Last reviewed 19 January 2022
One of the predictions for 2022 was that personalised learning experiences would become more frequent. The key to providing personalised learning experiences lies in using adaptive learning and, in particular, adaptive learning technologies. In the first of two articles, Judith Christian-Carter takes a look at adaptive learning, focusing on what it is and when it can be used to most benefit.
At the heart of adaptive learning is the idea that learners should have as personalised a learning experience as possible. Research has shown that adaptive learning does lead to better retention, content gets covered faster and individuals feel more engaged with their learning. Adaptive learning also reduces learning and development (L&D) professionals’ workloads and, therefore, saves the organisation’s resources.
Although adaptive learning has been around for a long time, the reason for the recent interest is due to the advances in artificial intelligence (AI) and machine learning. As a result of these advances it has become possible to develop and implement adaptive learning at scale.
What is adaptive learning?
People have asked, “Is it a new kind of technology?”, “Is it an instructional strategy?” and even, “Is it a buzzword?”. The answer to all these questions, is “Yes but only up to a point!”. So, based on the premise that adaptive learning must provide a personalised learning experience, the Learning Guild defines it thus:
It “… refers to a system in which learning experiences are tailored to the individual. Usually, when you hear about adaptive learning, it’s in the context of a computer-based system that adjusts the difficulty or format of learning material based on the learner’s mastery. However, adaptive learning technologies can also be applied by instructors. In fact, the impetus for adaptive learning technologies is often attributed to Benjamin Bloom (yes, that Benjamin Bloom), who in 1984 suggested that implementing tutoring and mastery learning could improve learning outcomes by two standard deviations. His problem? These approaches weren’t scalable. Not every student in a class could have their own personal tutor. This is referred to as Bloom’s ‘Two-Sigma Problem’.” (Learning Guild, 2021).
The technologies available today address the Two-Sigma Problem by providing each individual learner with personalised content and feedback, in much the same way as a human tutor would. However, unlike human tutors, technology-based learning is scalable, so, theoretically, every individual could learn as effectively as possible. The recent innovations in machine learning have reduced considerably both the time and cost of developing adaptive learning systems, which means that today, adaptive learning technology is more likely to be a good investment for organisations to use in the workplace.
Current use of adaptive learning
Some recent research showed that 44.2% of respondents did not personalise any of their learning provision for employees. However, 17.1% of respondents did use a learning platform (LMS or LXP) that provided content recommendations to individual learners, based on ongoing assessments of their performance during a course. Using a learning platform in this way is a good example of adaptive learning.
Some 34.2% of respondents said that their trainers adapted instruction to the individual needs of learners in real-time during a course. Not only is this an example of good instructors, it is also a quasi-form of adaptive learning. Likewise, building for a variety of learners and assigning them to courses based on a pre-test or other information about their existing knowledge or background, something which 23.1% of respondents said they did, is also a quasi-form of adaptive learning.
When it comes to who in organisations is able to access adaptive learning, the results are quite encouraging, with one in four organisations offering adaptive learning to everyone in the organisation, regardless of job role, responsibilities and location. Another one in four of organisations were going down the pilot route, limiting adaptive learning to specific audiences and test users/groups, in order to see how well this form of learning fits into the overall organisational learning strategy. The results also showed though that 44.7% of respondents said that no one in their organisation was receiving any adaptive learning. However, compared with 16.6% of organisations currently using adaptive learning, by the end of 2022 this figure should have increased to just over 40%, as 24.6% of respondents said that they had a timeline for implementing adaptive learning technology in their organisations over the next 12 months.
How does adaptive learning work?
There is not just one approach to adaptive learning, as there are a variety of technologies and instructional strategies that can be used to personalise learning. Furthermore, there are a variety of contexts and situation in which adaptive approaches can be used.
The simplest type of adaptive learning is decision trees, which, as the name implies, follow an “if/then” branching structure. Depending on a learner’s response the path taken through the content is determined. Decision trees are very resource-intensive to develop.
Rule-based adaptive learning systems can vary in complexity. A simple system would ask the learner to take a pre-test and, depending on how well they did (eg the rule is 75% or above), they may not be provided with the relevant content.
The most complex and modern adaptive learning systems use machine learning integrated with AI. Using machine learning, rules can be applied to a body of content to enable the latter to be automatically classified in terms of difficulty. Similar approaches can also be used with providing feedback, competencies and other aspects of learning.
When is adaptive learning most beneficial?
In deciding whether adaptive learning will benefit an organisation or not there are two important questions to ask at the outset, which are:
Is the organisation prepared to collect, manage and safeguard learner data?
Is the organisation prepared to repurpose, design or use to its maximum advantage learning content?
Answering “no” to the first question rules out adaptive learning because all adaptive learning systems require information about learners in order to personalise their learning experiences. Likewise, answering “no” to the second question also rules out adaptive learning because adaptive learning systems need content to personalise. In addition, organisations that have a requirement for fixed training hours, or an infrastructure that will not support a technological investment of this size, or do not function in well-defined domains where there are no clear right and wrong answers to questions, will find that adaptive learning does not work well.
However, adaptive learning is helpful in organisations that:
need individuals to learn specific skills and knowledge (as opposed to learning team tasks)
work in domains where there are clear “right” and “wrong” answers
need to get learners through a lot of repetitive content quickly
would benefit from freeing-up L&D resources
are prepared to make a financial investment in improving efficiency
manage content with clear levels of difficulty, ie easy, medium, hard.
Without a doubt, adaptive learning is an emerging trend in workplace learning. From the ideal, but often impractical situation, of seeing the effectiveness of a human being providing one-to-one tutoring, through to the far more practical use of adaptive learning technologies, perhaps adaptive learning has finally come of age.
The second of two articles will look at how to implement modern adaptive learning, starting with a litmus test for organisational readiness.