What the Evidence Says About Learning Styles

Despite massive global popularity, decades of rigorous cognitive science show that teaching students in their preferred "learning style" - such as visual, auditory, or kinesthetic - does not improve academic outcomes. While individuals absolutely have personal preferences for how they encounter information, tailoring instruction to match these preferences yields an effect size near zero. Educators and students are much better served by abandoning these restrictive labels and adopting evidence-based practices, such as active recall and multimodal instruction, which benefit all learners regardless of their supposed style.

The Enduring Appeal of a Neuromyth

If you have ever taken a quiz that told you that you are a "visual learner," or heard a teacher say they are adding a hands-on activity specifically for the "kinesthetic learners" in the room, you have encountered the learning styles theory. The most famous framework, the VARK model (Visual, Auditory, Reading/writing, Kinesthetic), suggests that people possess fixed, dominant ways of learning and that educators should match their instruction to these categories ¹¹².

The concept feels highly intuitive. We easily observe that some students gravitate toward diagrams while others prefer podcasts or building physical models. From these observable preferences, it feels like a logical leap to assume that teaching them in these preferred modalities will unlock their academic potential.

As a result, the idea has become deeply entrenched in global education. A 2020 systematic review involving 15,405 educators across 18 countries found that 89% of teachers believed that matching instruction to learning styles improves student outcomes ³⁴. This belief spans the globe: 93% of educators in the United Kingdom, 96% in the Netherlands and Greece, and 97% in Turkey and China endorse the concept ⁴⁶. Furthermore, social media platforms continuously propagate the idea, with educational organizations actively promoting professional development events based on this unsubstantiated theory ⁷.

Why does the myth refuse to die? Researchers suggest it offers an "appealing simplicity" ⁸. It caters to a deep-seated desire among parents and teachers to treat every child as a unique individual and to democratize the classroom ⁹¹⁰. Furthermore, it provides a convenient, guilt-free explanation for academic struggles: if a student is failing, it is less painful to blame a "mismatched learning style" than to address systemic gaps in foundational knowledge, insufficient study habits, or ineffective teaching ¹⁰. Bolstered by a commercial ecosystem of paid inventories, workshops, and textbooks, the learning styles concept has evolved from a harmless assumption into what cognitive scientists classify as a "neuromyth" - a misconception generated by a misunderstanding of scientifically established facts ⁶⁵.

Tracing the Origins of the Myth

To understand how the myth gained such traction, it is helpful to look at its origins. The concept of identifying individual learner differences is not new; even Aristotle recognized that children possess specific talents and skills ⁶. However, the modern iteration of learning styles began to take shape in the late 20th century, emerging alongside early advancements in cognitive psychology.

In the 1970s and 1980s, theorist David Kolb introduced his Experiential Learning Theory, which posited that learning is a cycle of concrete experience, reflective observation, abstract conceptualization, and active experimentation ²³. Kolb categorized learners into four types based on how they navigated this cycle: convergers, divergers, assimilators, and accommodators ³. Around the same time, researchers like Rita and Kenneth Dunn popularized their own learning style models, heavily advocating that teachers should arrange their instruction methods according to these styles ⁷¹⁴.

However, it was Neil Fleming's introduction of the VARK model in 1992 that truly captured the public's imagination. Due to its simplicity, VARK quickly became the dominant framework in classrooms worldwide ²⁶. Students were given self-report questionnaires asking what kind of learning experiences they enjoyed most (e.g., watching a video versus listening to a lecture), and were subsequently diagnosed with a definitive "style" ⁷⁸.

What began as a catalyst for reflection - a way to encourage students to think about how they study - was swiftly misinterpreted as a rigid neurological reality ⁹. The jump from "students have preferences" to "students learn better when taught exclusively in their preference" was made without empirical backing, setting the stage for decades of misguided educational policy.

Decoding the "Meshing Hypothesis"

To understand why cognitive scientists reject learning styles, we have to look at what researchers actually test. The core of the learning styles theory rests on what is known as the meshing hypothesis (or matching hypothesis). This hypothesis posits that learning is optimized when the instructional method meshes with a student's self-reported learning style ⁷¹⁴¹⁰.

In a seminal 2008 review published in Psychological Science in the Public Interest, psychologist Harold Pashler and colleagues established the strict criteria required to prove the meshing hypothesis. For the theory to be scientifically valid, a study must demonstrate a specific statistical phenomenon known as a "crossover interaction" ⁸¹¹¹².

