Last changed 3 June 2012 ............... Length about 10,000 words (69,000 bytes).
This is a WWW document maintained by Steve Draper, installed at You may copy it. How to refer to it.

Web site logical path: [] [~steve] [this page]

Learning styles (notes)

Stephen W. Draper


This document contains some notes on the notion of learning styles. These are only my personal views on some of the main issues, offered as an introduction to thinking about learning styles. Pointers to further reading are included.

  • Part 2 is notes on a sceptical position I put together about 1999.
  • More recently I'd just say read this, which is a 2008 sceptical position based on a psychology-based review:
    Harold Pashler, Mark McDaniel, Doug Rohrer, and Robert Bjork (2008) Learning Styles: Concepts and Evidence Psychological Science in the public interest vol.9 no.3 pp.105-119
  • Now I've thought of another angle: see Part 1.

    Part 1: Perspectives

    Stimulated by a student, Ailsa Darling, it occurred to me that it may be more insightful to consider the assumptions different researchers have made before they start. These assumptions are often not stated, and may not even be conscious. They are certainly not tested and supported by evidence. Here are 3? different angles from which people come to what perhaps should be a single field.

    1. Learning styles. The assumption is that learners are different (true) and that if teachers create different materials based on different presentational modes or media for different learners everything will be better. I.e. that the teacher should change, but not the learner (even though the entire point of education is to change the learner). Felder is the most cited person here.

    2. Sensory modalities and memory. It has long been known that mnemonists (people with exceptional abilities to memorise lists and other material) usually do so by assigning extra information to the material e.g. assign each item with a colour, a shape, an animal, a spatial position in an (often imaginary) room or journey. This is established for mnemonists; it is productive advice for students trying to memorise material for exams; and is consistent also with evidence about state-dependent memory e.g. how smells trigger recall of events and feelings associated in the past for you with that smell. Note that this is highly relevant to learning (even if not to the deep learning educationalists might wish to focus on), but the lesson is NOT to select a single medium different for each learner, BUT to encourage all learners to use all media to assist recall.

    3. Selecting the right medium for each piece of content to communicate. The dictum that "a picture is worth a thousand words" implies that it is always best to use pictures, always bad to use words. This goes against the evolution of language, common experience, the way you can follow TV news completely with the vision off, but can't "get" most of the content with the sound off (you can often tell which story it is from the pictures, but not what the story is today). However there are striking cases where a picture or diagram do communicate much better than words alone, as Tufte's books bring home. The right question here seems to be: what messages or content are best communicated by sentences, which by pictures or graphs? And then, how to optimise each is a skill that takes years to become expert in.

      This should be obvious to us. Textbooks for the last 50 years use copious expensive illustrations (but don't abandon the main text); many disciplines (most sciences, history of art, ....) routinely not only use but require both pictures and text in almost every paper.

    4. Using multiple integrated modalities / media in every communication. This is like the second position above; but focussing not on learner recall but on initial understanding (communication) (e.g. understanding the primary exposition such as a lecture or textbook). The idea here is not to have different materials for each learner, nor to select one optimum medium for the current content, but to expect good communication generally to use both text and pictures ... The assumption is not that different learners need different media but that all learners benefit from multiple media.

  • A sensible research programme should perhaps combine all four perspectives. it would have to study not just the learner (and teacher) but the relationship to particular material (learning content).

  • It would study media for learning materials, not as alternatives but as combinations.
  • It would study the methods students use to study (graphs are no use until a learner has had a lot of training in reading them).
  • It would lead to training students to use media, not pre-suppose that they have fixed traits, measurable as preferences. The latter is like testing an infant for reading skill and concluding that if they don't know how to read then they cannot be taught to; testing them for French and concluding they can't learn a second language; etc.

    Is this like Spontaneous Conceptions (often called "misconceptions")? Is it like the recurring error of thinking that technology causes learning?

    Part 2: notes on a sceptical position more than a decade ago.

    If you want my (then) personal conclusions, jump to the summary section in the middle.

    Contents (click to jump to a section)

  • A personal generalisation
  • Feminism, multiculturalism, and learning styles
  • Key references
  • Web sites for further reading
  • Other references
  • Pask references


    Students and others periodically ask me about learning styles. Personally I have tended to think them uninteresting; however interest in them recurs spontaneously. (Perhaps learning styles in education is a concept like the spontaneous misconceptions that recur in every subject, and are studied in the education literature: see for example Viennot, L. (1979) "Spontaneous reasoning in elementary dynamics" European journal of science education vol.1 no.2 pp.205-222.)

    They also all tend to believe something I think is quite wrong.

    (A better researched, but less sceptical, view is contained in this literature review: Coffield et al. (2004)).

    The basic intuition

    The intuition seems to be that people and therefore learners are different from each other, and that teachers should respond by creating different instruction for different kinds of learner. That's it. Is it sensible? Is it true?

