Should Course Evaluations Ask Whether Teaching Matched a Student's Learning Style?

In short: No. Asking students whether teaching matched their "learning style" builds a scientifically discredited construct into your quality data. Pashler, McDaniel, Rohrer, and Bjork (2008) reviewed the field and found virtually no rigorous evidence for the "meshing hypothesis" — the claim that matching instruction to a preferred style improves learning. An evaluation item premised on learning styles rewards instructors for a practice with no demonstrated benefit and mislabels a genuine problem (unclear teaching, poor task design) as a mismatch of style. Replace it with items about clarity, structure, and whether multiple representations were used.

What the research says

The modern case against learning styles rests on a landmark review commissioned by Psychological Science in the Public Interest. Harold Pashler, Mark McDaniel, Doug Rohrer, and Robert Bjork (2008) set out to find studies that met a specific, fair experimental standard: to justify tailoring instruction to a learning style, you need an experiment in which students are first classified by style (e.g., "visual" vs "auditory"), then randomly assigned to instructional methods, so that some learners are taught in their preferred style and some are not. The meshing hypothesis predicts a crossover interaction — visual learners should do best with visual instruction, auditory learners best with auditory instruction. The authors found that although the literature on learning styles is "enormous," the number of studies using this adequate design was vanishingly small, and those that did meet the standard generally produced evidence that contradicted the meshing hypothesis. Their conclusion was blunt: there is no adequate empirical justification for incorporating learning-styles assessments into general educational practice.

The finding has held up. Rohrer and Pashler (2012) reiterated in Medical Education that the absence of supporting evidence had not changed. Willingham, Hughes, and Dobolyi (2015), reviewing the scientific status of the theories in Teaching of Psychology, concluded that the theories do not have the empirical support their popularity implies. And Nancekivell, Shah, and Gelman (2020), publishing in the Journal of Educational Psychology, showed why the myth is so sticky: people hold essentialist beliefs about learning styles — treating "I'm a visual learner" as an innate, fixed, brain-based trait — which makes the belief resistant to correction. Crucially, surveys repeatedly find that a large majority of educators endorse learning styles, so the belief is not a fringe position; it is the default lay theory of how learning works.

What the evidence does not dispute is equally important. People genuinely have preferences for how they receive information, and some material is intrinsically better suited to one modality (you teach geography with maps, not spoken descriptions). What fails is the specific, testable claim that diagnosing a learner's style and matching instruction to it improves outcomes. Preference is real; the instructional payoff of matching is what the evidence rejects.

Why it matters for course evaluation in practice

Course-evaluation instruments are not neutral instruments; every item encodes a theory of good teaching. When a questionnaire asks students to agree or disagree with "The teaching methods suited my learning style," it does three damaging things at once.

First, it operationalises a myth as a quality criterion. An instructor who scores low is implicitly told to diagnose and cater to individual styles — advice that, if followed, wastes preparation time on an intervention with no demonstrated benefit and can crowd out practices that do work. Quality assurance should not create incentives to chase a debunked construct.

Second, it misattributes real problems. A student who found a course confusing has a legitimate grievance, but "it didn't match my learning style" is the wrong diagnosis. The underlying issue is usually unclear explanation, missing worked examples, poor sequencing, or a mismatch between assessment and teaching — all of which are fixable and all of which are better captured by low-inference teaching-behaviour items than by a style-mismatch item. Mislabelling the cause sends the improvement effort in the wrong direction.

Third, it degrades comparability. Because "learning style" is an ill-defined, essentialised belief, two students mean very different things when they invoke it. That is precisely the reference-bias problem: the item lacks a shared referent, so aggregating it produces a number that looks quantitative but is not comparable across students, cohorts, or courses.

The constructive move is to ask about the behaviour that the learning-styles myth is a folk-proxy for. The defensible version of "teach to different learners" is the well-supported principle of using multiple representations — presenting the same concept verbally, visually, and through worked examples — which benefits essentially all learners, not a subtype. So rather than "Did the teaching suit your learning style?", ask "Were ideas explained in more than one way (for example, in words and with diagrams or examples)?" and "When I was confused, I could find an alternative explanation." Those map onto what course evaluations should actually measure — clarity and structure that predict learning — instead of a preference-matching claim that does not.

