A REFERENCE FOR MAKING LEARNING MATERIALS

Principles for designing learning materials

Sixty of them, drawn from five fields — and an honest account of what each one is worth.

Five chapters, sixty numbered principles, drawn from the learning sciences, multimedia research, visual design, user experience and inclusive design. Each principle states what it claims, why it works, and where the claim comes from.

Every principle carries an evidence mark. Twenty-six are replicated findings. Twenty-three rest on a primary source narrower than the claim usually made of it. Seven are craft conventions with no empirical test. Four are contested: the current evidence does not support the way they are normally taught, or no primary source could be found at all.

WHY THIS EXISTS

A shared body of knowledge, and the words to argue in

Doctors have pharmacology. Teachers have pedagogy. Designers have a corpus of principles and the vocabulary to disagree in.

FIG 01 — FOUR BODIES OF KNOWLEDGE, ONE JOB

Making learning materials draws on at least four bodies of knowledge at once: pedagogy, learning science, visual design and experience design. Nobody arrives holding all four, and the work suffers less from ignorance than from having no shared terms in which to notice it.

Speaking a common language raises the quality of the conversation. It also makes the complexity of the work legible to people outside it, which is the difference between being consulted and being briefed.

WHAT IS ACTUALLY IN HERE

Not every principle here is the same kind of claim

Checking the sources behind these sixty principles found four with no evidence for the advice as written — and two of those with no primary source at all.

FIG 02 — ELEVEN OF SIXTY ARE NOT FINDINGS

These sixty are almost always presented as principles of equal standing. They are not. The Gutenberg diagram's attribution traces to hedged repetition rather than to any document; the Z-pattern has no eye-tracking study behind it; the F-pattern guidance points the opposite way from the research group that produced the F-pattern; and Miller's law, as it is used in interface design, is a documented misreading of Miller.

None of that makes them useless. A convention that has organised layouts for seventy years has earned some deference. It does mean a reference should say which is which.

HOW TO READ THE MARK ON EACH SLIDE

Four marks

REPLICATED Several peer-reviewed studies converge and the direction of the effect is stable
SUPPORTED A primary source exists, but narrower in scope than the claim made of it
CONVENTION A craft heuristic with documented provenance and no empirical test
CONTESTED The evidence does not support the advice as written, or no source could be found

REPLICATEDTwenty-six of sixty Multiple independent studies, consistent direction of effect, and in most cases a review or meta-analysis. The multimedia principles and the Gestalt grouping laws make up most of this group.

SUPPORTEDTwenty-three of sixty A real primary source exists and says something real, but less than the principle is usually asked to carry. Nielsen’s heuristics sit here: they were derived by factor analysis from real usability problems, which makes them a good instrument and not a set of findings.

CONVENTIONSeven of sixty Provenance is documented, evidence is absent. The rule of thirds was coined by John Thomas Smith in 1797 and explicitly offered as a rule of thumb. Colour harmony comes from painters. These are worth following; they are not worth citing as research.

CONTESTEDFour of sixty Either the current evidence contradicts the advice, or no primary source exists. Each of the four says so on its own slide, in its own panel, with what is known in its place.

The mark sits in the eyebrow of every principle, with the chapter letter and number. Where a mark is anything other than replicated, the panel says why in one line.

THE CLAIM

A shared language has to include the words for what a claim is worth

Otherwise a rumour repeated for seventy years and a finding replicated for seventy years sound exactly alike.

This is the one addition to the original material, and it is the reason the rest can be trusted. A profession that cannot mark the difference between a finding, a convention and a rumour has a vocabulary but not a body of knowledge.

Section A

How people learn

Nine strategies for sequencing a lesson, taken from the learning sciences.

A How people learn
B How media teaches
C How pages are seen
D How products are used
E Who the design leaves out

Nine strategies for sequencing what a learner does. They apply at the lesson design stage, before any screen exists.

DESIGN FOR EFFECTIVE LEARNING

Section A

Design for effective learning

Nine strategies for sequencing what a learner does, and when.

These principles belong at the lesson design stage, where the question is what order the activities go in. Read together they rule out one common shape: explain the idea, then set practice on it. Almost every principle here asks for something to happen before the explanation.

They were adapted by Manu Kapur from work that set out to codify instructional strategies and test which of them hold. Teaching remains a craft, but a craft with findings under it.

REPLICATED · Koedinger, Booth & Klahr, 2013; Koedinger, Corbett & Perfetti, 2012

A · 01 · REPLICATED

Activation

People learn better when relevant prior knowledge has been activated in ways that help them notice and encode novel material.

FIG 03 — SAME FIVE FACTS, TWO DIFFERENT PROSPECTS

Activating what a learner already knows gives the new material somewhere to attach. It also exposes the gaps, which is the more useful half: a learner who has just discovered they cannot predict what happens next is paying a different kind of attention.

Cognitive dissonance, comparative evaluation, prediction tasks, inducting a pattern from examples, productive failure and pretesting all do this job. So does any question the learner gets wrong before being taught the answer.

REPLICATED · Koedinger, Corbett & Perfetti, 2012

A · 02 · SUPPORTED

Expectations

People learn better when they expect that they will be tested after learning.

FIG 04 — ATTENTION HOLDS WHEN A DEMONSTRATION IS COMING

Knowing how information will be used changes how closely it is attended to. Self-determination theory offers the mechanism: intrinsic motivation is sustained by autonomy and by competence, and a test is an occasion to demonstrate competence rather than only to be measured.

The mark is supported rather than replicated because test expectancy on its own is a smaller and less consistent effect than the surrounding motivation literature it is usually bundled with.

SUPPORTED · Ryan & Deci, 2000; Niemiec & Ryan, 2009

A · 03 · REPLICATED

Comparison

People learn better when they are given a chance to compare and contrast multiple instances of a rule, law or principle.

FIG 05 — THE RULE IS WHAT THE CASES HAVE IN COMMON

A single worked example teaches the example. Several teach the rule, because the learner has to work out what the instances share, and that work is where the organising happens.

Contrasting cases are stronger than similar ones. The pairs that teach most are the ones that look alike and behave differently, which is also how expert discrimination is trained in every field that has studied it.

REPLICATED · Koedinger, Corbett & Perfetti, 2012

A · 04 · REPLICATED

Metacognition

People learn better when they are given a chance to reflect on their learning.

FIG 06 — TWO LOOPS THAT ARE ROUTINELY CONFUSED

Four things build it: frequent occasions to self-assess what is and is not known, help articulating one’s own thinking, explicit modelling of what reflection looks like, and a shared understanding of what the reflection is for.

Hold on to the distinction between learning to reflect and reflecting to learn. The first is a skill being taught; the second is a method for teaching something else. A prompt that does not say which it is usually achieves neither.

REPLICATED · Koedinger, Corbett & Perfetti, 2012

A · 05 · REPLICATED

Feedback

People learn better when feedback arrives during learning, and better still when it is immediate rather than delayed.

FIG 07 — AIMED AT THE SELF, FEEDBACK MOVES LEAST

Feedback has to be able to move a learner from where they are to where they are going, which means it has to know both. Hattie and Timperley separate four levels it can be aimed at — the task, the process, self-regulation and the self — and they do not work alike. Praise directed at the self is the weakest of the four.

In a resource this mostly becomes a question of latency. Feedback a learner reads while the attempt is still in mind does work that the same words cannot do a day later.

REPLICATED · Hattie & Timperley, 2007

A · 06 · REPLICATED

Concreteness fading

People learn better when learning is situated in multiple concrete examples and then faded over time towards abstraction and formalism.