    My position

    My position in summary is:

    "If you travel with us you will have to learn things you do not want to learn in ways you do not want to learn".
    [Doris Lessing; Nobel Laureate for literature; from a letter replying to a reader. Quoted in Alan Yentob's "Imagine" programme on Doris Lessing, broadcast Tues 27 May 2008, 10:35pm on BBC1 ]

    Stronger scepticism

    Daniel Willingham argues that what is true is that people differ in how good their memory for visual or auditory or kinesthetic information is. So on a test requiring learning of that kind of modality-specific information, they will differ. But this is of little relevance to most education: because most of that is about learning meanings, not sensory information. Someone whose style is visual is not helped by getting a visual representation of different tones of voice, or different birdsongs; and in fact they are not helped by getting visual representations of meanings.

    Another view

    A hint of a slightly different view comes from the two short chapters by Sternberg in Perspectives on thinking, learning, and cognitive styles R.J.Sternberg & Li-fang Zhang (eds.) (2001) (London : Erlbaum Associates). There's a spectrum of possibilities of what styles (if they exist) might be: The shift from neither having nor conceiving of a choice of learning method, to having and making such a choice must be a profound developmental shift, though perhaps different in different subject areas. In fact the stuck-ness observed in some learners may be of this kind. Certainly the realisation in problem-solving that not all problems can best be solved by top down approaches alone (or conversely by bottom up approaches alone) is a major milestone. To the extent that problem-solving and learning are similar, this implies that there can be no more deeply educational goal than to free learners from a single "style".

    Kinds of "style"

    There are different kinds of thing, any of which might conceivably be called "style", but all of which we should consider as they are all learner properties that differentially affect their achievements. Here is a list, probably not complete, defining terms in a way that may well differ from that used in some or all of the literature. The distinctions however are important, even if the terminology is disputable.

    Despite this, note that Cronbach & Snow (ch.11 p.375ff.) say that cognitive styles cannot logically (or do they mean methodologically?) be distinguished from traits or abilities. (All you can do is to measure how well people do on various specific tasks; and then see to what extent scores on one test/task go with (predict, correlate with) scores on another.)

    General learning ability

    Of course different learners benefit differently from the same material: firstly because some may already know more, secondly because some want to learn more, and so try harder, than others; but thirdly and above all because people vary considerably in their general ability and quickness at learning. IQ is one measure of this. Good learners will learn more than bad learners for almost all materials.

    Learners also differ from each other in more subject-specific aptitudes for learning e.g. some being better at verbal than numerical things, others vice versa. Subscales of IQ tests and other tests can show this. These differences between learners are not learning styles, and don't imply differences in how to create material, but that a given learner is often quicker at learning some kind of topic than others.

    Variable quality learning materials

    Of course a learner will do better with some materials than with others: some material is just worse than others, and even though clever students may still learn quite a lot from poor material, they will learn as much or more (and more quickly) from better material.

    Crossover effects:
    "Aptitude-treatment interactions"

    To show you have an effect that isn't just differences in general learning ability or general quality differences in the materials, you need an experiment that shows a crossover: where type A learners do well on material X but poorly on material Y, while at the same time type B learners do poorly with material X but well with material Y. On a graph, lines will cross. It is a sign of conditional (not universal) causality in operation (e.g. "if learning style A and teaching style Y, then low performance; ...").

    Illustrative graph These graphs show two equivalent views of the same imaginary 4 data points, representing the mean learning scores for 4 groups.

    This central point and others are elaborated carefully in: Cronbach,L.J. & Snow,R.E. (1977) Aptitudes and instructional methods: A handbook for research on interactions (New York: Irvington Publishers Inc).

    Informally, we could think of this as studying the interaction (or interference, good and bad) of alternative learning styles with alternative teaching styles. They call this "aptitude-treatment interactions", which is careful, technical, and general terminology. "Interaction" is a statistical term meaning that there is some crossover effect besides the "main effects" of differences in general learning ability in learners, and differences in teaching quality in different teaching materials. "Aptitude" refers to any kind of special learning ability that varies between individuals. "Treatment" refers to any teaching intervention in a study: whether different material, different tutor actions, etc.

    People get the impression from that book that no such effects have been found. However Gordon Pask did demonstrate such an effect. E.g. Pask, G. (1976) "Styles and strategies of learning" British Journal of Educational Psychology vol.46 pp.128-148. (A much longer list of publications by Pask is given in a later section.) A more recent study with clear crossovers is that by Shute (1993). She makes the point that the failure in many early studies to get a crossover was because they did classroom studies that didn't (couldn't) control other variables; but that if you use computer instruction for single learners (as Pask did and she did) then you can. This point cuts both ways: if you want to study such effects and the underlying causal factors, then that is the way to do the research; but on the other hand it equally implies that the effects are not big enough to have much effect in classroom education (though computer based learning and/or distance learning could be different in this respect).

