Learning styles is one of the most persistent myths in education. Teachers worldwide believe that students learn best when taught in their preferred modality, visual, auditory, or kinesthetic. Yet fifteen years of rigorous research consistently shows the same finding: matching instruction to a student's supposed learning style makes no difference to their grades or learning outcomes. In fact, the idea that the brain has separate "visual," "auditory," and "kinesthetic" processing channels is a neuromyth with no basis in cognitive science. This article explores why learning styles persists despite the evidence against it, and what evidence-based alternatives actually work.
Many teachers assess their students' learning styles, often using simple questionnaires or informal observation. A student who prefers images might be labeled a "visual learner," a student who enjoys discussion an "auditory learner," and a student who learns through movement a "kinesthetic learner." Based on these labels, teachers then attempt to match instruction to preference: videos for visual learners, podcasts for auditory learners, hands-on activities for kinesthetic learners. The logic seems sound. If students learn better through their preferred modality, shouldn't we teach them that way?
The intuitive appeal is strong. But beneath the surface lies a critical assumption known as the "meshing hypothesis"—the idea that academic performance improves when teaching methods match a student's learning style. Despite its widespread adoption, this assumption has never been supported by rigorous experimental evidence (Pashler, McDaniel, Rohrer, & Bjork, 2008).
The most comprehensive investigation into learning styles was conducted by Pashler and colleagues in 2008. They systematically reviewed over 70 published studies investigating the meshing hypothesis. Their conclusion was stark: "Although the literature on learning styles is enormous, very few studies have even attempted to examine the validity of learning styles theories, and there is no empirical support for the utility of assessing students' learning styles so that they can be matched to instruction." This landmark paper, published in *Psychological Science in the Public Interest*, has since been cited over 5,000 times by credible researchers worldwide.
The scope of the problem became even clearer when Coffield and colleagues (2004) conducted a systematic review of learning style models themselves. They examined 71 different frameworks for categorizing learning styles, not just VAK, but also Honey and Mumford, Myers-Briggs, and many others. Their finding was sobering: of the 71 models examined, not a single one met basic scientific criteria for reliability and validity. Some models were internally contradictory, with students sometimes scoring as both visual and kinesthetic simultaneously, or changing their "style" between testing sessions.
More recent research has reinforced these conclusions. Willingham, Hughes, and Dobolyi (2015) classified learning styles as a "neuromyth"—a misconception about how the brain works that has spread into education despite lacking any neuroscientific basis. The brain does not have separate visual, auditory, or kinesthetic processing channels that can be trained independently. When individual differences in learning do emerge, they reflect differences in prior knowledge, engagement, working memory capacity, and motivation, not learning style preferences.
Perhaps most tellingly, Husmann and O'Loughlin (2019) conducted a direct test with undergraduate anatomy students. They identified students' VARK learning styles, then examined whether students who studied in ways matching their style performed better than those who ignored their style. The result: no difference. Students whose study methods aligned with their learning style showed no advantage on anatomy exams compared to students whose methods did not align. The perceived benefit of matching style to instruction appears to be illusory.
Despite overwhelming evidence to the contrary, the vast majority of UK teachers continue to believe in learning styles. Newton (2015) found that 93% of practising teachers endorse learning styles theories. The persistence of a myth despite 15 years of contradictory research requires explanation, and understanding that explanation can help shift practice.
The first reason is confirmation bias. Teachers observe genuine individual differences among their students, some students do seem to enjoy visual materials, others prefer listening, and some thrive with hands-on activities. When a student who prefers visuals performs well, teachers attribute the success to visual learning style. When a student who prefers audio struggles, teachers assume they weren't taught auditorily enough. Teachers do not notice the occasions when their predictions fail, or when a "visual learner" thrives with audio instruction, because these disconfirming cases don't stand out. The mind sees what it expects to see.
A second reason is commercial interest. Learning style assessment tools, curricula designed around learning styles, and professional development programmes teaching learning styles theory are profitable enterprises. Publishers and consultants have financial incentives to promote the idea that personalizing instruction by learning style is essential. This commercial machinery has been extraordinarily effective at spreading the myth. See also our guide on Sherborne developmental movement.
A third reason reflects a genuine and laudable educational aspiration: the desire to treat each student as an individual. Teachers want to personalise learning. When learning styles theory appears to offer a straightforward framework for doing so, it is naturally attractive. The theory says: "Respect how your students learn best." That message resonates with good teaching values, even though the mechanism it proposes doesn't work.
If learning styles is not the answer, what is? Dunlosky and colleagues (2013) analysed over 300 experiments investigating learning techniques across a range of subjects, ages, and populations. Their meta-analysis identified a small number of strategies that have "high utility"—they work reliably across all learner types, all subjects, and all ages. These evidence-backed techniques far outperform learning styles interventions. For related guidance, see our article on whole-school policy resources.