FIG 08 — FADE THE SUPPORT, NOT THE CONCEPT

Bruner proposed three progressive forms: enactive, a physical model; iconic, a picture; symbolic, the notation. The sequence matters more than any one stage — starting at the symbol strands the learner, and staying at the object strands the idea.

A systematic review found the faded sequence outperforms concrete-only and abstract-only presentation across mathematics and science. The fading is the active ingredient, not the concreteness.

REPLICATED · Bruner, 1966; Fyfe, McNeil, Son & Goldstone, 2014

A · 07 · REPLICATED

Variability

People develop better fluency when they apply what they have learnt to varied instances.

FIG 09 — THE SCHEDULE THAT LOOKS WORSE IS THE ONE THAT LASTS

Varied practice mixes tasks so the learner is constantly meeting a novel application. It makes performance during practice worse and performance on a later test better, an effect known as contextual interference and demonstrated by Shea and Morgan in 1979.

The practical consequence is uncomfortable: the practice schedule that produces the best-looking lesson is not the one that produces the most durable learning. Blocked practice flatters everybody in the room.

REPLICATED · Shea & Morgan, 1979

A · 08 · REPLICATED

Testing

People develop better fluency if they are tested on material before they re-study it.

FIG 10 — RETRIEVAL BEATS REVIEW, AND THE GAP WIDENS

Retrieval practice treats a test as an occasion for learning rather than an instrument for measuring it. Roediger and Karpicke found re-reading holds up over minutes and collapses over a week, while being tested holds most of what it gained.

In a resource this argues for short mastery checks distributed through a topic rather than one assessment at the end. The check is the teaching, not the audit of it.

REPLICATED · Roediger & Karpicke, 2006a, 2006b

A · 09 · SUPPORTED

Scaffolding

People learn better when a sequence of instruction leads them through the zone of proximal development.

FIG 11 — THE SCAFFOLD SPANS THE MIDDLE BAND

Vygotsky defined the zone as the distance between what a learner can do alone and what they can do with guidance or with more able peers. Scaffolding is whatever spans that distance, and by definition it is temporary.

The mark is supported rather than replicated because the zone is a theoretical construct rather than an experimental result. It has organised a great deal of good practice; it has never been measured directly.

SUPPORTED · Vygotsky, 1978

Section B

How media teaches

Eleven principles for words, pictures and narration working together instead of competing.

A How people learn
B How media teaches
C How pages are seen
D How products are used
E Who the design leaves out

Eleven principles for words, pictures and narration. They presuppose that people learn better from words and images together than from words alone; what follows is how to combine them without the combination costing more than it pays.

REDUCE EXTRANEOUS PROCESSING

Section B

Reduce extraneous processing

Load created by the layout rather than by the material.

Extraneous processing is the work a learner does that serves no instructional purpose. It is created by us, in the layout, and it comes straight out of the same fixed capacity the material itself needs.

The five principles in this section all cut the same way: take out what is not doing a job. The tension worth naming is that novelty also holds attention, and stripping a resource to nothing is its own failure. Plain is not the same as coherent.

REPLICATED · Mayer, 2017; Sweller, 1988

B · 01 · REPLICATED

Coherence

People learn more deeply when extraneous material is excluded rather than included.

FIG 12 — EXTRANEOUS LOAD IS TAKEN OUT OF UNDERSTANDING

Seductive detail interferes three ways: it pulls attention off the instructional point, it disrupts the mental model being assembled, and it activates prior knowledge that turns out to be irrelevant. Every one of those costs come out of the same fixed capacity.

Cut irrelevant information, decorative animation and visual clutter. But plain is not automatically coherent — a resource stripped of all appeal fails a different way, by not being looked at.

REPLICATED · Mayer, 2017; Sweller, 1988

B · 02 · REPLICATED

Signalling

People learn more deeply when cues are added that highlight the main ideas and the organisation of the words.

FIG 13 — CUES MAKE THE STRUCTURE VISIBLE BEFORE IT IS READ

Signalling is how a reader finds the structure without reading everything first. Headings, weight, colour, spacing, arrows and consistent navigational cues all do it, and they work by telling the eye what to skip as much as what to read.

The cues have to mean something and go on meaning it. A bold that sometimes marks a key term and sometimes marks emphasis has stopped being a signal.

REPLICATED · Mayer, 2017

B · 03 · REPLICATED

Redundancy

People learn more deeply from animation and narration than from animation, narration and on-screen text together.

FIG 14 — THE EYE CANNOT CARRY THE WORDS TWICE

Narration does not need a transcript on screen beside it. Presenting the words in audio alone leaves the visual channel free for the graphic, instead of asking the eyes to read and watch at once.

Four exceptions are well established: no graphics on screen, a slow or learner-paced presentation, technical or unfamiliar terms, and text substantially shorter than the narration. Keywords on screen are not redundancy.

REPLICATED · Mayer, 2017

B · 04 · REPLICATED

Spatial contiguity

People learn more deeply when corresponding words and graphics are placed near rather than far from each other.

FIG 15 — A LEGEND CHARGES THE READER FOR EVERY LOOKUP

A legend forces the eye to travel between the label and the thing it names, hold one in memory across the trip, and repeat it for every element. That is extraneous load created entirely by the layout.

Put the label on the part. Where the geometry will not allow it, a leader line is the next best thing, because it at least removes the search.

REPLICATED · Mayer, 2017

B · 05 · REPLICATED

Temporal contiguity

People learn more deeply when corresponding graphics and narration are presented at the same time rather than one after the other.

FIG 16 — NO OVERLAP IN TIME, NO CONNECTION IN MEMORY

The spoken description and the thing it describes have to be in working memory together for the connection to be made. Narration that runs ahead of the animation, or explains a diagram that has already gone, asks the learner to hold one half while waiting for the other.

In practice this is a timing pass on the finished asset, not a design decision. It is also the defect most often introduced during revision, when copy is edited and timings are not.

REPLICATED · Mayer, 2017

MANAGE ESSENTIAL PROCESSING

Section B

Manage essential processing

Load created by the material itself, which cannot be removed, only staged.

Some complexity is the subject. You cannot simplify a hard idea by leaving parts of it out without teaching something false. What you can do is control when each part arrives.

Segmenting, pre-training and modality are three ways of spreading the same load across time and across channels, so that capacity is left over for the processing that actually produces understanding.

REPLICATED · Mayer, 2017

B · 06 · REPLICATED

Segmenting

People learn more deeply when a narrated animation is presented in learner-paced segments than as a continuous unit.

FIG 17 — THE BREAKS ARE WHERE THE LEARNER SETS THE PACE

Complex material cannot be made simpler by leaving steps out; that changes what is being taught. It can be broken into parts that carry one or two steps each, with the learner deciding when the next one starts.

The pacing is doing as much work as the chunking. A segmented video that plays straight through has kept the structure and given away the mechanism.

REPLICATED · Mayer, 2017

B · 07 · REPLICATED

Pre-training

People learn more deeply from a narrated animation when they already know the names and characteristics of the main concepts.

FIG 18 — MOVING THE NAMES EARLIER FLATTENS THE PEAK

Where the material would otherwise overwhelm, some of the processing can be moved earlier. Naming the parts before showing how they work means the learner is not learning vocabulary and mechanism at the same moment.

The practical version is unglamorous: put the glossary before the lesson rather than after it, and make the routines of a resource familiar so that no capacity is spent on the interface.

REPLICATED · Mayer, 2017

B · 08 · REPLICATED

Modality

People learn more deeply from graphics and narration than from graphics and on-screen text.

