The science
Your brain has a way it likes to learn. We teach to it.
Zavmo is not built on learning styles. It is built on a set of findings about how a human brain takes something in and keeps it, and on one older finding about how much better teaching gets when it is tuned to one person at a time.
This page sets out both. Every claim below names a paper, every paper is listed at the bottom with a link, and where we are describing a design decision rather than a research finding, it says so.
First, the claim we do not build on
You have probably been told you are a visual learner, or an auditory one. The first half of that idea is real enough: people do report a preferred channel. The second half, that teaching someone in their preferred channel improves how much they learn, is the part that matters and the part the evidence does not support.
Pashler, McDaniel, Rohrer and Bjork reviewed the literature for Psychological Science in the Public Interest. Testing the claim properly needs a particular experimental design, one that can show an interaction between a person’s style and the method that teaches them best. They found almost no studies that used it, and among the few that did, several produced results that contradicted the claim outright. Their conclusion was that there is no adequate evidence base for putting learning-styles assessment into general educational practice. Willingham, Hughes and Dobolyi reached the same place seven years later.
So there is no visual, auditory or kinaesthetic type anywhere in this product, and no promise that we will teach to one.
Second, the finding we are chasing
Bloom’s 1984 paper is the reason any of this is worth building. He compared students taught conventionally, in classes of about thirty, against students taught one to one with mastery learning, across replicated studies run by two doctoral researchers.
The result is the one the paper is named for. Using the standard deviation of the conventional class, the average tutored student finished about two standard deviations above the average of that class, which put the average tutored student above 98 per cent of the students in it. The spread changed as well as the average: about 90 per cent of the tutored students reached a level of achievement that only the highest 20 per cent reached under conventional instruction.
Bloom called the gap the two sigma problem, and he set the challenge as finding methods of group instruction that could get near one-to-one tutoring. That challenge is the reason Zavmo exists. The dimensions below are how we go at it: the things a tutor adjusts for a person, made adjustable at scale.
The eight dimensions we teach to
Working memory and cognitive load
How much you can hold at once, and it is smaller than you would think
The research: Cowan, N. (2001); Sweller, J. (1988); Unsworth, N., and Engle, R. W. (2007)The claim
Working memory holds a handful of things at a time, that handful is a different size in different people, and going past it does not slow understanding down so much as stop it.
What the research says
Cowan’s review is the standard reference for the size of the limit. Against the older folklore of seven items, he brought together evidence from a wide range of tasks for a much smaller capacity, and put it at three to five chunks. The paper is unusually careful about the conditions under which the limit can be seen at all, which is a large part of why it has held up: the number only means something when the material cannot be quietly recoded into bigger chunks or rehearsed into long-term memory.
Sweller’s cognitive load theory is the instructional half of the same idea. When material is presented so that a learner has to hold several interacting pieces at once, that holding consumes the capacity that would otherwise go into building the structure underneath. The important move in the theory is where it puts the blame: load is a property of the design, not of the person.
Unsworth and Engle are why we treat capacity as an individual difference rather than a constant. Their examination of simple and complex span measures and how they relate to higher order abilities sits in a large literature showing that working memory capacity varies substantially between people. Two people who would both describe themselves as perfectly ordinary can be working with very different room to think.
Why it matters for you
Most teaching ignores this and moves everyone at the same pace. If that pace sits above your capacity, you are not bad at the subject. You are overloaded, and trying harder does not create more room.
How Zavmo uses it We meter the load. Ideas arrive a few at a time, in an order that lets each one settle before the next one lands, and the pace follows what you are actually holding rather than a fixed syllabus.
Spacing and retrieval practice
Forgetting a little, then pulling it back, is what makes it stick
The research: Roediger, H. L., and Karpicke, J. D. (2006); Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., and Pashler, H. (2008); Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., and Willingham, D. T. (2013); Soderstrom, N. C., and Bjork, R. A. (2015)The claim
Two things reliably improve how much you keep: spreading practice across days rather than massing it, and retrieving an answer from your own head rather than rereading it.