The first is distributed practice or spacing. Teaching a concept, then returning to it a week later, a month later, and several months later produces far better long-term retention than massed practice (teaching the same concept repeatedly in one session). This effect is robust and large, and it works equally well whether the instruction is visual, auditory, or kinesthetic. Spacing works for all learners regardless of learning style preference.
The second is retrieval practice. Low-stakes quizzing, asking students to retrieve information from memory without high stakes, produces superior learning outcomes compared to passive reading or listening. Again, this effect is large and works universally. It is not style-dependent. Students who are quizzed perform better than students who are not, regardless of their learning style.
The third is multimodal instruction—presenting information through multiple modalities simultaneously, not individually. Show learners diagrams while explaining verbally and demonstrating physically. Provide both the visual image and the spoken explanation. This kind of dual coding (presenting content in visual and verbal form) enhances learning for all students, not just those with matching preferences. Importantly, the benefit is not because each student is learning "their way"—it is because the brain encodes information more richly when it receives multiple input channels at once.
A fourth strategy is connecting new learning to students' prior knowledge. Differentiation that actually works is differentiation by prior knowledge and challenge level. When a student has weaker foundational understanding, provide more concrete examples and stronger scaffolding. When a student has solid foundations, present more abstract and complex material. This kind of differentiation addresses genuine learning needs and has strong evidence behind it (Sweller, 1988; Rosenshine, 2012).
The Education Endowment Foundation (2019) differentiates instruction not by learning style, but by: clarity of explanation, worked examples, guided practice, and independent practice. These elements apply to all learners, regardless of learning style. A student learning to divide fractions benefits from a clear explanation, a worked example, guided practice, and then independent problems, delivered through whatever modality makes sense for the content.
The good news is that the evidence does not suggest we should stop differentiating for individual learners. It only suggests we should stop differentiating by learning style. Instead, differentiate by: (1) challenge level, (2) prior knowledge, (3) engagement and motivation, and (4) metacognitive scaffolding (how to think about thinking).
Challenge level is perhaps the most important. A student who finds algebra straightforward needs harder, more abstract problems. A student struggling with algebra needs simpler, more concrete examples, stronger scaffolding, and more guided practice. Neither approach is about learning style, it is about matching cognitive load to the learner's developing knowledge.
Prior knowledge is equally crucial. A student who struggles with algebra may have weak foundational understanding of place value or equations. That gap is not a learning style problem; it is a prior knowledge gap that needs targeted support. Addressing it requires diagnosis and intervention at the level of specific concepts, not at the level of visual/auditory/kinesthetic preference.
Engagement and motivation shape learning powerfully. Some students need choice and autonomy; others need clear structure and guidance. Some students are extrinsically motivated (grades, praise); others are intrinsically motivated (interest, mastery). These individual differences are real and worth accounting for, but they are not the same as learning styles, and they should not be conflated with them.
When speaking with school leaders or parents who believe in learning styles, validate their underlying goal, personalising learning, while gently redirecting toward what evidence supports. You might say: "I want to personalise learning too. Rather than matching learning style, I'm focusing on matching challenge level to prior knowledge, and using spacing and retrieval practice for all students. Research shows this approach works better for all learners, not just some."
| If learning styles don't work, why do teachers still use them? | Learning styles persist due to confirmation bias (teachers see individual differences and assume they map to VAK), commercial interests (learning style assessments and curricula are profitable), and intuitive appeal (the theory is easy to understand). Teachers also have good intentions, they want to personalise learning, and learning styles theory appears to offer a simple framework for doing so. |
| What about Honey and Mumford, Myers-Briggs, or other learning style models? | All learning style frameworks suffer from the same foundational problems. Coffield et al. (2004) systematically reviewed 71 different models and found that none met basic criteria for scientific reliability and validity. The problem is not specific to VAK; it is endemic to learning styles as a concept. |
| Doesn't research show that SOME students prefer visual learning? | It is true that some students express a preference for visual materials, others for audio, and others for hands-on activity. But preference is not the same as learning style. A student may prefer visual materials yet still learn equally well, or better, through audio. Preference and optimal learning modality are not the same thing. |
| How do I differentiate for different learners if not by style? | Differentiate by prior knowledge, challenge level, engagement, and metacognitive support. Use spacing, retrieval practice, and multimodal instruction. Provide clear explanations, worked examples, and guided practice. Adjust the complexity and abstractness of tasks based on where each student is in their learning journey, not on learning style preference. |
| Will my visual learner do worse if I use audio instruction? | No. A student with a stated visual preference will learn equally well, or better, when audio instruction is paired with visual diagrams (multimodal instruction). Dual-coded learning, where content is presented both visually and aurally, enhances memory encoding for all students regardless of preference. |
Learning styles is one of the most persistent myths in education. Teachers worldwide believe that students learn best when taught in their preferred modality, visual, auditory, or kinesthetic. Yet fifteen years of rigorous research consistently shows the same finding: matching instruction to a student's supposed learning style makes no difference to their grades or learning outcomes. In fact, the idea that the brain has separate "visual," "auditory," and "kinesthetic" processing channels is a neuromyth with no basis in cognitive science. This article explores why learning styles persists despite the evidence against it, and what evidence-based alternatives actually work.
