Recently we hosted a webinar featuring panelists who are committee leaders of the Open edX Adaptive Working Group. More about the working group can be found here. The panel consisted of leaders at prominent organizations like edX, McKinsey, and the MIT Lincoln Laboratory. We asked panelists a series of questions relating to adaptive learning, including their motivations for implementing adaptive learning projects, challenges they’ve seen along the way, and best practices for getting started.
Our goal was to bring attendees more than what they might find on Google and offer real, actionable tips from practitioners who are experimenting and implementing adaptive learning. In this post, I’ll shine light on some of the interesting discussion points. For the full experience, you should access the recording here, Getting Started with Adaptive Learning Webinar Recording.
Summary of Key Topics:
Motivations for Adaptive Learning
A commonly cited metric in the press is to drive greater course completion. However, Olivier Le Gouanvic, Engineering Manager at France Université Numérique, states that he’d like to use adaptive learning to “facilitate lasting memorization of material”. This moves the conversation to a more interesting metric that relates to the learner rather than the course. Additional learner centric metrics discussed included reducing time to reach competency for a given topic and improved assessment scores.
Differentiated vs Personalized vs Adaptive Learning
Since adaptive learning is a nascent topic, the definition varies considerably from group to group, and it is important to have a unified glossary of terms when embarking on an adaptive project. As Julia Mullen, Lead of Open edX Projects at the MIT Lincoln Laboratory, mentioned, “Differentiated is setting people on different paths, not necessarily adaptive paths, but perhaps starting them from different points. Adaptive uses data and algorithms to bring together a delivery method as well as a pace that is key to the way that the student has operated before. Personalized incorporates not just the data but something more about their learner profile”
Content is king and not all content is created equal
Adaptive learning experiences rely heavily on content, particularly content that is developed with adaptivity in mind. Many new courses will attempt to leverage existing content, but this isn’t always as useful as practitioners initially think. Specifically, existing offline and online content is linear in nature. Much of the existing content is long form and constructed in a linear progression. This poses two problems. The first is modularity. Adaptive learning needs small modular pieces of content. This requires organizations to spend time and money "cutting up" their content and rewriting it so it is not so dependent on consuming large chunks (i.e. not readily adaptable) of content. This adds cost, time, and complexity to projects. Second is linearity. Adaptive requires a learner to be able to jump around a knowledge graph which is not something existing content is designed for. These two items further the need for content massaging or greenfield content. “Most courses today are not designed with adaptivity in mind so you have to retroactively apply a structure and metadata that does support adaptive and...the cost of this effort could be prohibitive for many courses”, says Scott Dunn, Director of Engineering for Teaching and learning edX.
There is great promise for adaptive learning; however, as our panelists discussed, it is important to ground these expectations for the future with realities of today. Stakeholders might imagine an adaptive experience as capable as a supersonic jet while the technology can only support the equivalent of a 1920s biplane. To counteract this, as Furqan Nazeeri, Partner of ExtensionEngine states, “think about learning goals and outcomes first, then adaptive and any other toolset.” With this order in mind, you’ll stay in the realm of practical and only take on new technology when it can drive pedagogy.
Personalization is the thin end of the wedge to adaptive
One way to offer a more dynamic learner experience beyond the linear nature of most courses today is to look at personalized learner experiences, as in allowing learners different content or paths based upon their learner profile. For example, through taking a personal assessment up front, the learner could be placed in a “fast track” path or an “enriched” path if more instruction is needed. This can all be done without requiring advanced adaptive and machine learning algorithms. As Jennifer Gormley, Senior Director of Product at McKinsey Academy indicates, “Personalizing the experiences allows us to scale our digital content to any type of industry or level of tenure or level of exposure to course content which has been a challenge for us as we deliver these courses at a larger scale. For us the personalization [of content] is the thin end of the wedge to adaptive and is a manageable step forward.”
Future of adaptive learning
One of the questions asked during the panel was, “Where will Adaptive Learning be in 5 years?”. While the exact future is uncertain, the panelists offered a few interesting predictions. Ivan Ostrowicz, Cofounder of Domoscio, believes a considerable amount of learning content will become digitized. “Only 5% of content is digital, which is behind other industries. Once more and more is digital, the easier it will be to integrate and adapt the content.”
Badi Ibrahim, Director of Digital Learning at IONISX continues on a related point noting that once more content is digitized and adaptive algorithms are improved, machine learning will play an even larger role. While maybe not in 5 years, eventually machine learning will be “able to predict what learners what to learn and put learners on right path, before they even know they want to do it.”
Another prediction revolved around outcomes. Nazeeri believes with an increased amount of adaptivity and data, “We’ll be able to deliver more concrete outcomes, better learner experiences, and ultimately in the long term, cheaper learning experiences.”
Landscape of the adaptive learning industry
Adaptive learning is evolving over time and many technologies and companies are jumping in. One resource recommended was from Tyton Partners, a leading investment banking and strategy consulting firm. It wrote a Bill and Melinda Gates sponsored research report on the landscape of adaptive learning. The report contains 10 or so proprietary vendors you can use today. However, for a fully custom experience not limited by off the shelf technology you're best off having an exploratory call with ExtensionEngine. Schedule a call now
That’s it for the highlights, for the full experience you should access the recording here,Getting Started with Adaptive Learning Webinar Recording. Stay tuned for more updates from ExtensionEngine pertaining to adaptive learning and the open edX Adaptive Working Group.
In launching 70+ online programs for 40+ institutions in the last 18 months, we created an approach to learning unlike any other. We call it custom learning experiences. It's helped exceptional universities, corporations and non-profits deliver innovative online and blended learning experiences.
Why custom learning experiences? Elite institutions looking to offer unique and engaging learning experiences often struggle with what to do because solutions offered by most LMS and OPM vendors are undifferentiated, antiquated, boring or require expensive revenue sharing.
Custom learning experiences leverage the latest techniques in learning science and advances in technology that involve the learner, get them to collaborate, get them to act, thereby resulting in true learning. It creates a learning environment that is not bounded by off-the-shelf technologies or dated courses.
With a staff of over 200 experts, we help our clients with each aspect of building a custom learning experience, including project management, instructional design, UX/UI, development/QA, hosting/support and learner acquisition services. We operate on a fee-for-service basis, meaning the IP and revenue are yours to keep.