Among tablets and 3D TVs at CES, one-size-fits-all learning is facing a digital death knell.
Like schools of fish traveling in sync, the traditional education experience has been centered around the class, that near-random group of students following the teacher’s lead as they make their way through a course together, some succeeding brilliantly, others, not so much. But it looks like that model might just be given a run for its money with the advent of adaptive online learning.
Online learning startup Knewton and Arizona State University announced a partnership Thursday at the International Consumer Electronics Show in Las Vegas in which Knewton will apply its adaptive virtual learning platform to run two of the university’s primary introductory math courses and two remedial math courses starting this spring.
Knewton's deal with ASU, one of the largest public universities in the U.S., is a major test for the young company and for adaptive learning technology, with an expected 6-8,000 ASU student enrollments in the first year and an aggressive expansion plan.
With $21 million in venture capital funding from firms that include Accel Partners, Bessemer Venture Partners, and First Round Capital, New York-based Knewton has built up its adaptive learning system over the past two years mainly by offering online test prep courses -- SAT, LSAT, GMAT -- to schools and individual students.<!-- more -->
With the U.S. market for online learning expected to grow from $18.2 billion in 2010 to $24.2 billion by 2015 according to market research firm Ambient Insight, Knewton could very well be sitting pretty.
The company’s adaptive learning courses change in real time based on a student’s performance by collecting a significant amount of data as a student moves through course material and answers test questions.
“If a kid goes home and studies for three or four hours, that kid produces high stakes data. Right now, we just lose it all,” says Jose Ferreira, Knewton’s CEO and a veteran executive from test-prep behemoth Kaplan. “If you do your homework online, we can capture 100,000, 200,000 data points. And then we can create a personal textbook for you based on your weaknesses.”
While practically all American colleges have incorporated at least some online elements into their courses over the past decade or so, many systems are used to simply dump course content online for students to download.
To be sure, there are other companies that offer elements of adaptive learning in their online courses, and tests like the Educational Testing Services’ GRE ask questions that change based on student performance.
“The idea of offering different ways to go about learning material online pops up in a lot of other places. Pearson’s MyMathLab is one. The real question is the extent to which it does tailor to the student,” says Ben Miller, a policy analyst at Education Sector, a think tank based in Washington D.C.
As a student progresses through a Knewton course, the system gathers more and more information about the student; how he or she learns best, what bores them, what concepts they know well (or don’t know well), even at what time of day they perform best.
In an introductory algebra course, for instance, if a student is stronger with, say, linear equations and weak with exponents, the course will emphasize exponents as it progresses. If a student learns better with video, the course will add more video elements to future lessons.
Similar to the way that Google (goog) collects data based on its users’ search patterns, Knewton collects data from every student that has taken one of its courses and uses it to improve its courses. Eduational content that achieves better student results will then be ranked higher in the system and be used more often. Ideally, the system becomes smarter and better over time.
“It’s like a giant recommendation engine on steroids,” says Ferreira.
The data that the company collects could potentially prove useful to educational researchers.
“We’re still trying to map out the way people learn. Collecting more data rather than looking at course grades is a welcome change,” says Miller.
Several universities, including Virginia Tech and Carnegie Mellon, have begun to use online learning, especially for introductory courses. Introductory college courses at colleges across the country are often oversubscribed, putting significant resource strains on the institutions.
“We teach courses in freshman math groups of 30-50 students at a time. And the results are uneven across the board,” says Phil Regier, executive vice provost and dean of ASU Online, the university’s online education program.
Arizona State will be conducting its virtual math courses with two thirds of its freshman class, according to Regier. Its on-campus students will attend classes in a computer lab, with instructors available to assist students as they make their way through the courses at their own pace. They will also be able to access the course outside of the lab. ASU’s online students will take the courses in an entirely virtual environment.
At large public universities, a sizable portion of high school graduates are in need of remedial education before they can even begin to work at a college level, putting a strain on teachers and students. As a public university with approximately 70,000 students and a 90% admissions rate, ASU is part of this cohort of strained public universities.
“At any given moment, about 20% of the first time freshman at ASU are going to need a developmental math course,” says Regier.
ASU has been sending students to community college for remedial math, with less than stellar results. It will begin to offer the remedial courses with Knewton instead. The hope is that the virtual courses will be able to speed the remediation process along by quickly identifying the primary problem spots that an individual student faces and filling in the gaps before getting to the point that a student gives up altogether.
“We hope to be able to retain upwards of 500 more students this year as a result of this,” says Regier.
With its six-year graduation rate for full-time, non-transfer students at 56% as of 2009, ASU has a steep hill to climb. And virtual learning programs that focus on a student’s weak spots are only likely to be effective for those students who have at least some facility with the subjects to begin with.
“If you’re really so completely far behind that you’ve never learned it, my guess is that you still would not pass with this,” says Miller.
For several decades, if not longer, most conversations surrounding education centered on equal opportunity, whether along racial, economic class, religious, or gender-based lines. While virtual education is no silver bullet, it is coming of age at a time when educators and leaders are increasingly asking how to not only ensure that students are given a shot at receiving an education, but how to guarantee that they are educated well, a marked shift.
“We are starting to have a national conversation that looks at success instead of access,” says Miller.