Developmental Guide for Teachers

Stage–environment fit

When students move between school stages, their motivation and wellbeing often drop because schools do not always meet their developmental needs. Young people need more independence, stronger relationships with adults, and opportunities to make real decisions. This gap is not down to any weakness in the students. It shows most in early adolescence, when students become more self-conscious and more of their work is judged in public.

Do: audit each unit. Does this give students autonomy, a close adult, and hands-on decisions when they need them?

Eccles & Roeser, via Kuther (2023), p. 593–598.

The dual-systems brain & hot vs cold cognition

During adolescence, the brain's reward and emotion systems develop faster than the parts responsible for self-control. Teenagers can think like adults in calm, low-pressure situations, but struggle more than adults when things get emotional, rushed, or when peers are watching. Younger children generally have less self-control in all situations. Treat the two-systems picture as a useful simplification: researchers still debate how sharply the systems separate and how much the gap varies between individuals.

Do: manage the emotional heat at decision points, with cool-down time, drafts, and private rehearsal, rather than only adding thinking scaffolds.

Mills et al., 2014; Sisk, 2017; decision-making p. 588–590, via Kuther (2023).

Reasoning is patchy; scaffold the move to more than one variable

Advanced reasoning skills develop unevenly and depend on the subject. Many teenagers and adults cannot always use these skills when they need them, and often focus on just one idea at a time. Younger children think best about things they can see or touch. Do not expect students to juggle several ideas or perspectives at once.

Do: introduce one frame or variable, make it visible, then build the connection to the next one out loud.

Kuhn, 2013; formal operations p. 574–578, via Kuther (2023).

Working memory is limited; get it out onto the page

Working memory is limited. At age eight it is about half the size of an adult's, and it grows as children get older and learn more. If working memory is overloaded, students cannot keep thinking about new ideas. Background knowledge frees up space for new learning.

Do: put thinking on the wall and the page: concept boards, instruction in chunks, memory strategies taught on purpose, written planners, and background knowledge given before it is needed.

Simmonds & Luna, 2015; memory strategies p. 459–461, via Kuther (2023).

Perspective-taking competes with working memory

Seeing things from someone else's point of view means setting your own aside, and that takes a lot of mental effort. When students are dealing with new material, new partners, or new tasks, they tend to fall back on their own point of view, and group work gets harder.

Do: put group work and perspective-taking in the middle of a unit, once the ideas have settled, not at the overloaded start.

Selman; Kilford et al., 2016, p. 584–586, via Kuther (2023).

Adolescent egocentrism, stage of public exposure

Adolescents often feel that everyone is watching them. The name for this is the imaginary audience. They may also believe that consequences do not apply to them, which is called the personal fable. Both feelings are strongest in early to middle adolescence. Social media can make them feel even more real.

Do: stage public products by audience size and band: peers first, then a trusted teacher, then parents, then the whole school.

Elkind, 1967; Elkind & Bowen, 1979, p. 586–588, via Kuther (2023).

Variability is the rule; design for the spread

Children vary within themselves (concrete here, abstract there), between children of the same age (SES, prior knowledge, stress), and across contexts and cultures. A planner that flattens that variation is making a developmental error.

Do: design a floor / target / extension for every concept, plus a ladder of representations so you can meet a learner where they are without dropping the concept:

Variability & culture p. 456–465; representational modes after Bruner.

Authoritative teaching

Across most cultures and backgrounds studied, the best teaching combines high expectations, warmth, and support for student independence. It works better than being too controlling, too lenient, or uninvolved. A strong student-teacher relationship helps students grow.

Do: hold the bar high and stay warm and connected. Give real choice inside a clear structure.

Kuther (2023), p. 597–598.

Schooling builds readiness, don't age-gate

Cognitive development mostly comes from strong academic experiences. It is not something that has to be in place before learning can start. Measured reasoning and working memory grow faster during the school year than over the summer holiday, and children who have had an extra year of school outperform children of the same age who have not. Advanced thinking skills develop later in places with less schooling. It is better to help students get ready to learn than to wait for them to be ready.

Do: keep the concept and change the representation. Offer an intellectually honest version at every band (Bruner) instead of putting the idea off.

Flynn & Sternberg, 2019; Roberts et al., 2015, p. 460–465, via Kuther (2023).

We know students will bring different strengths and needs, so we plan for that before we meet them

There is no real student who matches the average in attention, background knowledge, language, and motivation all at once. When we plan a unit for the average, it rarely fits anyone well without changes. Universal Design for Learning starts from the idea that barriers often come from the design, not the student, so we can remove them when we plan instead of waiting until someone struggles.

Do: build options into the unit during planning, not after a student has trouble. When support comes only after a student struggles, it often feels too late and can make them feel singled out.

CAST, UDL Guidelines 3.0, udlguidelines.cast.org.

Three questions about the design, not three kinds of learners

Universal Design for Learning asks three questions about a unit. Engagement is the why: what will make this learning matter to students? Representation is the what: in how many ways can students connect with this idea? Action and expression is the how: in how many ways can students show what they understand?

These three questions help us plan the unit, not label students. When we use labels like visual learner or kinaesthetic learner, we risk teaching to the label instead of removing barriers for everyone.

Do: ask all three questions about the unit, and expect that every student will benefit from having more than one way in. Labels do not help us plan for real differences. See Neuromyths for why the label version fails.

CAST, UDL Guidelines 3.0.

Engagement is not decoration

Engagement is often treated as just a fun activity at the start, but research shows it is much more. Emotion guides thinking and helps students focus, make decisions, and remember. When a student does not care about a question, it is not just that they are less willing to think about it. They are also less able to.

Research also shows that a short pause for reflection during a busy lesson helps students make sense of what they have learned. These quiet moments after hard thinking are part of the learning, not just a break.

Do: treat engagement as something to plan for throughout the lesson, not just at the start. Make sure the question matters to students, and allow time for quiet reflection after challenging work instead of filling every minute.

Immordino-Yang & Damasio (2007), We Feel, Therefore We Learn; Immordino-Yang, Christodoulou & Singh (2012), Rest is not idleness, on reflective pauses; Immordino-Yang (2016), Emotions, Learning, and the Brain. See also Motivation & productive struggle.

What the three questions look like at each band

Access changes as students grow. The same idea leads to different choices depending on what students at each age can manage, sustain, and feel comfortable sharing in class.

Early Primary (K–2). Engagement runs on play, novelty and concrete materials. Representation should be multisensory and largely oral, since reading remains the bottleneck. Expression is often spoken, drawn or built long before it is written, and insisting on writing measures handwriting rather than understanding.

Upper Primary (3–5). Engagement responds to real choice inside a shared task. Representation benefits from worked examples and models to compare, which is also how generalisations get worked out. Expression can begin to carry structured writing, as long as the structure is given, not assumed.

Middle School (6–8). Engagement is bound up with autonomy and peer standing, so the stakes of being publicly wrong are at their highest. Representation should get working memory out onto the page, because reasoning and worrying about how you look are competing for the same limited space. Expression needs private rehearsal before public performance.

High School (9–12). Engagement responds to authentic audience and consequence. Representation should spell out the conventions of the subject rather than leave them for the already-advantaged to pick up. Expression can carry real choice of form, as long as the thinking demanded is the same whichever option they choose.