What the research says
Roediger and Karpicke ran the comparison directly. Students who took a test on material retained more of it over the long term than students who spent the same time restudying it, even though restudying produced better performance in the short run and felt more productive at the time. Testing is not only how you check learning. It is one of the ways you cause it.
Cepeda, Vul, Rohrer, Wixted and Pashler mapped the spacing side. The useful part of their work is that there is no single correct gap: the interval that maximises retention depends on how long the material has to survive, which is why a review schedule that is not tied to a horizon is guessing.
Dunlosky, Rawson, Marsh, Nathan and Willingham then rated ten study techniques for utility. Practice testing and distributed practice came out high. Elaborative interrogation, self-explanation and interleaved practice came out moderate. Summarisation, highlighting, rereading, the keyword mnemonic and imagery use for text came out low. The awkward part of that list is that the low-utility end contains the two things most people actually do.
Soderstrom and Bjork explain why the awkward part persists. Performance during a study session and learning that survives the session are different quantities, and the techniques that make a session feel fluent are often the ones that leave least behind.
Why it matters for you
It is why cramming feels productive on the night and has evaporated by the following week. A little, often, with some effortful recall in it, beats one heroic session, even though the heroic session feels considerably better while you are in it.
How Zavmo uses it We space your practice and bring ideas back around the point you would otherwise start losing them, and we ask you to retrieve rather than reread. It is the most under-used thing in education and it is built into how the product paces you.
Difficulty calibration
Difficulty is not the enemy, and easy is not the same as effective
The research: Soderstrom, N. C., and Bjork, R. A. (2015); Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., and Willingham, D. T. (2013)The claim
Conditions that make learning feel harder while you are in it often improve what you retain afterwards, which makes how well a session went a poor measure of how much you learned.
What the research says
Soderstrom and Bjork’s review is built around a distinction that sounds pedantic and is not: performance is what you can do during instruction, and learning is what remains once instruction has stopped. They can move in opposite directions. A change that makes practice go smoothly can leave you with less, and a change that makes it feel laboured can leave you with more. These are the difficulties Bjork called desirable.
The technique ratings back this up from another angle. Interleaving different kinds of problem, rather than blocking them so you already know which method applies, feels worse during practice and tends to do better on a later test, which is why Dunlosky and colleagues rate it as useful while noting that learners rarely choose it.
None of this licenses simply making things harder. The claim is narrower and more useful than that: the difficulty has to be pitched against the person, and the fact that it feels hard is not on its own evidence that it is wrong.
Why it matters for you
One pace cannot be right for everyone in a room. Sit below your level and you disengage; sit far above it and you stall. Repeated over years, that mismatch is how capable people quietly decide a subject is not for them.
How Zavmo uses it We move the difficulty with you, up as you get stronger and down when you are genuinely stuck, and we do not treat a session that went smoothly as proof that it worked.
Attention and absorption
Deep attention is a state that design invites, not a virtue you have or lack
The research: Engeser, S., and Rheinberg, F. (2008); Pekrun, R. (2006)The claim
The absorbed state where the work almost carries itself shows up when a real challenge meets real skill, and a learning design can make it more or less likely.
What the research says
Engeser and Rheinberg studied that state in real learning and performance settings and related it to how people actually did. Their paper is titled for the moderators, and that is the honest reading of it: the balance between challenge and skill matters, but how much it matters is conditioned by other things about the person and the situation rather than holding as a universal law.
Pekrun’s control-value theory supplies the other half. The emotions that show up in learning, including boredom at one end and anxiety at the other, follow from how much control a learner feels over the activity and how much they value it. Those emotions then shape where attention goes. Boredom and anxiety are not failures of character. They are downstream of a task that has been pitched wrongly or framed as not mattering.
Why it matters for you
Attention is not a moral quality you are failing at. Interruptions, a cluttered interface and badly pitched difficulty are what break it, and they are all things somebody chose.