    If you do such an experiment and get not a crossover, but two lines of very different slope, then statistically this is still an interaction (a conditional outcome that depend on both the learning and the teaching style). However it would no longer be of the same practical interest: e.g. suppose in the left graph, line A was horizontal i.e. point 3 was near point 2. That would mean that although teaching material B was only good for one kind of learner, material A was good for both, and it would no longer be sensible to have different materials for different learners: you would only want to use material A.

    (Val Shute told me that "ordinal interaction" is the proper term for this latter effect where the lines don't cross (one always better than the other); while what I called here a "crossover effect" is technically a "disordinal interaction". Details on this here (use your browser's Find command for "disordinal").)

    What should (learning) designers do?

    How should we respond, if/when it is demonstrated that a) learners do have different learning styles, and b) there are some learning materials that benefit different styles differently?

    Many people without thought assume we should produce different sets of learning materials, each optimised for each different learning style. This does not follow.

    Landauer on HCI and user variability

    In a book on HCI (Human Computer Interaction), Landauer refers to the literature on learning styles, and produces an argument about HCI that is of interest here.
    Landauer,T.K. (1995) The trouble with computers: Usefulness, usability, and productivity (MIT press; Cambridge, MA)

    In this book (in ch.12 and around p.292), Landauer argues that in many current user interfaces there is a big variation in user performance because it depends on certain mental skills. He says that many designers conclude that they should provide different designs for different users. He argues however that that is quite wrong. It is the case that many designers have unusual mental skills and have produced interfaces that require these skills, but that it is quite possible to produce designs that are good for everyone, not just a mental elite, and on which everyone performs as well as the best do on current interfaces.

    The analogous argument for education is, again, that we should produce materials that are optimal for everyone. An experiment that shows a crossover effect may establish that learning styles exist with an effect on performance for those materials. But this can not show that the effect will exist for all possible materials: instead, perhaps teachers (like user interface designers) should try to develop materials that are equally successful for all styles of learning.


    Egan (1985, 1988) argues that our preferred response should be to seek to develop "robust designs" that, unlike the ones giving a crossover effect, are "robust" with respect to the individual difference measured i.e. are good for all learning styles. He is able to describe a number of successful cases of this in the field of human computer interaction design.

    In this context, he lists as approaches:

    1. Robust designs: find a design good for all.
    2. User prototypes: classify all users into a few contrasting groups; and treat them differently. E.g provide a Spanish language version for Spanish speakers. This is the design approach pre-supposed without thought by many in this area. But usually, Egan observes, it is hard to discover how to tailor designs to the measured differences.
    3. Adaptive trainer systems. If the materials are digital, at least, then learner performance (e.g. on tests) can be used in theory to identify common types of error and trigger special remediation for that type of error. So there will be a single "resource" for all learners, but it will behave differently, and in a way to reduce differential learning: Automatic recognition of common errors, which then get remedial action by the software. The text version of this is textbooks with self-assessment questions, with optional extra sections for those having trouble with each question.
    4. Automated "mastery learning" [Bloom's term]. Instruction, test, targeted remediation, iterate.

    Learning methods: learning to learn

    Another issue to consider is that learners learn methods for learning, and that these make an important difference. If so, then even if you found a crossover effect, perhaps your response should not be to create several kinds of material to suit different learners, but to train learners on better methods of learning.

    It is almost certainly true that learning to learn makes a considerable difference to every learner: most can get better at learning, not by magically growing their IQ, but by improving their skills at learning. Furthermore it is likely that a given method of learning may allow a learner to do well at only certain kinds of material, while another learning method would be better for other types of material. It follows that you may well find a crossover effect because in your participants some had one learning method, others another, and your materials gave different advantages depending on the learning method applied. However if you retrained your participants so that they could all use either learning method and were able to select and apply the best one for a given material, then the crossover effect would vanish. Even if you find (crossover) evidence that learners currently have different learning methods/styles, this does not show that these couldn't be changed — perhaps quite easily.

    I believe that Pask thinks this is the case (but you should check this out).

    If true, this means that it does NOT follow from establishing a crossover effect that teachers should produce alternative versions of material. Instead, they should train learners to have more than one learning method, so that all can benefit from the same material, and indeed so that all can benefit from a range of styles of material.

    Cognitive style and exams

    There are some cognitive styles, such as field dependence/independence and the size of working memory, that have been shown (e.g. in work at CSE) to have large effects on the scores of students on test or exam questions. Unless the examiners actually intend to test for the size of working memory, rather than ability at the subject (e.g. chemistry), then this work can be interpreted as showing that many exam questions or school tests are in fact badly designed and invalid i.e. test something different from what they are meant to. It does mean educators should take cognitive styles seriously in this sense of trying to design tests that are independent of such styles; it does not mean they should design different tests for learners with different styles.


    One way of looking at this is that it is part of personalising learning materials for each learner. Perhaps part of the appeal for teachers is that this gives importance to their professional role, and seems the obvious response of a caring person. Against this might be argued:


    Thus it is my provisional opinion that the topic of "learning styles" is based entirely on a recurrent mistake that concludes from the fact that there are important and relevant differences between individual learners that educators should design different materials for different types of learner. My opinion is not based on a thorough reading of the recent literature. To form your own judgement, you should do that: starting points are provided below in the links to web sites.

