The Knewton Blog



Screen shot 2012-04-05 at 9.57.28 AM

When the idea of digital learning first began to gain traction, the focus was largely on two things: distribution and access. The internet, with the help of OER initiatives like MIT OpenCourseWare, allowed educational content to be disseminated more efficiently to students around the globe.

But that was a decade ago. Things evolve. Digital learning is no exception.

Last month I spoke to a room full of K-12 superintendents and higher education administrators at a conference for The Center of the Future of Arizona. It was obvious from our discussion that today’s school leaders realize that improving content distribution and access isn’t enough. What students need — and what technology now has the ability to give them — is content with context. Rather than simply using online learning as a knowledge-delivery device, schools can use it to personalize the learning process in ways that would be impossible in a traditional offline course.

Notably, MIT OpenCourseWare has now been eclipsed by a new initiative, MITx, which will launch this fall. MITx will not only provide access to educational materials, but also present learners with a certificate of completion upon proving mastery.

This ability, to receive feedback and prove mastery within an online course, is an important step. But the potential of digital learning is even more vast and impactful.

Ongoing advances in machine learning allow schools to harness the huge amounts of data generated by students in online environments. Instead of simply presenting students with the same content, Knewton can identify each student’s strengths and weaknesses and prescribe precisely what combination of lessons will be most effective for him or her. Reporting interfaces give instructors access to detailed performance feedback for individual students, groups of similar students, and whole classes, providing them a clear idea of how to direct their teaching.

The increasingly popular “course redesign movement” — the effort to focus attention on concept mastery and tangible student outcomes rather than graduation rates or seat time — is in large part a reaction to these new technological possibilities. Now, a school like Arizona State University can accomplish both their goal of increasing access, by allowing anyone who wants to to enroll in their online programs, as well as that of improving outcomes, by providing each student with a personalized learning experience particularly tailored for his strengths or weaknesses. A particularly advanced student might master the necessary concepts and finish the course early, while a student in need of remediation might use the whole semester to address the root causes of his confusion. Regardless, the system’s real-time, personalized learning recommendations greatly improve both students’ chances of success.

Great teachers have been putting content into context for generations. What’s different now is that teachers can use technology to scale this process without sacrificing quality — ultimately allowing both school and student success rates to soar.

Posted in Featured, Inside Knewton, Knewton | 5 comments



  • Nikos M.

    Do you have any statistics from the ASU course to show more concretely the improvements of the students taking the adaptive course?

    • http://www.knewton.com Knewton Team

      Hey Nikos,

      Yes, we do! Initial stats are really promising. In the first semester that Knewton Math Readiness was used at ASU (with over 5,000 students), course withdrawal rates dropped by 50% and pass rates went from 66% to 75%. Plus, half of the students completed the course 4 weeks early — a great sign that they were able to move through the material at their own pace without wasting time. 

      If you’re interested in reading more, check out this article in Forbes: http://www.forbes.com/sites/bruceupbin/2012/02/22/knewton-is-building-the-worlds-smartest-tutor/

    • David L.

      Hi Nikos, we have some early results from our work with ASU’s developmental math course (Fall 2011) that show mastery levels improving from 66% to 74.5% yr over yr, withdrawals being reduced from 13.6% to 5.6% yr over yr, and 50% of the students finishing the class 4 weeks early.  We’re excited, but we have a long way to go and will continue to improve upon these metrics.

  • Anthony Arroyo

    Great post! As someone who has taught (in an actual classroom filled with humans, imagine!) for 6 years, I’m sometimes worried that people working on exciting, ed-tech stuff think teaching and learning is purely about the effective transfer of information. It’s also about feedback, structure and motivation. I’m glad to see that you and the Knewton team are keeping this stuff in mind. 

  • ChristopherCadwell

    I know of institutions that use “tangible student outcomes” and varied learning drills, along with checklists and student tailored schedules (learn at your own pace); they have been using these techniques for decades with great success. It works. 

    What has not been done is to gauge that this student does better with this learning type drill – so we will do more of that with him because it works better with him. Well maybe Montessori has this built in to some degree.

    It will be interesting to see how this is able to be put into a digital cloud.

    Looking forward to it.