Universal Design for Learning comes up a lot in conversations about inclusive teaching. The basic idea is appealing enough. Instead of waiting for students to struggle and then adding support afterwards, we design learning so fewer barriers appear in the first place.
That’s a goal most teachers would agree with.
But the more I hear UDL discussed, the more I find myself wondering where it actually sits alongside what we already know about learning. Because many of the practices described under the UDL umbrella feel very familiar. And much of the research explaining why they work seems to come from somewhere else entirely.
I’m going to think out loud. What exactly is UDL adding?
What is Universal Design for Learning?
UDL is usually described as a framework for designing more inclusive learning environments. It was developed by CAST and takes its inspiration from universal design in architecture.
The idea there is pretty straightforward. If buildings are designed well from the start, they’re accessible to more people without endless adjustments later. Ramps. Wide doorways. Automatic doors. Features originally intended for accessibility often end up benefiting everyone.
UDL suggests curriculum can work in much the same way. Instead of waiting until students struggle and then adding accommodations afterwards, teachers should think about potential barriers during the design stage.
In truth, that probably doesn’t sound especially revolutionary to most teachers. When we plan lessons, we’re already trying to anticipate where students might get stuck. Where explanations might need to be clearer. Where a concept might need breaking down further. It’s also one reason many of us start new topics with some form of pre-assessment. If you know what students already understand, it’s much easier to see where the gaps might appear.
The framework is usually summarised through three principles:
- Multiple means of representation – how students access information
- Multiple means of action and expression – how they demonstrate understanding
- Multiple means of engagement – how they connect with and persist through learning
CAST grounds these three principles in neuroscience, aligning each one with a broad network in the brain: recognition networks for what we learn, strategic networks for how we act on it, and affective networks for why we engage with it in the first place.
That’s an interesting framing. We’ll come back to how well it holds up.
At the level of values, it’s difficult to disagree with any of this. Classrooms contain enormous variation in language, background knowledge, confidence, and learning history. Of course it makes sense to ask whether the design of a lesson might unintentionally block some students from accessing the intellectual challenge. We can all pretty much agree that this is a sensible starting point.
It’s also worth acknowledging where UDL came from. The framework was developed initially in response to students with disabilities, and that context matters. Its strongest case has always been for removing genuinely unnecessary barriers for students with specific access needs, whether physical, sensory, or linguistic. Keeping that origin in mind is useful, because it helps us think more clearly about which parts of UDL are doing real work and which parts have drifted rather far from home.
What parts of UDL are genuinely useful?
If you look at what teachers actually do under the banner of UDL, a lot of it’s perfectly sensible. In fact, much of it will feel very familiar:
- Clear explanations
- Well-designed materials
- Support that helps students get started
- Opportunities for students to show what they understand
None of that’s controversial. Most teachers would recognise these things as basic good teaching.
Multiple means of action and expression is the idea that students should have different ways to demonstrate what they know. Fair enough. Students might write explanations, present investigations, build models, run experiments, analyse case studies. If you work in something like the IB’s MYP, that probably doesn’t feel new. Performance-based tasks and varied assessments already exist.
Where things get messy is when this turns into unlimited choice.
Assessment still has to do a job. It needs to be valid. Comparable. And it needs to maintain cognitive demand. Some formats make it much easier for students to avoid the disciplinary thinking the task was designed to assess. Offering a student the choice between writing a lab report and making a poster is not always an act of inclusion. Sometimes it’s just a way of letting them sidestep the hard thinking.
Multiple means of engagement focuses on how students become motivated to participate and persist with challenging tasks. This can include:
- offering some choice
- connecting learning to meaningful contexts
- encouraging collaboration
- helping students regulate their own effort and progress
Teachers have always tried to design learning that captures interest and sustains attention, and in the MYP much of this is already built into the curriculum through inquiry, real-world contexts, and reflection on learning. But the research explaining why these things matter largely sits elsewhere. Self-determination theory (Deci and Ryan) explores how autonomy, competence, and relatedness influence motivation. Research on self-regulated learning (Zimmerman) examines how students monitor and manage their own effort. These are rich, well-developed bodies of work. UDL borrows from them without adding much to them.
Multiple means of representation means presenting ideas in different ways so learners can access the content more easily. Diagrams. Models. Highlighting key information. Breaking ideas into smaller steps. Again, perfectly sensible. But the explanation for why these things work doesn’t come from UDL.
