Few ideas in education have travelled as far on as little evidence. The claim that each person has a learning style, and learns better when taught in it, is repeated in teacher training, corporate onboarding and study guides worldwide — while the research that has tested it directly keeps failing to find the effect. Both halves of that sentence are worth understanding, because the underlying observation is not nonsense; it is the inference drawn from it that does not hold.
The Claim, Stated Precisely
A learning style is a preferred way of taking in information — visually, by listening, by reading, or by doing. The theory holds that these preferences are stable individual characteristics and that instruction matched to them produces better learning than instruction that ignores them.
Researchers call that second half the meshing hypothesis, and separating it from the first half is the whole game. Preferences existing is uncontroversial. Preferences determining what should be taught to whom is the testable, consequential, and largely unsupported claim.
Where the Idea Came From
The concept has many parents. Neil Fleming's VARK inventory, published in the early 1990s, is the version most people have met; Kolb's experiential learning model, Honey and Mumford's adaptation, and Dunn and Dunn's model all predate or parallel it. Coffield and colleagues counted 71 in 2004, which by itself says something about how loosely defined the construct is.
Notably, Fleming and Mills presented VARK as a catalyst for reflection — a prompt for teachers and students to think about how material is presented, not a diagnostic sorting people into fixed categories. The stronger version arrived later, in the hands of others.
Why It Spread So Fast
The idea is unusually appealing. It flatters everyone: no one is a bad learner, only mismatched. It hands teachers a concrete-sounding lever. It converts a difficult question — why is this hard for me — into a simple one with a comfortable answer. And it comes with a short quiz, which makes it easy to deliver.
It also arrived alongside a genuine and welcome push toward varied, multi-sensory teaching. That push was good practice for reasons that have nothing to do with individual styles, but the two got bundled, and the bundle proved durable.
What the Evidence Found
Testing the meshing hypothesis requires a specific design: identify learners' styles, randomly assign them to matched or mismatched instruction, and test everyone on the same material. If the theory is right, matched groups should outperform mismatched ones. Pashler and colleagues went looking for studies using that design in 2008 and found almost none — and the few that qualified produced results contradicting the hypothesis.
That is the crux, and do learning styles actually work goes through it in detail. The finding is not that studies came out mixed; it is that the claim had been adopted at enormous scale without the test that would have supported it ever really being run.
What Survives
Three things do survive, and they are worth keeping. Preferences are real and predict what people will choose and enjoy, which matters for motivation and persistence. Presenting material in more than one form helps most learners, because combining words with relevant images aids comprehension generally, not selectively. And the modality is usually set by the content — geography needs maps, music needs sound, surgery needs hands — regardless of who is learning it.
So a style result is best read as a description of what you find comfortable, not an instruction about what you should use. That is how the learning style test is worth taking, and how to study without learning styles covers the techniques that do have evidence behind them.