It’s good to see some serious R&D (with an emphasis on research) behind the “ed-tech” movement. It’s even better to see Pittsburgh serving as an incubator for these efforts.

The past few years have seen a veritable gold rush in ed-tech products—apps, programs, games, and other digital technology to foster learning in and out of the classroom. Venture capital investments in this area are on pace to be five times higher than in 2002, according to John Koetsier at VentureBeat. “All of education,” claims Koetsier, “is ripe for disruption.

Perhaps. But relatively few of these products to date have a grounding in child development or in an understanding of how kids learn firm enough to truly make a difference in classrooms.

There’s been such a rush into this realm that Frank Catalano of Geekwire sees early warning signs “that the hype could be outpacing the reality,” he wrote recently in “Here Comes Another Tech Bubble—in Education.”

He and others worry that “digital learning may be getting too popular among some entrepreneurs and investors … because said popularity may be increasingly for the wrong reasons, and have little to do with actually improving education.”

It’s time, in other words, for some serious research on what works and why. That’s where the National Science Foundation comes in. As Sean Cavanagh writes in Education Week, a new NSF initiative, Cyberlearning: Transforming Education, is spurring and supporting research on advanced learning technologies. The goal is to spark innovation in education technology, and test its effects, to enhance student learning.

Several of the projects that have been funded are in Pittsburgh. Researchers at Carnegie Mellon, for example, are exploring whether EEG readouts can help “intelligent tutoring” systems detect how and when students are learning and how to use that information to improve learning.

Also related to intelligent tutoring, Marsha Lovett, the director of the university’s Eberly Center for Teaching Excellence, and Christopher Genovese, a professor of statistics at Carnegie Mellon, are creating a high-tech dashboard so teachers can follow, in real time, how their students are advancing on certain lessons. Using artificial intelligence to exploit feedback data, they’re creating an intelligent tutoring system that feeds the teacher all sorts of helpful information to improve learning and teaching.

Still another Pittsburgh-based project is measuring how learning unfolds in maker spaces. The Children’s Museum of Pittsburgh’s MAKESHOP space is the site for the project’s design experiments. As Sean Cavanaugh writes in Education Week,

“The researchers hope to develop something akin to a ‘cyber-enabled critique tool’ that would allow students using maker spaces to work with a broader community and refine their work based on feedback from experts, and allow researchers to understand how that learning can be transferred to other math and science skills.”

There are at least 75 other projects underway through the NSF program.

We do know a lot about what works in education. We know a developmentally appropriate curriculum that challenges kids’ curiosity, builds on the essentials of reading, vocabulary, and math, and assess student progress regularly, can make a real difference in kids’ lives. We know that teachers who have the support (time, professional development, and feedback) to implement new strategies and new methods of teaching can have a huge impact. Studying what is most effective with digital media is the next stage in folding in smart, effective pedagogy.

Of course, the real test—and the perennial bugaboo—will be in going to scale with winning ideas. Ideas can work in one or two classrooms or under ideal conditions, but when they get introduced into the messy world of decentralized, widely divergent classrooms across the nation, many efforts break down. But with smart testing and a real understanding of what goes on in the classroom, this NSF initiative could be an important step toward truly innovative learning advances through technology.