The idea that you learn best as a 'visual' or 'auditory' learner is one of education's most stubborn myths: decades of studies find that matching lessons to a supposed style makes little difference to how much we actually learn

Almost everyone can name their learning style. Ask a room of people whether they learn best by seeing, hearing, or doing, and most will answer without hesitation.

That confident answer has a problem. The idea that teaching people in their preferred mode helps them learn more has been tested, and it keeps failing.

This is not a claim that we are all the same, or that preferences do not exist. Plenty of us genuinely prefer a diagram to a lecture. The narrower idea that has collapsed: that catering to that preference changes how much you actually absorb.

What the theory actually claims

The popular version, best known through the VARK model, sorts learners into four types: visual, auditory, reading and writing, and kinesthetic. Its ancestors go back to the 1970s, with Dunn and Dunn’s Learning Style Inventory, and VARK itself emerged in the early 1990s; the framework has been part of much teacher training ever since. Its central claim has a name in the research: the “matching” hypothesis. It says you learn better when the way something is taught matches your preferred style.

That claim is testable, which is what makes it scientific rather than just intuitive. To test it properly, a study has to teach the same material to different people in different modes, then check whether visual learners really did better with visual material and auditory learners really did better with audio. Each group winning in its own preferred mode is the whole test.

What the research actually found

In 2008, four cognitive psychologists were reviewed the evidence. Harold Pashler, Mark McDaniel, Doug Rohrer and Robert Bjork went looking for experiments built to test the matching idea the right way. They found almost none, and the few solid studies that existed pointed the other way. Pashler and colleagues wrote that “there is no adequate evidence base to justify incorporating learning-styles assessments into general educational practice.”

That is narrower than saying learning styles do not exist, and the authors were careful about the difference. It would be wrong to assume every possible version had been tested and ruled out. What they had shown was that the evidence needed to build teaching around learning styles simply was not there.

Later work sharpened the picture. A 2025 review in Educational Psychology Review pulled together 17 earlier meta-analyses. Only four of them tested the matching hypothesis directly. Those four produced an average effect close enough to zero to be indistinguishable from it. The other 13 looked more promising, but the review argued they confuse learning styles with learning strategies, which is why the myth keeps looking alive.

Not every result is a flat zero. A 2024 meta-analysis by Clinton-Lisell and Litzinger, covering 21 studies and 1,712 people, found a small benefit from matching. But  the authors said the effects were too small and too rare to build teaching around. Cognitive psychologists Cedar Riener and Daniel Willingham put it more bluntly, arguing that “there is no credible evidence that learning styles exist.”

Why the myth refuses to die

The evidence has been public for years, and belief has barely moved. A systematic review by Newton and Salvi pooled 37 studies covering 15,405 educators across 18 countries. On average, 89.1 percent said they believed matching instruction to learning styles works. Among trainee teachers it was higher still, at 95.4 percent. The people newest to the profession believed it most.

Part of the answer may lie in how deeply the belief is held. Research by Shaylene Nancekivell and colleagues found that many people, especially those teaching younger children, treat a learning style as something a child is born with, wired into the brain from the start. A belief framed that way is hard to shift with a study. It feels less like a teaching technique and more like a fact about who a person is.

There is also a real cost to getting it wrong. Nancekivell says her “biggest concern is that time is being spent teaching young children maladaptive strategies for learning.” Time spent labelling a child a visual learner, and steering them away from reading aloud or working with their hands, is time not spent on approaches that would help.

What the evidence does support

None of this means teaching should be uniform, or that variety is pointless. The finding is more useful than that. The mode that works best usually depends on the material, not on the person receiving it. The shape of a coastline is probably a visual task; how a word is pronounced is likely a listening task; both are true for the whole class, not just the students who filled in a questionnaire and got sorted into one box.

The practical shift the research points to: ask what the content needs, then teach it in the mode that fits, and use several modes, because most subjects have several kinds of thing to learn. What matters is the task, not the label.