FIG 19 — TWO CHANNELS ONCE EACH BEAT ONE CHANNEL TWICE

Words in audio and pictures in vision arrive through two channels rather than queueing in one. The gain disappears the moment the modalities compete instead of complementing — narration that repeats on-screen text is two channels carrying one message.

Keep words on screen when they are technical, unfamiliar or long. A term the learner will need to recognise later has to be seen, not only heard.

REPLICATED · Mayer, 2017

FOSTER GENERATIVE PROCESSING

Section B

Foster generative processing

The processing that turns material into understanding, and what motivates a learner to do it.

Generative processing is the organising and integrating that makes material into knowledge. Unlike the other two kinds of load it cannot be engineered directly, only invited, because it depends on whether the learner is willing.

These three principles are about the social surface of a resource: how it addresses the learner, what it sounds like, and whether anything on screen behaves like a person. Their effects are real and smaller than the eight before them.

SUPPORTED · Mayer, 2017

B · 09 · SUPPORTED

Personalisation

People learn more deeply when words are in a conversational style rather than a formal one.

FIG 20 — ADDRESSED TO SOMEONE, AND WORKED HARDER AT

This runs against the instinct that instructional writing should be spare and impersonal. Work on discourse processing finds people work harder to understand material when they feel addressed by someone rather than issued with information.

Second person, active voice, and feedback written to the learner rather than about them. The effect is real and modest; conversational style will not rescue material that is wrong.

SUPPORTED · Mayer, 2017; Beck, McKeown, Sandora, Kucan & Worthy, 1996

B · 10 · SUPPORTED

Voice

People learn more deeply when narration is spoken in a standard-accented human voice than in a machine voice.

FIG 21 — A HUMAN VOICE, AND A REGISTER THAT FITS THE LEARNER

Human narration outperforms synthesis, and smooth tone outperforms strain. The original finding also reported a penalty for foreign-accented narration; that result has aged poorly and is better read as a familiarity effect than a property of accents.

Register matters alongside voice: polite, indirect phrasing helps learners new to the material, while confident learners do better with direct instruction.

SUPPORTED · Mayer, 2017

B · 11 · SUPPORTED

Embodiment

People learn more deeply when an on-screen agent behaves in a human-like way — gesture, movement, expression, gaze.

FIG 22 — A STATIC AGENT IS WORSE THAN NO AGENT

A pedagogical agent that moves like a person helps retention and transfer; a static one is worse than no agent at all. Drawing a diagram while explaining it beats narrating a diagram that is already finished, for the same reason.

The evidence base here is thinner than for the eight principles before it, which the original material also noted. Treat it as a promising direction rather than a settled one.

SUPPORTED · Mayer, 2017

Section C

How pages are seen

Fifteen principles of visual design, from the grid underneath to the grouping the eye does on its own.

A How people learn
B How media teaches
C How pages are seen
D How products are used
E Who the design leaves out

Fifteen principles for what a reader sees before they have read anything. This is the chapter where the checking found most of the trouble: four of the fifteen are conventions with no test behind them, and three are contested.

GRIDS

Section C

Grids

The structure underneath a layout, and the two rules that fall out of it.

A grid is what keeps a page consistent with the page before it. Used across a whole resource it does more work than any single decision made on any single screen, because coherence is a property of the set rather than the item.

Two familiar rules follow from thinking in grids: the rule of thirds, and the discipline of leaving space empty. Both are craft conventions with long provenance and no experimental test, which is not a reason to abandon them — only a reason to say so.

CONVENTION · Smith, 1797; Samara, 2011

C · 01 · CONVENTION

Rule of thirds

Divide the frame in thirds both ways and place points of interest on the lines or their intersections.

FIG 23 — THE CENTRE IS NOT ONE OF THE FOUR STRONG POINTS

The horizontal lines are useful places for titles and captions. The four intersections are where a subject sits without looking either centred or accidental. Dead centre is the one position the rule specifically does not recommend.

John Thomas Smith coined the phrase in 1797 and was careful about it: he offered it as a rule of thumb, with the aside “if I may be allowed so to call it”. No study has since shown that thirds-based composition measurably changes how a viewer responds. Follow it as craft.

CONVENTION · Smith, 1797

C · 02 · CONVENTION

Sufficient whitespace

The space between elements is an element. It does not have to be white.

FIG 24 — FOUR LINES AND A HIERARCHY BEAT TEN AND NONE

Whitespace runs from the margin down to the space inside a letterform. Used deliberately it focuses attention, raises perceived quality and makes text measurably easier to read; used as leftovers it is the first thing sacrificed when content arrives late.

The related discipline is the grid itself. Consistency across a set of pages does more for coherence than any single page can, because coherence is a property of the set.

CONVENTION · Samara, 2011; Lupton & Phillips, 2015

VISUAL HIERARCHY

Section C

Visual hierarchy

Arranging elements so that the order the eye takes is the order you intended.

Hierarchy is the difference between a page a reader can enter and one they have to solve. The eye separates elements by four characteristics — colour, size, alignment and character — and every one of them is a lever you already control.

The two principles here are the two most often got wrong: scale, which is relational rather than absolute, and type, where the choice of face matters less than the contrast it is set against.

CONVENTION · Lupton & Phillips, 2015

C · 03 · CONVENTION

Size and scale

Size is absolute; scale is the relationship between elements. Only the second builds a hierarchy.

FIG 25 — FOUR CHARACTERISTICS, ONE ELEMENT PULLED OUT OF EACH ROW

The eye separates elements by four characteristics: colour, size, alignment and character. Any one of them will pull an element out of a row of equals, and using two at once rarely doubles the effect — it usually just costs you a level of the hierarchy.

Set the focal point first and let everything else be measured against it. A page where three things are large has nothing large in it.

CONVENTION · Lupton & Phillips, 2015

C · 04 · CONTESTED

Effective typography

Legibility comes from size, spacing and contrast — not from a specialist typeface.

FIG 26 — THE TYPEFACE WAS NEVER THE VARIABLE

The original advice here was that some fonts are easier for children to read than others, and that dyslexia-specific typefaces help. Testing that claim, Kuster and colleagues found purpose-built dyslexia fonts gave no benefit to children with or without dyslexia. The gains people observe come from larger type and looser spacing, which those fonts happen to ship with.

What survives: set type large enough, space it generously, and give it real contrast against its background. Dark text on light remains easier than light on dark for extended reading.

CONTESTED · Kuster, van Weerdenburg, Gompel & Bosman, 2018; Rello & Baeza-Yates, 2013

READING PATTERNS

Section C

Reading patterns

What people are supposed to do with their eyes, and what the evidence actually shows.

This is the part of the original guidance that did not survive checking. Three reading patterns are usually taught together as though they had the same standing. They do not.

The Gutenberg diagram has no primary source. The Z-pattern has no research behind it of any kind. Only the F-pattern comes from eye-tracking work, and the group that produced it now argues it is a symptom of unformatted text rather than a template to design to.

CONTESTED · Nielsen, 2006; Pernice, 2017

C · READING GRAVITY · CONTESTED

The Gutenberg diagram

Four quadrants, a diagonal sweep, and no primary source anyone has been able to produce.

FIG 27 — AN ATTRIBUTION THAT TERMINATES IN HEDGED REPETITION

The diagram claims the eye sweeps in horizontal passes from a primary optical area at the top left to a terminal area at the bottom right, leaving two fallow corners. It is repeated everywhere, and the attribution to the typographer Edmund Arnold is hedged in identical language wherever it appears — a sign of one source copying another rather than two confirming each other.

