Skip to content

Cognitive Styles

Principle · All principles
Evidence · 24 claims (2 for, 17 mixed, 5 against) · 17 studies (8 review, 5 causal, 3 theoretical, 1 associational), q1–q3 · 4 of 17 report an effect size · 22 claims rest on one study

Conditional relationship

This page owns the question of what to do with a learner's characteristic way of approaching a task: seeing a whole before its parts or analysing parts separately (field dependence and independence), answering quickly or deliberately (impulsivity and reflectivity). Two uses were proposed on the earlier page: give learners tasks that suit their style, or deliberately build the approach a learner lacks. The model here is conditional on the task, not the person. For a learner whose attempts at a task stall in a way that looks like a habit of approach rather than missing knowledge (a holistic learner who cannot organise a lab report, a quick responder who makes careless errors in multi-step problems, a deliberate learner who freezes in a brainstorm), explicit modelling and supported practice of the approach the task requires, given to every learner who needs it and faded across sessions, is expected to produce better immediate-performance on that kind of task than assigning tasks or formats to match a measured style. Whether the habit itself changes, and whether a gain lasts to delayed-retention, are separate questions the evidence here does not settle.

No claim in the wiki tests either use of cognitive styles with learners. The nearest evidence is against the first: matching the format of instruction to a measured learning style (visual, auditory, Kolb's assimilator and accommodator) has failed to produce the predicted style-by-instruction interaction (Learning Styles Matching Does Not Improve Learning [-S]; No significant interaction between learning mode and learning style on gain scores, while the learning-mode main effect was significant [-M]). Those studies tested learning styles, not field dependence or reflectivity, so carrying them here is a transfer the earlier page's own distinction between the two constructs invites caution about. The second use, building balance, rests on theoretical argument only: styles are argued to be partly socialised and so modifiable (Intellectual styles are partly socialized and therefore modifiable to some degree [+W]), and heavy reliance on one style is argued to make the opposite harder to adopt (More stylistic information processing makes individuals' viewpoints narrower and more rigid, reducing ability to adopt opposite styles [~W]). A review reports that a Kolb style measure did not change over a year of varied instruction (Learning styles measured by Kolb's inventory showed no significant change over one year of varied instruction [~M]), which limits how much change to expect in the measured style, though it says nothing about task performance.

The relationship is also conditional on what a style measure is measuring. The embedded figures test, the common measure of field independence, correlated relatively highly with an intelligence test in a reviewed study (Embedded figures test performance differed by sex and correlated relatively highly with intelligence [~M]), and measured styles differ by cultural and linguistic group, partly because tests are given in English (Cultural and linguistic background influences measured learning and cognitive styles, and should be considered in style assessment [~M]). So a low "field-independence" score can be a low score on an ability or language test, and a design that sorts on it can sort on ability or background.

This page sits beside Personalization, which owns what to tailor to each learner and on what basis, and Adaptive Learning, which owns choosing the next task from responses; both already advise against tailoring on learning styles, and this page does not repeat their models. Cognitive Flexibility owns revisiting content across cases and perspectives; building the approach a learner lacks is a narrower move about one learner's habit on one kind of task.

Default design, while the relationship is untested

The earlier page's guidance, kept as a concrete default a designer can act on and revise. Each step is a proposal; its evidence status is stated beside it.

