Learning Styles, Re-examined

If you are involved in any remote way with education, you've almost certainly heard of the VARK questionnaire. It stands for Visual, Aural, Read/write and Kinesthetic, four styles of learning. Some of us are visual learners; some are aural learners, preferring to hear a lecture; some learn best by reading; and others are kinesthetic, or hands-on learners. The sixteen questions in the VARK system will help you nail down which of these four you are. This allows teachers to tailor the curriculum to your particular learning style, and then — the theory goes — you'll learn more effectively. Sounds like a wonderful deal, right? The only problem is it doesn't work.

The promise of learning styles is that there is an easy way to make education more effective: to break students down into groups, each of which has its own learning style; and then to teach to each group in that style which will work best. One group may be visual learners, so we'll show them videos. One group may be auditory learners, so we'll give them lectures. Another group gets the hands-on lessons. And then, at the end of the day, each student will have learned the same lesson much more effectively than they could have with a one-size-fits-all approach. Same effort, better outcome. And as a side benefit, each student got to spend the day doing what they enjoy best, so it works out great for everyone.

With an idea having so much promise as that, we've (somewhat predictably) seen an enormous amount of research go into it, hoping to find that perfect model. And by "enormous amount" I mean that the four styles proposed in VARK are just one of at least 71 different theories that have been proposed — that number coming from the most often cited survey of the subject, a 2004 paper by Coffield et al. titled Learning styles and pedagogy in post-16 learning: A systematic and critical review. Coffield and his co-authors found that most of these theorized models were dichotomies; like are you "this" type of learner or "that" type? As a brief sample, these include:

• convergers vs. divergers
• holists vs. serialists
• activists vs. reflectors
• assimilators vs. accommodators
• globalists vs. analysts
• initiators vs. reasoners

...and on and on. Some of those sound pretty odd, and they are; by and large, nearly all learning styles systems in use rely on three of the four VARK styles: visual, auditory, or kinesthetic. And it turns out that a solid majority of educators believe that learning styles are real, and that using them will improve educational outcomes. A 2017 paper in Frontiers in Psychology summarized:

A 2012 study demonstrated that 93% of schoolteachers in the UK agree with the statement "Individuals learn better when they receive information in their preferred Learning Style (e.g., auditory, visual, kinaesthetic)" ...A 2014 survey reported that 76% of UK schoolteachers 'used Learning Styles' and most stated that to do so benefited their pupils in some way... A study of Higher Education faculty in the USA showed that 64% agreed with the statement "Does teaching to a student's learning style enhance learning?". A recent study demonstrated that current research papers 'about' Learning Styles, in the higher education research literature, overwhelmingly endorsed their use...

And as you might expect, the market has responded. All kinds of educational materials and products are sold to teachers, to students, to parents, and even to corporate workplaces. Nearly all publishers and other companies devoted to education technology and resources promote learning styles as if they are a given, and their product catalogs reflect that accordingly.

The problem is that no sound evidence suggests that learning styles work. For decades, major research projects have looked into this question and come up empty handed. Perhaps the most often cited is by Pashler et al. in Psychological Science in the Public Interest from 2008:

We conclude... that at present, there is no adequate evidence base to justify incorporating learning-styles assessments into general educational practice.

The paper in Frontiers in Psychology stated this rather bluntly:

The empirical evidence is clear that there is currently no evidence to support the use of Learning Styles.

But these are just the tip of the iceberg. Any reasonable review of just a small percentage of the academic work on learning styles gives you the same answer: there's no evidence that they work, despite widespread belief in them among educators. One paper even quoted our dear friend, the late James Randi: "No amount of belief makes something a fact."

So let's have a look at why they don't work.

First of all, if you're a long-time Skeptoid listener, this idea of dividing people into two groups based on dichotomies may sound familiar. Remember episode #221 on the Myers-Briggs Personality Test? This personality indicator separates people into 16 groups, each based on their classification into each of four dichotomies. What are those dichotomies? Well, no big surprise, you will also find them to be four of the 71 learning style theories identified by Coffield's team:

• extroverts vs. introverts
• sensing vs. intuition
• thinking vs. feeling
• judging vs. perceiving

This segues us right into the first problem that critics of learning styles often raise, and this is that people don't fit neatly into one of two groups, or even one of three or four groups. While it's true that most of us can probably answer a simple question like whether we prefer to learn by watching a video or by reading a textbook, that answer might change day to day based on how we feel or what we're in the mood for; it might change based on the subject matter; and the answer might be no preference or only a slight preference. You can ask people a question and if you only offer two answer checkboxes, well, you'll get two groups; but when we ask much finer graded questions you'll find people's answers cover the full spectrum. Forcing everyone into just a few compartments may yield a small number of compartments, but it does not accurately represent any individual person's actual place on a spectrum.