Arnold was real and his books on newspaper design are real. No page citation anywhere leads into them. Treat the diagram as a layout convention with an unverified origin story, and do not cite it as research.

CONTESTED · no primary source located; the attribution appears only in tertiary works

C · 05 · CONTESTED

Z-pattern

A composition rule for sparse pages, often presented as a research finding. It is not one.

FIG 28 — A COMPOSITION RULE, NOT A RESEARCH FINDING

The pattern — across the top, diagonally down to the left, across the bottom — is a reasonable way to arrange a page with a headline, a mark, some copy and one action. It works because it puts the important things where a reader would look anyway, which is a fact about layout rather than about eyes.

There is no eye-tracking study of a Z-shaped scan path. The group that produced the F-pattern research has never reported a Z-pattern equivalent. Use it as a compositional device and say that is what it is.

CONTESTED · no empirical source located

C · 06 · CONTESTED

F-pattern

Real, well documented, and pointing the opposite way from the advice usually attached to it.

FIG 29 — THE F IS A SYMPTOM OF MISSING CUES, NOT A TEMPLATE

Nielsen’s 2006 eye-tracking work found readers scanning content in an F: a full pass across the top, a shorter second pass, then a vertical sweep down the left. The finding is solid. The conclusion drawn from it in most design guidance is not.

Updating the study eleven years later, Pernice argues the F-pattern emerges when text lacks formatting cues, that it is one of several scanning shapes, and that good design suppresses it — headings, lists and bold produce a layer-cake pattern instead, which serves the reader better. Design against the F, not to it.

CONTESTED · Nielsen, 2006; Pernice, 2017

COLOUR

Section C

Colour

What colour does to a reader, and how to choose more than one of them.

Colour is the fastest signal on a page and the one people remember. It is also the one where the design literature and the research literature are furthest apart: the wheel is a compositional tool from painters, not a finding.

Keep the palette to four or fewer, and never let colour be the only thing carrying a distinction — a reader who cannot separate two hues then loses the information entirely.

SUPPORTED · Elliot & Maier, 2014; Itten, 1961

C · THE WHEEL · CONVENTION

Harmony is geometry

Complementary, analogous and triadic schemes are three ways of choosing positions on a twelve-point ring.

FIG 30 — TWELVE POSITIONS, THREE WAYS TO CHOOSE FROM THEM

The wheel is a compositional instrument from painters — Itten and Albers — not a finding from perception research. What it gives you is a rule for picking a second and third colour that will not fight: opposite for contrast, adjacent for calm, evenly spaced for variety inside a set.

Limit a palette to four. Beyond that the scheme stops being a scheme and colour stops carrying meaning.

CONVENTION · Itten, 1961; Albers, 1963

C · 07 · SUPPORTED

Colour psychology

Colour affects mood, attention and judgement — in context, and less dramatically than the popular accounts claim.

FIG 31 — ONE SWATCH, TWO READINGS, AND NEITHER IS THE SWATCH

Elliot and Maier’s review is the careful version: effects are real, they are context-dependent, and the same hue can carry opposite meanings in different settings. The neat table of colours and emotions that circulates in design writing is not what the research supports.

Two things are dependable. Colour is the fastest way to make something findable, and it is perceived relationally — a colour is read against what surrounds it, not on its own.

SUPPORTED · Elliot & Maier, 2014; Palmer & Schloss, 2010

C · 08 · CONVENTION

Colour harmony

Any two positions on the wheel can be made to work; the scheme is a shortcut, not a guarantee.

FIG 32 — FOUR COLOURS CAN EACH BE GIVEN A JOB; EIGHT CANNOT

Complementary pairs sit opposite and give maximum contrast. Analogous schemes take neighbours and read as quiet. Triadic schemes space three evenly and stay lively without clashing. More elaborate schemes vary hue, value and saturation together.

Never let colour be the only carrier of a distinction. A reader who cannot separate two hues loses the information entirely, which is why the inclusive design chapter asks for colour plus shape or label.

CONVENTION · Itten, 1961

GESTALT

Section C

Gestalt

Seven ways the eye groups things before anyone has decided anything.

Gestalt grouping happens whether or not it was designed for. Elements near each other are read as belonging together; elements that look alike are read as a set; a shape with a gap in it is read as closed. None of this is optional for the viewer.

Which makes these principles a checking tool as much as a composing one. The question to ask of a finished layout is not only whether the intended grouping reads, but whether an unintended one does.

REPLICATED · Wagemans et al., 2012a, 2012b

C · 09 · REPLICATED

Simplicity

Given a busy or ambiguous figure, the eye settles on the simplest reading it can find.

FIG 33 — SEVEN GROUPINGS THE EYE PERFORMS WITHOUT BEING ASKED

Also called the law of Prägnanz. Confronted with a complicated shape, the eye reorganises it into simpler components or into a simpler whole. This is not a preference; it is what perception does before anything reaches deliberation.

The seven groupings that follow are all consequences of the same tendency, and all of them operate on your layout whether or not you designed for them.

REPLICATED · Wagemans et al., 2012a, 2012b

C · 10 · REPLICATED

Closure

The eye completes an incomplete figure, given enough of it to work with.

FIG 34 — ENOUGH MISSING TO COMPLETE, NOT ENOUGH TO BREAK

Closure is the glue: parts are combined into a whole and the missing information is supplied. It is why a logo can be drawn with gaps and still read, and why an interrupted border still groups what it surrounds.

The balance is narrow. Too little information and the parts stay separate; too much and there is nothing left for the reader to complete, which is where the pleasure of the device came from.

REPLICATED · Wagemans et al., 2012a

C · 11 · REPLICATED

Similarity

Shared colour, shape or size makes a set, even across distance.

FIG 35 — LIKENESS GROUPS MORE STRONGLY THAN POSITION

Colour, shape, size, texture — any shared characteristic will do it. The grouping happens across distance, which means similarity can quietly override the arrangement you intended.

The direct application is consistency of interactive elements. Links and controls must look like each other and unlike ordinary text, everywhere in the resource, or the learner has to test each one.

REPLICATED · Wagemans et al., 2012a

C · 12 · REPLICATED

Continuity

Whatever falls along a common line or curve is read as belonging together.

FIG 36 — FOUR SEGMENTS MEET, AND NOBODY SEES FOUR

Once the eye starts along a path it keeps going until something stops it. Elements that fall on that path join the group; elements just off it do not, however close they are.

A second consequence is that shapes are perceived as continuing past where they end, which is what makes a partially hidden object read as whole rather than as cut.

REPLICATED · Wagemans et al., 2012a

C · 13 · REPLICATED

Proximity

Elements placed close together are read as a group, whatever else they have in common.

FIG 37 — TWELVE MARKS, THREE GROUPINGS, SPACING ALONE

Proximity is the strongest of the grouping principles and the cheapest to apply: nothing has to match, nothing has to be drawn, only the spacing has to be right. Elements need only be closer to each other than to anything outside the group.

It is also the one most often violated by accident. A caption equidistant from two images belongs to neither, and no amount of styling will fix what the spacing has already said.

REPLICATED · Wagemans et al., 2012a

C · 14 · REPLICATED

Symmetry

The eye seeks symmetrical arrangement around a centre, and takes precedence over proximity when it finds one.

FIG 38 — SYMMETRY IS FOUND BEFORE PROXIMITY IS

Symmetry reads as order and solidity, which is why it is found so quickly. A composition does not have to be symmetrical to be balanced, but where symmetry exists it will be seen before other groupings are.

