Is Adaptive Learning the Way to Get Schools into Learning Games?

Adaptive learning might be the hottest catch phrase in ed tech at the start of 2016, and some of the biggest education heavyweights are gearing up to pitch their own versions of the learning technology to schools.

Pearson announced last month they intend to steer their products toward more “adaptive, personalised ‘next generation’ courseware,” especially in STEM subjects “where enrollments are growing and which lend themselves to this approach.” The British company released a series of case-studies touting the benefits of their MyLab series and its real-time adaptive learning feedback which they say leads to higher exam scores. These announcements arrived in the aftermath of a 10 percent cut in the company’s workforce.

High-profile adaptive learning startups like Knewton have attracted more than $100 million in international venture capital investments for its own “big data” platform. And some ed tech experts have said the adaptive capabilities of web-based programs like McGraw-Hill’s ALEKS could replace textbooks altogether.

Meanwhile, Seattle-based adaptive learning platform Enlearn is using game-based learning to try and carve out their own space in an increasingly crowded adaptive learning market. They’re hoping a recent $3 million grant from the Bill & Melinda Gates Foundation will help lift them out of the pilot phase, and get the platform commercially available and into classrooms across the country in time for the 2016-2017 school year.

Popovic with his original game FoldIt.

Popovic with his original game Foldit.

Enlearn founder Zoran Popović is director of the Center for Game Science at the University of Washington, and is perhaps best known for co-creating Foldit, a biochemistry game that inspired a community of players so devoted to tinkering with complex proteins they collectively mapped the structure of an HIV protein vital to the disease’s replication process for the first time, giving researchers new ideas for drug treatments.

In the wake of Foldit’s success, Popović had even bigger ambitions in mind for an adaptive learning platform that used a unique algorithm in concert with gameplay to engage students — and revolutionize the way we teach our kids.

“I had many people propose to start companies with me, but I didn’t want to start something if I couldn’t make a dent in the universe,” Popovićh told gamesandlearning.org.

Since 2012, Popović and Enlearn CEO John Mullin have received just over $7 million in grants from the Bill & Melinda Gates Foundation College Ready program, which has been the financial backbone behind fours years of tests, research and new relationships with educational assessment providers and game developers, all in support of their version of adaptive learning which they’ve dubbed “Generative Adaptation.”

Popović adapted a version of the award-winning DragonBox math series, combined it with Enlearn’s platform and piloted an extensive series of competitive math programs in Washington, Minnesota and Norway.

The initial results were impressive: In less than two hours, 93 percent of kids across K – 12 who used Dragonbox Adaptive achieved mastery rates in basic algebra concepts that usually take about a month to teach in the average 7- or 8-grade class.

Popović said he believes the promising data doesn’t only come from Dragonbox’s colorful drag-n-drop algebra gameplay, but the platform’s ability to adjust to the player’s learning needs in real time by giving them an infinite amount of individualized content based on how they’re performing. This adaptation, Popović said, gives teachers more freedom to escape from behind the podium and provide more one-on-one help to students.

“If we just focus on students by themselves, turns out, and I’ve done many studies on this, you can’t make nearly as much difference as when you integrate teachers and the entire classroom in the process,” Popović said. “Which means you need engagement. I need them to wake up in the morning and say, ‘I have an idea. I want to go and try this,’ which is very different than the standard badges and other things that we see today.”

The promising data generated by Enlearn’s Dragonbox Adaptive trials has spurred new relationships with additional content providers, and not just in mathematics. In March, Enlearn announced a partnership with curriculum and assessment provider Voyager Sopris Learning to integrate English Language Arts content into their platform.

The nonprofit has also looked overseas, partnering with India-based EkStep to develop two as yet unnamed math games for 200 million first generation learners for a March release date.

Popović described one project as a counting game in which players learn basic math skills by loading and unloading chickens. The deal represents a new scale of work for the firm, he said, adding the Indian sales team hopes to sell upwards of half-a-million games per grade.

“We’re not doing a small intervention. This is five years of material. [Dragonbox Adaptive] was one month of material. We’ll get pre- and post-data and see how it goes,” he said.

A Problem of Definition

Even though adaptive learning is attracting hundreds of millions of dollars in venture capital and generating promising data in test trials, there’s still a lot of uncertainty about how adaptive learning platform will work together with learning apps and how schools will integrate the two in class.

One approach is to completely change how schools operate. Schools like AltSchool have fully-embraced “big data” in the classroom to develop and pilot eight schools in San Francisco and New York. Founder Max Ventilla believes their lack of school bells and report cards combined with a unique approach to student data collection and real-time feedback will revolutionize education and eventually cater to students from underprivileged communities in addition to the tech-elite areas in which AltSchools currently reside.

