I had a conversation the other day with NAU colleague Larry Gallagher that inspired me to return to something I’d been thinking and writing about a while back: the astonishing persistence of the learning styles notion. You may work with someone like Larry: he’s the kind of colleague whom you can talk to for hours about the issues you’re both fired up about. Larry leads NAU’s Faculty Professional Development program, and is unusual in that he also has a vast professional background in K-12, where the learning styles idea is particularly tenacious. Let’s be grateful for colleagues like this! My thoughts on the current state of learning styles, and what cognitive psychology has to contribute, are in the piece below.
Years back, when I first started giving talks about learning and technology, I used to go fairly deeply into the sensory learning styles or “VAK” theory. You’ve likely heard of this notion in one form or another – the gist of it is that individuals can be neatly sorted into visual, auditory, or kinesthetic learners. It follows that people learn information best when it’s presented in their dominant sense, be that visual or auditory, or through physical manipulation in the case of the kinesthetic folks.
If it were true, this idea would be an important one for us to take into account as we create learning environments. Especially in the case of online materials, where we have to make a lot of design choices about the multimedia we are going to use, it would be critical to find out and match the styles of our learners. After all, favoring one modality to the exclusion of others would be tantamount to presenting to a bilingual class in one language only: privileging some, while putting others in constant catch-up mode. And it would be crucial to pre-assess the difference before trying to teach students anything, perhaps even then to create separate tracks for the different styles.
But the idea isn’t true, and here’s why I feel confident making that blanket statement. Even before we had any systematic studies on matching style, VAK theory seemed fishy to cognitive psychologists, based on what we know about how the mind and brain handle information from the senses. When you take in something in one sense, it doesn’t just stay in the system for processing that sense. We can and often do recode stimuli into different modalities, and this recoded form can be similar or identical to the mental representation that would have been generated had the stimulus been presented that way to begin with. Take mental visualization, the modality that we know the most about from neuroscience studies. When we visualize, we activate the same components of the brain that we use when we actually look at things “out there” in the world. This holds even when we created that mental representation some other way, as in reading a description or hearing a spoken one.
The neural areas that process information from our senses – like most parts of the brain – also swap information freely, so once something comes in to one area it bleeds over into others fairly readily. Given that information can easily start out in one spot and end up someplace completely different, it seems odd to think that learners are stuck with whatever modality the teacher is using. Couldn’t visual learners, for example, just move information over into their preferred system, rather than needing it to be visually presented in the first place? Then, there is the strange case of text processing. Is it visual? It comes in through the eyes, but because it’s language, it activates a ton of processing mechanisms linked to speech. To decode it, the visual and auditory systems do a furious amount of collaboration. Something similar happens when we manipulate objects in “kinesthetic” learning – the eyes and the hands are primed to collaborate, so processing will be spread over all kinds of brain systems to get the task done.
All this played out in the line of research that gets cited all the time as the one that tore down the theory. The work took aim specifically at the hypothesis that matching presentation modality to learning style would result in improved retention. The authors argued, and I agree, that the VAK theory clearly predicts such a matching effect – so if matching doesn’t pan out, the whole theory doesn’t hold water. Of course, the matching effect didn’t pan out in any of the series of studies they reported. In light of this, I think it’s hard for any educator to argue that matching learning styles is a good use of anyone’s time or design expertise.
Blessedly, VAK seems to be well on its way out in higher ed circles these days, and people seem less offended when I criticize the theory in talks, compared to how it was just a few years back. There are lots of articles (here, here, and here are some examples) that have explained quite eloquently what’s wrong with the theory. This is perhaps a success story for evidence-based teaching: We did the research, got the word out to practitioners, and practice changed for the better.
In the K-12 world, though, this may not be happening. I’ve heard from teachers in the field that VAK theory is still going strong, and that some school districts are still paying good (and in the case of public schools, very limited) money for seminars on how to assess and match VAK styles. This article lists VAK as one of several “neuromyths” that persist in K-12, along with others like left-brain/right brain theory and the idea that we use 10% of our brains.
Why does VAK have such staying power? One reason is probably the well-known human tendency to gravitate toward typology schemes – the same psychological dynamic that gets people to believe in horoscopes. We love being put into a categories that simultaneously explain our uniqueness and our similarities to others, and we are highly predisposed to see ourselves as fitting labels once they are bestowed.
A phenomenon called the P.T. Barnum or Forer effect is a classic demonstration of how credulous we can be when we’re handed a description, no matter how vague or completely unconnected to evidence. Like spotting faces in clouds, we pick up on patterns that aren’t there. So legions of students told that they are one type of learner, or other, are unlikely to have rejected or even questioned the labels given them. Rather, they are likely to have eagerly adopted them, perpetuating the idea that VAK offers something important and valuable.
VAK’s enduring appeal reflects our blind spots, but I think part of it is the fault of research psychologists – because we have offered little in its place. Clearly, teachers and learners want and need a framework for understanding why people are different in learning situations. A straightforward, robust heuristic for figuring out who will thrive in different subjects, different kinds of classroom, with different materials: This would be really valuable, if cognitive science could only offer it.