Do: choose the option that matches what students at that age can manage and remember, and check that it fits their stage of development. See Cross-cutting principles.

Access is not lowered demand

A common mistake is to use Universal Design for Learning as a reason to lower expectations. In fact, the main idea is to keep the goal the same for every student, but allow different ways to reach it and show it. If an option changes what students are expected to understand, it is not access. It is a different unit.

Universal Design for Learning is not the same as differentiated instruction. Differentiation changes content, process, or product for individual students. Universal Design for Learning plans the environment so fewer changes are needed later. Both approaches are useful and can work together, but it is important to know which one we are using.

Do: keep the main goal the same for everyone, vary how students get there and how they show their understanding, and make sure each option still asks for deep thinking.

CAST, UDL Guidelines 3.0; Tomlinson (2014) for the differentiation contrast.

Readiness, interest and learner profile: respond without lowering the goal

Students arrive at different points (readiness), care about different things (interest) and work best in different ways (learner profile). Responding to these is differentiation. The essential understanding stays the same for everyone; what changes is the route, the support and the pace. Every version of a task should be respectful: as engaging, as focused on the generalisation and as demanding of thought as any other. Learner profile here means how students prefer to work, not learning styles, which the evidence does not support.

Do: write the baseline and ceiling case studies from the same generalisation; check each version against the same task success criteria; teach up, so the ceiling is the class's aim and the baseline is the way in.

Tomlinson, 2014; Pashler et al., 2008 on learning styles.

Plan support from goals, for every learner

Inclusion is a shared responsibility: every student belongs in the main classroom, and support travels with them rather than pulling them out. Start with the barriers that the design can remove for everyone, then add targeted support where an individual goal needs it, and track growth towards that goal.

Do: for any student with an individual goal, name the goal, the support in this unit, and the checkpoint where you will see growth.

Jung, 2018; Jung, 2022; CAST, UDL Guidelines 3.0.

Letting students choose how they show their understanding gives them agency, not an easier path

Universal Design for Learning aims to give students real choices in how they show their learning. Agency depends not just on what options exist, but on how we offer them. For example, a recorded explanation instead of an essay is empowering when it is presented as a real way to communicate in the subject, not as an easier option for students who cannot do the main task.

This is especially important from the middle years on, when students care about earning respect and meeting high expectations. Adolescents are willing to work hard for adults who set a high standard and make it clear they believe students can reach it. They also notice when expectations are lowered, and may see it as a judgement about their ability.

Do: offer choices that are truly equal in challenge, explain why each is a valid way to communicate in the subject, and let students decide. When giving critical feedback, say clearly that you are doing so because you hold a high standard and expect students to meet it.

Yeager (2024), 10 to 25, on the mentor mindset and on status and respect; CAST, UDL Guidelines 3.0, which names learner agency as the goal. See also Motivation & productive struggle.

Sometimes we confuse language difficulty with difficulty in understanding the concept

A multilingual student may understand the idea well but not have enough English to show it. If assessment does not separate language from understanding, it may look like the student does not know the concept, and teaching may focus on the wrong issue.

The generalisation is where this bites first. It is the most compressed sentence in the unit, usually the most abstract, and the one students have to reproduce from memory, so a student who holds the concept can still fail to hold the wording. Writing it plainly removes most of the problem before any support is needed. See Say it plainly, or students cannot carry it.

Do: give students sentence frames to help with reasoning, not just content. Allow students to use their home language for thinking and discussion, even if the final product is in English. Assess the concept and the language separately so both are clear.

García & Wei (2014), Translanguaging; Zwiers & Crawford (2011), Academic Conversations.

Teach students the specific moves they need for academic discussion, instead of hoping discussion will happen on its own

Academic conversation is made up of skills students can learn, not just something some students do naturally. We can teach students to clarify, support ideas with evidence, build on others' ideas, challenge ideas respectfully, and reflect on the conversation. Naming these moves helps quieter or less confident students join in, especially in the middle years.

Do: post these discussion moves in the classroom, model them, and use their names during discussion. Assess whether students use these moves, not just whether they speak.

Zwiers & Crawford (2011), Academic Conversations.

One equity check at the design table

It is easy to skip access questions when planning under time pressure. Taking a moment to look at the unit and ask who it serves well and whose experience is missing, whether in examples, contexts, or ways of showing understanding, can make a real difference.

Do: ask once per unit, "who benefits from this design, and whose experience is missing from it?", then change one decision in response. One acted-on answer beats a checklist nobody reads.

What this is not

This is not about learning styles, which are not supported by evidence. It is not a list of accommodations for students with diagnoses, because that approach designs for labels and leaves out students without them. It is also not a reason to lower expectations. If a unit becomes easier for everyone, it is not more accessible. It is just less challenging.

See also Neuromyths, what not to build on.

Engagement, the why

Open with a provocation that leaves a gap students can feel, not a pleasant starter. Two examples that seem to contradict each other will do it, because the contradiction makes the resolution worth having, and the resolution is the generalisation you want them to reach.

Let students choose which contrasting case they investigate. Choice inside a shared task hands over real control while keeping the class heading for the same understanding. That is what makes it workable rather than chaos.

Then say out loud that the confusion is the work and that you expect them to come through it. Productive struggle without that assurance looks and feels exactly like failing, and a student who reads it as failing will stop.

Do: build the gap on purpose, offer the choice inside the shared task rather than instead of it, and name the struggle as the work before it starts.

See Motivation & productive struggle and Yeager (2024), 10 to 25.

Representation, the what

Never one example. A generalisation cannot be worked out from a single case, so plan at least two contrasting examples and at least one non-example. The non-example is often the sharpest teacher in the set, because it shows where the idea stops rather than giving another instance of it.

Vary the surface and hold the structure constant. What should survive across the cases is the conceptual relationship, not the topic, and students only see that when the topics differ enough to be distracting.

Put the comparison in something they can see all at once, a matrix or a side-by-side chart. Otherwise working memory is spent holding case one while analysing case three, and the comparison never happens.

Spell out the conventions of the discipline rather than leaving them to be picked up. Picking things up works for the students who arrived already knowing, and those are exactly the students access is not for.

Do: two contrasting examples and one non-example as the minimum, laid out where they can be seen together.

The examples-first sequence is covered in Conceptual precision & logic.

Action and expression, the how

Let the generalisation be stated in speech, in writing, as a diagram or as a demonstration. The test is whether it names a relationship that transfers, not whether it arrived as a paragraph. A student who can say it precisely and cannot yet write it precisely holds the understanding, and the writing is a separate goal worth teaching separately.

Supply frames for the reasoning move, not for the content, so the scaffold carries the sentence and the student carries the idea. A frame that supplies the answer has taken away the thinking it was meant to support.

Build in private rehearsal before any public commitment. This matters most in the middle years, where being publicly wrong costs more than being quietly unsure.

Then ask for the understanding to be applied to a new case, one they have not seen, in whatever form. Transfer is the evidence. Everything before it is preparation.

Do: accept any form that names a transferable relationship, and require transfer to an unseen case as the evidence rather than a restatement of the taught one.

CAST, UDL Guidelines 3.0, which names learner agency as the goal. See also Choice of form is agency, not leniency.