  1. Name what each task demands, not what each learner is. For each main task, write which approach it needs: part-by-part analysis (a lab report, a multi-step procedure), a holistic overview (summarising a text's argument, reading a map), deliberate checking (multi-step mathematics, proof-reading), or quick spontaneous response (brainstorming, cooperative turn-taking, team sport). Untested proposal, from the earlier page's account of which tasks suit which style.
  2. Give every learner the format the content needs. Choose diagrams for spatial structure, speech for pronunciation, text for an argument, for everyone; do not sort learners into parallel materials by a style inventory. Against matching: Learning Styles Matching Does Not Improve Learning [-S]; in one virtual-reality experiment both Kolb style groups gained most from the same guided mode (Assimilator learners gained significantly more in the VR (guided exploration) mode than in VR (non-guided exploration) or Non VR modes [-M]; Accommodator learners gained significantly more in the VR (guided exploration) mode than in VR (non-guided exploration) or Non VR modes [-M]).
  3. Observe before describing a habit, and describe the habit on a task. Before deciding that a learner has a characteristic approach, look at how they tackle at least three tasks with different demands over two or three weeks (an untested proposal for the earlier page's "multiple contexts"). Record "on the lab report, listed observations without grouping them", not "field-dependent". Do not use a style test to decide this alone: its score overlaps with general ability and with language and cultural background (embedded figures and intelligence [~M]; cultural and linguistic background [~M]).
  4. Where a learner stalls on an analytic task, teach the structure explicitly and fade it over four to six sessions. Untested proposal; the earlier page's "explicit help organising and analysing academic tasks". Session 1: model the structure for 5–10 minutes on a worked case (for a lab report: question, variables, method, result, claim), thinking aloud about where each observation goes. Sessions 2–3: the learner fills a printed frame with the headings, with a check-in at the end of each section. Sessions 4–5: the frame becomes a checklist the learner ticks after writing. Session 6 on: a blank page, with the checklist used only to self-check. Move on when two consecutive tasks are organised correctly without the frame. Modifiability is argued, not tested: Intellectual styles are partly socialized and therefore modifiable to some degree [+W].
  5. Where quick answers produce avoidable errors, add a stop-and-check step and fade it. Untested proposal, from the earlier page's account of impulsivity in mathematics and reading. Before submitting each multi-step problem, the learner spends 1–2 minutes estimating the answer or checking it by a second route, and marks the check on the page. Keep the visible check for three or four sessions; drop the written mark once the learner's error rate on unaided problems has fallen on two consecutive sets, and keep the check itself.
  6. Where deliberate learners hold back on a task that needs spontaneity, give short timed rounds with judgement deferred. Untested proposal, from the earlier page (creative writing, cooperative partnering). Two or three rounds of 2–3 minutes per session in which ideas are written or said without evaluation, followed by a separate round that selects among them; repeat for three or four sessions, then let the learner set the round length. Do not grade the generating rounds.
  7. Keep more than one route of expression open where the goal allows it, and let the learner choose. Where the objective does not fix the form (a diagram, a written explanation or an oral account of the same concept), allow any of them against the same criterion. Untested proposal, from the merged page "Teach students to make the best of their intellectual styles rather than forcing one mandated mode". The merged page's concern that schools reward one way of working is argued, not tested: Schools overwhelmingly reward executive styles and confuse style with level of intellect [~W].
  8. Do not keep a learner in their preferred approach. Pair step 7 with at least one task per unit that needs the approach the learner uses least, supported as in steps 4–6. Untested; argued in More stylistic information processing makes individuals' viewpoints narrower and more rigid, reducing ability to adopt opposite styles [~W].
  9. Where the habit does not shift, compensate rather than remediate. If after six sessions the learner still needs the frame or checklist, let them keep it as a permanent job aid on assessed work where the goal is the product, not the unaided structuring. Untested proposal, from the merged page's "compensation mechanisms may be needed instead". A review reports a measured style unchanged after a year of varied instruction: Kolb's inventory stable over one year [~M].

Across a term: weeks 1–3, observe approaches on varied tasks with everyone given the content's format; weeks 3–8, explicit structure or check routines for the learners who stall, faded as in steps 4–6; around week 8, re-observe on a fresh task of the kind that stalled, with no frame, and record what changed. When a condition is not met: if nobody can observe a learner across tasks, build the structure or check into the task for everyone and let those who do not need it skip it.

Fitting the design to a situation

The three facts that most change the decision:

  • What the task demands, and whether the learner's difficulty is about approach or knowledge. A learner who cannot organise a lab report may lack the structure, or may not understand the content. If unstated, ask: can the learner explain the content aloud when the organisation is not required, and do they stall on the same kind of task across subjects?
  • What the style information is and where it came from. An inventory score, a teacher's impression and a record of several task attempts are different evidence; a test score can reflect ability or language. If unstated, ask: what was the learner asked to do, in what language, and was it compared with their performance on the actual tasks?
  • How long and how often you will see the learner. Building an approach needs several sessions with fading; a single session can only give everyone the structure the task needs. If unstated, ask: how many sessions, how long, and who looks at each attempt?
If the brief says… Then change… Basis
The brief proposes sorting learners by a style inventory and giving each group matched materials Do not split materials by style. Give everyone the format the content needs; use the time saved for structure and checks on the tasks where learners stall Learning Styles Matching Does Not Improve Learning [-S] (learning styles, not cognitive styles; fifth graders and adults); No significant interaction between learning mode and learning style on gain scores [-M] (one VR experiment, Kolb styles)
Learners are novice on the task, or advanced Novice: give the structure or worked model to all, regardless of style. Advanced: drop the frame and check the approach on one unaided task; prior knowledge, not style, sets how much guidance to give Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~S] (cognitive-load studies; does not address styles)
Learners are child (primary age) Never tell a child, or their parents, what style they "are". Teach the structure or check as a tool for a kind of task ("for lab reports we use this frame"), with shorter rounds (5 minutes of modelling) untested proposal
Learners are adult in a work setting (adult-workplace) Ask how they already tackle this kind of task at work; build the structure or check into their own job aid or form rather than a classroom frame, and let them keep it untested proposal
The task needs part-by-part analysis and some learners see only the whole Model the structure with a worked case, then a printed frame, then a checklist, then a blank page (step 4 of the default design) untested proposal; modifiability argued in styles socialized and modifiable [+W]
The task needs quick spontaneous response and some learners hold back Timed 2–3 minute rounds with judgement deferred, not graded (step 6) untested proposal; stylistic rigidity [~W] (argument, no data)
The brief reports careless errors on multi-step problems Use one of the learners' actual wrong answers as a worked contrast: show the answer, then the 1-minute check that would have caught it, then add the stop-and-check step to every problem for three or four sessions untested proposal
A style or embedded-figures score is the only evidence about a learner Treat it as partly an ability and language measure; observe the learner on two or three real tasks before acting, and do not place a learner in a group or track on it Embedded figures and intelligence [~M] (one study in a 1985 review)
Learners are culturally or linguistically diverse, or the style measure is in their second language Do not use the measure for placement; give instructions non-verbally where possible and judge the approach from task attempts Cultural and linguistic background influences measured learning and cognitive styles [~M] (Native and non-Native samples in Canada and Alaska, reported in a review)
Learners have low confidence or low literacy Present the frame or check as something experts use, not a fix for a weakness; keep the frame's language simpler than the task; let a learner show the structure by sorting cards before writing it untested proposal
online-self-paced, no teacher present No style quiz at entry. Offer the frame or checklist as an optional download beside every analytic task and a built-in check step on multi-step problems; let the learner hide them once they no longer need them untested proposal; matching evidence carried (Learning Styles Matching [-S])
Mixed modes: some learners join remotely while others are in the room, or work on a phone only Share the frame as a single on-screen checklist everyone uses, not a printed sheet; run timed rounds in a shared document or chat so remote learners contribute in the same round untested proposal
One teacher with a large group, or scarce facilitator time (volunteers, one tutor checking in briefly) Do not diagnose individuals. Build the structure and the check into the task for everyone; spend the brief check-in only on learners who stalled on the last task, looking at one section of their work untested proposal
A single session only Give every learner the structure or check the task needs, model it once (5 minutes), and leave it in place; do not try to change a habit or claim it changed untested proposal
A fixed short time budget (for example 30 minutes) 5 minutes modelling the structure on a worked case, 20 minutes on the task with the frame, 5 minutes self-checking against the checklist; one product, revised once untested proposal
The brief names its own categories, rubric or routine (a school's style framework, its own report template, its own check-in) Keep it as given. Use its categories as descriptions of approaches to tasks, not labels on learners; change only what this page settles: no one restricted to matched tasks, and one supported task per unit in the less-used approach untested proposal
A learner's style profile or reflections on how they work are private or sensitive The learner decides who sees any profile or notes on their approach; record observations about tasks, not traits, in shared records untested proposal
The goal is a whole-task performance the brief wants kept whole, or a product for a real audience Do not split the performance into style-matched parts; keep it whole and put the frame or check inside it. The audience's needs, not the learner's style, set the product's form untested proposal
Stakes are high (graded assessment, certification) Assess every learner on the same capability and criterion; allow a choice of form only where the objective does not fix it; check the rubric does not reward one way of working the goal does not require untested; concern argued in Schools overwhelmingly reward executive styles [~W]

Eight of the nineteen rows cite a claim. Only the first rests on comparisons of matched and unmatched instruction, and those tested learning styles, not field dependence or reflectivity; the novice row carries a cognitive-load finding to this question; the measurement rows rest on studies reported in one 1985 review; the rest are untested proposals.

Observation, state and explanation

What can be observed is an attempt on a task under stated conditions: the task, the support available (a frame, a checklist, a time limit), what the learner produced, how long it took, and which errors appeared. "Field-dependent", "impulsive" and "reflective" are inferred states, and so is "lacks the knowledge". A style label is an interpretation of several such attempts, not an observation.

Response Plausible explanations Observation that could change the interpretation
Contributes well in group work, produces a disorganised solo lab report A holistic approach without a structure for analysis; weak grasp of the content; reading or writing load; prefers working with others Ask the learner to explain the experiment aloud; give the same task with a frame; give an equivalent task with less writing.
Fast answers with many errors on multi-step problems A habit of quick response; does not know the procedure; time pressure or anxiety; low interest in the task Give an untimed problem with a required check and ask the learner to explain each step; compare errors with and without the check.
Slow and accurate on analytic work, produces few ideas in a brainstorm A deliberate approach; fear of being judged; little content knowledge to draw on Run a round where ideas are anonymous and ungraded; give a topic the learner knows well.
Low score on an embedded-figures or style test An analytic approach not habitually used; lower general ability; unfamiliar test format or language Compare with the learner's attempts on real analytic tasks; repeat with non-verbal instructions.
The learner says "I'm a visual learner" and disengages from text A stated preference; a label taught earlier; a real difficulty with reading Offer text and diagram versions of matched content and compare learning on matched items; check reading separately.