How Zavmo uses it We design for the state rather than demanding it: the right challenge for where you are, friction stripped out of the path, and a rhythm you can drop into instead of reorienting to every time. The system tries to become invisible.
Autonomy and self-direction
You go further when the choices are genuinely yours
The research: Deci, E. L., and Ryan, R. M. (2000); Zimmerman, B. J. (2002)The claim
The sense that what you are doing came from you, rather than being imposed on you, is one of the better evidenced drivers of motivation and of learning that lasts.
What the research says
Deci and Ryan’s self-determination theory puts autonomy alongside competence and relatedness as a basic psychological need, and draws the line that matters here: between doing something because you chose it and doing the same thing because you were made to. The two produce different quality of engagement and different persistence, even when the activity is identical.
Zimmerman’s account of self-regulated learning describes what that looks like in practice. Learners who set their own goals, choose their strategies and monitor their own progress are doing something the learner in a fixed sequence never gets the chance to do, and the difference shows up in how far they get.
Why it matters for you
Being marched through a sequence someone else designed drains the motivation that learning runs on, even when the content itself is perfectly good. Ownership is not a nice-to-have here. It is part of the mechanism.
How Zavmo uses it We hand you the wheel: real say over direction, pace and what you go deep on, with the structure kept sound underneath so the freedom does not cost you coherence.
Learning by teaching
Explaining it to someone else is how you find out whether you understand it
The research: Fiorella, L., and Mayer, R. E. (2013); Chase, C. C., Chin, D. B., Oppezzo, M. A., and Schwartz, D. L. (2009)The claim
Producing an explanation for another person improves your own grasp of the material, and the benefit comes from doing the explaining rather than from expecting to.
What the research says
Fiorella and Mayer set the two apart deliberately. There is a benefit to preparing while expecting to teach, and there is a benefit to actually teaching, and their paper is about the relative size of the two. That distinction is the whole reason we build in moments where you have to say the thing out loud rather than merely promising you will be asked to later.
Chase, Chin, Oppezzo and Schwartz found the effect from a different direction, with teachable agents. Learners working on behalf of a character they were responsible for teaching put in more effort than learners working for themselves, which is the pattern usually called the protege effect.
Why it matters for you
If you have ever only really understood something at the moment you had to explain it to somebody, that is not a quirk of yours. It is the effect. Learning entirely alone leaves that lever unpulled.
How Zavmo uses it We build in moments to explain, teach back and think out loud, with the companion and, where it fits, with other people, so you are not only receiving ideas.
Emotion and safety
Very little is learned in a state of fear
The research: Immordino-Yang, M. H., and Damasio, A. (2007); Pekrun, R. (2006)The claim
Emotion is part of the machinery of thinking rather than interference around it, and the wrong emotional conditions block the processes learning depends on.
What the research says
Immordino-Yang and Damasio make the argument from neuroscience, and the title of the paper is the argument: we feel, therefore we learn. The aspects of cognition most useful in education, including decision making and meaningful memory, are not separable from emotion in the way the folk model assumes.
Pekrun’s work gives the mechanism a shape you can design against. Anxiety and shame narrow attention onto the threat and away from the material, and they arise predictably when a learner feels no control over an activity that matters. Psychological safety is not a soft add-on to this. It is a precondition.
Why it matters for you
It is why being made to feel stupid is such efficient anti-teaching, and why the learners who feel safest are often the fastest.
How Zavmo uses it We keep it safe to be wrong. The companion is encouraging without flattery, a mistake is treated as information rather than as a verdict, and there is no audience to perform for, so your attention can go to the work instead of to protecting yourself.
Words with pictures
It lands better when the words arrive with something to picture
The research: Clark, J. M., and Paivio, A. (1991); Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., and Willingham, D. T. (2013)The claim
Pairing a verbal explanation with a well-chosen image or worked example gives memory two routes in instead of one. That is a fact about instruction, not a fact about you.
