Tips, Tricks and Resources for Teachers 4 Easy Ways to Help Students Catch AI Mistakes TeacherTalk.site

4 Easy Ways to Help Students Catch AI Mistakes

Four proven ways to teach primary students to check AI output: plant an error, rerun and compare, ask one question every time, and build knowledge first.

Teaching in the Digital Age · AI Literacy

4 Proven Ways to Spot AI Mistakes in Your Classroom

AI hands your students a confident answer every time. The skill worth teaching is knowing when to doubt it.

⏱ 6 min read · By Matt Faulkner · Pedagogy / Gen AI

Last week’s look at the PISA device myth ended on one finding worth a second look. Students taught to check AI output outperformed students who used AI to write their drafts. Same tool, different job for the student.

That finding raises a practical question. What does it look like to teach a Year 5 class to check AI output on a Tuesday afternoon? Most advice stops at “teach critical thinking.” That gives you a goal, not a lesson.

AJ Juliani’s recent piece, Teach the Frontier, offers a sharper starting point. Every serious plan for keeping AI honest assumes someone has the knowledge and judgement to say “that’s wrong.” By 2040, those people will be the students sitting in your room right now. Here are four ways to start building that judgement in a primary classroom.

Way 1: Build Knowledge Before You Check AI Output

Start with the step most AI advice skips. A student has nothing to check an answer against without knowing the topic first.

Daniel Willingham made this case long before chatbots arrived. Critical thinking, he argued, is not a free-floating skill you teach once and apply everywhere. It depends on what you know about the subject in front of you. In other words, a student who thinks critically about fractions might be helpless with a question about volcanoes.

Picture a Year 4 HASS class asking an AI tool about the First Fleet. The answer says the convicts settled at Botany Bay. A student who has studied the unit knows the fleet moved on to Sydney Cove. A student who hasn’t will copy the error into their book, neatly and confidently.

A student only catches a wrong answer about something they already know. Knowledge is the checking tool.

Juliani puts it plainly: judgement sits on top of knowledge. As AI tools get more capable, background knowledge matters more, not less. For this reason, a knowledge-rich unit is your first and best protection against AI errors.

Way 2: Plant an Error, Then Check AI Output

The simplest place to start is an AI answer you have checked yourself. Before the lesson, run a prompt, read the response, and make sure it contains one or two errors. Some will appear on their own. Otherwise, edit one in.

Run the prompt yourself rather than handing students a login. Many AI tools set age limits above primary school. The Australian Framework for Generative AI in Schools also asks schools to protect student data. Displaying the output on the board avoids both problems.

For example, ask an AI tool to explain to a Year 5 student how to compare decimals. Then plant a classic misconception: “0.25 is bigger than 0.5 because 25 is bigger than 5.” Tell the class there are two errors, and give pairs five minutes to find them. The misconception you plant is often the one half the room already holds. As a result, the error hunt doubles as a diagnostic.

Way 3: Rerun the Prompt to Check AI Output

Next, run the same prompt twice and put both answers side by side. Ask students one question: where do these answers disagree?

Disagreement is a signal. If one answer gives a river’s length in one figure and the second gives another, at least one is wrong. Students then go to the atlas, the textbook, or a trusted website to settle it. That trip to a second source is how students learn to check AI output.

Even so, teach the other half of the lesson too. Two AI answers that agree can still both be wrong, because they come from the same tool. Agreement tells you the tool is consistent. Only a source outside the tool tells you it is right.

Way 4: Make One Question a Routine

Juliani asks his own children one question every time AI gives them an answer: “How would you know if it was wrong?” His hope is that one day they ask it without him.

That repetition is the point. A routine question works the same way as the higher-order questioning you already use, except students eventually take it over. It also pushes back on the problem we covered in the AI confidence trap. AI tools tend to agree with the user, so students need a habit of disagreeing with the tool.

🔍 Five Questions for Every AI Answer

How would you know if this was wrong? The anchor question. Ask it every time.
Where else could you check this? Name a source outside the AI tool.
Does this match what we learnt in class? Sends students back to their own knowledge.
Which part sounds sure but shows no evidence? Confident tone is not proof.
What did it leave out? A harder one for Years 5 and 6.

Print the list as a bookmark or a poster near the board. In practice, the first question carries most of the weight. The other four give older students somewhere to go next.