The second problem is also one shared with Myers-Briggs: a preference for something is not the same thing as an aptitude for it. This is a stark problem with the VARK questionnaire. Every one of the VARK questions asks what you would like, what you would prefer; do you want the doctor to show you a plastic model of your heart, to give you a pamphlet to read, to explain your heart, or to show you a diagram? This is fundamental, and it's a crucial fact. A preference is not an aptitude. You may love plastic models, but that absolutely does not mean you're going to remember its structure better than if you heard a well worded and well organized lecture on how and why the heart works.

A third problem with learning styles is that the vast majority of the models proposed are junk science. What I mean by that is that they fail to satisfy basic criteria that qualifies them as scientifically valid. One 2016 paper titled "Stop propagating the learning styles myth" is among the most widely cited of all this vast library criticizing learning styles. The author, Paul Kirschner, took the 13 most common models analyzed by the Coffield team and matched each with the four most basic criteria: internal consistency, test-retest reliability, construct validity, and predictive validity. We'd like to see all the papers satisfy all four of these criteria, which any decent scientific theory would. However, of these 13 most common learning styles models, the average number of criteria satisfied was 1.5. Three of them met none at all, and only one met all four — and that one was about cognitive styles, not really learning styles. Kirschner made four conclusions:

1. The premise that there are learners with different learning styles and that they should receive instruction using different instructional methods that match those styles is not a 'proven' fact, but rather a belief which is backed up by precious little, if any, scientific evidence.
   
   
2. There are a lot of very fundamental problems regarding measuring learning styles.
   
   
3. The theoretical basis for the assumed interactions between learning styles and instructional methods is very thin.
   
   
4. Significant empirical evidence for the learning-styles hypothesis is almost non-existent.

When we wonder why teachers — whom we presume to be thoroughly knowledgable when it comes to pedagogy — would buy into a theory so thoroughly beaten down by academia, we find a familiar old friend at the root of it: Confirmation bias. From a 2010 article in Change: The Magazine of Higher Learning:

...Learning-styles theory has succeeded in becoming "common knowledge." Its widespread acceptance serves as an unfortunately compelling reason to believe it. This is accompanied by a well-known cognitive phenomenon called confirmation bias. When evaluating our own beliefs, we tend to seek out information that confirms our beliefs and ignore contrary information, even when we encounter it repeatedly.

I'd like to sum up this episode by offering my thoughts on the question of whether Learning Styles is a pseudoscience. My answer is that it's usually not, but sometimes is — and those cases that I'm thinking of are where companies try to sell some Learning Styles based product or method or service to either parents, students, or teachers. These products always list the research they claim supports the effectiveness of their product; and by and large, those are deliberately misrepresented in order to promote the product. This misuse of research toward a for-profit goal is definitely pseudoscience.

But for the rest of it — the massive body of research constituting the 71 models found by Coffield et al. — it's more fair to characterize them as genuine, well-intentioned scientific research directions that led to dead ends. Such dead ends are not only a fundamental of the scientific method, they represent the vast majority of all scientific experiments! Ask any molecular biologist how many of the countless petrie dishes she's taken out of the incubator in the middle of the night that turned out to be the cure for cancer. Precious few of them; probably none. In science, a million ideas lead nowhere for every one that ends up working. Half a century of beating the bushes looking for a way to make education better was not a pseudoscientific pursuit, despite the great big zero that it's given us. It's pretty clear by now that there's nothing to it. This doesn't mean that tomorrow someone won't overturn everything and find the miraculously easy solution to education, but it does mean that's a lot less likely now than it was 50 years ago.

So as we see with so many other offerings life throws our way, magically easy solutions to complicated problems are much scarcer than we wish. The promise of improving a one-size-fits-all teaching method by splitting it into two, three, or even four variants may sound like a great idea and was certainly worth the experiment, but it's been tried and it hasn't worked. Whenever you see a product or service offered with a pitch about learning styles, pass.

By Brian Dunning
Follow @BrianDunning