That ranking is the useful part: if a layout needs to communicate at a glance, symmetry will do it faster than spacing will.

REPLICATED · Wagemans et al., 2012b

C · 15 · REPLICATED

Synchrony

Elements that move together are read as belonging together, however far apart they sit.

FIG 39 — DIRECTION GROUPS WHAT DISTANCE DOES NOT

Also called common fate, and the only grouping principle that depends on time. Elements sharing a direction or a destination are grouped even when they differ in every other respect and are widely separated on screen.

In an interactive this is a strong and underused signal: things that change together are understood to be one thing, which makes coordinated motion a way of teaching structure without labelling it.

REPLICATED · Wagemans et al., 2012a

Section D

How products are used

Ten laws drawn from psychology, and ten heuristics for finding what is wrong with an interface.

A How people learn
B How media teaches
C How pages are seen
D How products are used
E Who the design leaves out

Twenty principles for resources the learner has to operate. The ten laws are drawn from psychology and vary widely in standing; the ten heuristics are an inspection instrument for finding what is already wrong.

LAWS OF UX

Section D

Laws of UX

Ten maxims from psychology, of rather different standing.

These are the principles to reach for when designing an interface someone has to operate — an interactive, a player, anything with controls. Four of the ten are replicated experimental findings. Three are industry adages with no study behind them. One is a well-documented misreading of the paper it names.

They are collected under the name Jon Yablonski gave them, and the collection is useful. The marks on each slide are there so the collection is not mistaken for a body of evidence.

CONVENTION · Yablonski, 2024

D · 01 · SUPPORTED

Aesthetic-usability effect

An attractive design is judged easier to use than it is, and its small faults go unnoticed.

FIG 40 — BEAUTY PREDICTS PERCEIVED EASE, NOT ACTUAL EASE

Kurosu and Kashimura had 252 participants rate 26 variations of a cash-machine interface for beauty and for ease of use. Perceived beauty correlated more strongly with perceived ease than actual ease did.

Worth knowing that the famous study is a two-page extended abstract, so the evidence is thinner than its reputation. The practical warning survives either way: an attractive prototype will test better than it works, which is how real usability problems get missed.

SUPPORTED · Kurosu & Kashimura, 1995

D · 02 · REPLICATED

Fitts’s law

The further away and the smaller a target is, the longer it takes to hit.

FIG 41 — DISTANCE AND SIZE TRADE AGAINST EACH OTHER

Fitts showed in 1954 that movement time rises with distance and falls with target size, and that fast movements at small targets buy speed with errors. It is one of the most replicated results in human performance.

For a learning resource this is mostly about buttons and hit areas, and it matters most for the youngest learners, whose fine motor control makes every target effectively smaller than it looks.

REPLICATED · Fitts, 1954

D · 03 · REPLICATED

Hick’s law

Each doubling of the options adds a fixed amount to the time it takes to choose.

FIG 42 — EACH DOUBLING OF THE CHOICES COSTS THE SAME AGAIN

Hick and Hyman established the relationship independently in 1952 and 1953. Because the growth is logarithmic rather than linear, going from two options to four costs about what going from four to eight costs — which is why long lists are less catastrophic than they feel, and why splitting one list into two often saves nothing.

The useful moves are breaking a complex task into steps and marking a recommended option, which lets a learner stop deciding early.

REPLICATED · Hick, 1952; Hyman, 1953

D · 04 · CONVENTION

Jakob’s law

Learners arrive trained by every other app they use, and expect this one to behave the same way.

FIG 43 — A FAMILIAR PATTERN STARTS WHERE A NOVEL ONE MUST CLIMB

Familiar patterns are cheaper to learn because they have already been learnt. For a learning resource this argues for behaving like the applications students already use, and for keeping the same patterns across a whole set of materials.

The mark is convention: this is Nielsen’s own formulation of accumulated observation, not a study. The reasoning is sound and the evidence for it is indirect.

CONVENTION · Yablonski, 2024

D · 05 · CONTESTED

Miller’s law

Chunking works. The seven-item interface rule attached to it is a misreading of the paper.

FIG 44 — MILLER WROTE ABOUT CHUNKING, NOT ABOUT MENU LENGTH

Miller’s 1956 paper is about channel capacity and about how chunking raw items into meaningful units lets people hold more. It is not a claim that an interface should offer between five and nine options, and Miller made no such recommendation.

What survives is the better half anyway: group related content into meaningful units, and the count stops being the constraint. Design navigation from the structure of the material, not from a number.

CONTESTED · Miller, 1956 — widely misapplied in interface guidance

D · 06 · REPLICATED

Serial position effect

The first and last items in a sequence are the ones that stay; the middle fades.

FIG 45 — THE MIDDLE IS WHERE THINGS GO TO BE FORGOTTEN

Primacy and recency together produce a bowed recall curve. Ebbinghaus demonstrated the phenomenon in 1885; Murdock’s 1962 free-recall experiments established the curve and the name in the modern literature.

Put the actions that matter at the ends of a navigation, and never bury the one thing a learner has to take away in the middle of a list.

REPLICATED · Murdock, 1962; Ebbinghaus, 1885/1913

D · 07 · CONVENTION

Tesler’s law

Some complexity belongs to the task itself. It can be shifted between maker and user, never removed.

FIG 46 — THE RESIDUE DOES NOT REDUCE; IT ONLY MOVES

Also called the law of conservation of complexity. Larry Tesler’s argument was about where the residue should sit: an engineer spending an extra week to absorb it saves a great many users a minute each, every time.

For learning materials the trade is between simplifying for the learner and the cost of doing so in the build. Some difficulty belongs with the learner — desirable difficulty is a real effect — but it should be the difficulty of the subject, not of the interface.

CONVENTION · Tesler, n.d. — stated by its originator, never formally published

D · 08 · REPLICATED

Von Restorff effect

Make one item unlike its neighbours and that is the item people recall.

FIG 47 — EMPHASIS IS A FIXED BUDGET

Hedwig von Restorff showed in 1933 that an item isolated by contrast within a list of categorically similar items is recalled better. It is among the most reliable findings in memory research.

Which makes visual distinctiveness a scarce resource. Every element made prominent reduces the prominence of the others, so a page with six highlights has none.

REPLICATED · von Restorff, 1933

D · 09 · SUPPORTED

Zeigarnik effect

A task left unfinished stays in mind longer than one that was completed.

FIG 48 — THE VISIBLE REMAINDER IS THE HOOK

Bluma Zeigarnik reported in 1927 that waiters recalled unsettled orders better than settled ones, and reproduced the finding experimentally. Progress indicators exploit it: a task visibly incomplete stays in mind and is more likely to be finished.

The mark is supported rather than replicated because the replication record is uneven, with effect sizes varying considerably across conditions.

SUPPORTED · Zeigarnik, 1927

D · 10 · CONVENTION

Pareto principle

A small share of the causes usually accounts for most of the effect: the familiar 80/20.

FIG 49 — THE FIRST FIFTH CARRIES FOUR FIFTHS

Get the core learning content, the central interactions and the main path right before building anything else. That much is sound project judgement.

The provenance is a good story. Joseph Juran coined the name, then published a retraction in 1975 saying he had misattributed it: Pareto observed a distribution of wealth in economics and never proposed a universal. The 80/20 rule is Juran’s own, by his own account.

CONVENTION · Juran, 1975

USABILITY HEURISTICS

Section D

Usability heuristics

Ten questions to ask of an interface that already exists.

