When we published Urban Myths about Learning and Education over a decade ago, we devoted an entire chapter to the seductive appeal of learning styles. Back then, it already felt like a necessary myth-busting exercise. The idea that students learn better when you match instruction to their preferred learning style—visual, auditory, kinesthetic, and so on—was everywhere. And yet, the evidence was thin at best, and often outright absent.

Fast forward to 2025, and John Hattie and Timothy O’Leary have just published a thorough review in Educational Psychology Review that confirms what we argued back then: the matching hypothesis still doesn’t hold up. What is the effect of actually aligning teaching with students’ preferred styles? A tiny effect size of d = 0.04. Statistically negligible. Practically meaningless. And yet, learning styles are back in the spotlight—again.

Hattie and O’Leary do something clever. They distinguish between studies that test the matching hypothesis—does tailoring instruction to someone’s learning style help?—and those that merely look at correlations between self-reported styles and outcomes. And that distinction matters. The first group shows no real effect. The second group sometimes finds small relationships but says little about cause and effect. Worse: they often confuse learning styles with learning strategies or general learner characteristics. Precisely the kind of conflation we warned against in our book.

Still, belief in learning styles remains remarkably resilient. Surveys show that more than 90% of teachers believe that students learn better when taught according to their preferred style. Even among award-winning educators, the myth persists. The review lists reasons for this, from the appeal of individualisation to the desire to see each student as unique, from the seductive simplicity of the VAK framework to the commercial ecosystem of tests, training, and materials that keep it alive. Add brain-based jargon or a colourful infographic, and suddenly a discredited idea sounds like scientific insight.

The article is handy because it doesn’t stop pointing out the problem. It urges us to move the conversation forward: not to how students prefer to learn, but to what strategies actually help learning, depending on the task, the context, and the prior knowledge of the learner. Not fixed styles, but flexible approaches. Not labels, but tools.

That shift matters. Because a teacher who sees a student as a “visual learner” may overlook the real reason that student struggles—or may miss the chance to teach a more effective approach. Because telling students they have a fixed style doesn’t just fail to help them—it can limit their beliefs about their own potential. And because the best teachers know it’s not about matching instruction to preferences, but to content. You don’t teach grammar with a dance routine, and you don’t teach physical movement with a poem. The medium follows the message, not the learner’s declared style.

This new review is long, thorough, and at times damning. It reinforces what we wrote all those years ago and adds more evidence, nuance, and urgency. Learning styles aren’t just unsupported—they’re a distraction. They take time, energy, and professional development space away from things that actually work. And while the idea may feel intuitive or even caring, it’s not. It’s a well-meaning illusion.

We need to stop asking, “What kind of learner is this student?” and start asking, “What kind of learning does this student need right now?” Because learning isn’t about being visual or auditory. It’s about becoming better at learning itself.

Abstract of the double meta-analysis:

The persistence of learning styles as a concept in educational discourse and research is paradoxical, given the overwhelming evidence discrediting the matching hypothesis, the notion that aligning teaching methods with students’ preferred learning styles enhances achievement. This paper examines the resurgence of learning styles across meta-analyses and proposes an explanation for their enduring appeal. Drawing on 17 meta-analyses, we distinguish between studies testing the matching hypothesis (effect size d = .04) and correlational studies (average correlation r = .24), revealing that the latter often conflates learning styles with learning strategies. Much of the research is flawed, and the argument is that there needs to be a shift away from matching learning styles toward teaching students adaptable and effective learning strategies that align more closely with task complexity and learning goals.