Keep Most Tasks AI-Free, and Say Why

All four ways share one condition. Students need plenty of work where AI plays no part at all. Juliani’s advice is to keep most tasks AI-free and to tell students the reason. The hard thinking stays with the student, because that thinking builds the knowledge they will later check AI output against.

This links straight back to productive struggle. A student who hands the hard part to a chatbot skips the struggle, and skips the learning with it. Telling them so is honest, and most Year 6 students will follow the logic.

The Framework backs this up. Guiding statement 1.2 asks schools to teach students how AI tools work, including their limitations and biases. Statement 1.4 asks for AI use that supports critical thinking rather than restricting it. Statement 1.5 asks every task to state clearly whether AI belongs in it. The four ways above give you a classroom version of all three.

🗓️ A Four-Week Starter Plan

Week 1: Teach the content. No AI. Tell students why.
Week 2: Plant two errors in an AI answer on that content and run the hunt.
Week 3: Rerun one prompt twice and settle the disagreements with a second source.
Week 4: Hand pairs a fresh AI answer and the five questions. Step back and watch.

Your turn

Try the planted-error hunt with your class this term. Which error did your students catch first, and which one slipped past everyone? Tell us in the comments.

References

Australian Government Department of Education. (2023). Australian framework for generative artificial intelligence (AI) in schools. education.gov.au

Juliani, A. J. (2026, September 18). Teach the frontier. Next Gen Schools. ajjuliani.beehiiv.com

Juliani, A. J. (2026, September). What the PISA results really show us about technology and curriculum. AJ Juliani. ajjuliani.com

Willingham, D. T. (2007). Critical thinking: Why is it so hard to teach? American Educator, 31(2), 8–19.

5 Proven Ways to Beat the PISA Device Myth

Teaching in the Digital Age · Tech-Scepticism

5 Proven Ways to Beat the PISA Device Myth

The PISA data doesn’t say what the headlines say it does. Singapore and Estonia prove the story was never really about the devices.

⏱ 6 min read · By Matt Faulkner · Teaching with Technology / Pedagogy

Every few months, the PISA device myth does the rounds again: PISA scores are falling, and screens are to blame. Ban the phones, lock up the laptops, go back to exercise books. It is a tidy story, and it is wrong.

The decline is real. Between 2015 and 2025, OECD average reading scores dropped by 28 points and maths by 22. In the United States, reading fell to 490. Something happened. But when you look at which countries actually held their ground, the “devices did this” theory falls apart fast.

Singapore and Estonia are two of the most tech-saturated school systems in the world, and both are still at or near the top of the table. If screens caused the decline, they should have crashed hardest. They did not. That is the thread worth pulling, and it leads somewhere more useful than a device ban.

The Decline Is Real. The Explanation Isn’t

It is tempting to draw a straight line from “more devices in classrooms” to “lower test scores,” because the timing lines up. Correlation is doing a lot of heavy lifting in that argument. Plenty of things changed for students between 2015 and 2025: attendance patterns, post-pandemic disruption, curriculum shifts, teacher shortages. Devices were one variable among many, not the obvious cause.

Jared Cooney Horvath’s The Digital Delusion makes a useful case against uncritical edtech adoption, and it is worth reading. But even Horvath’s critics point out the same weakness that undermines the device-blame headlines: the data is correlational. It shows devices were present during a decline. It does not show devices caused it.

Singapore and Estonia Broke the PISA Device Myth

Here is the part the headlines skip. Singapore posted 535 in reading and 563 in maths, among the strongest results anywhere, in a school system built around one-to-one devices and heavy digital integration. Estonia sits in the top ten in Europe on the same measures, with a similarly tech-saturated approach.

If device density were the cause of decline, these two systems should be the worst-hit, not the best-performing. They are not exceptions to explain away. They are the evidence that breaks the PISA device myth wide open.

It was never a question of whether schools use technology. It is a question of what students are actually doing with it once it is in their hands.

It’s Not Whether. It’s How

This is where the research gets genuinely practical, because it stops being about policy and starts being about pedagogy. AJ Juliani’s analysis of the PISA data draws a sharp line between learning-use and leisure-use of devices. Students using devices for moderate, purposeful learning tasks outperformed both heavy users and non-users. Students using devices mainly for browsing, gaming or passive content consumption did not.

The same split shows up in how students use AI. Students who used AI to evaluate and critique their own draft work outperformed students who used AI to generate the draft for them. We wrote about that exact pattern in the AI confidence trap, where students who let AI do the thinking lost the very judgement PISA is trying to measure.