    A more moderate view is that of Pask, who says in the conclusion of a review paper (Pask 1988) "It seems evident that distinctive learning strategies exist. ... There are also certain distinct styles, or dispositions to adopt classes of strategy. ... However I am still perplexed by the extent to which some individuals persist in using the same strategy across contexts. ... Also, some persons seem more flexible, versatile, and context-sensitive than do other persons."

    The core issue for teaching methods and materials is going to be how easy or hard it is to change:

    But in addition, we cannot neglect (as I used to think we could) the whole issue of learning and cognitive styles and traits, because it has now been shown that many of the tests and exam questions we produce in fact perform differently for different learning styles, thus measuring something different from what they are meant to.

    Finally learning styles are just one part of a more fundamental issue. Should teachers adapt to learners, or learners to teachers? The answer is "both"; and the concept to think of is that of learning communities. All (institutional) learning can be thought of from a wholly social perspective, as one of the learner joining a community, and becoming enculturated. From that point of view, the learner needs to do the adapting, and the more they do so, the more they gain access to that subculture and its knowledge.

    The complementary viewpoint is that teachers should adapt, not so much to individuals, as to the broadest audience possible; to make their material accessible to the most people. The best writers can do this — indeed, learning to write is largely the process of overcoming "egocentricity", and learning how to make your private thoughts comprehensible to others with different experience. That is part of a teacher's job too.

    Some different kinds of "learning style"

    There are many kinds of learning style. Follow the web links below to see more of them. Here are just a few.


    Aural vs. visual medium (for example).
    Verbal vs. pictorial medium.

    Caveats here include:

    Holist / serialist

    Pask's two learning styles were holist vs. serialist.

    Bernice McCarthy's 4MAT system

    See this page for an introduction.

    Short term memory; field dependence

    Recent work at the centre for science education here has showed important differences in student performance depending on their working memory capacity and their degree of field dependence. However a main implication of this is that examiners in, say, chemistry should decide whether they want to test for chemistry knowledge or for the size of STM; and redesign questions so that they can be equally well answered by students with a range of scores on working memory and field dependence tests (see, for example, Witkin et al., 1977).


    It has been shown that vocabulary is a significant problem in teaching in HE, with tutors assuming students understand terms that they in fact do not. Consequently, students with a larger prior relevant vocabulary will probably do better than those without it. Again, the responsible conclusion is that teachers should teach the vocabulary they are using (it is really part of the subject matter being taught), instead of penalising students differentially for the teacher's poor practice.
    (For more information see my notes.)

    Cognitive styles vs. learning styles

    Below are some messages referring to researchers who believe in a distinction between cognitive and learning styles.

    A possible logical distinction might be:

    Despite this, note that Cronbach & Snow (ch.11 p.375ff.) say that cognitive styles cannot logically (or do they mean methodologically?) be distinguished from traits or abilities. (All you can do is to measure how well people do on various specific tasks; and then see to what extent scores on one test/task go with (predict, correlate with) scores on another.)

    (The notes below are taken from messages to ITFORUM.)

    by: "Laura J. Little-Reynolds" :
    You're right...the research on learning styles and cognitive styles is a bit confusing. If I recall correctly, learning styles refer to ways that people learn information, and cognitive styles are more global, referring to the way that people see the world around them and interact with it. A fine distinction, but it's there.

    An excellent resource is Jonassen and Grabowski's Handbook of Individual Differences in Learning and Instruction. It defines learning styles and cognitive styles (my explanation is adapted from it), and gives an overview of the different styles that have been identified, and summarizes research related to each style.

    Here's a web site from the Maisie Center that has a good list of resources:

    by Steve Tripp:
    I would urge great caution as you review the learning and cognitive styles literature. Valerie Shute has used branching based on prior knowledge in a CBT package to teach statistics. There's also a body of research that suggests that novices who possess less domain knowledge do not benefit from menu choices as much as competents or experts. I think Ruth Colvin Clark addresses this in her new book, Building Expertise, published by the International Society for Performance Improvement. I am unfamiliar with research showing that other measures of cognitive styles such as thinking inventories or learning styles (elaborative versus reflective) can account for significant amounts of variance in domain-specific performance. If you find anything, please let me know. It is also important to note that there are a lot of people marketing their own learning or cognitive style instruments. I would be leery of them, unless you know how they were normed and what their validity and reliability are. Moreover, a person could still create a valid and reliable instrument that really doesn't account for a lot of variance in domain level performance. I would look for research that provides this link. The current revival of designing computer systems that adapt to learning and cognitive styles harkens back to the Richard Snow and Lee Cronbach's 1976 work in aptitude-treatment interaction research. While they were unable to find aptitudes (such as learning styles) that accounted for significant variation in performance, they said it may be possible. Snow (mid-1990s) has an inventory of abilities, some of which may lend themselves to variations in computer system design.
    Cognitive Preference
    by Bruce Jones:
    As I recall, Jerome Bruner proposed a theory of cognitive preference.