Memory is the residue of thought

Students remember what they think about, not what you cover or how engaging the activity felt. A lively task where the thinking is all about the packaging (cutting, colouring, the game rules) leaves the concept unthought about, and so unremembered.

Do: for every task, ask "what will they actually be thinking about?" and make sure the answer is the generalisation, not the craft around it.

Willingham, 2009; Deans for Impact, 2015.

Working memory is narrow and long-term memory is vast (cognitive load)

New information goes through working memory, which can only hold a few things for a short time. Long-term memory, on the other hand, can store much more. If a task is too demanding for working memory, learning stops.

There are two loads to manage, not three. Intrinsic load comes from the material itself. Extraneous load comes from how it is presented and what the task makes students do besides think. Germane load is often described as a third kind, the effort of building understanding, but Sweller revised that: it does not add to the total. It redistributes working memory from extraneous demands to the material itself. That is why cutting clutter does not just free up room. It moves the room to where the learning is.

Do: chunk instruction, remove decorative "seductive details", and build background knowledge so more of the concept fits in working memory at once. Treat every bit of clutter you cut as room moved to the concept, not just room saved.

Sweller, van Merriënboer & Paas (2019), Cognitive architecture and instructional design: 20 years later, Educational Psychology Review, 31(2), 261–292, which supersedes the 1998 three-load account still in wide circulation. See also Kirschner, Hendrick & Heal, 2022.

Difficulty is not a property of the material

How hard something is depends on how many pieces a learner must hold at once and relate to each other, which depends on what they already know. Sweller's example: the word "characteristics" is a single element for a fluent reader, retrieved whole, and a dozen interacting squiggles for a child learning to read. Same word, different load.

This matters when planning. A task is not difficult or easy in itself. It is difficult or easy for a particular learner on a particular day. Sweller puts it bluntly: measures of complexity that ignore the learner's knowledge are largely useless.

Do: judge difficulty against the students in front of you, not against the task. When work is too hard, ask which pieces they are holding separately that a more expert learner would hold as one, and teach that chunk first.

Sweller, van Merriënboer & Paas (2019); element interactivity.

Scaffolding has an expiry date

Worked examples, heavy guidance and step-by-step support are among the best-evidenced moves in teaching, and they stop working. As expertise grows the same support becomes unnecessary, and then it gets in the way: students spend working memory squaring guidance they no longer need with knowledge they already hold. This is the expertise reversal effect, and it is why a strategy that worked in September can hinder in March.

Much of the pressure to keep a scaffold comes from the students. A class that asks for the graphic organiser in every unit is not reporting a need, it is reporting a habit, and it feels productive while it is being filled in, which is exactly the judgement Self-regulation & metacognition warns against trusting. An organiser is structure on paper, and the point is for the structure to end up inside the student. It should run the same course as self-talk: given, then partly given, then student-generated, then unnecessary. A class still asking for it in the final year is a sequence that stalled at the first step, with the sheet mistaken for the thinking.

What sits at the far end is not the absence of support but support that has changed in kind. A final year Design class can be given nothing but "find a problem worth solving and show me how a designer works", with everything after that supplied as questions rather than answers, as provocations, as the teacher thinking aloud, as mentorship, and as the explicit message that the standard is high and they are capable of meeting it (see High standards, high support, said out loud). That is more demanding of the teacher than a worksheet, not less. It also rests entirely on the years of structure behind it. The same lesson handed to beginners is unguided discovery, and it fails hardest for the students with the least prior knowledge.

Do: plan the withdrawal at the same time as the support. Fade worked examples into completion problems and then into independent problems, vary the shape of any organiser so students cannot pattern-match to the form instead of the reasoning, and treat a student who has stopped needing the scaffold as the signal to remove it rather than as proof it works.

Kalyuga et al. (2003), the expertise reversal effect; Renkl & Atkinson (2003), guidance fading; both reviewed in Sweller, van Merriënboer & Paas (2019).

Inquiry is not the absence of instruction

The argument between inquiry and direct instruction is usually set up as a line with a virtuous end at each side, which makes every real lesson look like a compromise. It is not a line. Inquiry is the frame a unit is built in, and direct instruction is one of the moves made inside it. A move is not a concession. An inquiry is not diluted by the teaching it contains, and withholding a necessary explanation does not make it purer, it makes it fail.

Two things trigger direct teaching. The first is that a convention cannot be induced. Names, notation, units, tool technique and safety procedure are not sitting inside the examples waiting to be found, so there is nothing for a student to work out, and holding them back is a guessing game, not rigour. The second is load: knowledge that could be worked out should still be shown or told when working it out would swamp a beginner, which is the case for worked examples above.

Classroom procedure is the clearest example of the first trigger, not an exception to the principle. Nothing in the situation lets a student work out that the guard stays down on the machine, or how a class moves between benches with tools in hand. In a workshop the cost of treating that as an inquiry is not a wasted lesson, it is an injury. Told, rehearsed, non-negotiable.

The risk is that needed quietly becomes convenient, and the frame collapses into ordinary teaching wearing an inquiry label. Two questions keep it honest, and they are asked about the same move at different moments. First: did the teaching unblock the inquiry, or pre-empt it? Instruction may supply anything except the destination. The term, the tool, the procedure and a worked example of a sub-skill all clear the path; the generalisation the students were about to reach is the one thing it may not hand over. Then, weeks later: is it still needed? Support that overstays becomes pre-emption on its own, because students who no longer need the structure stop working things out and start following. The same explanation can unblock on Tuesday and pre-empt by Friday.

Do: before teaching something directly, ask whether the cases could carry it. If they could not, teach it without apology. If they could, ask whether this class could carry it today. Then decide when you will stop, because a scaffold that is working is the hardest kind to remove.

Dewey (1938), Experience and Education, which criticises the progressive failure as sharply as the traditional one and holds that the adult's greater experience creates a responsibility to guide rather than a licence to withdraw. On the second question see Scaffolding has an expiry date above.

Prior knowledge is the biggest lever

We learn best when new ideas connect to what we already know. Having strong background knowledge helps free up working memory for thinking and problem-solving. This benefit is not just for young children. It applies to all learners.

Do: bring out and check prior knowledge before demanding tasks. Teach the background knowledge that makes the task possible first, rather than assuming it.

Deans for Impact, 2015; Willingham, 2009.

Retrieval practice beats re-reading

Practising recalling information, like with low-pressure quizzes, brain dumps, or asking "what did we learn last lesson?", helps students remember it much better than just reviewing or re-reading. Reviewing feels helpful, but it mostly improves familiarity rather than memory. This is called the testing effect: testing helps students learn, not just show what they know.

Do: open lessons by recalling prior learning. Use frequent, low-stakes recall of the unit's key generalisations rather than end-of-unit cramming.

Roediger & Karpicke, 2006; Dunlosky et al., 2013; Agarwal & Bain, 2019.

Space it and mix it, and it powers transfer

Spreading study sessions over several days (spacing) and mixing different types of problems (interleaving) help students remember and use what they learn for longer, even if it feels more difficult at first. These helpful challenges, called "desirable difficulties", are built into this programme by revisiting and mixing examples over time.

Do: revisit each key generalisation across the unit, not once. Mix contrasting cases so students must decide which idea applies, not just apply the one just taught.