What the research says
Clark and Paivio set out dual coding theory for education: a verbal system and a nonverbal, imagery-based system that can each carry part of the same idea, with additive effects when both are engaged. It is one of the more durable findings in the field, and it is routinely misread.
The misreading is worth naming, because it is the learning-styles claim wearing a lab coat. Dual coding says that words and images together help people. It does not say that some people are picture people and should be given the picture instead of the words. Those are different claims and only the first one has the evidence.
It is also not the same as telling yourself to form mental images while you read. Dunlosky and colleagues rated imagery use for text as low utility as a study technique, which is a separate question from whether well-designed instructional graphics alongside an explanation help, and worth keeping separate.
Why it matters for you
It is why one vivid worked example can do more than pages of definition, and why understanding something in theory but not in practice usually means it was never made concrete in the first place.
How Zavmo uses it We pair explanation with visuals and worked examples, and we anchor ideas in your actual work rather than in the abstract, so the same idea arrives through more than one door. That last part is a design decision about relevance, not a finding we are attributing to the papers above.
What this is not
- This is not a study of Zavmo. Every finding on this page comes from someone else’s published work, cited below, and none of it was collected on Zavmo learners.
- None of these papers tested this product. They are the principles it is designed around. They do not establish that our implementation of them works, and we have no published effect size of our own to put next to them.
- Bloom’s two sigma figures describe one-to-one human tutoring with mastery learning in the studies he reviewed. That is the target we are aiming at. It is not a result we are claiming to have reproduced.
- This is not a learning-styles model and there is no visual, auditory or kinaesthetic type anywhere in it, because the evidence does not support teaching to one.
- Eight dimensions is a working set rather than a complete account of human learning, and several of the boundaries between them are ours rather than the literature’s.
- No figure appears on this page unless it was read in the source it is attributed to. Where a paper is used as a frame rather than as data, the note under the reference says so.
References
Every reference below was checked field by field against the publisher record for its identifier before it was written here, and every one links out so you can read it yourself. Where a paper is used as a frame rather than as data, the note under it says so.
-
(1984). The 2 sigma problem: the search for methods of group instruction as effective as one-to-one tutoring . Educational Researcher, 13(6), 4 to 16.
Both figures quoted on this page, read from the paper itself. Comparing conventional classes of about thirty students against one-to-one tutoring with mastery learning, the average tutored student finished about two standard deviations above the average of the control class, which put the average tutored student above 98 per cent of the students in that class. About 90 per cent of tutored students reached the level of achievement reached by only the highest 20 per cent under conventional instruction.
-
(2008). Learning styles: concepts and evidence . Psychological Science in the Public Interest, 9(3), 105 to 119.
The reason this page opens by refusing the learning-styles claim. The review found virtually no studies using the design needed to test it, and concluded there is no adequate evidence base for putting learning-styles assessment into general educational practice.
-
(2015). The scientific status of learning styles theories . Teaching of Psychology, 42(3), 266 to 271.
The later review reaching the same conclusion, cited so the position does not rest on a single paper.
-
(2001). The magical number 4 in short-term memory: a reconsideration of mental storage capacity . Behavioral and Brain Sciences, 24(1), 87 to 114.
The working memory capacity limit behind the first dimension. The review argues the real limit is three to five chunks rather than the seven of the older folklore, and it is careful about the conditions under which the limit can be observed at all.
-
(1988). Cognitive load during problem solving: effects on learning . Cognitive Science, 12(2), 257 to 285.
Cognitive load theory: the instructional half of the same dimension. Cited for the principle that load is a property of how material is presented, not a property of the learner.
-
(2007). On the division of short-term and working memory: an examination of simple and complex span and their relation to higher order abilities . Psychological Bulletin, 133(6), 1038 to 1066.
Why we treat working memory capacity as an individual difference to design around rather than a constant to assume. Frame, not data: no figure from this paper appears on this page.
-
(2006). Test-enhanced learning: taking memory tests improves long-term retention . Psychological Science, 17(3), 249 to 255.