    1. As we develop, we prefer the presentation of information in a progression from psychomotor or kinetic to audio and visual to abstract representations. The implication is that for the very young, the most effective presentation of information would be psychomotor.

    2. Then we develop a general preference for one or more of the three categories of information presentation. The implication is that some prefer their encounters with the world in a highly kinetic way: surgeons, athletes, short-order cooks, etc. Or, if we are abstract takers of the world we may insist on reading the book about tennis to learn the sport. [joke]

    3. An interesting interpretation of these ideas is that the first point may apply to a person regardless of chronological age but according to "novelty of the subject matter to an individual." This you might refer to a relative age. The implication is that a person with an abstract preference who is middle-aged may be better served with a psychomotor presentation of information if it is highly novel to him or her. In other words, a person with a literary preference may still be better served, as would a child, with a psychomotor presentation of breadmaking. At the same time, knowing that the individual has an abstract bent, an instructor can supplement the early learning with articles on dough kneading techniques.

    Disclaimer: I may be off in my recollection of the theory.

    Try this site as well:

    Laurillard; personal vs. abstract; abstractions vs. examples

    As theories such as the Laurillard model make clear, good teaching should embrace multiple aspects of a given topic, but most does not. Again, the implication is to improve teaching to cover all aspects properly. In the absence of that, differences in students may show up as "style preferences". For example, all teaching should cover both the public, conceptual aspects (what the theory is and what it is called) and the personal aspect: how this topic connects with each student's own experience.

    Another similar issue is that all generalisations should be conveyed by both specific examples or cases and a statement of the generalisation. Everyone needs both, though some may be more tolerant than others of the omission of one or the other.

    Kolb's learning styles could in my view be similarly regarded: as a set that all learners should cover and which should all be equally though separately supported in properly designed material. (For an sketch, see this.)

    Turkle and Papert: concrete and formal thinking

    Turkle,S. & Papert,S. (1991) "Epistemological pluralism and the revaluation of the concrete" ch.9 pp.161-191 in Harel,I., & Papert,S. (eds.) Constructionism (Ablex: Norwood, N.J.)

    Turkle & Papert (1991) argue that there are 2 styles of thinking and learning: abstract and concrete. P.175 they say that for Piaget "mature thinking is abstract thinking ...[but for Turkle and Papert] formal reasoning is not a stage, but a style." And they make clear that this can also be important for learning e.g. LOGO was designed to offer a way into mathematics for concrete thinkers who have no other way in at all. But I myself think that this interesting and important contrast is neither about developmental maturity nor a style, but identifies two kinds of thinking both of which should be supported in any well designed learning and teaching about a topic. They describe, no doubt correctly, the strong preferences for one over the other that many people exhibit. But for me, when a person with a broken toe exhibits a strong preference for resting their weight on the other foot, this is not an argument against healing the toe or expecting them to walk one-legged for the rest of their life. When teachers push only one aspect then this is like forcing learners to hop one-legged and is equally unbalanced and perverse. For me, the challenge is to identify what the two apsects are for each topic, and how to teach them. Thus computing science lectures almost entirely on the abstract pre-planning approach and fails to teach what they all know is important in reaility: how to debug programs, how to use debugging tools, how to develop programs "middle out" from a simple but complete working version to a complex working version.

    A personal generalisation

    Underneath this issue of learning styles and how to construct materials, I think is a fundamental situation for all learning, represented by the profound analogy of learning/teaching as a conversation.

    In a normal 1:1 face to face conversation, both parties "coordinate" and adapt the language they use to an extent researchers are only beginning to notice and document. On the other hand, we have discovered in the last two millennia that language can be used successfully to transmit ideas even when frozen by writing. It is not as easy to read Plato as to listen to a good personal philosophy tutor, but it is quite possible. So how much of the effort of meeting half way should a) the teacher b) the learner do?

    Part of the issue is cost and teaching ratios. (If only 1/40th of the teacher has been paid for the benefit of each individual learner, then the learner may have to do most of the adaptation.)

    But more fundamentally something else is going on. Part of learning a subject, e.g. "physics" or "medicine" is learning the language used by experts: so if the learner doesn't do the adapting by learning how to join that language community, they won't really learn the subject. This is a situativity perspective: learning as joining a community. In the end, a teacher can't do this for a learner. (See also Feynman on how, in the end, to understand physics you have to grapple with the math: it cannot be translated into a human natural language, even by an outstanding communicator doing lectures specially constructed for a non-specialist audience. R.Feynman "The character of physical law: a series of lectures" 1965 — ch.2 I think for the relationship of math to physics.)