This is the part the PISA device myth gets backwards. Put simply: a device in a student’s hand is not the variable. What the task asks them to do with it is the variable. That distinction should sound familiar, because it is the same one we made when spotting fake engagement in the classroom. Busy is not the same as thinking. Present is not the same as using well.

England Didn’t Need to Ban Screens

England offers the clearest counter-example to “fewer devices fixes it.” England’s PISA gains over the same period came from a curriculum-first, phonics-and-fundamentals strategy, not a device rollback. Schools sharpened what students were taught and how it was sequenced. Technology was neither the villain nor the hero. It was mostly beside the point.

That should reassure anyone feeling pressure to rip out every laptop cart before term four. The lever that actually moves outcomes is instructional quality: what you ask students to do, how you sequence it, and whether the task demands real thinking. We have made this case before around higher-order questioning, and the PISA data backs it up again from a completely different angle.

🎯 Five Questions Before You Blame the Tablet

Purpose: what is this device actually for in this lesson, right now?
Thinking versus completion: does the task demand judgement, or just clicking through?
Interaction: is the student engaging with content, or just consuming it?
Opportunity cost: what would students be doing instead, and is that better?
Professional judgement: would you defend this use of tech to a sceptical colleague?

These five questions, adapted from Amy Kelly’s edtech audit, work as a quick check before any lesson that leans on a screen. If you can answer all five with confidence, the device is probably earning its place. If you are stuck on two or three, that is worth a rethink, not a ban.

What This Means for Monday Morning

None of this is an argument for more screen time, and it is not a reason to keep repeating the PISA device myth in the staffroom. It is an argument for better screen time, and for resisting the urge to solve a complex problem with a blunt policy. Nine times out of ten, the fix is not “less tech.” It is “clearer task.”

Before you write your next unit plan, ask whether the device is carrying the thinking or the student is. If the answer is the device, the task needs redesigning, whether or not you keep the laptops in the room.

Your turn

What is one task in your current unit where a device could be doing the thinking instead of your students? Drop it in the comments, we would love to help you redesign it.

References

Horvath, J. C. (2025). The Digital Delusion: How Classroom Technology Harms Our Kids’ Learning, and How to Help Them Thrive Again. Penguin Random House.

Juliani, A. J. (2026, September). What the PISA results really show us about technology and curriculum. AJ Juliani. ajjuliani.com

Kelly, A. L. (2026, July). 5 questions elementary teachers can ask to evaluate their edtech use. Edutopia.

5 Proven Ways to Spot Fake Engagement in Your Classroom

Pedagogy · Edtech Habits

5 Proven Ways to Spot Fake Engagement in Your Classroom

The app says every student is engaged. It has no way of knowing whether anyone is thinking.

⏱ 5 min read · By Matt Faulkner · Pedagogy / ICT

A classroom of students each looking down at a phone at their own desk, none talking to each other

Duolingo has a problem, and the problem isn’t the one you’d expect. Neuroscientist Jared Cooney Horvath calls it “all engagement, little retention.” Students tap through lesson after lesson, the streak climbs, the app feels alive. Ask them to hold a conversation in the language a week later, though, and the words are gone.

That gap between engagement and thinking isn’t limited to language apps. Horvath’s The Digital Delusion pulls together 398 meta-analyses covering 21,000 studies and lands on an average effect size of +0.29, well short of the +0.40 threshold researchers treat as a meaningful gain. PISA data tell a sharper story: students using computers six or more hours a day scored 66 points lower than non-users, a slide from the 50th percentile to the 24th. TIMSS 2019 found daily computer users 41 points behind in maths and 51 points behind in science, compared with occasional users.

None of that means every app wastes a lesson. So the signs of clicking, not thinking, stay easy to miss in a busy classroom. Edutopia’s Amy Kelly built a five-question audit for exactly this problem. Here’s what each question looks like as a symptom in your own room.

Sign 1: Nobody Can Tell You Why

Watch what happens when a student asks why the class is using a particular app today. If the honest answer is “we always do this on Fridays,” the tool has slipped from a teaching decision into a habit. Kelly’s first question, drawn from UNESCO’s 2023 Global Education Monitoring Report, is blunt: what is this helping students do? Not what does it cover, or what does the timetable say. What does it do for learning, this lesson, for these students?

