Carnegie Mellon University will lead a five-year, $5 million project funded by the National Science Foundation to improve educational outcomes for teachers and students.
Computer scientists will build LearnSphere, an online database designed to store information on learning. Researchers will be able to use the data to study how students learn, while educators can find out how to create better courses in content and delivery.
Project leader and CMU professor Ken Koedinger said LearnSphere will help eliminate what he calls the “expert’s blind spot.”
“Since they’re designing their courses based on their conscious access to their expertise, which, by the way, is estimated to be maybe about 30 percent of what they know, they’re not going to get ideal courses from that intuition, from that conscious reflection,” he said.
By filling in the gaps, Koedinger said LearnSphere will not only improve teachers’ abilities to educate, but will lead to better, more efficient learning by the students.
“We think we can help people make courses that help students learn say two times better or two times faster or both,” he said.
The data storage system will include more than 550 data sets collected from online educational tools such as interactive tutoring systems, games and massive open online courses, also known as MOOCs.
While about 95 percent of people who sign up to take a MOOC eventually drop out, Koedinger said there still exists a vast amount of minable information.
“The volume of data is still quite high,” he said. “Even though you may start with 100,000 students poking around, if 10,000 people finish [a course], that’s a huge amount of data.”
CMU will partner with MIT, Stanford University and the University of Memphis on the project.