Bjork & Bjork, 2011; Dunlosky et al., 2013.

Elaboration: make students ask and answer "how" and "why"

When students explain how and why something works, and connect new ideas to what they already know, they build stronger and more useful memories than if they just repeat facts. Asking questions like "why is that true?" or "how does this connect to what we learned before?" helps turn information into real understanding.

Do: have students put the generalisation in their own words, link it to a prior unit, and answer "why?" and "how does this connect?", not only "what?".

Dunlosky et al., 2013; Weinstein, Sumeracki & Caviglioli, 2018.

Concrete examples for abstract ideas, then vary them

Students learn and remember abstract ideas better when they are linked to clear, specific examples. Understanding grows when students see different examples and figure out what they have in common. This is why this programme uses varied cases, so students do not just remember one story, but understand the bigger idea.

Do: teach each generalisation from two or more contrasting concrete cases and have students name what stays the same underneath. Examples first, not a single worked instance.

Weinstein, Sumeracki & Caviglioli, 2018; Willingham, 2009; Gentner, Loewenstein & Thompson, 2003, on comparing cases to build a transferable schema.

Worked examples first, then fade (the expertise-reversal effect)

Beginners learn best by studying complete examples, rather than trying to solve problems on their own, because too much information at once can overwhelm them. As students become more skilled, too much support can actually get in the way, so it is important to gradually remove help as they improve.

Do: model the full example, then fade the scaffold across the sequence towards independent application, matching guidance to where the learner is, not giving every learner the same support.

Sweller, Ayres & Kalyuga, 2011; Rosenshine, 2012.

Two channels, used well (dual coding, and its limits)

Combining words with a helpful diagram is more effective than using either one alone, because they use different parts of working memory. However, if the picture is just for decoration, or if the text on the screen repeats what is being said, it can actually make learning harder instead of easier.

Do: pair concept diagrams with spoken or written explanation. Cut repeated text and decorative images, and don't read a dense slide aloud word for word.

Mayer, 2009; Weinstein, Sumeracki & Caviglioli, 2018.

Feedback only works if there is time to act on it

Good feedback tells students where they are headed, how they are doing, and what to do next. It should focus on the work, not the person. Feedback is only useful if students have time and a chance to act on it. The learning happens when they respond to it.

Do: build feed-forward and dedicated response time into the unit, so students redraft or re-attempt against the feedback rather than filing a marked piece away.

Hattie & Timperley, 2007; Wiliam, 2011.

Emotion knowledge comes first

Emotional literacy, being able to identify, express, and understand your own feelings and those of others, is the foundation for other skills. Emotions drive social skills, and being able to manage emotions starts with naming them. The right way to show emotion depends on culture, family, and community, so expect real differences among students and see them as diversity, not problems.

Do: name emotions as they arise, build a feelings vocabulary alongside the content vocabulary, and model your own regulation aloud ("I'm frustrated, so I'm pausing before I answer").

Hargraves, 2020 (The Education Hub); on culture, Psychology of Classroom Learning (2009).

Regulation is a skill to teach, not a temperament to wait for

Children have different temperaments. Those who are more likely to feel sad, scared, angry, or frustrated, or who act impulsively, need extra support and teaching. Good self-regulators use strategies like asking for help, looking at the problem in a new way, changing how they see the situation, taking a break, or distracting themselves. Sometimes, they can even use their emotions in a positive way, like turning anger into motivation.

Do: teach the strategies openly and rehearse them in calm moments so they are there in hot ones. This is the hot-versus-cold principle above, in classroom form.

Hargraves, 2020.

Social competence runs on perspective-taking

Skills like cooperating, taking turns, negotiating, and solving social problems all depend on understanding the situation and seeing things from another person's perspective. This ability develops over time and can be difficult when students are mentally overloaded. These skills grow through regular practice, not just by telling students to use them.

Do: run regular structured routines: Newsball for active listening, Humans of Your School for meeting unfamiliar perspectives, and build-and-make challenges for teamwork under constraint (all free from Harvard's Making Caring Common project).

Selman, via Kuther (2023); Making Caring Common, Harvard GSE.

Self-management is an executive function at work

Skills like paying attention, ignoring distractions, sticking with tasks, planning, and thinking before acting all use the same mental abilities as working memory. These skills develop through practice and support. In early childhood, play is one of the best ways to build them. Sometimes, a child knows what to do, like "use your words", but cannot do it when upset. This is a normal part of development, not a moral failing.

Do: put self-management out in the open the way you do working memory: visible routines, checklists, planned pause points, and asking for help taught as a move, not a failure.

Hargraves, 2020; Galileo Educational Network (early-years play and executive function).

Emotion is part of the machinery of learning

Stress and anxiety make it harder for students to think, pay attention, solve problems, and interact with others. Students who struggle to have positive social experiences may avoid difficult tasks. Over the long term, social skills in kindergarten can predict wellbeing, job success, and mental health as adults.

Do: treat the social-emotional work as part of the unit, not an add-on: calm entry routines, predictable structures, and emotional safety before public exposure (see the staging principle above).

Hargraves, 2020; Denham & Brown, 2010; Jones, Greenberg & Crowley, 2015.

The psychosocial stakes by band (Erikson)

At each stage of school, students face different social and emotional challenges that affect how they respond to feedback. Young children are learning to take initiative and want to know if they can make things that work, so real experiences of success matter more than praise. This focus lasts until about age twelve. In adolescence, the main question becomes "Who am I becoming?" This is why public feedback and peer opinions have such a strong impact during these years.

Do: in primary, build in real mastery, real products, and visible progress. In adolescence, give safe ways to try on positions and roles inside the work, and grade the work, never the self.

Erikson, 1950/1968.

Empathy grows by habit and matures into intellectual empathy

Caring develops through regular practice and gradually being exposed to different viewpoints. For older students, this grows into intellectual empathy, the ability to understand and fairly consider ideas they may not agree with, which is a key trait of critical thinking according to Paul and Elder. The line from taking turns as a small child to weighing opposing arguments fairly is one continuous line, and the next section picks it up.

Do: keep one caring routine running all year rather than staging a one-off unit. Move exposure to unfamiliar perspectives from near (a classmate's story) to far (community elders, other cultures), and use brief mindfulness moments to support the self-regulation underneath it.

Making Caring Common, Harvard GSE; Paul & Elder (criticalthinking.org).

Clean Language, questions that keep the child's meaning intact

Clean Language, created by David Grove and introduced to classrooms by Caitlin Walker, uses the child's own words in neutral questions like "what kind of X?", "is there anything else about X?", or "what would you like to have happen?" This approach means adults do not add their own interpretation. These questions show careful listening, help shift the class from judgement to curiosity, and give students space to express what they think and feel in their own time. This builds emotional literacy and perspective-taking through real conversations.

Do: when emotion runs high, ask "what would you like to have happen next?" rather than prescribing a fix; use "what kind of...?" and "is there anything else about...?" to help a child describe a feeling before anyone tries to solve it.

Grove (Clean Language); Walker, From Contempt to Curiosity (2014); training and practice resources at cleanlearning.co.uk.

