The Magical – and Flawed? – DARPA Analogy
President Obama’s 2012 budget proposes to create an Advanced Research Projects Agency for Education—also known as ARPA-ED—to address what the administration says is an under-investment in learning technology, with the ultimate goal of transforming education through the creation of digital tutors as effective as personal tutors, courses that improve the more students use them, and educational software that is as compelling as the best video game.
Having all of these elements at hand in a transformed ecosystem of learning would certainly be a game changer. The question, though, is if creating ARPA-ED will catalyze this transformation, even as it does bolster the country’s paltry spending on education research and development, which today only accounts for 0.2 percent of the country’s K-12 expenditures on education.
ARPA-ED is modeled after the Defense Advanced Research Projects Agency, which is more popularly known as DARPA. DARPA commissions advanced research for the Department of Defense and has taken on a bit of a mystique based on its early funding of projects that have led to the creation of such blockbuster innovations as the Internet and GPS, which have not just transformed the military, but also commerce, travel, and our daily lives.
ARPA-ED “will fund projects run by industry, universities, or other innovative organizations” that have the potential to transform teaching and learning. The Department of Education has a great primer on the thinking behind and the plan for ARPA-ED here.
The challenge with this initiative—and others such as the creation of the Advanced Research Projects Agency-Energy (or ARPA-E)—is that creating agencies to spark innovation modeled on the “best practices” of DARPA may very well fail not because they are implemented unfaithfully, but because the circumstances in which each operate are starkly different. We have written in Chapter 8 of Disrupting Class about the significant dangers of following “best practices.” The potential pitfalls are easy to see here.
DARPA essentially works by funding projects that have a three- to five-year time horizon and address specific technology challenges that are important to the Department of Defense—its customer. One of the key reasons DARPA works well is because its customer’s needs are understood clearly; the Department of Defense is a consumer of these research projects. If DARPA funds something that solves a challenge, the military is nearly certain to adopt it, so the market functions well. To put it in economic terms, there is a demand from the military (which is funded by the federal government), and the federal government funds the supply side as well. The two match up.
ARPA-ED, like ARPA-E, however, does not have a clear-cut connection between supply and demand. Although some may point out that public school districts—the customers for the vast majority of educational products and services in this country—are government funded (unlike the consumers for energy in many cases), so it should be simple to align supply and demand, it’s anything but.
Because districts receive funds from so many disparate sources at the state, local, and federal level with various strings attached, and all have different local political considerations at hand, in practice it is quite hard to predict demand. In her book Educational Economics, Marguerite Roza shows how districts don’t even realize that their strategy, in practice, is very different from what their intended one is, as their actual spending and resource allocation process do not align at all to their stated goals. It’s not that the market is fragmented per se, but that districts act in seemingly quirky ways and often define the requirements of their jobs to be done differently from each other; there are long buying cycles in education; and districts often break contracts when they have changes in their leadership, all of which make it hard for innovative products and services to penetrate the market.
The reason, then, that companies don’t invest lots of dollars in basic R&D in education isn’t because they don’t want to fund risky projects per se—one stated reason for ARPA-ED—it’s more because the market won’t support this (more on this below).
Now there still may be a case for the government investment of dollars into truly basic, risky research, as the government does in the sciences. The paltry level of overall investment in research and development combined in education suggests there is a need for this, but ARPA-ED won’t be funding risky, basic research that may have long time horizons and little market applicability. It will purposely invest in what it calls “directed development” instead–the other half of R&D. Directed development, the Department of Education says, “provides the ability to pursue a small number of high-impact projects, from concept through demonstration or prototyping. Directed development projects begin with a specific end goal, rather than the aim to increase broad areas of knowledge, and generally include a defined time period and path forward. Directed development focuses on advancing beyond the state-of-the-art such that the activities of the field are unlikely to produce the desired outcome in the prescribed time frame.”
This means that the federal government is taking aim at supposedly market-oriented projects—the kind that companies won’t fund because there is no market. The federal government is immune to market pressures, but it’s also notoriously bad at predicting district demand itself. This means it may fund the creation of some neat things, they very well may not get adopted widely, and therefore the initiative will transform little about the education system.
To be fair, the Department of Education seems to have anticipated some of this problem. It has a small section in its ARPA-ED plan about nurturing demand, as it suggests that “the Subcommittee will use a combination of policy advocacy and ‘pull’ mechanisms such as Advance Market Commitments and working with the Department of Defense Education Activity (DoDEA) schools to serve as platforms for innovation and as early ‘customers’ for advanced effective education technology. For example, in the recently finalized Presidential Study Directive on Military Families, DOD committed to making DODEA a ‘leader in the use of advanced learning technologies.’”
This last thing is quite exciting, as DODEA represents the only place where the federal government has direct control of the school system, has a $1.8 billion annual budget, and therefore could potentially catalyze some interesting demand, but the reality remains that beyond this, anticipating and stoking demand is critical for this initiative to work and a huge question mark and difference from DARPA. And the federal government has not historically been good at altering the demand landscape in K-12 education—and the answer that Race to the Top and i3 have changed this is far from a certainty.
If ARPA-ED does move forward and, after a few years, the Department of Education suggests that the initiative is a success because of the investment it has seeded independent of the success in actual adoption, this will be a sham. As I’ve written before, investments in and of themselves are not a sign of success. ARPA-E, for example, has fallen back on this and claimed that it shows that this model can work outside of the defense industry because it has managed to award nearly $400 million since April 2009 to over 100 projects. Why it’s a shock that researchers and organizations would be willing to submit papers and projects in exchange for money is beyond me. Real success would instead mean that the investments in these initiatives have changed widespread practice, just as DARPA has helped change how the military conducts wars.
There are several reasons to wonder if they can do so.
– Michael B. Horn