The testing effect: taking a test on material beats restudying it for long-term retention. Behind the retrieval half of the second dimension.
-
(2008). Spacing effects in learning: a temporal ridgeline of optimal retention . Psychological Science, 19(11), 1095 to 1102.
How the best gap between study sessions depends on how long the material needs to be retained. Behind the spacing half of the second dimension, and the reason our review intervals are not a fixed number of days.
-
(2013). Improving students’ learning with effective learning techniques . Psychological Science in the Public Interest, 14(1), 4 to 58.
The utility ratings quoted on this page. Practice testing and distributed practice were rated high utility; elaborative interrogation, self-explanation and interleaved practice moderate; summarisation, highlighting, the keyword mnemonic, imagery use for text and rereading low.
-
(2015). Learning versus performance: an integrative review . Perspectives on Psychological Science, 10(2), 176 to 199.
Desirable difficulties, and the distinction between performance during study and learning that lasts. It is the source for the claim that a session going smoothly is a poor guide to how much of it you will keep.
-
(2008). Flow, performance and moderators of challenge-skill balance . Motivation and Emotion, 32(3), 158 to 172.
Challenge and skill balance, and the fact that its relationship with the absorbed state is moderated rather than universal. It replaced a citation to a popular book that had no resolvable record behind it.
-
(2006). The control-value theory of achievement emotions: assumptions, corollaries, and implications for educational research and practice . Educational Psychology Review, 18(4), 315 to 341.
Achievement emotions and their control and value antecedents: how much control a learner feels and how much they value the task shape the emotions that follow, and those emotions shape attention.
-
(2000). The "what" and "why" of goal pursuits: human needs and the self-determination of behavior . Psychological Inquiry, 11(4), 227 to 268.
Self-determination theory, and specifically the autonomy need: the reason a learner having real say over direction is treated here as a driver of learning rather than a courtesy.
-
(2002). Becoming a self-regulated learner: an overview . Theory Into Practice, 41(2), 64 to 70.
Self-regulated learning: goal setting, strategy choice and self-monitoring as things a learner does rather than things done to them. The second half of the autonomy dimension.
-
(2013). The relative benefits of learning by teaching and teaching expectancy . Contemporary Educational Psychology, 38(4), 281 to 288.
The distinction this page rests the teaching dimension on, between the benefit of actually teaching and the benefit of merely expecting to teach.
-
(2009). Teachable agents and the protégé effect: increasing the effort towards learning . Journal of Science Education and Technology, 18(4), 334 to 352.
The protege effect: the effort learners put in on behalf of someone they are teaching. The second half of the same dimension.
-
(2007). We feel, therefore we learn: the relevance of affective and social neuroscience to education . Mind, Brain, and Education, 1(1), 3 to 10.
Why emotion is treated here as part of the machinery of thinking rather than noise around it, and therefore as part of the learning design rather than a soft extra.
-
(1991). Dual coding theory and education . Educational Psychology Review, 3(3), 149 to 210.
Dual coding, and the reason the last dimension exists. Cited for what it actually says, which is that words and images together help everyone, and not as support for a preferred channel. It replaced a citation to an experiential learning model with no resolvable record behind it.
This is what we mean by neurogogy
Teaching has a lineage of names for who is in front of it. Pedagogy is the teaching of children. Andragogy names the teaching of adults, who turn up with experience behind them and a reason of their own for being there. Heutagogy names learning that is genuinely self-determined, where the learner sets the direction rather than following one that was set for them. Neurogogy is our founder’s term for the next question in that line: not who the learner is, but what the particular brain in front of us is actually like.
That is a lineage of ideas and we cite it as one. Pedagogy, andragogy and heutagogy are not the evidence for anything on this page, and none of them appears in the reference list. The evidence is the eight dimensions above and the papers underneath them.
Which leaves the plain version. Not teaching to an imaginary average brain. Teaching to the real one in front of us, on the dimensions the research actually supports. The first step is knowing yours.