Research on cognitive load (Sweller) explains why well-structured materials matter. Clear layouts, well-integrated diagrams, and manageable steps reduce unnecessary cognitive load so students can focus on the concept itself. But here’s a wrinkle worth pausing on: the same research also explains why multiple representations can backfire. If visuals and text are presenting the same information redundantly, the result is increased load, not decreased. The split-attention effect shows that poorly integrated materials can actually make learning harder. So “more ways of representing something” isn’t automatically better. It depends entirely on whether the different formats are genuinely complementary rather than just decorative.
Allan Paivio’s work on dual coding points in a similar direction. Combining words with meaningful visuals can strengthen understanding because the information is processed through two distinct but interconnected cognitive systems: verbal and nonverbal. The key distinction is that this is about the structure of the content, not the preference of the learner. It doesn’t matter whether a student considers themselves a “visual person.” What matters is whether the visual and verbal information are genuinely complementary, each carrying something the other can’t. One has a robust evidence base. The other is essentially folklore.
Rosenshine’s Principles of Instruction bring together three separate bodies of research: cognitive science, observations of master teachers, and studies of cognitive supports. They converge on a consistent set of practices:
- clear and well-sequenced explanations
- modelling expert thinking so students can see how problems are approached
- scaffolding through worked examples and guided prompts
- regular checking for understanding before students move to independent work
These practices have a coherent, well-tested explanation for why they work. They reduce unnecessary cognitive load and build schema gradually.
UDL describes similar practices. But describing something and explaining it are very different things.
Which raises the obvious question. If these ideas are already well established elsewhere, what exactly is UDL adding?
What does UDL actually add?
One possibility is that UDL mainly acts as an organising framework. It gathers together practices teachers already recognise and places them under a shared language of accessibility and learner variability.
That shift in language isn’t meaningless. Instead of asking why some students struggle, it nudges teachers to ask a different question: are there barriers in the way the learning has been designed? That reframing is useful. It moves the problem away from the learner and towards the design of the environment.
But describing practices isn’t the same thing as explaining why they work.
The strategies associated with UDL already have strong explanations in cognitive science and instructional research. That research looks at the mechanisms of learning:
- Why certain explanations work better than others
- Why some forms of practice strengthen understanding and others don’t
- Why reducing extraneous load frees up working memory for the thinking that actually matters
UDL doesn’t really operate at that level. It’s more a way of thinking about the design of learning environments than an explanation of how learning itself happens.
So if a framework draws on practices from cognitive science and instructional research, can we really apply them well without understanding the science behind them?
- How do we know which representation to choose?
- Which scaffold is actually helpful?
- Which explanation will make the idea clearer rather than more confusing?
Those decisions depend on understanding how learning works. Without that, strategies like scaffolding or worked examples can easily become surface features of lesson design rather than tools used deliberately to support thinking. We laminate the keyword wall. We add the diagram. We tick the box. And nothing much changes.
Seen in that light, UDL gathers together practices already supported by research on curriculum, cognitive science, and instruction, and frames them through a lens of accessibility and learner variability. That framing can be helpful. But the explanations for why those practices work sit elsewhere.
And if we already understand those explanations, the question still lingers.
What does UDL add?
**cough** Marketing, obviously.
What happens when the framework is implemented in schools?
Like many educational frameworks, UDL rarely stays exactly the same as it moves into everyday classroom practice.
Ideas usually start out as complex research frameworks. But by the time they travel through policy documents, professional development sessions, and guidance materials, they tend to get simplified. Complex ideas become bullet points. We’ve all played the game of telephone. We know what comes next.
Once that happens, people fill in the gaps with ideas they already recognise.
This is where things start to drift.
One place you can see this quite clearly is in the persistence of learning styles thinking. The idea that students learn best when teaching matches their preferred sensory modality feels intuitive, which is probably why it has proved so difficult to kill off. Harold Pashler and colleagues conducted a thorough review of the literature and found no credible support for the so-called meshing hypothesis:
- Students may have preferences about how they receive information
- But those preferences don’t predict how well they learn
- Preference doesn’t equal learning
UDL itself doesn’t promote learning styles. But when it’s enacted without a clear understanding of the learning science behind the practices it draws on, the language of multiple means of representation can easily slide into the familiar frame of visual, auditory, and kinaesthetic learners.
It’s worth being precise about the terminology here. The phrase “lethal mutation,” borrowed from educational research and popularised by Dylan Wiliam, describes something specific: a practice that is grounded in sound evidence but implemented so differently from the original that it loses all effectiveness. Strictly speaking, learning styles doesn’t qualify as a lethal mutation, because it was never well-evidenced to begin with. It’s more of a myth that found a comfortable home inside UDL’s language about variability. The word “mutation” implies something that was once healthy. Learning styles was always just wishful thinking dressed up in a questionnaire.
The result is the same either way. The language of learner variability remains, but the underlying science disappears.