The heuristics are an evaluation instrument, not a set of findings. Nielsen derived the current ten by factor analysis of several hundred real usability problems, choosing the smallest set that explained the most of them.

That is exactly what makes them useful and exactly what limits them: they tell you where to look, not what is true. They have not changed since 1994.

SUPPORTED · Nielsen, 1994a; Nielsen & Molich, 1990

D · N01 · SUPPORTED

Visibility of system status

Show people what the system is doing, soon enough that they never have to guess.

FIG 50 — THE WAIT IS THE SAME LENGTH; THE DOUBT IS NOT

Feedback is how a system earns trust. When people can see the state it is in, they can predict what it will do and choose their next move with confidence.

In a resource: show that an answer was received, that a file is loading, that progress was saved. Silence is read as failure.

SUPPORTED · Nielsen, 1994a, 2024

D · N02 · SUPPORTED

Match between the system and the real world

Use the words and ideas people already have, and order things the way the real task runs.

FIG 51 — A TERM WITH NOTHING ON THE OTHER SIDE

Interfaces that follow real-world conventions demonstrate that somebody thought about the person using them. The related idea is isomorphism: people interpret what they see through their own experience, so an unfamiliar metaphor costs more than a plain label.

For school resources this means the learner’s vocabulary, not the curriculum document’s.

SUPPORTED · Nielsen, 1994a, 2024

D · N03 · SUPPORTED

User control and freedom

Wrong clicks are normal, so every step needs an obvious way back, with undo and redo.

FIG 52 — AN EXIT AT EVERY STEP MAKES TRYING FREE

The emergency exit has to be visible before it is needed, not discoverable afterwards. Undo is the strongest form because it removes the cost of experimenting.

In a learning resource the specific case is letting a learner go back and reconsider an answer. A path that only runs forwards turns every activity into an assessment.

SUPPORTED · Nielsen, 1994a, 2024

D · N04 · SUPPORTED

Consistency and standards

One word, one meaning: the same thing should look and behave the same everywhere.

FIG 53 — FOUND WITHOUT LOOKING, OR FOUND BY SEARCHING

Consistency has an internal and an external half. Internally, the same information appears in the same place, in the same format, on every screen. Externally, the resource follows the conventions of the platform it runs on.

An established design system is the cheapest way to get both, because the decisions have already been made and written down.

SUPPORTED · Nielsen, 1994a, 2024

D · N05 · SUPPORTED

Error prevention

Prevent the error rather than explaining it afterwards, through constraints, suggestions and forgiving defaults.

FIG 54 — THE WRONG TURN REMOVED, NOT REGRETTED

The phrase “user error” assigns the blame wrongly. If it was easy to do the wrong thing, the design made it easy. Constraints that make the wrong action unavailable are better than warnings that make it inadvisable.

Distinguish slips from mistakes: a slip is a right intention badly executed and is fixed by the interface, while a mistake is a wrong intention and is fixed by the explanation.

SUPPORTED · Nielsen, 1994a, 2024

D · N06 · SUPPORTED

Recognition rather than recall

Showing people something they can recognise beats asking them to retrieve it from nothing.

FIG 55 — THE CONTEXT DOES THE RETRIEVING

Recognising a face is easy; producing the name is hard. Interfaces that keep options, instructions and previously entered information visible remove that second, harder operation entirely.

This is the one heuristic that pulls against a learning principle. Retrieval practice deliberately asks for recall — but of the subject matter, never of how the resource works.

SUPPORTED · Nielsen, 1994a, 2024

D · N07 · SUPPORTED

Flexibility and efficiency of use

Give experts shortcuts that beginners never need to see, so one design serves both.

FIG 56 — TWO ROADS, NOT TWO PRODUCTS

Provide more than one route to the same outcome and let people tailor the actions they repeat. The constraint is not to gold-plate: options nobody uses are cost without benefit, and user research is how you tell the difference between an assumed need and a real one.

SUPPORTED · Nielsen, 1994a, 2024

D · N08 · SUPPORTED

Aesthetic and minimalist design

Anything on screen that is not needed pulls attention away from what is.

FIG 57 — NOTHING WAS TAKEN AWAY; THINGS WERE ADDED AROUND IT

Minimalism here means essential rather than sparse. Colour and visual cues are worth having as long as each one means something and goes on meaning it.

Guard against the conceit that a learner will be as interested in a feature as its designer was. Feature creep in a learning resource costs attention that was supposed to go to the subject.

SUPPORTED · Nielsen, 1994a, 2024

D · N09 · SUPPORTED

Help people recognise, diagnose and recover from errors

Say what went wrong in plain language, and say what to do next.

FIG 58 — THE TWENTIETH ALERT IS NOT READ AT ALL

People will click through anything they see often enough — alert fatigue is well documented. A message earns attention by being specific, by avoiding codes, and by being constructive rather than accusatory.

A good error message tells the learner the nature of the problem, what caused it, and the concrete step that recovers from it.

SUPPORTED · Nielsen, 1994a, 2024

D · N10 · SUPPORTED

Help and documentation

Better if it is not needed — and where it is, make it searchable, task-focused and concrete.

FIG 59 — A FIRST TASK IS FINISHED; A TUTORIAL IS ABANDONED

Documentation should be findable at the moment of need and written around what someone is trying to do, in concrete steps.

The temptation with something new is a tutorial. Most common tutorial patterns are ineffective; the alternative worth studying is how games onboard players, which is by giving them something to do in a safe first level rather than something to read.

SUPPORTED · Nielsen, 1994a, 2024

Section E

Who the design leaves out

Five imperatives for designing to the range of learners rather than to the middle of it.

A How people learn
B How media teaches
C How pages are seen
D How products are used
E Who the design leaves out

Five imperatives. Most of the differences they address are situational rather than permanent, which is why designing for them helps a far wider group than the one each was written for.

INCLUSIVE DESIGN

Section E

Inclusive design

Designing for the range of learners rather than the middle of it.

Design for the average and you have designed for nobody. Learners differ in vision, hearing, dexterity, reading proficiency and in the conditions they are working in — a bright room, a shared device, no headphones — and most of those differences are situational rather than permanent.

The five imperatives here are cheaper to meet at the start than to retrofit, and each of them helps a much wider group than the one it was written for.

SUPPORTED · W3C, 2024; CAST, 2024; Connell et al., 1997

E · 01 · SUPPORTED

Make it easy to read

Vision varies, and so do the rooms, screens and lighting conditions a resource is used in.

FIG 60 — THREE THRESHOLDS ON A SCALE THAT RUNS TO 21:1

Short-sightedness, colour vision deficiency, sensitivity to glare, a bright classroom, an old device with a dim screen — the causes differ and the remedy is the same. Choose colours with ample contrast, use colour together with shape and text rather than alone, and set type that is comfortable at arm’s length.

WCAG gives the numbers: 4.5:1 for body text, 3:1 for large text and interface parts, 7:1 for the stricter level.

SUPPORTED · W3C, 2024; CAST, 2024

E · 02 · SUPPORTED

Make it easy to hear

Hearing varies, and so does the room. A learner with no headphones in a shared space has the same problem as one with hearing loss.

FIG 61 — BOTH BARS POINT THE SAME WAY

Caption everything and provide transcripts. Gernsbacher’s review found captions benefit comprehension, attention and memory for hearing viewers as well as deaf ones, which makes this the clearest case in the chapter of an accommodation that helps everybody.

Use sound effects and music where they carry meaning rather than as decoration, and keep narration paced for someone who is following rather than someone who already knows.