    One level of abstraction up, education in general is trying to connect learners to the now enormous body of socially distributed knowledge. Teaching them to be able to access and mine that for themselves, using whatever communicative conventions that body uses, is the real job. Producing specially translated digests may get them some special information, but doesn't help the big goal. I understand that the best current evidence (as seen from the UK) is that small class sizes have a demonstrable benefit for the first 3 years of school, but ONLY for that (even though 'most everyone feels small class sizes are naturally desirable for everyone all the time). I interpret that as the induction period, when it is vital to get diverse children aligned with the education system, and teachers need to make big efforts with personal services. Again, aligning the learners is important for them and their future, rather than aligning teachers and materials for the benefit of unchanging individual characters.

    So I think that holding out for materials that are good for almost all learners, rather than going for a tower of Babel solution of multiple differently adapted materials fits in with the fundamental requirements of education and forging a WW community of knowledge. Now within all that there is still a huge and difficult brief for teachers: scaffolding learners as they try to connect with a given knowledge community, scaffolding their personal learning work in finding personally meaningful connections between that and their personal experience, and finding how to create materials that "work" for many types of learner at once, rather than only for the teacher or the subset of learners most like the teacher. But note how this is essentially the same as the work done by both "great" and commercially successful novelists, painters, film makers: they find how to create things that mean something to a wide range of different individuals. One of the big lessons I feel I've learned about novels, is how the best ones mean such different things to different people, who agree only in that it was worth reading. (See the preface to the second edition of "The Golden Notebook" by Doris Lessing for one fascinating illustration of this.)

    And a rejoinder by Clark Quinn

    These are goals I fundamentally agree with. In many contexts, exemplified by educational institutions. However, there are other contexts. I would argue that there are times when we want to develop materials that are adapted to the individual's learning characteristics.

    For example, those learners who are well past, and did not have the benefit of, those first three years of small class sizes. Are they to be excluded from task communities because they can't use the existing materials to acquire the terminology? Or can communities become more open by leveraging educational research to facilitate integration? I'm reminded that the medical education community is using new representations to facilitate learners acquiring diagnostic skills (model-based reasoning) that not all got from existing materials. Maybe dynamic and spatial visualization is not as mutable as we would like.

    Or those situations with immediate training needs where performance outcomes need to be maximized and training times need to be minimized. Such as, hypothetically, training rescuers for survivors of devastating earthquakes.

    Or material that needs to be known by all learners, not just most. Such as the social prohibition on murder, or how to protect against AIDS.

    I haven't exhaustively researched all the characteristics, but I believe some learning characteristics are malleable, some may be amenable to coping strategies, and there are likely to be some things that cannot be overcome (at the extreme: blindness and deafness). For instance, despite the controversy, there do seem to be quantitative differences on cognitive abilities between individuals. I'm consequently inclined to believe there should also be some redundancy in the materials. Let me restate that as flexibility in the materials.

    One issue is definitely cost, whether adapting the teacher or adapting the materials. In the technology case however, you might amortize the cost, grabbing the best representations from eclectic sources, making it affordable. If you first specify (and yes, we need research) what the things are that make up a good design, including the minimum multiplicity needed, you can design to it. If you balance the breadth of representation with some real practicality in development costs, can you not hit both targets?

    With as many textbooks there are in any given area, certainly the content and different representations are probably already available somewhere. Can we make a mechanism to characterize and pull together relevant different representations? See, for example, Tom Murray's 1998 paper on this topic: 'A Model for Distributed Curriculum on the World Wide Web' in the Journal of Interactive Media in Education.

    So, I agree that there are and will be many situations where we want to ensure learners have the greatest ability to self learn, which includes Steve's goal of 'aligning the learners', and that we will need to develop a design that works well for most learners. But I will also argue that we should also strive to understand and learn how to develop learning materials that work for ALL learners.

    Clark Quinn    24 Aug 1999
    Knowledge Universe Interactive Studio

    Feminism, multiculturalism, and learning styles

    Here's an extract from a paper: "Issues of feminism and multicultural education for educational technology" Martin,D.J., Lucek,L.E., & Fuentes,S. (1999) ["Published" on ITFORUM on 8 Nov 1999.]

    "We are always on shaky ground when considering cultural differences. It is vital to examine how culture may influence learning and achievement in school, but the danger lies in overgeneralizing its effects" (Nieto, 2000, p. 140). To illustrate the "shaky ground," we discuss one attribute of learning styles, the characteristics of field dependence (more recently, this is referred to as field sensitivity) and field independence. Bennett (1995) indicates that learners with a more field dependent style tend to have a more global view, are more sensitive with "highly developed social skills," and are extrinsically motivated. Field independent learners may be better able to perceive discrete parts, are more individualistic, and are more intrinsically motivated. Shade (1997) summarizes that African-Americans tend to be more field dependent whereas Euro-centric students tend to be more field independent. Bennett (1995) also indicates that "Mexican Americans tend to be relatively field dependent or global in orientation" (p. 168).