Spotting difficulty early

Look for outward signs of difficulty, like tantrums, aggression, or not following rules, as well as quieter signs, like withdrawal, lack of emotion, sadness, or ongoing anxiety. Most problems get better when adults adjust the environment and their own actions, rather than focusing on the child alone. If problems continue, involve family and professional support early.

Do: adjust your practice and the environment first, and bring in others deliberately for children at real risk.

Hargraves, 2020.

Never hand over the understanding students are building

If the "correct" generalisation is going to arrive from the teacher anyway, there is no reason to put in the hard work of writing one. The reveal takes away both the need to think and the reward of having thought. Students meet the conceptual question, never its answer.

Do: let students' own wording stand. Use the strand's conceptual question as the exit ticket and judge accuracy from what they wrote, not from how close they came to a version you then supply.

Yeager, 2024; Yeager, Dahl & Dweck, 2018.

Meet inaccuracy with questions, not corrections

While students build their generalisations, deal with inaccuracy by asking further guiding questions that steer the meaning-making, and give plenty of chances to bring misunderstandings into the open and fix them along the way. A question that walks a learner from noticing to acting leaves the insight theirs. A correction hands ownership back to the teacher. Every student is allowed their eureka moment, and that moment is the earned prestige that pays for the next struggle.

Do: prepare the question ladder in advance (concept formation → factual → debatable → conceptual). When a generalisation is wrong, offer the case that breaks it rather than the fix.

Yeager, 2024; Crone & Dahl, 2012.

High standards, high support, said out loud

Questioning instead of correcting only motivates when it reads as respect, not as holding back. Adolescents scan feedback for status information: is this teacher judging me or investing in me? Critical feedback lands when it carries the clear message that the standard is high and the teacher believes the learner can reach it.

Do: name the standard and the belief together ("this generalisation needs to survive a harder case, and I am asking because your first version shows you can get there"), then ask the next question.

Yeager, Purdie-Vaughns et al., 2014; Yeager, 2024.

Belonging protects effort

Students who worry that difficulty means they do not belong stop trying when work gets hard. Framing early struggle as normal and temporary, and showing that everyone in the room is expected to reach the standard, lets the struggle stay productive.

Do: at the first hard checkpoint, tell students that this is where most people find it hard and that it gets easier with work, then show a route forward.

Walton & Cohen, 2011; Yeager, 2024.

Growth mindset: build self-belief through the work, not the slogan

Believing that ability can grow may help students persist, but the evidence is modest. Meta-analyses find small average effects, near zero in the most rigorous studies. A national pre-registered trial found small gains for lower-achieving students, mainly where peer norms supported taking on challenge. Posters and praise for effort alone do little. Self-belief grows when students succeed at hard work with support, and can see the growth.

Do: make growth visible (first draft beside final, earlier generalisation beside revised); praise the strategy and the improvement, not the person; pair every high standard with the support to reach it.

Dweck, 2006; Sisk et al., 2018; Yeager et al., 2019; Macnamara & Burgoyne, 2023.

Three parts, and they are not interchangeable

Self-regulated learning has three components. Cognition is the strategy for doing the work: how to memorise, how to solve, how to draft. Metacognition is monitoring and directing that work: choosing the strategy, noticing it is failing, switching. Motivation is the willingness to spend either.

The order matters. A student cannot be metacognitive about strategies they do not have, which is why "think about your thinking" lands as an empty instruction so often. There has to be something to think about first.

Do: teach the strategy before asking a student to evaluate their use of it. If they cannot name two ways to approach the task, the reflection question has nothing to work on.

Education Endowment Foundation (2025), Metacognition and self-regulated learning: guidance report (v2.4.0). London: EEF.

What a learner brings before they start

Students arrive at a task already holding three kinds of knowledge about learning: knowledge of themselves as a learner, knowledge of the strategies available, and knowledge of the task. A gap in any one of them looks the same from the outside, a student who starts badly, so it is worth knowing which is missing before you step in.

Do: when a student stalls, ask which of the three is thin. Not knowing where to start is a task problem. Knowing where to start and not how is a strategy problem. Knowing both and not beginning is usually neither.

Plan, monitor, evaluate is a loop, not a sequence

The regulation cycle is planning the approach, monitoring it while working, and evaluating how it went. It is tempting to timetable these as three stages, one at the start of a unit and one at the end. The evidence describes something continuous: each pass updates the student's subject knowledge and their knowledge of themselves, strategies and tasks, which changes how they plan the next attempt.

Do: build the three moves into the middle of tasks, not only around them. A monitoring prompt halfway through is worth more than a reflection sheet after the work is finished and nothing can change.

Teach it inside the subject, not beside it

This is the finding most likely to contradict what a school currently does. Metacognitive strategies can be introduced generically, but they are applied to particular content and are best taught that way. The report is blunt about why: it is very hard to have knowledge about learning without the relevant subject knowledge, and there is little evidence for teaching these skills in isolation.

So a skills programme in tutor time, a poster of learning behaviours, or a standalone lesson on planning will do less than the same content taught inside the history essay or the titration. The skill is not general even when its name is.

Do: attach every learning skill to the unit where it is actually needed and teach it there. If a skill cannot be named against a specific task in a specific unit, it is not yet being taught.

EEF (2025), Recommendation 2.

Model your own thinking out loud

Expert thinking is invisible, and students cannot copy what they cannot see. The remedy is the teacher narrating decisions while working: what they considered, what they rejected, what made them change course. Not the finished answer, but the weighing up.

The report sets out a gradual handover: bring out what students already know, teach the strategy openly, model it, have students commit it to memory, practise it with guidance, practise it on their own, then reflect on it in a structured way. Support is withdrawn across that sequence rather than removed at once, and steps are revisited when a class is not ready.

Do: narrate one genuine decision per lesson, including a wrong turn. A model that only shows the correct path teaches that experts do not struggle, which is the opposite of what you want believed.

EEF (2025), Recommendations 2 and 3.

Can this be taught at all? A disagreement worth knowing about

Two strong sources part company here, and it is worth seeing where.

Sweller distinguishes biologically primary knowledge, which we evolved to acquire without instruction, from biologically secondary knowledge, which is why schools exist. Speech, face recognition, general problem-solving, planning, and regulating your own thinking all fall on the primary side. His conclusion is uncomfortable: attempts to teach generic skills "tend to fail, not because the skills are unimportant but because they are of such importance to humans that we have evolved to acquire them automatically without instruction". Applied to self-regulation, the same argument implies that it relies largely on primary knowledge and so may not be teachable at all.

The evidence for the first half of that is strong, and the conclusion in the second half does not follow. What the trials show is that generic skills are not passed on by direct instruction, taught as a subject, on a timetable. That is not the same as saying they cannot be developed. Sweller's own architecture says so: his borrowing and reorganising principle holds that the vast bulk of what sits in long-term memory came from other people, and that acquiring from others is itself a biologically primary skill. His randomness-as-genesis principle adds that a learner with nobody to borrow from falls back on random generate-and-test, the least efficient method available.

So the position this guide takes is Vygotsky's: children grow into the intellectual life of those around them. The capacity is not delivered by instruction. It grows through long stretches of working alongside people who already have it, thinking out loud where it can be overheard. That is mentorship rather than a course, and it explains both findings at once, why the skills programme fails and why the apprentice improves.