SUPPORTED · Gernsbacher, 2015; W3C, 2024

E · 03 · SUPPORTED

Make it easy to interact with

Dexterity varies. So does the desk, the device and how much of it is free to use.

FIG 62 — THE FINGER IS BIGGER THAN THE BUTTON

Younger learners are still developing fine motor control, and every learner is occasionally working one-handed on a crowded table. Design for alternative means of interaction — touch and gesture as well as a pointer — and make targets large and well spaced.

Fitts’s law is the quantitative version of this imperative, which is why the two chapters keep meeting.

SUPPORTED · W3C, 2024; Fitts, 1954

E · 04 · SUPPORTED

Make it easy to understand

Reading proficiency varies more within a class than between classes.

FIG 63 — THREE ROUTES LOSE FEWER PEOPLE THAN ONE

Plain language and short sentences. Present the same information more than one way — text and picture, example and rule — so that a learner who misses one route has another. Structure long passages so they can be scanned before they are read.

This is the same instruction the multimedia chapter gives for different reasons, and where two chapters agree from different starting points the advice is usually safe.

SUPPORTED · CAST, 2024; Rello & Baeza-Yates, 2013

E · 05 · SUPPORTED

Make it easy to customise

Respecting a learner’s own choice is the part of inclusion that no amount of anticipation replaces.

FIG 64 — ONE SETTING CANNOT FIT A DISTRIBUTION

Let people change text size, spacing, contrast, background and zoom, and let those choices persist. Customisation is the learner adjusting the resource; personalisation is the resource adjusting itself. The first is more reliable, because it does not have to guess.

It is also the cheapest of the five to build if it is designed in, and close to impossible to retrofit into a layout that assumed one size.

SUPPORTED · CAST, 2024; Connell et al., 1997

Twelve books and where the field publishes

Six of the books are now in newer editions than the ones usually cited. These are the current ones.

BOOKS

Worth owning

Lidwell, Holden & Butler · Universal Principles of Design, 3rd ed. 2023
Clark & Mayer · e-Learning and the Science of Instruction, 5th ed. 2023
Hanington & Martin · Universal Methods of Design, rev. 2019
Lupton & Phillips · Graphic Design: The New Basics, 2nd ed. 2015
Weinschenk · 100 Things Every Designer Needs to Know, 2nd ed. 2020
Garrett · The Elements of User Experience, 2nd ed. 2011

PUBLISHED CONTINUOUSLY

Worth following

Nielsen Norman Group · usability research and eye-tracking
Interaction Design Foundation · long-form design literature
Smashing Magazine · practice and technique
Inclusive Design Research Centre · the inclusive design guide
W3C Web Accessibility Initiative · the standard itself
Laws of UX · the collection, with its own sources

The full reference list at the foot of this document carries every source the principles rest on, with a link to the publication where one exists.

Where a work has been superseded, the current edition is listed rather than an earlier one: Universal Principles of Design (3rd ed., 2023), e-Learning and the Science of Instruction (5th ed., 2023), Universal Methods of Design (revised, 2019), Multimedia Learning (3rd ed., 2021), 100 Things Every Designer Needs to Know About People (2nd ed., 2020), and Laws of UX (2nd ed., 2024).

Sixty principles, five chapters, one vocabulary.

Eleven of the sixty are not findings, and now say so.

A shared language is worth having. A shared standard of evidence is what makes it worth using.

END

Use it, argue with it, mark it up

The marks are a first pass. Where a principle deserves a different one, the argument for changing it is exactly the kind this document is for.

WEB geraldajam.com
WCAG 2.2 w3.org/TR/WCAG22
UDL GUIDELINES 3.0 udlguidelines.cast.org
NIELSEN NORMAN GROUP nngroup.com

Source trail

References

Every source the sixty principles rest on. Entries follow APA 7th edition; links point to the DOI, publication or standards page. Each was checked against a publisher record, a DOI registry or the standards body itself while this document was assembled. Where a principle carries a contested mark, that is because the checking found no primary source, or found one that says something other than the advice attached to it.

