Adapting to Learner Attributes

PART 1 OF 3

In the previous newsletter, we looked at some different ways to sense the state of learner attributes, including just asking learners how they feel about such things as:

• The worries they might have about taking a course
• The personal appropriateness of the content
• How confident they are in their learning success
• How they feel about e-learning (or whatever approach is being used)
• Using a perpetual adjustable slider to reflect their current feelings

Now, the question becomes how to respond. If the course structure has no conditional branching capabilities, but is instead a linear presentation, well, adapting the learning experience for optimal outcomes isn’t going to be easy.

Artificial Intelligence in e-Learning

While interesting work in artificial intelligence promises adaptive responses on the fly, it appears those most truly in tune with the state of the art are dubious about many near-term expectations.

One of the top of the concerns is curbing large language models’ tendency toward hallucinations. It is apparently much more difficult than I would have expected and requires significant architectural modification still in research.

Additionally, Large Language Models come with a hidden cost. They consume an enormous amount of energy and water. Really! I know, but...

• A single ChatGPT query uses 25x more energy than a Google search.
• A back-and-forth ChatGPT conversation uses about a half-liter of water to cool the servers running all the software. 

Does it matter? What can we do in the interim as this gets sorted out? Can we prepare for efficient, affordable, and effective AI integration without having to wait or having to throw out what we may have done while waiting?

The answer is yes.

Microlearning

Once again, microlearning yields advantages. If you’re going to adapt, you need some means of branching (the smarter the branching algorithms, the better). And if you’re going to branch, you need things to branch to. That’s where modularity becomes essential.

If you’ve encountered any of my writings or presentations over recent years, the CCAF framework is already familiar to you. (And you can skip over this brief introduction to it, especially if it’s already a framework you’re using.) If not, please take a close look at this:

CCAF stands for the four critical components of interactive learning:

• Context
• Challenge
• Activity
• Feedback.

These four identified and well-defined components help course authors create effective microlearning experience modules in a straightforward and efficient manner.

The CCAF components are shown here as interlocking jigsaw puzzle pieces because they interconnect with each other to create interactive (thus the big “i") learning experiences. They build on each other to form microlearning modules.

Context sets forth an instance in which the skills learners will build are of particular value. As such, it helps learners see the personal value or relevance of the content. Your buyer is objecting to the recent price hike. You are a paramedic at a crash site. Your managers complain new hires take too long to get up to speed. The common failure to set context weakens learning activities, makes them seem academic, and sometimes causes learners to wonder what relevance or value this learning will have for them. Context is critical.

Challenge is a call to action. It generally appears in the form of a problem to be solved: The buyer is about to cancel a standing order. The patient is having trouble breathing. New employees are anxious about fitting in. The challenge level can be increased or decreased depending on what aids are made available, how much time is allowed, and/or how much detail is given versus must be ferreted out by the learner.

Activity is the learner’s response to the Context and Challenge. What would you do? How would you fix this problem? Instead of asking learners questions about what should be done, perhaps by selecting A) B) C) or D), we want to see learners taking steps to resolve the problem. Multiple steps are often required to solve real world problems. Providing the close guardrails of minimal listed choices and giving immediate feedback makes activity less authentic. It’s a particularly valuable experience when learners discover their errors and can back up to correct them without instructional intervention.

Feedback is the time and place in the CCAF model for instruction, instead of at the opening of a module or course when little is known about each learner. Feedback generally appears in one or a combination of three forms:

1. Nonjudgmental consequences of each action taken by the learner
2. Explanations and/or demonstrations of more effective actions
3. Scores as in games efficiently indicate successes and failures

Again, note that the traditional practice of explaining or demonstrating effective actions before confronting learners with a challenge does not appear in this model. Only when we have evidence of what would be appropriate for an individual learner do we delivering instruction.

Learner activity generates data about the learner’s current skill levels and misunderstandings, providing the basis for presenting or skipping instruction as well as the practice that may be required to reach confident proficiency.

Within Module Adaptation

In a subsequent newsletter we’ll talk about how CCAF modularity facilitates adaptation and expedites authoring. But let’s look now at how cycles within a CCAF module can easily adapt learning experiences to learners in empathetic and responsive ways.

Think of CCAF as both a structure of modular building blocks each of which provides a cyclical flow that is optimal for adapting to learners and also practicing to mastery. Some learners will work through multiple iterations of an adjusting module while others will move on to other modules after only a single cycle.

In making mistakes or performing flawlessly, novice learners reveal for themselves (and us) what instruction they need or don’t need. Through feedback, we can decide what to do based on both performance and how the learner is feeling.

Here’s a sample logic table that determines what comes next by considering whether the learner was successful in meeting the challenge, expresses confidence in the ability to perform successfully, and asked for help along the way. in order what to do next. Of course, additional data can be factored in.

When recycling, the Context and Challenge can remain the same or, as learner proficiency is measured, either or both of these components can be modified. For example:

1. In the initial cycle, various helps, tips, and resources can be available to all learners as we're just getting to know the learner and don't want the initial challenge to be too intimidating.
2. In subsequent cycles, these aids can be enhanced or removed, or learners can just be discouraged from using them unless they feel they are absolutely necessary.

Generally, when it’s time to change the Context, it’s time to move on to another CCAF microlearning module.

CCAF Is Robust

Because of its inherent ability to adapt to the range of learners’ needs and because instruction is offered in response to identified needs rather than thrust insensitively and in identical form on all learners, the model is widely applicable. And as such, CCAF can expedite the design of highly instructional programs while providing engaging and maximally effective learning experiences.