These rows are proposals, untested with learners as diagnostics. A probe is also practice and can change what it measures, so record each one and keep the explanations that remain open.

Evidence and qualifications

No claim in the wiki tests building a cognitive style, or matching tasks to field dependence or reflectivity. The claims below bear on parts of the relationship.

Learners, measures and settings differ across these claims (fifth graders on comprehension tests, adults in virtual reality, adolescents learning English vocabulary, reviews of style inventories), so do not rank them by effect label. None measured delayed task performance after teaching an approach.

The evidence by question

  • Does assigning tasks or formats to match a learner's measured style improve learning more than giving every learner the same well-chosen task? Learning Styles Matching Does Not Improve Learning [-S]: its table gathers the studies behind it (4 distinct studies as it now counts them), all on learning styles rather than field dependence or reflectivity.

Further evidence, not yet read against this model

Claims about cognitive and learning styles that were read for this rewrite but are not part of the model. None tests a design decision this page makes.

Objective and learner-valued goal

The designer's objective is usually a capability on a kind of task (an organised lab report, accurate multi-step problem solving, a set of ideas in a brainstorm), unaided or with a stated aid, at a stated horizon. Elicit separately what the learner wants: to work with friends, to finish quickly, to avoid being wrong in public, to be good at a subject they already like. A learner who values working with others may resist solo analytic tasks for that reason, not for want of an approach; a learner who has been told they are "a hands-on learner" may value keeping that identity. Agreement and divergence are both observations.

For example, a field-dependent-seeming secondary student may want to do well in history, which they enjoy, while the designer's objective is an organised science lab report. Teaching the report frame serves the designer's objective; showing that the same frame organises a history source analysis connects it to what the learner values. Record both, so success on the lab report is not read as the learner's goal met.

What would revise this model?

The expectation against matching should weaken if a study using the crossover design, with learners classified by field dependence or reflectivity and randomly assigned to matched or mismatched tasks, found that matched learners learned more on a common measure. The "build the missing approach" proposal should be dropped or narrowed if supported practice in the less-used approach produced no better task performance than ordinary practice, or if gains appeared only while the frame or check was present. If style measures turned out to be mostly ability or language measures for a given population, the observation step should rely on task attempts alone. Do not protect the model by calling every failure "the habit is too fixed" after the fact.

Matching, building an approach, compensating with a job aid, and a change in the measured style are separate claims. The present evidence establishes none of the session counts, timings or fading points above.

Claims

Examples

Key Sources

  • Witkin, H. A., Moore, C. A., Goodenough, D. R., & Cox, P. W. (1977). Field-dependent and field-independent cognitive styles and their educational implications. Review of Educational Research, 47(1), 1-64. doi:10.3102/00346543047001001
  • Zhang, L. F., & Sternberg, R. J. (2005). A threefold model of intellectual styles. Educational Psychology Review, 17(1), 1-53. doi:10.1007/s10648-005-1635-4
  • Zhang, L. F., & Sternberg, R. J. (2006). The nature of intellectual styles. Lawrence Erlbaum Associates.
  • Evans, J. St. B. T. (2004). Rationality and reasoning. Psychology Press.
  • Pritchard, A. (2005). Ways of learning: Learning theories and learning styles in the classroom. David Fulton Publishers.
  • Loo, R. (2004). Kolb's learning styles and learning preferences: Is there a linkage? Educational Psychology, 24(1), 99-108. doi:10.1080/0144341032000146476
  • Stahl, S. A. (2002). Different strokes for different folks? A critique of learning styles. American Educator, 26(3), 29-49.
  • Arduini-Van Hoose, N. (2020). Learning differences. In Educational psychology. Retrieved from https://edpsych.pressbooks.sunycreate.cloud. CC BY-NC-SA 4.0.
  • Presseisen, Barbara Z.; Sternberg, Robert J.; Fischer, Kurt W.; Knight, Catharine C.; Feuerstein, Reuven. (1990). Learning and Thinking Styles: Classroom Interaction. National Education Association. https://eric.ed.gov/?id=ED327322
Page metadata
Field Value
type principle
id cognitive-styles
aliases teach-to-intellectual-styles-flexibly
title Cognitive Styles
description For a learner whose attempts stall in a way that looks like a habit of approach (holistic or analytic, quick or deliberate), teaching the approach the task requires to everyone who needs it, faded across sessions, is expected to help more than matching tasks or formats to a measured style; no claim in the wiki tests either use of cognitive styles, and the matching evidence, from learning styles, runs against it.
status review
canonical True
generated by claude/unspecified on 2026-10-15
trust unverified — no one has confirmed this page's content
sources 1 — presseisen-1990

Frontmatter is the page's machine-readable record — see the Schema & Guide.