Limitations and honest caveats

Intellectual honesty requires flagging what this argument does not establish. The Pashler review is a critique of an evidence base, and absence of evidence is not proof of impossibility; it remains logically possible that a well-designed future study could find a narrow meshing effect for some material and some learners. The stronger, defensible claim is the practical one: there is currently no warrant for building instruction — or evaluation — around learning styles.

Second, the critique targets the meshing hypothesis specifically, not every idea that travels under the "learning styles" banner. Adapting instruction to prior knowledge, interest, or self-regulation skill is well supported; those are not modality-based styles and should not be dismissed by association.

Third, the evidence on learners' preferences is not in dispute — students really do prefer certain formats, and preference can affect motivation and engagement even where it does not affect learning per modality-match. An evaluation programme may legitimately want to know about satisfaction and engagement; the caution is only against treating a preference-match as a learning-quality criterion. Finally, most of the classification-and-random-assignment studies were conducted in specific settings (often lab or lecture contexts); generalisation to every discipline and delivery mode should be made with appropriate humility. None of these caveats rescues the meshing hypothesis for evaluation use, but a critical reader deserves to see them stated.

How Koji incorporates this

Koji is designed to keep debunked constructs out of the evaluation instrument and to surface the real, fixable causes underneath a complaint.

• Construct-checked question design. Koji's structured question types (open_ended, scale, single_choice, multiple_choice, ranking, yes_no) are used to author items around evidence-supported constructs — clarity, structure, use of multiple representations, alignment of assessment with teaching — rather than around style-matching. The platform makes it easy to replace a legacy "suited my learning style" item with behaviourally specific scale items that a critical reader would accept as valid.
• Conversational probing that finds the real cause. Because Koji runs an AI-moderated conversational interview rather than a static form, when a student says a course "didn't suit how I learn," the follow-up probe asks what specifically was hard to follow — an unclear derivation, a missing example, pacing — converting an essentialised, uncomparable complaint into an actionable observation. This is designed to mitigate the reference-bias and misattribution problems the myth creates, not merely to record the surface phrase.
• Automatic thematic analysis with a clarity lens. Open-text responses are analysed thematically so that comments clustering around "confusing," "no examples," or "went too fast" are grouped and quantified — giving QA staff a clarity-and-design signal instead of a modality label.
• Bias-aware reporting. Koji frames results as evidence to interpret, flagging when an item rests on a contested construct, so that committees do not act on a style-mismatch number as though it were a validated learning measure. Koji's core research platform at koji.so applies the same AI-moderated interview engine to product and customer research, where the discipline of separating stated preference from actual behaviour is equally central.

As always, Koji is designed to mitigate — not eliminate — the pull of folk theories in student feedback; the instrument still depends on institutions choosing evidence-based items.

Related Resources

• Teacher Clarity Predicts Learning Better Than Charisma: What Course Evaluations Should Measure
• Stop Asking "Was the Lecturer Clear?": The Case for Low-Inference Teaching-Behaviour Items
• Why "I Learned a Lot" Can't Be Compared Across Courses: Reference Bias
• Students Rate the Classes Where They Learn the Most Lower: The Feeling-of-Learning Gap
• What Do Student Evaluations Actually Measure? Marsh, the SEEQ, and the Case for Multidimensional Feedback
• What Cognitive Load Theory Says Your Course Evaluation Should — and Shouldn't — Ask

References

• Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008). Learning Styles: Concepts and Evidence. Psychological Science in the Public Interest, 9(3), 105–119. https://doi.org/10.1111/j.1539-6053.2009.01038.x
• Rohrer, D., & Pashler, H. (2012). Learning styles: Where''s the evidence? Medical Education, 46(7), 634–635. https://doi.org/10.1111/j.1365-2923.2012.04273.x
• Willingham, D. T., Hughes, E. M., & Dobolyi, D. G. (2015). The Scientific Status of Learning Styles Theories. Teaching of Psychology, 42(3), 266–271. https://doi.org/10.1177/0098628315589505
• Nancekivell, S. E., Shah, P., & Gelman, S. A. (2020). Maybe they''re born with it, or maybe it''s experience: Toward a deeper understanding of the learning style myth. Journal of Educational Psychology, 112(2), 221–235. https://doi.org/10.1037/edu0000366