Proving a crossover interaction requires a highly specific experimental design: 1. A group of learners must be assessed and divided by their preferred style (e.g., visual vs. auditory). 2. Learners within each group must be randomly assigned to different instructional methods (e.g., half the visual learners get visual instruction, half get auditory instruction; half the auditory learners get visual, half get auditory). 3. All learners must take the exact same objective achievement test. 4. The results must show that visual learners performed significantly better with visual instruction than with auditory instruction, AND auditory learners performed significantly better with auditory instruction than with visual instruction ¹⁴¹²²⁰.

If visual instruction simply makes everyone score higher (perhaps because a diagram clarifies a complex geometric concept better than a verbal description), that is not evidence of learning styles. That is simply evidence that a visual aid was the objectively superior pedagogical tool for that specific subject matter ¹⁴¹². Only a full crossover interaction validates the meshing hypothesis.

What the Latest Meta-Analyses Reveal

When researchers rigorously test the meshing hypothesis using the crossover interaction standard, the theory systematically collapses.

In a landmark 2025 review published in the Educational Psychology Review, researchers John Hattie and Timothy O'Leary analyzed 17 meta-analyses involving over 100,000 students. Their conclusion was definitive: the effect size of matching teaching to students' preferred learning styles was d = 0.04, which is statistically negligible ⁸⁹¹³.

To put this in perspective, an effect size of 0.40 is generally considered the "hinge point" or average impact for standard educational interventions ²⁰. Strategies supported by cognitive science yield massive returns, while the learning styles matching hypothesis offers virtually nothing. Data indicates a stark contrast in efficacy: while matching learning styles yields an insignificant d = 0.04, providing students with actionable feedback yields an impressive d = 0.73, and direct, explicit instruction yields d = 0.59 ⁸⁹¹⁴.

A separate 2024 meta-analysis by Clinton-Lisell and Litzinger yielded similarly dismal conclusions for the matching hypothesis. Out of 42 learning outcome measures analyzed across 21 eligible studies, only 11 (26%) showed the crossover interaction required to support the theory ¹¹¹⁵. Furthermore, the researchers noted that the few studies that did show a crossover interaction generally failed to meet the rigorous quality standards of the What Works Clearinghouse (WWC), frequently suffering from a lack of reliability statistics and poor methodological design ¹¹¹⁵. Hattie and O'Leary echoed this sentiment, noting that learning styles research stands uniquely as the most methodologically flawed area in educational psychology, plagued by calculation errors, low validity, and commercial bias ⁸.

The Conflation of "Preference" and "Strategy"

If the evidence is so poor, why do some correlational studies claim to find a link between learning styles and student success? Hattie and O'Leary (2025) point out a critical flaw in the broader literature: the conflation of learning styles with learning strategies ⁸⁹.

When researchers look at correlations (referred to as "r-studies" rather than experimental "d-studies"), they sometimes find moderate relationships (average r = 0.24) between a student's self-reported style and their academic achievement ²⁹²⁰. For example, a review of Turkish experimental theses by Kanadli (2016) found an overall correlation of r = 0.46 between learning style scores and academic achievement ²²⁰.

However, these correlational studies do not test the matching hypothesis. Instead, they often accidentally measure a student's use of highly effective study strategies ²⁸. For instance, a student who identifies as a "reading/writing" learner likely scores high on inventory tests because they regularly take detailed notes, summarize texts, and rewrite concepts. These are evidence-based, deep-processing learning strategies that would help any student succeed. The correlational studies mistake the effectiveness of the strategy (note-taking) for an inherent neurological trait (a reading/writing "style") ².

Subjective Feelings vs. Objective Reality

Another reason the myth persists is that studying in a preferred modality simply feels better to the learner. A study by Knoll et al. (2017) demonstrated that a student's learning style is strongly associated with their "Judgments of Learning" (JOLs) - a subjective measure of how well they think they have learned something ¹⁶.

In the study, participants with strong "visualizer" preferences reported higher JOLs when studying pictures, while "verbalizers" reported higher JOLs when studying words. However, when the researchers actually tested the participants' recall accuracy, there was absolutely no association between their preferred style and their objective performance ⁸¹⁶. People consistently feel as though they are learning better in their preferred style, but this subjective comfort does not translate into cognitive acquisition ⁸.

Summary of Learning Concepts

To clarify the debate, it is essential to distinguish between preferences, strategies, and the meshing hypothesis.