    Field dependent learners tend to favor a "spectator approach" to learning and field independent learners tend to favor "inquiry" approaches (Bennett, 1995). Is it possible that a particular theory of instruction, such as Reigeluth's Elaboration Theory, is appropriate for designing instruction for some minority groups that have a more field dependent learning style? If some African-Americans tend to be more social and relational in learning styles (field dependent), they may learn more productively with interactive, collaborative situations, but not be as successful with inquiry/Socratic learning situations and with competitive educational methods. Euro-centric students may learn more successfully in inquiry learning situations and individual-based situations, but have more difficulty with collaborative situations.

    In her case study of computer use, Chisholm (1996) discusses problems of computer access, but goes beyond that to note learning style differences among a culturally diverse group of young students. Chisholm identified the following cultural themes that emerged in the use of computers:

    The students whose cultures value cooperation and interdependence, such as the Mexican-Americans and the African-Americans, could work and share with others. Those whose cultures value independence and self-reliance, such as the white culture, could work alone. Whereas those whose native culture tends to look at the world holistically, such as the Mexican-Americans, could explore and learn through play, those from cultures valuing analytic thinking could learn in a step-by-step deductive fashion (p. 171).

    These propositions are not intended to highlight cultural "deficiencies," but to highlight strengths. We are familiar with the literature that indicates the importance of using a variety of learning styles and teaching styles. The argument, however, is that education in the United States has tended to focus on learning styles for the Euro-centric students' competitive, inquiry-driven, and independent work. A vital caution — where we are standing on shaky ground — is in the "misapplication of learning style theories" (Nieto, 2000, p. 143). Nieto summarizes studies in which teachers made incorrect assumptions. For example, in one study, Flora Ida Ortiz indicated that teachers assumed Hispanic students would not want to assume leadership roles in the class activities; thus teachers did not provide the Hispanic students with opportunities they provided to non-Hispanic students. Nieto indicates there is particular promise with Howard Gardner's multiple intelligence theory "in challenging current assessment practices that focus almost exclusively on logical-mathematical and linguistic intelligence" (p. 144).

    Web sites for further reading

    Here are a few web links that give short introductions to the topic, and/or short and useful-looking reading lists.

    CSE publications list: this has a subsection called "learning styles".

    Biblography at the "learning styles network"

  • 7 minute video on why learning styles don't exist

    Learning styles: reading list, page:

    Another useful page:

    Another introductory page:

    Learning styles page:

  • Liz Brown's references

    Learning strategies database

    Learning Styles

  • Richard Felder's website with his papers and instruments. He too assumes teachers must change, not learners: as if education should not change learners, and the job of teachers is to preserve them from having to change or to learn new methods.

    Pask references

    Why read Pask? My two reasons are:
    1. He demonstrated a crossover effect. Who else has?
    2. Laurillard is my hero. In her book she says Gordon Pask is one of her heros (along with Ference Marton).

    The book jacket to Pask's 1975 Cybernetics book says he has over 80 papers, besides books; and has quite a lot of refs to himself. Obviously this is only a selection.

    Pask, G. (1957) "A teaching machine for radar training" Auomation progress April pp.214-217 [Earliest ref. in Pask 1975).]

    Pask, G. (1961) An approach to cybernetics (London: Hutchinson)

    Pask, G. (1965) "Comments on the organisation of men, machines, and concepts" in Education for information science (ed.) L.Heilprin et al. p.133 (London: Macmillan)

    Pask, G. (1967) "The control of learning in small subsystems of programmed educational system" IRE trans. human factors election vol.8 pp.88-

    Pask, G. (1968) "A cybernetic model for some types of learning and mentation" in Cybernetic problems in bionics (eds.) H.C.Oestreicher & D.R.Moore p.531-585 (London: Gordon and Breach)

    Pask, G. & Lewis,B.N. (1968) "The use of a null-point method to study the acquisition of simple and complex transformation skills" Brit.J.Math. statistical psychology vol.21 pp.61-83

    Pask, G. (1969) "Strategy, competence and conversation as determinants of learning" Programmed learning Oct. pp.250-

    Pask, G. (1969) "Adaptive machines" Proc. NATO symposium major trends in programmed learning research Oct. pp.251- (Paris: Dunod)

    Pask, G. (1969) "Interaction between a teaching machine and the student's attention directing systems" Proc. 16th int. congress applied psychology Oct. pp.269- (Amsterdam: Swets & Zeitlinger)

    Pask, G. (1970) "Computer assisted learning and teaching" Paper at NCET conf. on computer based learning, Leeds.

    Pask, G. (1970) "Fundamental aspects of educational technology (illustrated by the principles of conversational systems)" Proc. IFIP world conference on computer eductation, part 1 (ed.) B.Scheepmaker

    Pask, G. (1971) Behavioural cybernetics (?)

