I'll start with a confession, because it's the honest way into this one. For years, I used the learning styles myth to help sell video.

It was the early 2010s and the idea was everywhere. Some people are visual learners, some are auditory, some are kinaesthetic, and you get the best results when you match the material to the person. It was a gift to someone in my position. If a good chunk of any workforce is "visual," then of course they need video, and here I am, a man who makes video. I never had to think too hard about whether it was true. It pointed at the thing I was already selling.

It is not true. Despite being promoted for decades, the central claim has never developed a reliable evidence base.

I want to be careful about what that does and does not mean, because the correction gets misheard constantly, and the misunderstanding is arguably doing as much harm as the myth.

What people believe

The claim, in its usual form, is that each of us has a dominant learning style, most often sorted into visual, auditory, reading/writing and kinaesthetic, the four that make up the VARK model. Identify your style, the theory goes, and you should seek out material that matches it. Teach a visual learner with diagrams, a kinaesthetic learner with hands-on tasks, and everyone does better.

It is an appealing idea. It flatters the learner, it gives trainers a tidy framework, and it comes with a questionnaire, which always makes something feel more scientific than it is. It has been taught on teacher training courses, printed on L&D induction slides, and quietly built into how a lot of organisations buy training for the best part of thirty years.

What the research found

Here is the problem. When researchers actually test the core prediction, the effect vanishes.

The prediction is specific and testable: people given material in their preferred style should outperform people given the same material in a different style. This is called the "meshing hypothesis," and it's the part that would need to be true for any of the practical advice to work.

In 2008, a group of respected psychologists reviewed the field for the journal Psychological Science in the Public Interest. They were looking for studies that tested the prediction properly, meaning they sorted learners by style, taught them in matched and mismatched ways, and compared the results. They found that studies meeting that standard were almost entirely absent, and that several of the suitably designed studies actively contradicted the theory. Their conclusion was blunt: there was no adequate evidence base to justify building teaching around learning styles [1].

Later work has been less kind still. A 2019 study followed anatomy students, gave them the VARK questionnaire, and then checked whether those who studied in ways matching their supposed style did any better. They did not. Most students weren't even studying in line with their own results, and those who did showed no performance advantage [2].

It's worth being fair to the other side of the ledger here, because the title of this piece promises what the evidence actually says, not just the convenient half of it. A 2024 meta-analysis did find a small average benefit from matched instruction. But the specific pattern needed to support the theory appeared in only around a quarter of the outcomes studied, and the authors concluded that the benefits were too small and inconsistent, and the underlying research too weak, to justify organisations adopting learning-style matching [4]. In other words, even the evidence that looks most favourable to learning styles doesn't hold up as a reason to build training around them.

And this isn't a case of a promising idea starving for funding. A large UK review identified 71 competing learning-styles models and examined 13 of the most influential in detail. It found serious problems across much of the field, including concerns about reliability, validity and how the models were being applied in practice, while noting how much money and training time was being spent on them anyway [3].

Decades of looking, seventy-one models, and still nothing solid to stand on. If the effect were real and as strong as the industry behaves as though it is, it would not be this hard to find.

The part everyone gets wrong

Now the important bit, and the reason I was careful earlier.

Learning styles being a myth does not mean you should pick one format and stick to it. It means very nearly the opposite.

When I put this to Professor Joe Devlin, a cognitive neuroscientist at UCL, for our recent whitepaper, he was quick to head off exactly this confusion:

"It's really tricky, because when you say learning styles are a myth, people think you mean that you shouldn't teach people in different formats. Quite the contrary. You absolutely need to use multiple formats."

His own father, an English teacher, was offended when Joe first told him the theory was bunk, until Joe pointed out that his dad had been doing the right thing his whole career. Of course you use pictures and text and discussion and activity. You just don't do it because Sally is a visual learner and Ben is a kinaesthetic one. You do it because thoughtfully combining complementary formats can help people understand, connect and remember the material.

The reason is mechanical, and once you see it the whole thing makes more sense than the myth ever did. Each relevant channel you use, seeing, hearing, reading, discussing or doing, can build another useful association around the same piece of knowledge. More ways into a memory means more ways back out to it when someone needs it later. The brain is multimodal by default. It is not sorting your team into four tribes, and it never was.

One caveat worth stating plainly, because it's the mirror image of the myth we're busting. More formats are not automatically better. Piling channels on top of each other without thought introduces redundancy and distraction, and the research on multimedia learning is clear that good design, including cutting information that isn't pulling its weight, matters as much as the number of channels [5]. The lesson is not "add more." It is "combine the right things well."

So the useful half of the idea survives. Use lots of formats. The damaging half, that you should match a format to a person, is the part with no evidence behind it.

Why it still matters commercially

You might reasonably ask who cares, if the practical advice ("use several formats") comes out roughly right anyway.

Here is who cares. The myth does its real damage in procurement, and it does it quietly.

When an organisation believes people have fixed learning styles, it starts treating formats as alternatives. Buy the video or the workbook or the workshop, because different people learn differently, so we'll pick the one our team responds to and save on the rest. I have sat in those conversations. I have, to my shame, nodded along in some of them.

They should not automatically be treated as alternatives. Used well, they become complementary layers, and the combination is often where deeper learning happens. Strip a course down to a single channel because you've decided your audience is "more visual" and you haven't tailored it to them. You may simply have removed several useful routes into the material and kept one.

That is the expensive version of this myth, and it's still being paid for across the industry today.

The short version

Learning styles, as fixed categories that tell us how material should be delivered to each person, are not supported by reliable evidence, and testing the idea properly has repeatedly come up empty. This doesn't mean people are all the same. They have preferences, strengths, prior knowledge, disabilities and needs that all matter enormously. What a VARK label can't do is tell you how a given person must be taught. But the instinct underneath the myth, that people need variety, is sound. People often learn better when complementary formats are combined thoughtfully, not because of who they are, but because of how memory works.

Stop labelling learners. Start layering formats. That's the whole lesson.