The Dangers of Pigeonholing: Why "Matching" Harms

Believing in learning styles is not just a harmless quirk; it actively damages the educational ecosystem. Cognitive scientist Stephen Chew offers a helpful analogy for parents and educators: children will naturally prefer candy and soft drinks over milk and fruit, but a responsible parent does not feed them a purely sugar-based diet just because it is preferred ¹⁷. Giving students only what they prefer robs them of the opportunity to develop cognitive adaptability and resilience ¹⁷.

The failure of the meshing hypothesis makes sense when viewed through the lens of modern neuroscience. The implicit assumption of the VARK model is that information delivered via one sensory modality (e.g., visual) is processed independently from other modalities in the brain ¹⁰. However, brain imaging shows massive cross-modal processing and interconnectivity. Input modalities in the brain are always interlinked; it is impossible for a learner to rely strictly on a single sensory pathway to build complex conceptual understanding ⁷¹⁰.

Furthermore, memory is rarely tied to the sensory format in which it was learned. When you learn that Paris is the capital of France, your brain stores the semantic meaning of that fact, independent of whether you heard a teacher say it, read it in a book, or pointed to it on a map .

Bias and Essentialism in the Classroom

Worse, labeling children creates fixed mindsets and unconscious biases. A 2023 sociological study by Sun et al. revealed that parents and teachers who endorsed the learning styles myth consistently judged children labeled as "visual learners" to be smarter and more capable in core academic subjects (math, language, and social sciences) than children labeled as "hands-on" or kinesthetic learners ⁸¹⁸. Conversely, they predicted that kinesthetic learners would only excel in non-core subjects like gym, music, or art ¹⁸.

These value-laden categories shape expectations, potentially limiting the academic opportunities offered to students deemed "tactile" ⁸. This stems from an "essentialist" view of learning styles - the false belief that a student is biologically hardwired to learn in only one way, severely restricting their perceived potential ¹⁵²⁸.

The Opportunity Cost for Educators

The persistence of the learning styles myth also exacts a massive toll on teacher workload. Educators spend countless hours administering unvalidated learning style inventories and subsequently attempting to design four different versions of a single lesson (one visual, one auditory, one reading, one kinesthetic) to accommodate their diverse classrooms ²⁹.

This is a tremendous opportunity cost. Time spent trying to perfectly "mesh" instruction to 30 different students' preferences is time diverted away from evidence-based practices that actually work - such as providing high-quality, actionable feedback, designing formative assessments, or scaffolding difficult material ²⁹¹⁹³¹.

Surprisingly, there is even evidence that matching instruction to a preferred style can penalize learning. A 2026 study examining digital biology lessons found that for complex subjects, students sometimes performed better when the instructional format intentionally mismatched their preferred learning style ²⁸. This suggests that a mismatch can act as a "desirable difficulty," forcing the brain out of cognitive autopilot and requiring deeper processing to understand the material ²⁸³².

Cultural Nuances: Preferences vs. Stereotypes

The debate over learning styles often intersects with cultural dynamics, particularly in international education. Global studies show that cultural factors - such as communication norms, student-teacher power dynamics, and societal values - significantly impact how students behave in a classroom ³³²⁰. However, it is vital to separate cultural classroom behaviors from inherent cognitive learning styles.

For example, students from Asian nations with a Confucian Heritage Culture (CHC) - such as China, Vietnam, Korea, and Japan - are frequently stereotyped in Western literature as "passive," rote-memorization learners who rely heavily on reading and writing styles ³⁵. They are often characterized as unwilling to ask questions, challenge authority, or engage in active dialogue ³⁵³⁶.

However, sociological research reveals this is a misinterpretation of cultural values, not a neurological learning style. The perceived "passiveness" is often a reflection of deep-seated cultural respect for teachers, a collectivist focus on group harmony over individual disruption, and specific assessment structures, rather than an inability to learn dynamically ³⁵³⁶. When pedagogical structures change - such as in well-designed collaborative projects - students from collectivist cultures often excel in peer-learning and active group problem-solving ³⁶.

Similarly, research into differentiated instruction in Indonesia demonstrated that when teachers moved away from rigid monolingual lecturing and incorporated varied, multimodal activities (visual, auditory, and kinesthetic elements combined), students' scientific communication skills significantly improved ⁹. The success of these interventions is frequently misattributed to "accommodating learning styles," when in reality, the success is due to transitioning from passive lectures to highly engaging, active learning strategies ⁹. Distinct learning style patterns do not neatly map onto specific cultural, national, or racial groups, and educators must avoid using the learning styles myth to stereotype diverse populations ²⁰²¹.