Many teachers assess their students' learning styles, often using simple questionnaires or informal observation. A student who prefers images might be labeled a "visual learner," a student who enjoys discussion an "auditory learner," and a student who learns through movement a "kinesthetic learner." Based on these labels, teachers then attempt to match instruction to preference: videos for visual learners, podcasts for auditory learners, hands-on activities for kinesthetic learners. The logic seems sound. If students learn better through their preferred modality, shouldn't we teach them that way?
The intuitive appeal is strong. But beneath the surface lies a critical assumption known as the "meshing hypothesis"—the idea that academic performance improves when teaching methods match a student's learning style. Despite its widespread adoption, this assumption has never been supported by rigorous experimental evidence (Pashler, McDaniel, Rohrer, & Bjork, 2008).
The most comprehensive investigation into learning styles was conducted by Pashler and colleagues in 2008. They systematically reviewed over 70 published studies investigating the meshing hypothesis. Their conclusion was stark: "Although the literature on learning styles is enormous, very few studies have even attempted to examine the validity of learning styles theories, and there is no empirical support for the utility of assessing students' learning styles so that they can be matched to instruction." This landmark paper, published in *Psychological Science in the Public Interest*, has since been cited over 5,000 times by credible researchers worldwide.
The scope of the problem became even clearer when Coffield and colleagues (2004) conducted a systematic review of learning style models themselves. They examined 71 different frameworks for categorizing learning styles, not just VAK, but also Honey and Mumford, Myers-Briggs, and many others. Their finding was sobering: of the 71 models examined, not a single one met basic scientific criteria for reliability and validity. Some models were internally contradictory, with students sometimes scoring as both visual and kinesthetic simultaneously, or changing their "style" between testing sessions.
More recent research has reinforced these conclusions. Willingham, Hughes, and Dobolyi (2015) classified learning styles as a "neuromyth"—a misconception about how the brain works that has spread into education despite lacking any neuroscientific basis. The brain does not have separate visual, auditory, or kinesthetic processing channels that can be trained independently. When individual differences in learning do emerge, they reflect differences in prior knowledge, engagement, working memory capacity, and motivation, not learning style preferences.
Perhaps most tellingly, Husmann and O'Loughlin (2019) conducted a direct test with undergraduate anatomy students. They identified students' VARK learning styles, then examined whether students who studied in ways matching their style performed better than those who ignored their style. The result: no difference. Students whose study methods aligned with their learning style showed no advantage on anatomy exams compared to students whose methods did not align. The perceived benefit of matching style to instruction appears to be illusory.
Despite overwhelming evidence to the contrary, the vast majority of UK teachers continue to believe in learning styles. Newton (2015) found that 93% of practising teachers endorse learning styles theories. The persistence of a myth despite 15 years of contradictory research requires explanation, and understanding that explanation can help shift practice.
The first reason is confirmation bias. Teachers observe genuine individual differences among their students, some students do seem to enjoy visual materials, others prefer listening, and some thrive with hands-on activities. When a student who prefers visuals performs well, teachers attribute the success to visual learning style. When a student who prefers audio struggles, teachers assume they weren't taught auditorily enough. Teachers do not notice the occasions when their predictions fail, or when a "visual learner" thrives with audio instruction, because these disconfirming cases don't stand out. The mind sees what it expects to see.
A second reason is commercial interest. Learning style assessment tools, curricula designed around learning styles, and professional development programmes teaching learning styles theory are profitable enterprises. Publishers and consultants have financial incentives to promote the idea that personalizing instruction by learning style is essential. This commercial machinery has been extraordinarily effective at spreading the myth. See also our guide on Sherborne developmental movement.
A third reason reflects a genuine and laudable educational aspiration: the desire to treat each student as an individual. Teachers want to personalise learning. When learning styles theory appears to offer a straightforward framework for doing so, it is naturally attractive. The theory says: "Respect how your students learn best." That message resonates with good teaching values, even though the mechanism it proposes doesn't work.
If learning styles is not the answer, what is? Dunlosky and colleagues (2013) analysed over 300 experiments investigating learning techniques across a range of subjects, ages, and populations. Their meta-analysis identified a small number of strategies that have "high utility"—they work reliably across all learner types, all subjects, and all ages. These evidence-backed techniques far outperform learning styles interventions. For related guidance, see our article on whole-school policy resources.