    Pask, G. (1971) "Teaching machines" in Modern trends in education (ed.) B.Rose (London: Macmillan)

    Pask, G. & Scott,B.C.E. (1971/2) "Learning and teaching strategies in a transformation skill" Brit.J.Math. statistical psychology vol.24 pp.205-229

    Pask, G. (1971) "A cybernetic experimental method and its underlying philosophy" IJMMS vol.3 pp.279-337

    Pask, G. (1972) "A fresh look at cognition and the individual" International Journal of Man-Machine studies vol.4 pp.211-216

    Pask, G. & Scott,B.C.E. (1972) "Learning strategies and individual competence" International Journal of Man-Machine studies vol.4 pp.217-253

    Pask, G. & Scott,B.C.E. (1973) "CASTE: a system for exhibiting learning strategies and regulating uncertainties" International Journal of Man-Machine studies vol.5 pp.17-52

    Pask, G. (1975) Conversation, cognition and learning: a cybernetic theory and method (New York: Elsevier) [Pask says this is the basic one]

    Pask, G. (1975) The cybernetics of human learning and performance (London: Hutchinson)

    Pask, G. (1976) Conversation theory: applications in education and epistemology (New York: Elsevier) [Opaque; follow-on from Conversation, cognition and learning]

    Pask, G. (1976a) "Conversational techniques in the study and practice of education" B.J.Educ.Psych. vol.46 pp.12-25

    Pask, G. (1976b) "Styles and strategies of learning" B.J.Educ.Psych. vol.46 pp.128-148

    Pask, G. (1983) Knowledge and innovation of decision makers: final technical report (London: US Army research institute)

    Pask, G. (1984) "Review of conversation theory and a protologic (or protolanguage) Lp" Educational communications and technology journal vol.32 no.1 pp.3-40

    Pask, G. (1988) "Learning strategies, teaching strategies, and conceptual or learning style" ch.4 pp.83-100 (New York: Plenum) in R.R.Schmeck (ed.) Learning strategies and learning styles ch.4 pp.83-100 (New York: Plenum) [Seems the best overview]

    Key references

    Pask 1976, 1988

    Other/main references

    Bahar, Mehmet (1999) Investigation of biology students' cognitive structure through word association tests, mind maps and structural communication grids (Ph.D. dissertation, Centre for Science Education, University of Glasgow). This has some useful references and descriptions about cognitive styles and how to test them.

    J.B. Carroll Human Cognitive Abilities: A survey of factor-analytic studies [Clark Quinn says: provides an alternative take than Jonassen & Grabowski]

    Coffield, F., Moseley, D., Hall, E. & Ecclestone, K. (2004) Learning styles and pedagogy in post-16 learning: A systematic and critical review (Learning and Research Skills Centre). Or try this link, or this.

    Cowards et al. (2012) "Examining the Effect of Gender and Presentation Mode on Learning from a Multimedia Presentation"

    Cronbach,L.J. & Snow,R.E. (1977) Aptitudes and instructional methods: A handbook for research on interactions (New York: Irvington Publishers Inc).

    Egan,D.E. & Gomez,L.M. (1985) "Assaying, isolating and accommodating individual differences in learning a complex skill" vol.2 ch.? pp.207 R.F.Dillon (ed.) Individual differences in cognition (Academic Press: New York)

    Egan,D.E. (1988) "Individual differences in human-computer interaction" ch.24 pp.543-568 M.Helander (ed.) Handbook of Human-Computer Interaction (North-Holland: London)

    David H. Jonassen & Barbara L. Grabowski (1993) Handbook of Individual Differences, Learning and Instruction Lawrence Erlbaum Assoc; ISBN: 0805814132) [I haven't read this, but several people recommend it.]

    Landauer,T.K. (1995) The trouble with computers: Usefulness, usability, and productivity (MIT press; Cambridge, MA)

    Martin,D.J., Lucek,L.E., & Fuentes,S. (1999) "Issues of feminism and multicultural education for educational technology" ["Published" on ITFORUM on 8 Nov 1999.]

    Pashler,H., Mark McDaniel, Doug Rohrer, and Robert Bjork (2008) Learning Styles: Concepts and Evidence Psychological Science in the public interest vol.9 no.3 pp.105-119

    Pask, G. (1976) "Styles and strategies of learning" B.J.Educ.Psych. vol.46 pp.128-148.

    Pask, G. (1988) "Learning strategies, teaching strategies, and conceptual or learning style" ch.4 pp.83-100 (New York: Plenum) in R.R.Schmeck (ed.) Learning strategies and learning styles ch.4 pp.83-100 (New York: Plenum)

    V.J.Shute (1993) "A comparison of learning environments: All that glitters ..." in S.P.Lajoie & S.J.Derry (eds.) Computers as cognitive tools pp.47-74 (Erlbaum: Hillsdale, New Jersey) PDF

    Witkin,H.A., Goodenough,D.R., Moore,C.A. & Cox,P.W. (1977) "Field-dependent and field-independent cognitive styles and their educational implications" Review of educational research vol.47 no.1 pp.1-64

    Witkin,H.A. & Goodenough,D.R. (1981) Cognitive styles: essence and origins of field dependence and field independence (New York: International university press).

    Web site logical path: [] [~steve] [this page]
    [Top of this page]