Those choices about how to represent ideas should be driven by the structure of the knowledge being taught, not by assumptions about learner preferences. A diagram of the carbon cycle isn’t there because some students are “visual learners.” It’s there because the spatial relationships between processes are genuinely easier to see than to describe.
Does focusing on learner variability distract from curriculum design?
There’s another question sitting underneath all of this.
UDL begins with learner variability. Students differ in language, background knowledge, confidence, and experience. Learning environments should therefore be flexible enough to accommodate that variation.
That seems sensible. But curriculum research often starts somewhere else entirely. Instead of beginning with the learner, it begins with the knowledge. How should ideas be organised so that they actually become learnable?
Much of what looks like learner variability in classrooms is often better explained by differences in prior knowledge. Students aren’t necessarily learning in different ways. They simply know different things when they walk into the room.
That matters. My gosh, it REALLY matters.
Graham Nuthall’s meticulous classroom research, documented in The Hidden Lives of Learners, makes this vivid. Using recordings of student talk, pre-testing, post-testing, and follow-up interviews, his team mapped what individual students actually learned from lessons with remarkable precision. The key findings:
- Students need at least three substantive encounters with the complete set of information about a concept before it’s likely to stick. Not three mentions. Three genuine engagements with everything needed to understand it.
- Using observation methods no classroom teacher could replicate, Nuthall’s team could predict whether individual students had actually learned something with 80 to 85% accuracy. The implication isn’t “count to three”. It’s that learning is far less visible than we assume.
- It applied to everyone, not just struggling students.
The variation in outcomes was explained most strongly by differences in what students already knew coming in, and by whether they actually engaged with the complete information during the lesson. Students who already had relevant background knowledge could piece things together from fewer, less complete exposures. Those who lacked it needed more.
The implication is important. What looks like a “different type of learner” is very often just a student with a different starting point.
Those repeated encounters don’t have to look identical. An idea might first appear through explanation, then through a diagram, then through a worked example or application task. Seeing the same idea from different angles builds a more secure understanding. But this is driven by the structure of the knowledge and what it takes to understand it fully. It’s not the same logic as learning styles, which assumes that some students need a visual version and others need an auditory one. The variation comes from the demands of the content, not from assumed preferences.
From this perspective, one of the most powerful things teachers can do isn’t providing endless pathways through content. It’s designing a curriculum where knowledge is carefully sequenced and revisited over time. Checking for understanding becomes crucial, because it allows teachers to see what students have actually grasped and where misconceptions are forming. That information drives genuinely adaptive teaching.
When curriculum is coherent and teaching is responsive in this way, many of the access issues frameworks like UDL try to address begin to shrink on their own.
Inclusive learning doesn’t begin with multiplying pathways through content. It begins with carefully structured knowledge and teaching that responds to what students actually understand.
So how should we think about UDL?
UDL has done something useful. It reminds us to think about barriers created by the way learning is designed. Too often in education we assume the problem sits with the learner rather than with the curriculum or the explanation. Asking whether something in the design of learning is getting in the way is a helpful shift in perspective. And for students with genuine access needs, whether physical, sensory, or linguistic, the proactive thinking UDL encourages is genuinely valuable. That was always its strongest suit.
But when you start unpacking the practices associated with UDL, most of them aren’t new:
- Clear explanations
- Carefully designed materials
- Modelling
- Scaffolding
- Checking for understanding
All of those ideas are already well supported by research on teaching and learning, and the mechanisms explaining why they work sit outside the framework itself, in cognitive science, instructional research, and curriculum theory.
And the neuroscience framing that CAST uses to justify the framework?
A 2024 independent review examined the neuroscience sources CAST cites to justify its framework. It found:
- of the 1,442 unique sources CAST had cited, just 1% were neuroscience studies
- 75% of the UDL guidelines had no neuroscientific sources at all
- most of the cited research more broadly didn’t actually measure learning outcomes
- most didn’t involve offering learners a choice
The confident neuroscience language, it turns out, is doing considerably more work than the evidence behind it can support.
Which brings us back to the earlier question.
What exactly is UDL adding?
If teachers understand how learning works, they can design explanations, examples, and practice in ways that help students build knowledge over time. They can revisit ideas, check understanding, and adapt their teaching when misconceptions appear.
That’s where inclusive learning really starts.
Not with multiplying pathways through content, but with carefully structured knowledge and teaching that responds to what students actually understand.
In many ways, that might already be the closest thing we have to universal design.
Good curriculum.
Good teaching.
Done thoughtfully.
— and aren’t we all trying to do that already? —
Discover more from Edu'Flections
Subscribe to get the latest posts sent to your email.