Multimodal Learning: A Superior Paradigm

If learning styles do not work, how should educators address the undeniable diversity in their classrooms? The answer lies in multimodal learning.

While learning styles theory categorizes the student, multimodal learning diversifies the instruction. Research clearly indicates that learners retain more information when words and pictures are presented simultaneously, rather than words alone ⁶. This aligns with "Dual Coding" theory, which posits that your brain stores images and words through separate channels. Pairing a verbal explanation with a visual representation (like sketching a flowchart while explaining a sequence of events) gives the brain two independent, interconnected paths back to the same memory ³²³⁸.

To understand the difference for a general audience, consider the analogy of cooking a pasta bolognese ⁶. If the ingredients symbolize the different ways we interact with content (visuals, discussions, physical models, texts), a learning styles approach would force you to serve a bowl of plain crushed tomatoes to one guest, a plate of dry pasta to another, and raw ground beef to a third, simply based on their "favorite ingredient." A multimodal approach combines all the ingredients into a rich, complex sauce, recognizing that the integration of multiple senses creates a vastly superior experience for everyone at the table ⁶²².

Instructional modalities should be chosen based on the nature of the content, not the preference of the learner. You cannot effectively teach the sound a cow makes without an audio clip, just as you cannot effectively teach someone the physical mechanics of swimming purely through a textbook, regardless of whether a student self-identifies as a "reading/writing" learner ²³.

Evidence-Based Strategies That Actually Work

By abandoning the learning styles myth, educators and students can redirect their time and energy toward evidence-based teaching practices (EBTs) that have proven, massive effects on learning and retention ²⁹²⁴. These methods capitalize on how the human brain actually encodes and retrieves information.

Active Recall (The Testing Effect)

Most students rely on passive study methods like re-reading textbooks, highlighting text, or copying notes ³². While these activities feel productive, they rely on simple recognition rather than recall, making them highly ineffective for long-term memory ²⁵.

Active recall involves closing the book and forcing your brain to retrieve the information from memory. Whether through flashcards, practice problems, or blank-page recall, the mental effort required to search your memory physically strengthens the neural pathways associated with that information ³²²⁵. The harder the brain works to retrieve the fact, the stronger the memory becomes.

Spaced Repetition

Cramming for an exam might work for the next morning, but the information vanishes rapidly. Spaced repetition combats the brain's natural "forgetting curve" by reviewing material at gradually increasing intervals (e.g., one day, three days, one week, one month) ³²²⁵²⁶. By strategically spacing out review sessions, you force your memory to work harder just as it is beginning to forget the concept, making the memory highly durable ²⁵²⁶.

Interleaving

Rather than studying one topic in a massive block for hours (massed practice), interleaving involves mixing different but related topics within a single study session ¹⁴³². For example, instead of practicing 20 identical division problems, a student might mix division, multiplication, and subtraction problems. This forces the brain to continuously identify which strategy to apply, improving high-level problem-solving abilities and preventing cognitive autopilot ³²³⁸.

Relational Reasoning and Analogies

Instead of trying to match a student's "style," educators should help students build conceptual frameworks using analogies. Analogical reasoning allows learners to map novel, complex concepts onto well-understood systems ⁴⁴. For example, teaching students that "epigenetics is like a Lego set" or that nuclear fission is like "striking pool balls" provides a cognitive framework to understand structural relationships ⁴⁴²⁷. This builds higher-order thinking skills that transcend sensory modalities.

Universal Design for Learning (UDL)

Rather than classifying students into fixed categories, Universal Design for Learning (UDL) advocates for designing flexible instructional environments from the outset. UDL provides students with multiple means of engagement, representation, and action ³¹²⁸. It gives learners autonomy to choose how they demonstrate their knowledge - whether through an essay, a presentation, or a project - without assuming their choice is dictated by an immutable biological learning style ²⁸.

Bottom line

The idea that we learn best when instruction is strictly tailored to a specific "learning style" is a well-meaning illusion. Decades of cognitive science confirm that while we all have personal preferences for how we consume information, matching instruction to those preferences yields an effect size near zero and can inadvertently create harmful, essentialist biases about a student's intelligence. Instead of asking what "type" of learner someone is, educators and students should focus on what the task demands - utilizing multimodal instruction, active recall, dual coding, and spaced repetition to build lasting, adaptable knowledge.