Skip to content

Adaptive Learners, Not Adaptive Learning

Some variation of adaptive or personalized learning is rumoured to “disrupt” education in the near future. Adaptive courseware providers have received extensive funding and this emerging marketplace has been referred to as the “holy grail” of education (Jose Ferreira at an EdTech Innovation conference that I hosted in Calgary in 2013). The prospects are tantalizing: each student receiving personal guidance (from software) about what she should learn next and support provided (by the teacher) when warranted. Students, in theory, will learn more effectively and at a pace that matches their knowledge needs, ensuring that everyone masters the main concepts.

The software “learns” from the students and adapts the content to each student. End result? Better learning gains, less time spent on irrelevant content, less time spent on reviewing content that the student already knows, reduced costs, tutor support when needed, and so on. These are important benefits in being able to teach to the back row. While early results are somewhat muted (pdf), universities, foundations, and startups are diving in eagerly to grow the potential of new adaptive/personalized learning approaches.

Today’s technological version of adaptive learning is at least partly an instantiation of Keller’s Personalized System of Instruction. Like the Keller Plan, a weakness of today’s adaptive learning software is the heavy emphasis on content and curriculum. Through ongoing evaluation of learner knowledge levels, the software presents next step or adjacent knowledge that the learner should learn.

Content is the least stable and least valuable part of education. Reports continue to emphasize the automated future of work (pfdf). The skills needed by 2020 are process attributes and not product skills. Process attributes involve being able to work with others, think creatively, self-regulate, set goals, and solve complex challenges. Product skills, in contrast, involve the ability to do a technical skill or perform routine tasks (anything routine is at risk for automation).

This is where adaptive learning fails today: the future of work is about process attributes whereas the focus of adaptive learning is on product skills and low-level memorizable knowledge. I’ll take it a step further: today’s adaptive software robs learners of the development of the key attributes needed for continual learning – metacognitive, goal setting, and self-regulation – because it makes those decisions on behalf of the learner.

Here I’ll turn to a concept that my colleague Dragan Gasevic often emphasizes (we are current writing a paper on this, right Dragan?!): What we need to do today is create adaptive learners rather than adaptive learning. Our software should develop those attributes of learners that are required to function with ambiguity and complexity. The future of work and life requires creativity and innovation, coupled with integrative thinking and an ability to function in a state of continual flux.

Basically, we have to shift education from focusing mainly on the acquisition of knowledge (the central underpinning of most adaptive learning software today) to the development of learner states of being (affect, emotion, self-regulation, goal setting, and so on). Adaptive learners are central to the future of work and society, whereas adaptive learning is more an attempt to make more efficient a system of learning that is no longer needed.

8 Comments

  1. Mike C. wrote:

    This is part of what I’m trying to get at with the recent “choral explanations” piece. It’s admittedly about content, but the key to choral explanations is that the adaptive algorithm is in the learner — it’s just that the environment supports a more exploratory algorithm than does a textbook. (I’m sure you’ve read this 7,000 word monster post, but if you haven’t, here you go)

    In fact, that’s what I’m most obsessed about with choral explanations, is the way on stack exchange or quora a reader can use the multiple presentations to level-up an understanding bit by bit by finding a unique and personal path through the materials.

    It provides a middle way in between “You’ll be using the internet a lot, so just get used to bad signal/noise ratios” and “Here is the perfect pre-selected content for you.”

    Wednesday, July 20, 2016 at 3:26 pm | Permalink
  2. This is a crucial distinction, and for me plays into the conversation about the value of liberal arts degrees. If the process goals like learning how to think through ambiguous issues and solve them in a group setting are valuable skills, the meaty issues involved in the liberal arts (broadly conceived) seem particularly relevant. One might even say that messy or somewhat ambiguous learning is important learning, and that too smooth of a personalized pathway to content knowledge without struggle actually inhibits the development of certain traits like resilience and interdependence. Thanks for the post, George.

    Saturday, July 23, 2016 at 8:49 am | Permalink
  3. Rory McGreal wrote:

    Learning IS about content and skills. This content and those skills facilitate becoming a learner. How can you become a learner without learning something? The process goal of learning how to think is useless unless there are content and skills involved. I believe I know how to think (I have a PhD to prove it!{:-]), but this does not help me much to solve a problem in the viscosity of oils or any other subject that I have not learnt. I would argue that my learning how to think is the result of learning content and skills and that it is impossible to learn thinking skills without focusing on content and skill development. In fact, knowing how to think is only of importance when knowledge and skills are mastered. Knowing how to think does not help you learn Russian, but knowledge of Russian opens up new ways of thinking for you. As, knowledge of and skill in a birth language becomes the first step in a baby’s thinking process.
    All the best.
    Rory

    Sunday, July 24, 2016 at 3:42 pm | Permalink
  4. Roy Williams wrote:

    Process, process, process …
    You can learn to copy a process(aka ‘product’ or skill, and you might or might not learn, en passant, how to think about it at the same time.

    However …

    1. The transfer of skills from the one to the other cannot be assumed, and doesn’t always happen. As ‘educators’ we have to do more than teach skills (however intricate).

    2. The purpose of higher education (in particular) is to produce (that word, ‘product’ again!) people who can do both, and who can question and challenge existing skills and products, and build, extrapolate, and tease out, innovation and entrepreneurship (both financial and social) from a set of learnt and certified skills.

