Learning styles
Claim assessed: Pupils learn better when teaching matches their learning style (visual, auditory, etc.).
Published · how we work
In short
The idea of learning styles says that each person has a preferred way of taking in information, for example visual, auditory or hands-on, and learns better when teaching matches that style. People do have preferences, but rigorous tests have not found that matching teaching to them helps them learn more. A 2024 meta-analysis found a small average advantage, but only a quarter of the results showed the pattern the hypothesis would require, and mostly in low-quality studies. The idea nonetheless remains very widespread among teachers.
What it says
The idea. There are over 70 systems for classifying styles, according to Newton (2015). The best known are VAK or VARK (visual, auditory, read/write, kinaesthetic) and Kolb’s model. They all start from the same promise: find out the pupil’s style, teach in that style, and they will learn better. In research, this promise is called the matching (or meshing) hypothesis. The idea doesn’t come from any one study but from educational practice and commercial questionnaires.
How it is tested. Pashler et al. (2008) set out what a proper test should look like, according to Aslaksen and Lorås (2018) and Rogowsky et al. (2020):
- each participant’s style is measured first;
- participants are randomly given teaching that matches or doesn’t match their style;
- everyone takes the same test;
- the result must be a “crossover”: visual learners learn better from visual teaching, and auditory learners from auditory teaching. It isn’t enough for one method to be better for everyone.
Pashler et al. found, as quoted by Rogowsky et al., that “very few studies” had used a design capable of testing the hypothesis. We haven’t read the article; we describe it from the sources above.
The rigorous tests. Aslaksen and Lorås (2018) went through over 1,200 articles. Only one study fully met the criteria: Rogowsky et al. (2015), with adults. In the studies that applied these criteria, the effects were very small and not significant. Rogowsky et al. (2020) repeated the design with about 120 fifth-graders: they found no link between the preferred style, visual or auditory, and how much the children understood from a text they read or listened to. Many children didn’t actually have a clear style: of 107, 47 had equal visual and auditory scores, and another 26 differed by only one or two points.
The meta-analyses.
- Clinton-Lisell and Litzinger (2024) combined 21 studies with 1,712 participants. On average, matched teaching gave slightly better results (g = 0.31), and the analysis showed no sign of missing unpublished studies. But only 11 of 42 results (26%) had the crossover pattern the hypothesis requires, and these came mostly from studies that did not meet quality standards. Results varied greatly from study to study. The authors’ conclusion: the benefits are “too small and too infrequent” to justify widespread adoption.
- Hattie and O’Leary (2025) reviewed 17 meta-analyses on learning styles. Those that actually tested matching gave an effect of d = 0.04, practically zero. The others found a link between styles and results (r = 0.24), but they were correlational and, the authors note, often confused styles with learning strategies.
Why it persists. Newton (2015) found that 89% of recent research papers on learning styles in the ERIC and PubMed databases endorsed them, explicitly or implicitly. According to the same article, 93% of the UK teachers surveyed by Dekker et al. (2012) believed that people learn better in their preferred style. A teacher looking for evidence therefore finds mostly favourable texts.
Example
A pupil fills in a questionnaire at the start of the year and learns they are “auditory”. From then on they avoid diagrams and charts and listen to recorded lessons, even in geometry. Rigorous tests say the label doesn’t help them learn more; a good diagram in geometry would probably help, whatever their style. The example is an illustration, not a study cited here.
How to apply it
The steps below are a practical way to apply it that we propose, based on the studies cited.
- Don’t spend time on style questionnaires. Their results don’t tell you how to teach or learn better.
- Choose the format by content, not by learner. A map is learned from an image, a foreign language also by listening. Clinton-Lisell and Litzinger note that teaching in several formats seems preferable to matching.
- Use the strategies that have evidence for everyone. Spaced repetition and testing your own memory help regardless of “style”.
- Treat a preference as a preference. Liking a format may matter for motivation, but it doesn’t mean you learn more from it.
Limits and nuances
- Preferences exist. The studies don’t deny that people prefer a format or that they differ in abilities. What didn’t hold is the next step: that matching teaching to the preference increases learning.
- Not a clean zero. The 2024 meta-analysis found a small positive average effect, with no sign of missing unpublished studies. The authors attribute it to weak studies and large variation, but the effect isn’t zero. And the rigorous tests are small: Rogowsky et al. (2020) could compare directly only about 40 children with a clear style. There is no large preregistered replication.
- Many systems, few tested. Most tests concern sensory styles (visual, auditory). For other classifications, such as Kolb’s, we have fewer data tested in the matching form.
- What we read. We read in full Aslaksen and Lorås (2018), Rogowsky et al. (2020), Newton (2015) and Clinton-Lisell and Litzinger (2024), from PubMed Central. Hattie and O’Leary (2025) is open access, but the publisher’s site blocked our download; we read the page text through an automated tool and use only the figures from the abstract. Pashler et al. (2008), Rogowsky et al. (2015) and Dekker et al. (2012) we describe from the sources we read.
Sources
- Harold Pashler, Mark McDaniel, Doug Rohrer, Robert Bjork (2008). Learning styles: Concepts and evidence
- Beth A. Rogowsky, Barbara M. Calhoun, Paula Tallal (2015). Matching learning style to instructional method: Effects on comprehension
- Philip M. Newton (2015). The learning styles myth is thriving in higher education
- Karoline Aslaksen, Håvard Lorås (2018). The modality-specific learning style hypothesis: A mini-review
- Beth A. Rogowsky, Barbara M. Calhoun, Paula Tallal (2020). Providing instruction based on students' learning style preferences does not improve learning
- Virginia Clinton-Lisell, Christine Litzinger (2024). Is it really a neuromyth? A meta-analysis of the learning styles matching hypothesis
- John Hattie, Tim O'Leary (2025). Learning styles, preferences, or strategies? An explanation for the resurgence of styles across many meta-analyses
See also: Interleaving, Spacing and retrieval practice