But most traditional schools are struggling with the growing pains of modernizing classrooms and it’s still unclear how and when adaptive learning will shape content given to students.

A 2015 survey found half of school administrators said they make adaptive learning tools available in their schools while only about a third of teachers said they actually use them in the classroom.

Part of the challenge may be explaining what is a fairly complex idea in an elevator pitch, admitted Mullin.

“We haven’t found the easy way to get these messages across in 30 seconds or less,” Mullin said, even though the Enlearn and Dragonbox Adaptive earned a fair amount of attention after releasing the results of their initial pilot program.

It’s hard for a consumer to understand the nuances between the different types of adaptive learning. I think there are shades of definition of what that means, and I think some companies genuinely believe their products are adaptive. It doesn’t mean they’re wrong or that there’s an intent to mislead. But would I be shocked if there are some companies who use the term ‘adaptive learning’ to help market their educational products? No.
— John Mullin, Enlearn


A Fancier Skill and Drill?

And some of those who do understand the difference aren’t fans of the idea. One game-based learning veteran has called adaptive learning platforms the “reindustrialization of American schools.”

“If there’s truly to be a revolution, it won’t be through Generative Adaptation or any kind of adaption. We need to start over,” Robert Clegg recently told gamesandlearing.org.


Some critics worry adaptive learning could become simply a new form of old schooling.

Some critics worry adaptive learning could become simply a new form of old schooling.

Clegg is CEO of educational game developer Knoverse and has produced corporate training games for the NFL, Nike and the Bill & Melinda Gates Foundation. His Halo-style educational game DimensionU has tried to teach STEM principles to players since he helped create and launch the game and its multiplayer network a decade ago.

According to Clegg, the biggest problem with adaptive learning platforms is they don’t harness the fundamental reason why people love games: motivation. He argues that one of adaptive learning platforms biggest draws for administrators— the ability to generate an infinite amount of content to individually match the learning needs of any student — is worthless unless you can get kids excited to learn in the first place.

You all have these kids that have learned math by sitting there and doing equations until their eyes bleed. That’s the worst kind of learning. It’s like getting the cheat guide to a game. If you look at concepts like real-time hinting, it takes out motivation and the concept of discovery out of solving a puzzle.
— Robert Clegg, Knoverse

Clegg believes the best games teach almost as a side-effect of playing. However, games like Popović’s Foldit — which Clegg believes is a work of art — are an effective teacher of biochemistry because the people who play it are already motivated by their interest in biochemistry, not by folding proteins to solve big research problems.

“The moment you talk about developing [academic] content to these foundations and venture capitalists, they want one game that helps everybody and a platform that scales world-wide. Nobody wants to fund ten different games, they just want one,” Clegg said.

Although adaptive learning critics like Clegg and the creators of Enlearn’s might disagree on the technology’s efficacy, both Popović and Mullin are quick acknowledge the role motivation plays in learning, their platform is not a singular fix for revolutionizing the classroom, and effective learning consists of more than helping students become exceptional plug-n-chuggers.

“I completely agree,” Mullin said. “100 percent of what we’re doing is trying to look at the whole cognitive and meta-cognitive conditions of the classroom. We’re not looking at just problem solving, we’re looking at their motivation, engagement, persistence and their mindset, to understand the whole student. Learning is a contextual and social thing.”

Mullin said they’re working with “one of the largest testing agencies in the U.S.” to expand the platform’s assessment capabilities, college level remediation and professional teacher development.

“We need to built up the core of what we can provide, allowing [Enlearn] to go to college level remediation, as well as professional development of teachers. We’re only halfway there. One of the mistakes in [learning research] is you say, ‘Let’s isolate for curriculum or the social aspect.’ And then you put the improved thing into the class. But that’s not the environment. Classes change from moment to moment. And that’s why robot teachers don’t make any sense.”

Despite the challenges, both Popović and Mullin remain optimistic about Enlearn’s potential while acknowledging shortfalls.

“It’s true that to be maximally accessible to a kid, you can’t get away with one game, you probably need to have several games to catch a lot of students,” Popović said. “When we held the [algebra pilot test] challenges, the key measure of the entire challenge was the percentage of engagement.”

Tags: , , , ,

Christopher B. Allen Christopher B. Allen is a contributing editor for gamesandlearning.org, as well as a radio producer and anchor for Montana Public Radio. He graduated with a Bachelor of Arts in Journalism from the University of Montana, and won 1st place in the 2014 National Hearst Journalism Awards for radio broadcast. Chris is also an Air Force veteran.