects. The current ARPA-like pilot at IES (called the Accelerate, Transform, and Scale, or ATS, initiative) is putting $30 million into high-impact, high-potential education R&D. The Senate version of the NEED Act authorizes “such sums necessary” for NCADE, and would allow for some discretion to scale the program as its budget permits.
Does NCADE need to be a new Center within IES?
The current proposal is designed to ensure that it has the structure, flexibility, and talent needed to be true to the ARPA model. However, it is worth discussing a different model if it would:
Bring in the brightest scientific and technological minds by hiring and paying program managers outside of the traditional civil service rules, like at other ARPAs.
Allow NCADE project selection to operate outside of IES’ standard peer review process to give program managers significant autonomy – a critical tenet of the ARPA approach.
ARPA-like models work well in industries that are steeped in hard sciences, like medicine, defense, and energy. Why would it work in a field like education?
The truth is, the U.S. has never tried to apply the successful ARPA model broadly to education.
While education is a people-driven field, technological advances in artificial intelligence (AI), large language models (LLMs), and VR and AR, are creating unprecedented opportunities to develop new approaches to teaching and learning in education settings. To do that effectively, however, developers need opportunities to test their ideas and learn from their failures. NCADE programs will invest in approaches and technologies that have the potential to transform education in new ways that we can’t even imagine today.