Do: stop treating these as content to be delivered and start treating them as practices to be joined. The teacher who reasons aloud in front of students, admits a wrong turn and revises in public is doing the thing that works. The assembly on learning behaviours is not.

Sweller, van Merriënboer & Paas (2019) for primary and secondary knowledge and the borrowing principle, with the quotation on generic skills at p. 271 and the passage on self-regulation at p. 284; Vygotsky (1978), Mind in Society, for internalisation. The EEF's self-talk continuum, teacher prompt to paired talk to inner voice, is this same process described from the classroom end.

Talk is where strategies get tested

Metacognitive strategies improve when they are said out loud and met with disagreement. Prompts do most of the work: what do you already know, what makes you think that, could it be something else. Over time the questions move from the teacher's mouth to the pair to the student's own head, which is the point.

Do: post the prompts and use them by name, so students can eventually use them without you. Assess whether the moves were used, not just whether talking happened.

EEF (2025), Recommendation 4. See also Access & UDL on teaching discourse moves explicitly.

Challenge is what makes it possible, not the reward

A task a student can finish on autopilot produces no metacognition, because nothing forces a choice of strategy. Difficulty is what creates the moment where a student must decide how to proceed, notice it is not working, and change. Removing struggle removes the occasion for the thinking.

The limit is working memory. Difficulty that comes from the concept is productive. Difficulty that comes from holding too much at once is just overload, and it drags everything down. The distinction is between a demanding idea and a cluttered task.

Do: pitch tasks so students must choose an approach rather than run a known one, and strip out the load that is not the point, unclear instructions, cluttered layout, four things to remember at once, so the difficulty that remains is the difficulty you intended.

EEF (2025), Recommendation 5. See also Motivation & productive struggle.

What this looks like at each band

The cycle is the same throughout. Who voices it changes.

Early Primary (K–2). The teacher is the metacognition, said aloud on the children's behalf: "I am going to check whether that worked." Expect strategies to be named and used, not chosen. A choice between two known approaches is realistic. An open choice is not.

Upper Primary (3–5). Students can hold a plan across a task and evaluate against it if the criteria are supplied. Monitoring is the weakest of the three and needs prompting mid-task rather than trusting.

Middle School (6–8). Capable of really choosing a strategy, and at the same time most exposed to being seen choosing wrongly. Self-talk goes underground here. Private planning and low-stakes rehearsal matter more than at any other stage.

High School (9–12). Can run the full cycle independently, and will still misjudge how well it went. The work shifts from teaching the cycle to correcting the self-assessment inside it.

Do: match the demand to who can carry it. Asking a seven-year-old to select a strategy, or a seventeen-year-old to follow one without question, both waste the same opportunity.

Students are poor judges of their own learning

The most useful finding for senior students. Learners routinely misjudge how well they have learned something and which strategy was responsible. They underestimate what a task will take, a version of the planning fallacy, and they prefer massed practice, cramming, over spaced practice, because cramming feels more productive while it is happening.

This means self-evaluation cannot be trusted as feedback on its own. It has to be checked against something external, and the gap between the two is itself worth teaching.

Do: have students predict a result before they see it, then compare. The gap teaches more than the mark does, and it is the only reliable way to improve the judgement itself.

EEF (2025), Recommendation 6; Zimmerman (2002), cited there. See also How learning happens on spacing and retrieval.

What this is not

It is not a learning-to-learn course, a skills carousel, or a display of learning behaviours. Those are the isolated teaching the evidence does not support. It is not a reflection sheet at the end of a unit, which arrives after every decision it might have informed. And it is not a substitute for teaching the subject well: there is nothing to regulate until there is something being learned.

Critical thinking, the working definition

Critical thinking means reflecting on your own thinking to improve it. It involves being self-directed, disciplined, aware, and willing to correct yourself. There are two parts: skills and character. The skill side is about checking if your reasoning is clear, accurate, relevant, logical, and fair. The character side includes traits such as humility, courage, empathy, integrity, and trust in reason, as well as awareness of personal and group biases. The skills alone are not enough. Students also need these traits to be real critical thinkers.

Do: teach the standards as questions students ask of their own work, not as a rubric applied after the fact.

Paraphrased from Paul & Elder, Foundation for Critical Thinking (criticalthinking.org).

What critical thinking looks like by band

For grades K–2, critical thinking means noticing, sorting, and answering "how do you know?" with something they can see or count. In grades 3–5, students give reasons with evidence, check claims against sources, and spot one-sided stories. In grades 6–8, they evaluate sources, consider two perspectives or variables at once (with support), and start to critique their own first ideas. By grades 9–12, students assess their own reasoning without prompts, fairly reconstruct arguments they disagree with, and use intellectual empathy for systems as well as people.

Do: keep the same standards from K to 12 and raise the demand. The questions stay, and what counts as a good answer rises.

Paul & Elder; developmental staging after Kuhn, via Kuther (2023).

Ethics, the working definition

Ethics asks, "Of all the ways I could act, which one should I choose?" The answer comes from three ideas: values (what is good and worth protecting), principles (what is right and how we should act), and purpose (the reason behind our actions). Ethics means always thinking about, discovering, and standing up for these three, even when it is difficult.

Do: when a unit touches an ethical tension (fairness, sustainability, a design trade-off), name which of the three is in play before asking students to judge.

After The Ethics Centre (ethics.org.au).

How moral reasoning develops (Kohlberg, read with Gilligan)

Kohlberg's research describes moral development as a series of stages. In the pre-conventional stage, common in young children, right and wrong are based on avoiding punishment or making fair trades. The conventional stage, typical in teenagers and most adults, is about meeting expectations and following rules. The post-conventional stage, which is rare, involves thinking about social contracts and personal principles. These stages rank types of reasoning, not people. Gilligan added that caring about relationships is another valid way to reason about morals, not a lower stage.

Do: pitch ethical discussion one step above where students currently reason, "what would a fair rule be?" before "what principle is at stake?", and credit care reasoning as fully as justice reasoning.

Kohlberg, 1981; Colby et al., 1983; Gilligan, 1982.

Questions that make ethics teachable

Ethical reasoning becomes concrete when it is put as questions and translated for the age. K–2: "Is it fair? Is it kind? What if everyone did it?" 3–5: add the sunlight test, "Would I be happy for everyone to know what I chose, and why?" 6–8: add consequences and character, "who is helped, who is harmed?", "what habit does this build in me?" 9–12: add the universal and the non-negotiable, "would this reason hold if everyone in my position used it?", "is a line being crossed that no outcome justifies?"

Do: put one question on the wall per unit and return to it at the landing, so students judge the unit's ethical tension with the same question they met on day one.

Question set adapted from The Ethics Centre's decision guide.

What a concept is, and the two kinds

A concept is a mental representation of a category: a class of objects, events, or relations that share common properties. It transfers, it is largely timeless, and it is usually one or two words. Before defining one, decide which of two kinds it is, because that changes how you define it. Categorical concepts have membership criteria (mammal, acid, democracy) and are defined by listing critical attributes. Relational concepts are structures that hold between things (equilibrium, interdependence, predator) and are defined by naming the relation and its terms, not by listing features. Putting an attribute checklist on a relational concept is a common cause of fuzziness.