    3. In fairly open-ended research methods (biographical narratives) when students were asked ‘what they learnt that was important to them’ never mentioned skills or content – not that they had not learnt them, and not that they did not value them.

    3.1 Another however … what these students did explore, in detail and in depth, was how they coped (or didn’t cope) with learning how to “function with[in] ambiguity and complexity”, and how difficult it was to acquire and establish the skills of “metacognitive, goal setting, and self-regulation”. (https://learning-affordances.wikispaces.com/Project+Report)

    Learning the skills was obviously something they had to (and did) achieve – in various measures – but the “agency” skills: learning to become a self-initiating / self-regulating learner/practitioner – that was the important stuff they wanted to explore and document.

    Tuesday, July 26, 2016 at 4:34 am | Permalink
  5. In a self-study system, the most intelligent part of the system is the learner. No algorithm can compete with a human intelligence for adaptability and flexibility (at least so far).

    While I think a lot of what you referring to is higher level learning, I wrote a blog post a while back about this exact issue at the micro level.

    http://info.alleninteractions.com/overcoming-one-size-fits-all-learning

    Tuesday, July 26, 2016 at 1:32 pm | Permalink
  6. Antoinette wrote:

    Hi, I have to agree with Rory to a degree. In 1996 South Africa switched to an OBE (Outcomes Based Education) curriculum. The main difference with OBE was that we dumbed down the ‘knowledge’ content of the curriculum and added Skills and Attitudes/Attitude to the assessment criteria. The idea being that the student’s learning outcome had to prove that he/she had learnt something about the process of learning, they had learnt a new skill or attitude. 20 Years later we have ‘lost’ generation with no viable work skills, low knowledge content and an attitude of entitlement. Through learning content your mind is open to new ideas and experiences which in turn causes you to question and have further debate, both internal and otherwise.
    Students can quite easily become conceited, they have so little knowledge they are not even aware of what they don’t know.

    Thursday, July 28, 2016 at 5:14 am | Permalink
  7. Rory McGreal wrote:

    Roy, Julia, Antoinette
    Thanks for these stimulating propositions.

    1. “Do more than teach skills”. Yes, that is fine as long as we teach the skills and the knowledge to apply the skills. Perhaps students develop their higher order learning as a result of learning skills.
    2. There are many and variate “purposes” of HE. Challenging etc. is among them.
    3. This is simply not true. Many foreign students who learn using English claim that learning English has been more important than the subjects they learnt. So, sometimes students do mention skills and content. Moreover responses from students, although important are not the whole story. Objective evidence that what they say is true is also important. Their perceptions can be wrong.
    3.1 supports my argument that learning skills/knowledge is the way to develop the higher order skills. Yes, the “agency skills” are important, but they do not need to be specifically taught. These are self learnt by the student while acquiring the skills.

    Julie. “No algorithm can compete with human intelligence for adaptability and flexibility”. With all due respect, this statement is patently untrue. No human can fly a modern jet plane. It is flown by algorithms that are much more adaptable and flexible than any human. The pilots job is more one of monitoring or standing by in case of a malfunction. Moreover, most humans cannot even do a pilot’s job.

    Antoinette: I agree with you. What use are high level skills and positive attitudes if you don’t know anything and have no productive skills. And, if a student passes a traditional test on knowledge and skills, that proves more than anything that they learnt something about the process of learning. This emphasis on process rather than on skills and knowledge should be challenged. If your OBE curriculum does not emphasise skills and content knowleddge, then it is NOT OBE. Outcomes are specific skill sets based on knowledge learnt.

    All the best to all.
    Rory

    Thursday, July 28, 2016 at 8:36 am | Permalink
  8. Thanks for this interesting post and discussion offering a few angles to add comments, especially about skills and content knowledge, but it strikes me that what is also at issue is that there may be limits to the way one can learn in institutionalized settings.

    Rory’s language learning example reminds me of my own path from being told in high school that I would never learn how to speak a foreign language to speaking English and French to the extent that many people forget that these are not my mother tongue.

    How did that happen? The skills and content knowledge of my mother tongue and that I learned during language classes helped but they didn’t take me too far. What helped was the fact that once I landed in a foreign country there was no easy way back and that I was willing to fly by the seat of my pants.

    The adaptive nature of learning and its curve in a foreign environment when one is immersed in it 24/7 is astounding, not only for a child but also for an adult. The thing is that as adults we can rarely afford or have the courage to take this kind of plunge and see it through no matter the consequences.

    I agree that we need adaptive learners instead of adaptive learning, but a bit of skills and content knowledge helps to get off the ground. And I know this even when after 20 years I still have not applied the knowledge about oil rigs that I learned in my English text book but I’ve had many conversations about issues that I didn’t know anything about and learned in the process learning a language.

    I doubt that everyone can be or learn to be an adaptive learner without being willing to take risks and sacrifice what’s familiar and feels safe. Institutions can certainly offer a sense of safety for students to explore unfamiliar terrain beyond the acquisition of basic skills and content knowledge. So how can we find and encourage the students who are willing to fly by the seat of their pants?

    Thanks for indulging me and all the best,
    Matthias

    Friday, August 12, 2016 at 4:48 pm | Permalink