People’s minds do differ, but cognitive psychologists don’t have a good language for describing these differences. Personality psychologists, by contrast, picked up a long time ago on how to describe – if not in infinite nuance, at least general terms – what makes one person different from another. There are, of course, the intra-disciplinary quibbles about which framework is best, but the major ones are backed up with data and, unlike VAK, do a good job of predicting behavior and performance in the real world.
Personality psychologists have had their struggles with pseudoscience to be sure. The Myers-Briggs personality test, for example, has enormous popular appeal but shaky empirical support. This test is the one that gives you a four-letter code for your personality type; it’s easy to do, and people (including myself) tend to find their typology spot on correct and are endlessly fascinated by implications, such as famous people who share the same code. But, people’s typology doesn’t hold up over repeated tests as stably as it ought to, and there are other issues that call its value into question. The Myers-Briggs will probably never go away – indeed, there are a few holdouts in personality psychology who use it in research. But that hasn’t stood in the way of developing other schemes that are better, and that advance our understanding of human diversity.
Why can’t cognitive psychologists do something similar? We sometimes wriggle out of the question by pointing out, accurately enough, that our science focuses more on determining general principles that underlie thought, thus we are much more oriented toward research that seeks similarities among different minds, not differences. We’re a younger branch of psychology than most others, so perhaps it’s understandable that we are still at a point of discovering basic principles that are common to thought across individuals.
But it’s not true that we shy away from individual differences altogether. In fact, a number of influential research efforts use exactly this approach to test theories about cognition. By dividing up participants into groups – into high- and low- working memory, for example – we can see how those differences play out in different cognitive tasks, which in turn tells us about how different mechanisms of thought interrelate. Even some of my own work uses this tactic to illuminate memory and language processes and how those change over the lifespan.
One of my favorite lines of research on false memory did something similar. Researchers typed their subjects according to whether they were predisposed to unwittingly fabricate memories that never happened, or conversely, to forget information when it’s encountered in a new context. These typologies, done on the basis of laboratory memory tests, predicted patterns of recalling childhood sexual abuse. These findings helped reconcile some very contradictory ideas, helping to explain why recovered memory could be possible, even though false memories clearly exist. At the same time, it demonstrates that peoples’ memory systems do differ in some systematic, testable ways.
The false memory/forgotten memory paradigm is great for illuminating the recovered memory controversy, but it doesn’t say much for classroom learning. So are there other typology schemes in cognitive science that could be relevant? My perspective is that these are still fairly thin. In memory theory, there’s been the idea that individuals differ in terms of specific, distinct components of working memory (e.g., a component for syntax, a component for word meanings, another for visuospatial information and so on). Individual differences in capacity for each of these stores could, in theory, predict performance in different kinds of learning situations or subjects. In fact, tests of visuospatial capacity have been used to test aptitude for certain jobs – perhaps not the vision of personalized learning that many of us would ideally have, but it’s a start.
Phonological working memory capacity, as well, has some important predictive capabilities that perhaps could be better exploited for teaching and learning. PWM has a fascinating history as a subject of cognitive research; although it was the subject of hundreds of research projects, including the famous word length effect that linked memory to verbal rehearsal, its actual purpose within cognition was misunderstood for years. It turned out to (probably) be a mechanism for creating mental representations of new words, essentially a way for memory to buy time while setting up a long-term memory linking the new sequence of sounds with a meaning.
Phonological working memory is one of those memory capacities that differ significantly across individuals, and there is some research linking a larger PWM capacity to the ability to excel at foreign language learning. So this is one case where we could test and sort individuals to predict success in different learning tasks, something I talked about in this short article about helping students develop strategies for memorization. Perhaps researchers could tackle some other ways to harness the multiple capacities idea to steer students into the subjects and learning strategies that will work best for them.
There are some other evidence-based, theoretically sophisticated learning styles schemes out there, notably in work by Robert Sternberg and collaborators. These have to do not so much with memory or varied capacities as the way in which people prefer to weigh and analyze information on a more global level. I have to confess though, that I find these fairly opaque myself, and perhaps this is why they’ve failed to catch on the way VAK did.
And in that way, perhaps VAK is just a version of the same problem scholars run into time and again, the pull of catchy-but-wrong ideas against the nebulous, unsatisfying ones that are closer to the truth. I know I grapple with this one in my work translating cognitive science for audiences of teachers and ed tech professionals, and I am sure I haven’t always gotten it right. It gets exhausting to to pack huge, messy concepts into a slide or two, while preserving some semblance of nuance, but it is something – as many have pointed out – that we in the academy have got to do if we want our work to make an impact outside our tiny specialized worlds.
It’s time for the last vestiges of VAK to go, especially in K-12 where the needs are so pressing. But we can acknowledge that VAK-mania grew out of two good impulses, one, the wish to acknowledge that students are individuals and to use their individual capabilities to the fullest. It also shows how open educators across disciplines have become to incorporating cognitive science into their work, something that has not always been the case.
It could take a while to develop a framework that really does describe cognitive individuality in a useful way, and – crucially – to promote its use among front-line practitioners. And like much applied work, it will probably get a little or no praise from inside the disciplinary mainstream. But I hope that someone will take up this task in a serious way, because VAK is going to leave a vacuum that will be filled either with the next gimmick, or something of substance and quality.