Every learner is unique. With differing needs, they each approach learning in a unique way. In the field of adaptive learning, we try to understand these needs such that learning can approach the learner in a unique manner, tailored specifically for them and their needs.
There is a resource that can help adaptive learning designers that often goes underutilised:
The learners themselves.
By carefully preparing for a learning session, closely monitoring their thoughts during the session and evaluating their performance and progress after the session, a learner can further their own self-understanding and their understanding of the best strategies for learning. For the next learning session, they can implement this new understanding and, by doing this again and again, iteratively approach a way of learning tailored specifically for them and their needs (Metcalfe, 2009).
Learners often do not have complete control over their way of learning. Traditionally, teachers set certain work to be covered in a certain way, particularly with younger learners. By using an adaptive learning system, learners also give up some degree of self-regulation with the trust that – to an extent – the system can regulate their learning better than they can. Sometimes, particularly if the learner hasn’t chosen to use the adaptive system – as is often the case in schools and universities – this trust must be earned.
There are often misalignments between learners’ beliefs and the theories and evidence used to design curricula and to build adaptive systems. For example, learners may erroneously believe that cramming is better than spacing their learning sessions out. Spacing is best for long-term retention (Kornell, 2009). In these cases, the relationship of control over learning must be carefully navigated. As designers of adaptive learning systems, we can ensure that the system facilitates effective study behaviour, but we must also encourage learners to pursue this effective study behaviour. They are an essential component of the learning system. Going forward, this blog will focus on this optimal learning behaviour, and how to foster it. First, we will need to navigate the muddy waters of definitions. Self-regulation, metacognition and self-regulated learning are often used interchangeably, but they have distinct meanings:
Self-Regulation, broadly, is managing (regulating) our own behaviour and our reactions to others’ behaviour. When it comes to learning, this is all about control. Control of our study strategies, of our schedules and of a million and one other things that can affect learning: Getting enough sleep, managing our learning environments, anything.
Metacognition is all about our knowledge of our own cognitive processes and anything to do with them. Being aware and analysing our own thinking. Practically, this comes down to how we plan for an activity, how we monitor ourselves while performing it, and how we evaluate our performance and progress once it’s done.
So, self-regulation is monitoring and controlling our behaviour, and metacognition is monitoring and controlling our thoughts. Self-Regulated Learning is applying both of these things to learning.
Finally, an important concept to understand is the Zone of Proximal Development. This is the area between the things which can be done and those that cannot be done yet. This zone is where a learner can do something with help. For adaptive learning systems, this help often takes the form of some mechanism, like optimised spacing, to facilitate or aid learning. In this zone, learning has the biggest impact on what a learner can do. Things in this zone are close (proximal) to being mastered.
Each learner has a unique Zone of Proximal Development, but it can be very tricky to pin down exactly what lies within it. The borders of the zone are blurry, even to the learner. An adaptive algorithm like the one we use at MemoryLab aims to present questions of a “desirable difficulty” to the learner. These things are difficult, but not impossible – they are inside the zone. A key part of our algorithm is presenting the right question at the right time. Questions need to be repeated a few times until we can be sure that the learner can remember the answer for a good chunk of time. If you repeat a question too close together, the learner can easily remember the answer. If you wait too long, they will have forgotten. Based on how fast and accurately a learner answers questions, our algorithm can find that threshold of forgetting, per question, per learner. Right on the edge of that threshold is the perfect time to present them with the question again, when the answer is difficult to remember, but not too difficult. Here, it is at a desirable level of difficulty – it is in the zone. If a learner answers quickly and correctly on many repetitions of the question, we can be sure that they will remember the answer for a while, and we consider that question “mastered.” Once it has been mastered, it is no longer inside the Zone of Proximal Development. To maintain mastery over a question, a user must redo the exercise some time later, when the algorithm considers them likely to forget the answer. As their memory decays, we can use the same information to work out that it has re-entered the zone.
Similarly, by employing metacognitive strategies, a learner can try to identify what lies within their zone. It is important to consider that they may not always be correct – Bjork, et al., (2013), explore the issues with learners’ self-beliefs and point out that learners typically make some assumptions about learning that do not line up with metacognitive theory: That differences in performance are predominantly the result of innate abilities; that mistakes should be avoided; and that learning should be easy. In this blog we will address some of these assumptions, but it is good to know that they’re largely inaccurate. Performance is less dependent on innate ability than it is on effort and practice. Accordingly, mistakes and difficulties are important opportunities to put in effort, resulting in stronger performance.
By encouraging learners to use metacognitive strategies, and providing them with adaptive learning sessions, we can move towards richer understandings of mastered information, and more efficient learning sessions in general. If learners themselves choose what content to focus on, an adaptive learning system could point them in the right direction with feedback, perhaps some analytics of their previous performance.
The next blog in this two-part series looks at how we encourage students to use metacognitive strategies.
Bjork, R. A., Dunlosky, J., & Kornell, N. (2013). Self-Regulated Learning: Beliefs, Techniques, and Illusions. Annual Review of Psychology, 64(1), 417–444. https://doi.org/10.1146/annurev-psych-113011-143823
Kornell, N. (2009). Optimising learning using flashcards: Spacing is more effective than cramming. Applied Cognitive Psychology, 23(9), 1297–1317. https://doi.org/10.1002/acp.1537
Metcalfe, J. (2009). Metacognitive Judgments and Control of Study. Current Directions in Psychological Science, 18(3), 159–163. https://doi.org/10.1111/j.1467-8721.2009.01628.x