Do: label each concept categorical or relational before writing its definition, and define relational concepts by the relation they name.

After Erickson & Lanning (structure of knowledge); concept as category after Ashcraft & Radvansky; concept attainment after Taba and Bruner.

Three standards: clarity, accuracy, precision

These are three different tests, and teams often pass two while failing the third. Clarity: can you state the concept's meaning at all? Accuracy: does that meaning match how the discipline actually uses the term? Precision: is it specific enough to exclude near-misses? Energy defined as "a stuff that gets used up" can be perfectly clear and precise and still fail accuracy, because the physics concept is a conserved quantity that is transferred. Accuracy is the standard nobody's checklist catches, and it answers to the discipline's logic, not to what the team agrees on.

Do: check a working definition against the discipline (what an economist, an ecologist, a historian would recognise), not against agreement in the room.

Intellectual standards after Paul & Elder, Foundation for Critical Thinking.

Precision through contrast, and the removal test

Precision is built by contrast. For a categorical concept, name the critical attribute and the near-miss non-example that lacks it (predator versus scavenger: both feed on other organisms, only one kills). For a relational concept, the test splits: name the terms of the relation and ask what breaks when one is removed. Predation fails when there is no prey, and that failure of the structure, not a list of attributes, is the non-example.

Do: for every relational concept in a unit, run the removal test, "what breaks in the system when this is taken away?", and use the answer as your non-example.

Contrast technique after concept attainment (Taba, Bruner); relational framing after Erickson & Lanning.

A worked example: predator and ecosystem

Predator is relational, a role rather than a taxon: the same organism can be predator and prey. Definition: an organism that kills and consumes other organisms (prey) to obtain energy and matter. Critical attribute: obtains energy by killing and consuming another living organism. Non-examples: scavenger, parasite, decomposer. Ecosystem is a systems concept, defined by flows rather than an inventory: all organisms in an area plus the abiotic environment, bound by energy flow and matter cycling. The significance of predator to ecosystem is what makes them inseparable: the thing that defines a predator is itself a working part of an ecosystem. "Obtains energy by consuming prey" is a trophic link, the mechanism by which energy moves up the system and populations are kept in check. Remove the predator role and the system property changes (a trophic cascade). The resulting generalisation, "Predators regulate the flow of energy and the size of populations, sustaining the dynamic balance of an ecosystem," survives the truism check because top-down versus bottom-up control is a real scientific debate.

Do: before writing a generalisation, state the relation between its two concepts, then check the sentence against its opposite. If no reasonable person could disagree, one of the concepts is doing no work.

Trophic cascades after Paine (keystone species) and Estes (sea otters); worked for planning use.

Say it plainly, or students cannot carry it

A generalisation is the one sentence students have to hold, restate and take into a situation you have not seen. If they cannot say it back from memory in their own words, it will not travel, however precise it looks on the unit plan.

The rule about relationship verbs pulls against this without meaning to. The verbs it rewards tend to be academic ones, mediates, regulates, governs, accelerates, while the verbs it rules out are the everyday ones. Left alone, that pushes the writing towards density exactly where density costs most. The repair is small: the strong list also holds plain verbs, shapes, builds, limits, protects, links, leads, reduces, prevents. Use those unless no plain verb names the mechanism as exactly.

This lands hardest on multilingual students, and it is easy to misread. A student can hold the concept completely and still be unable to reconstruct fourteen words carrying three abstract nouns and a subordinate clause. What you then see looks like a student who does not understand the idea, so the teaching goes to the wrong problem. See Access & UDL on separating language from understanding.

Sophistication is not sentence length. Levels of sophistication raise the conceptual demand, not the grammar. An advanced generalisation can and usually should be written in plain words. A hard sentence is a drafting problem, not evidence of depth.

Do: before a generalisation goes on the wall, ask a student to say it back without looking. If they cannot, shorten the sentence before you simplify the idea. Those are different edits, and only one of them costs you the understanding.

Definition drift, the alignment check

Drift is the meaning of a concept quietly changing across a unit: the sense it carries in the Statement of Inquiry is not the sense in the generalisations, which is not the sense the assessment rewards. Each use looks fine on its own, so no single read of the unit plan catches it. In assessment terms this is a construct validity problem: the thing assessed is no longer the thing taught. In planning terms it is the most measurable sign of misalignment you have.

Do: take every occurrence of a key concept across the unit (inquiry statement, generalisations, learning experiences, assessment) and confirm that one agreed definition survives all of them.

Construct validity after Messick; alignment as a planning check.

What conceptual precision looks like by band

For grades K–2, precision means sorting by one attribute and answering "what makes it a ___?" with something they can see. In grades 3–5, students separate examples from non-examples and can say why the odd one out does not belong. In grades 6–8, they relate two concepts and use the removal test to explain how one depends on the other. By grades 9–12, students evaluate a definition against the discipline's use and can notice when a term has shifted meaning inside their own work.

Do: keep the same move from K to 12 and raise the demand. The contrast stays, and the sophistication rises.

Developmental staging consistent with the bands above; after Kuhn via Kuther (2023).

Etymology as a doorway: the gap is the lesson

A word's roots are a neutral anchor from outside the room: no one owns them, so they take the heat out of teachers arguing for their own private sense of a term. Often the root hands over the critical attribute and shows the concept's kind (predator traces to Latin praeda, "prey", which is why it is relational; concept traces to con- plus capere, "to grasp together", the categorising move itself). The trap is the etymological fallacy: origin is not current meaning (atom meant "uncuttable", energy "being at work", decimate "kill one in ten"). Treat the root as a guess about the core, to be tested against how the discipline uses the word now. The richest part is the gap. Where root and current meaning part company, students watch an idea being corrected by evidence, a live demonstration that understanding gets rebuilt as new information updates what we assumed.

Do: use the four beats: root and original sense, current meaning in the discipline, what changed and why, so what. And check the word history against a real source, because a made-up etymology used to teach revision would defeat itself.

Origins via standard etymological reference; fallacy caution standard; revision framing after Paul (intellectual humility).

Prior knowledge and the diagnostic pre-test

Defining a concept precisely is where prior knowledge gets dealt with, on two levels. The teacher's own starting point comes first: most teachers do not know the etymology of the concepts they define, so the roots show up the teacher's gap and make them a co-learner. The students' starting point is what the precision work is built to probe. They arrive with everyday ideas (a predator is a scary animal, energy is a stuff that gets used up) that differ from the disciplinary concept, and that gap is what an ungraded diagnostic pre-test exposes. The critical attributes and non-examples you just wrote are the misconception probes: "is a vulture a predator, and why?" tests the kill-versus-scavenge attribute. Keep it ungraded on purpose. Its job is to reveal thinking, not to grade it, and a grade destroys the honest starting point. The result sets the case studies: choose and order cases that confront the ideas students actually hold, then probe again at the landing so the growth, understanding rebuilt, can be seen.

Do: open the lesson cascade with a short diagnostic drawn from the concept's non-examples, keep it ungraded, and let what it reveals choose your case studies.

Prior knowledge and diagnostic assessment after standard formative-assessment practice (Wiliam); reveals-thinking stance consistent with the assessment sections above.