The first is distributed practice or spacing. Teaching a concept, then returning to it a week later, a month later, and several months later produces far better long-term retention than massed practice (teaching the same concept repeatedly in one session). This effect is robust and large, and it works equally well whether the instruction is visual, auditory, or kinesthetic. Spacing works for all learners regardless of learning style preference.
The second is retrieval practice. Low-stakes quizzing, asking students to retrieve information from memory without high stakes, produces superior learning outcomes compared to passive reading or listening. Again, this effect is large and works universally. It is not style-dependent. Students who are quizzed perform better than students who are not, regardless of their learning style.
The third is multimodal instruction—presenting information through multiple modalities simultaneously, not individually. Show learners diagrams while explaining verbally and demonstrating physically. Provide both the visual image and the spoken explanation. This kind of dual coding (presenting content in visual and verbal form) enhances learning for all students, not just those with matching preferences. Importantly, the benefit is not because each student is learning "their way"—it is because the brain encodes information more richly when it receives multiple input channels at once.
A fourth strategy is connecting new learning to students' prior knowledge. Differentiation that actually works is differentiation by prior knowledge and challenge level. When a student has weaker foundational understanding, provide more concrete examples and stronger scaffolding. When a student has solid foundations, present more abstract and complex material. This kind of differentiation addresses genuine learning needs and has strong evidence behind it (Sweller, 1988; Rosenshine, 2012).
The Education Endowment Foundation (2019) differentiates instruction not by learning style, but by: clarity of explanation, worked examples, guided practice, and independent practice. These elements apply to all learners, regardless of learning style. A student learning to divide fractions benefits from a clear explanation, a worked example, guided practice, and then independent problems, delivered through whatever modality makes sense for the content.
The good news is that the evidence does not suggest we should stop differentiating for individual learners. It only suggests we should stop differentiating by learning style. Instead, differentiate by: (1) challenge level, (2) prior knowledge, (3) engagement and motivation, and (4) metacognitive scaffolding (how to think about thinking).
Challenge level is perhaps the most important. A student who finds algebra straightforward needs harder, more abstract problems. A student struggling with algebra needs simpler, more concrete examples, stronger scaffolding, and more guided practice. Neither approach is about learning style, it is about matching cognitive load to the learner's developing knowledge.
Prior knowledge is equally crucial. A student who struggles with algebra may have weak foundational understanding of place value or equations. That gap is not a learning style problem; it is a prior knowledge gap that needs targeted support. Addressing it requires diagnosis and intervention at the level of specific concepts, not at the level of visual/auditory/kinesthetic preference.
Engagement and motivation shape learning powerfully. Some students need choice and autonomy; others need clear structure and guidance. Some students are extrinsically motivated (grades, praise); others are intrinsically motivated (interest, mastery). These individual differences are real and worth accounting for, but they are not the same as learning styles, and they should not be conflated with them.
When speaking with school leaders or parents who believe in learning styles, validate their underlying goal, personalising learning, while gently redirecting toward what evidence supports. You might say: "I want to personalise learning too. Rather than matching learning style, I'm focusing on matching challenge level to prior knowledge, and using spacing and retrieval practice for all students. Research shows this approach works better for all learners, not just some."
| If learning styles don't work, why do teachers still use them? | Learning styles persist due to confirmation bias (teachers see individual differences and assume they map to VAK), commercial interests (learning style assessments and curricula are profitable), and intuitive appeal (the theory is easy to understand). Teachers also have good intentions, they want to personalise learning, and learning styles theory appears to offer a simple framework for doing so. |
| What about Honey and Mumford, Myers-Briggs, or other learning style models? | All learning style frameworks suffer from the same foundational problems. Coffield et al. (2004) systematically reviewed 71 different models and found that none met basic criteria for scientific reliability and validity. The problem is not specific to VAK; it is endemic to learning styles as a concept. |
| Doesn't research show that SOME students prefer visual learning? | It is true that some students express a preference for visual materials, others for audio, and others for hands-on activity. But preference is not the same as learning style. A student may prefer visual materials yet still learn equally well, or better, through audio. Preference and optimal learning modality are not the same thing. |
| How do I differentiate for different learners if not by style? | Differentiate by prior knowledge, challenge level, engagement, and metacognitive support. Use spacing, retrieval practice, and multimodal instruction. Provide clear explanations, worked examples, and guided practice. Adjust the complexity and abstractness of tasks based on where each student is in their learning journey, not on learning style preference. |
| Will my visual learner do worse if I use audio instruction? | No. A student with a stated visual preference will learn equally well, or better, when audio instruction is paired with visual diagrams (multimodal instruction). Dual-coded learning, where content is presented both visually and aurally, enhances memory encoding for all students regardless of preference. |