Albers, J. (1963). Interaction of color. Yale University Press.
Anderson, S. P. (2011). Seductive interaction design: Creating playful, fun, and effective user experiences. New Riders.
Beck, I. L., McKeown, M. G., Sandora, C., Kucan, L., & Worthy, J. (1996). Questioning the author: A yearlong classroom implementation to engage students with text. The Elementary School Journal, 96(4), 385–414. https://doi.org/10.1086/461835
Bruner, J. S. (1966). Toward a theory of instruction. Belknap Press of Harvard University Press.
CAST. (2024). Universal design for learning guidelines version 3.0. https://udlguidelines.cast.org
Clark, R. C., & Mayer, R. E. (2023). e-Learning and the science of instruction: Proven guidelines for consumers and designers of multimedia learning (5th ed.). Wiley.
Connell, B. R., Jones, M., Mace, R., Mueller, J., Mullick, A., Ostroff, E., Sanford, J., Steinfeld, E., Story, M., & Vanderheiden, G. (1997). The principles of universal design (Version 2.0). Center for Universal Design, North Carolina State University. https://design.ncsu.edu/research/center-for-universal-design/
Duarte, N. (2008). Slide:ology: The art and science of creating great presentations. O’Reilly Media.
Ebbinghaus, H. (1913). Memory: A contribution to experimental psychology (H. A. Ruger & C. E. Bussenius, Trans.). Teachers College, Columbia University. (Original work published 1885)
Elliot, A. J., & Maier, M. A. (2014). Color psychology: Effects of perceiving color on psychological functioning in humans. Annual Review of Psychology, 65(1), 95–120. https://doi.org/10.1146/annurev-psych-010213-115035
Fitts, P. M. (1954). The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47(6), 381–391. https://doi.org/10.1037/h0055392
Fyfe, E. R., McNeil, N. M., Son, J. Y., & Goldstone, R. L. (2014). Concreteness fading in mathematics and science instruction: A systematic review. Educational Psychology Review, 26(1), 9–25. https://doi.org/10.1007/s10648-014-9249-3
Garrett, J. J. (2011). The elements of user experience: User-centered design for the web and beyond (2nd ed.). New Riders.
Gernsbacher, M. A. (2015). Video captions benefit everyone. Policy Insights from the Behavioral and Brain Sciences, 2(1), 195–202. https://doi.org/10.1177/2372732215602130
Hanington, B., & Martin, B. (2019). Universal methods of design, expanded and revised. Rockport Publishers.
Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77(1), 81–112. https://doi.org/10.3102/003465430298487
Hick, W. E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology, 4(1), 11–26. https://doi.org/10.1080/17470215208416600
Hyman, R. (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology, 45(3), 188–196. https://doi.org/10.1037/h0056940
Inclusive Design Research Centre. (n.d.). The inclusive design guide. OCAD University. Retrieved August 28, 2026, from https://guide.inclusivedesign.ca/
Itten, J. (1961). The art of color: The subjective experience and objective rationale of color (E. van Haagen, Trans.). Reinhold.
Juran, J. M. (1975). The non-Pareto principle; mea culpa. Quality Progress, 8(5), 8–9.
Koedinger, K. R., Booth, J. L., & Klahr, D. (2013). Instructional complexity and the science to constrain it. Science, 342(6161), 935–937. https://doi.org/10.1126/science.1238056
Koedinger, K. R., Corbett, A. T., & Perfetti, C. (2012). The Knowledge-Learning-Instruction framework: Bridging the science-practice chasm to enhance robust student learning. Cognitive Science, 36(5), 757–798. https://doi.org/10.1111/j.1551-6709.2012.01245.x
Koffka, K. (1935). Principles of Gestalt psychology. Harcourt, Brace.
Köhler, W. (1947). Gestalt psychology: An introduction to new concepts in modern psychology (Rev. ed.). Liveright.
Kurosu, M., & Kashimura, K. (1995). Apparent usability vs. inherent usability: Experimental analysis on the determinants of the apparent usability. In Conference companion on human factors in computing systems (CHI ’95) (pp. 292–293). ACM Press. https://doi.org/10.1145/223355.223680
Kuster, S. M., van Weerdenburg, M., Gompel, M., & Bosman, A. M. T. (2018). Dyslexie font does not benefit reading in children with or without dyslexia. Annals of Dyslexia, 68(1), 25–42. https://doi.org/10.1007/s11881-017-0154-6
Leborg, C. (2006). Visual grammar. Princeton Architectural Press.
Lidwell, W., Holden, K., & Butler, J. (2023). Universal principles of design (Updated and expanded 3rd ed.). Rockport Publishers.
Lupton, E., & Phillips, J. C. (2015). Graphic design: The new basics (2nd ed., rev. and expanded). Princeton Architectural Press.
Mayer, R. E. (2017). Using multimedia for e-learning. Journal of Computer Assisted Learning, 33(5), 403–423. https://doi.org/10.1111/jcal.12197
Mayer, R. E. (2021). Multimedia learning (3rd ed.). Cambridge University Press. https://doi.org/10.1017/9781316941355
Mayer, R. E., & Fiorella, L. (Eds.). (2021). The Cambridge handbook of multimedia learning (3rd ed.). Cambridge University Press. https://doi.org/10.1017/9781108894333
Microsoft. (2016). Inclusive: A Microsoft Design toolkit. Microsoft Design. https://inclusive.microsoft.design/
Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(2), 81–97. https://doi.org/10.1037/h0043158
Murdock, B. B., Jr. (1962). The serial position effect of free recall. Journal of Experimental Psychology, 64(5), 482–488. https://doi.org/10.1037/h0045106
Nielsen, J. (1994a). Enhancing the explanatory power of usability heuristics. In Proceedings of the SIGCHI conference on human factors in computing systems (CHI ’94) (pp. 152–158). ACM. https://doi.org/10.1145/191666.191729
Nielsen, J. (1994b). Heuristic evaluation. In J. Nielsen & R. L. Mack (Eds.), Usability inspection methods (pp. 25–62). John Wiley & Sons.
Nielsen, J. (2006, April 16). F-shaped pattern for reading web content (original study). Nielsen Norman Group. https://www.nngroup.com/articles/f-shaped-pattern-reading-web-content-discovered/
Nielsen, J. (2024, January 30). 10 usability heuristics for user interface design. Nielsen Norman Group. https://www.nngroup.com/articles/ten-usability-heuristics/
Nielsen, J., & Molich, R. (1990). Heuristic evaluation of user interfaces. In Proceedings of the SIGCHI conference on human factors in computing systems (CHI ’90) (pp. 249–256). ACM Press. https://doi.org/10.1145/97243.97281
Niemiec, C. P., & Ryan, R. M. (2009). Autonomy, competence, and relatedness in the classroom: Applying self-determination theory to educational practice. Theory and Research in Education, 7(2), 133–144. https://doi.org/10.1177/1477878509104318
Palmer, S. E., & Schloss, K. B. (2010). An ecological valence theory of human color preference. Proceedings of the National Academy of Sciences, 107(19), 8877–8882. https://doi.org/10.1073/pnas.0906172107
Pernice, K. (2017, November 12). F-shaped pattern of reading on the web: Misunderstood, but still relevant (even on mobile). Nielsen Norman Group. https://www.nngroup.com/articles/f-shaped-pattern-reading-web-content/
Rello, L., & Baeza-Yates, R. (2013). Good fonts for dyslexia. In Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility (pp. 1–8). Association for Computing Machinery. https://doi.org/10.1145/2513383.2513447
Restorff, H. von. (1933). Über die Wirkung von Bereichsbildungen im Spurenfeld. Psychologische Forschung, 18, 299–342. https://doi.org/10.1007/BF02409636
Reynolds, G. (2019). Presentation Zen: Simple ideas on presentation design and delivery (3rd ed.). New Riders.
Roediger, H. L., III, & Karpicke, J. D. (2006a). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. https://doi.org/10.1111/j.1467-9280.2006.01693.x
Roediger, H. L., III, & Karpicke, J. D. (2006b). The power of testing memory: Basic research and implications for educational practice. Perspectives on Psychological Science, 1(3), 181–210. https://doi.org/10.1111/j.1745-6916.2006.00012.x
Rose, D. H., & Meyer, A. (2002). Teaching every student in the digital age: Universal design for learning. Association for Supervision and Curriculum Development.
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78. https://doi.org/10.1037/0003-066X.55.1.68
Samara, T. (2011). Graphic designer’s essential reference: Visual elements, techniques, and layout strategies for busy designers. Rockport Publishers.
Shea, J. B., & Morgan, R. L. (1979). Contextual interference effects on the acquisition, retention, and transfer of a motor skill. Journal of Experimental Psychology: Human Learning and Memory, 5(2), 179–187. https://doi.org/10.1037/0278-7393.5.2.179
Smith, J. T. (1797). Remarks on rural scenery. Nathaniel Smith.
Swan, H., Pouncey, I., Pickering, H., & Watson, L. (n.d.). Inclusive design principles. Retrieved August 28, 2026, from https://inclusivedesignprinciples.info/
Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285. https://doi.org/10.1207/s15516709cog1202_4
Tesler, L. (n.d.). The law of conservation of complexity. Retrieved August 28, 2026, from https://www.nomodes.com/larry-tesler-consulting/complexity-law
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes (M. Cole, V. John-Steiner, S. Scribner, & E. Souberman, Eds.). Harvard University Press.
Wagemans, J., Elder, J. H., Kubovy, M., Palmer, S. E., Peterson, M. A., Singh, M., & von der Heydt, R. (2012a). A century of Gestalt psychology in visual perception: I. Perceptual grouping and figure–ground organization. Psychological Bulletin, 138(6), 1172–1217. https://doi.org/10.1037/a0029333
Wagemans, J., Feldman, J., Gepshtein, S., Kimchi, R., Pomerantz, J. R., van der Helm, P. A., & van Leeuwen, C. (2012b). A century of Gestalt psychology in visual perception: II. Conceptual and theoretical foundations. Psychological Bulletin, 138(6), 1218–1252. https://doi.org/10.1037/a0029334
WebAIM. (n.d.). Contrast checker. Utah State University. Retrieved August 28, 2026, from https://webaim.org/resources/contrastchecker/
Weinschenk, S. (2012). 100 things every presenter needs to know about people. New Riders.
Weinschenk, S. (2020). 100 things every designer needs to know about people (2nd ed.). New Riders.
Wertheimer, M. (1923). Untersuchungen zur Lehre von der Gestalt. II. Psychologische Forschung, 4, 301–350. https://doi.org/10.1007/BF00410640
World Wide Web Consortium. (2024). Web content accessibility guidelines (WCAG) 2.2 (W3C Recommendation, 12 December 2024). https://www.w3.org/TR/WCAG22/
Yablonski, J. (2024). Laws of UX: Using psychology to design better products & services (2nd ed.). O’Reilly Media.
Zeigarnik, B. (1927). Das Behalten erledigter und unerledigter Handlungen. Psychologische Forschung, 9, 1–85.

End of the guide

Back to geraldajam.com