The logic we want students to develop

The aim is not to hold the facts flat but to look for the structure that organises them. A student running the intended logic treats facts as instances, asks what the instances share (forming a category, a concept), asks how categories relate and what the relation explains (forming a generalisation), and tests it against new cases. That is the discipline's own organising logic, causal and systemic in science, made their own. The concept and the generalisation are the endpoints the team fixes and holds constant to prevent drift. The logic is the road the student walks to earn them. Hand over the organised structure and you delete the thinking you were trying to grow.

Do: organise the facts so the intended pattern can be found, and guide with further questions. Let each student reach the generalisation rather than receive it.

Facts-to-structure logic after Paul (logic of ideas) and Erickson & Lanning; schema construction after Oakley, Rogowsky & Sejnowski.

Learning styles (VAK: visual / auditory / kinaesthetic)

Myth

Every child has a fixed learning "style", and matching your teaching to it raises achievement.

Reality: students have learning preferences, but research shows that matching teaching to a supposed learning style does not improve results. What matters is matching the teaching method to the subject, for example, using maps for geography because it is a spatial subject, not because a student is a "visual learner".

Pashler, McDaniel, Rohrer & Bjork, 2008; Willingham, 2009; Tokuhama-Espinosa, 2018.

Left-brained vs right-brained learners

Myth

People are logical "left-brained" or creative "right-brained" types, and teaching should match the dominant side.

Reality: the two sides of the brain have some special roles, but they work together on almost every task. Brain scans show there is no overall "dominance" that divides people into types. The idea is a metaphor, not a scientific fact.

Nielsen et al., 2013; Howard-Jones, 2014; Tokuhama-Espinosa, 2018.

"We only use 10% of our brain"

Myth

Most of the brain lies dormant, waiting to be unlocked.

Reality: we use almost all parts of our brain throughout the day. Brain scans show activity everywhere, and damage to any area has an effect. There is no unused 90% waiting to be activated.

Tokuhama-Espinosa, 2018; OECD, 2007.

"Digital natives" who learn better while multitasking

Myth

Growing up with screens makes young people able to attend to several streams at once and learn well while multitasking.

Reality: everyone has limited attention. What we call "multitasking" is actually quickly switching between tasks, which makes it harder to remember and understand information. There is no special "digital native" way of thinking; everyone faces the same attention limits.

Kirschner & De Bruyckere, 2017.

"It's all decided by age three" (critical-period overreach)

Myth

Miss the early windows, and the door to learning closes.

Reality: there are sensitive periods for some skills, like learning the sounds of a first language or developing binocular vision, but the brain remains flexible and learning can happen throughout life. Most things taught in school do not have a strict biological window, and learning later on is still effective.

Howard-Jones, 2014; Tokuhama-Espinosa, 2018; OECD, 2007.

Brain Gym, "water crosses the midline", and "boys' vs girls' brains"

Myth

Specific movements rewire the brain for learning; a drink of water "crosses the midline"; boys' and girls' brains make them suited to different subjects.

Reality: there is no solid evidence that Brain Gym works as claimed, and average brain differences between boys and girls do not decide who can succeed in maths, languages, or art. Drinking water, moving, and getting enough sleep do help learning, but for normal physical reasons, not because of these specific claims.

Howard-Jones, 2014; Tokuhama-Espinosa, 2018.

Multiple intelligences as planning categories

Myth

Each student has a dominant intelligence (musical, kinaesthetic, interpersonal, etc.), and units should be planned around it.

Reality: researchers, and even Gardner himself, agree there is no strong evidence for eight separate intelligences. Abilities tend to be linked (the g factor), and planning lessons by intelligence type has the same problems as learning styles. The idea that everyone can contribute and succeed in different ways is a valuable belief, but it is not a scientific basis for planning instruction.

Waterhouse, 2006; Sternberg, 1994; Gottfredson, 2004.

What a milestone is, and what it is not

A milestone sits at each developmental stage of the unit where students are expected to form a generalisation. It checks conceptual understanding: can the student state how the concepts in this pairing relate, and back it with the cases? It is not a test of transfer. Far transfer waits for the summative task and the conclusion of the unit.

A milestone only earns its place if there is time left to act on what it shows. Plan the next lesson so that the feed-forward gets used.

Do: one milestone for each concept pairing, each with its evidence, its criterion and the feed-forward comments you will reach for. Write the comments while you plan, not while you mark.

See Feedback only works if there is time to act on it and Students are poor judges of their own learning.

Early Primary (K–2)

Milestone evidence. The idea said, drawn, sorted or built, never only written. A child sorts pictures into two groups and tells you the rule, or points to the case that does not belong and says why. Listen for "because". Record what was said, since the talk is the evidence.

Feed-forward sentence starters. Keep it to one next step, said aloud and shown.

• "You told me ... . Next time, show me which picture proves it."
• "You noticed ... . Now find one more that fits your rule."
• "Try saying it with 'because': ... because ... ."

Far transfer. The same idea met in a new, concrete setting: a story, an object or a place the class has not used. "Does our idea work here too? Show me."

Upper Primary (3–5)

Milestone evidence. A sentence or two linking the two concepts, built from a frame if one is needed, and checked against two cases. Comparing cases side by side is where the generalisation gets worked out, so collect the comparison as well as the sentence.

Feed-forward sentence starters. One step, tied to one criterion the student can see.

• "Your sentence links ... and ... . Add a case that shows how."
• "This works for one case. Test it on the second case and change a word if you need to."
• "Look at the criterion. Your next step is ... ."

Far transfer. A new case the class has not studied, with the same concepts at work. Students explain whether their generalisation holds and what they would change.

Middle School (6–8)

Milestone evidence. A generalisation that says how one concept affects the other, supported by evidence from at least two cases. Give private rehearsal first (a notebook, a draft, a quiet pair) before anything is said in public, because being seen to be wrong costs more at this age than at any other.

Feed-forward sentence starters. Given privately, framed as the next move rather than a verdict.

• "You have shown that ... . Now say how ... changes ... ."
• "Your evidence fits case one. What does case two add, or challenge?"
• "Your next step against the criterion is ... . Try it in your draft before we share."

Far transfer. An unfamiliar context, possibly from another subject, where students decide for themselves which concepts apply and test their generalisation against it.

High School (9–12)

Milestone evidence. A qualified generalisation (for example, using "often", "can" or "under these conditions"), tested against a counter-case and written in the discipline's conventions. Students assess their own draft against the criteria first, then compare it with yours.

Feed-forward sentence starters. Aimed at the gap between the student's self-assessment and the evidence.

• "You judged this at ... . The evidence shows ... . What would move it?"
• "Your generalisation holds for ... . Find the condition where it breaks, and qualify it."
• "Against the criterion, the next step is ... . Show it in the next draft."

Far transfer. A new, authentic problem with a real audience, where students choose the concepts, justify the choice and say where their generalisation stops working.

What does not change across bands

The criteria and the standard are the same for every student. A younger or less confident student shows the same understanding in a different form, not a smaller understanding. Grades are decided against each programme's own criteria at the summative point, where students get both feedback on the task and feed-forward for the next unit.

Do: use these starters as hints when you write the milestone comments, then make each one specific to your pairing and your cases.

Credit: Black and Wiliam (1998); Sadler (1989); Hattie and Timperley (2007); Wiliam (2011). Summary in our own words.