The fix takes ten seconds. Before you open the app, finish this sentence out loud: “We’re using this because it helps you ___.” If you can’t finish it, neither can your students, and the fifteen minutes probably belongs to something else.

Sign 2: Finishing Fast Is Fake Engagement, Not Learning

A finished activity and a learned one aren’t the same thing, and fast finishers are the first clue. Horvath’s “all engagement, little retention” line describes exactly this pattern: students clear level after level, the badges pile up, and none of it survives contact with a test or a conversation a week later. Kelly’s second question separates completion from thinking, asking whether students analyse, explain, create, discuss, or connect ideas, rather than click and select their way to the next screen.

In practice, the swap is small. Replace “did you finish?” with “what changed in your thinking?” and watch how many students go quiet. That silence is the data you were missing.

Try This Tomorrow

The moment the app closes, ask one question that needs an explanation, not a score: “What was the hardest part, and why?” A student who can’t answer probably wasn’t thinking. A student who can just showed you the app worked.

Sign 3: The Room Goes Quiet in the Wrong Way

Some quiet means deep focus. Other quiet means twenty-five students in headphones, each alone with a screen, learning nothing from each other. Kelly’s third question checks whether a tool increases interaction or reduces it, because learning grows through discussion, feedback, questioning, and shared experience with other people, not through a device alone.

By contrast, a well-chosen tool creates something worth talking about. If your students haven’t turned to a neighbour in the last twenty minutes, the screens have taken over a job that belongs to the room.

Sign 4: You Can’t Name What You Skipped to Fit It In

Every block of screen time replaces a block of something else. Horvath’s advice, cited directly in Kelly’s audit, asks teachers to name “what may be displaced along the way”: discussion, hands-on work, sustained reading, handwriting, peer collaboration, productive struggle. None of that shows up on the lesson plan. All of it shows up in what students can actually do six months later.

At the same time, this is the easiest sign to catch, because it only needs one honest question before you add a tech block to the week: what am I trading this for, and is that a fair trade?

“All engagement, little retention.” Horvath’s verdict on lessons that feel alive in the moment and leave nothing behind by Monday.

Sign 5: You Forgot You Still Get a Say

Mandatory platforms and district logins can make tech use feel automatic, like a decision someone else already made. Kelly’s fifth question pushes back on that. Teachers still choose when students collaborate versus work independently, when a conversation replaces a digital response, and when the best move is to close the laptops and step away from the screen entirely.

Even so, that choice only gets used if you remember it exists. The first four signs tell you something has gone wrong. The fifth one reminds you that you’re allowed to fix it.

The Five-Second Version

Before your next tech-based lesson, run it past five questions: What’s the purpose? Thinking or completion? More interaction or less? What’s it replacing? And where’s my decision still mine to make? Five signs, five questions, one habit worth building.

None of this argues for pulling the plug on every app in the building, any more than the AI confidence trap argues for banning AI outright. Singapore and Estonia sit at the top of the PISA tables while running some of the most tech-saturated classrooms in the world, so the device itself isn’t the villain here. Next time, we’ll look at what the PISA headlines actually say about technology and test scores, and what they leave out.

References

Kelly, A. L. (2026). 5 Questions Elementary Teachers Can Ask to Evaluate Their Edtech Use. Edutopia.
Horvath, J. C. (2026). The Digital Delusion: How Classroom Technology Is Making Our Kids Dumber.
UNESCO (2023). Global Education Monitoring Report.

4 Proven Moves to Stop AI’s Confidence Trap

Generative AI will agree with your students before they’ve earned it. Here are four classroom moves that build the confident humility AI can’t fake for them.

Pedagogy · Mindset

4 Proven Moves to Stop AI’s Confidence Trap

AI will always agree with your students. Confident humility is what happens when they learn not to need that.

⏱ 6 min read · By Matt Faulkner · Pedagogy / Gen AI

Primary school students working through their own thinking in class, the kind of moment that builds confident humility rather than borrowed certainty.

In the last post in this series, we looked at what makes a growth mindset real for students rather than a poster on the wall. This post picks up the same thread from a different angle. A student can believe they’re capable of learning and still be wrong about what they currently know. In other words, that gap, between confidence and accuracy, is where confident humility lives. And right now, it’s under more pressure than it’s ever been.

Educator John Spencer named the skill well: confident humility means staying confident in your ability to learn while staying aware that what you currently know might be incomplete. It sounds like a small distinction. In a classroom shaped by generative AI, it’s becoming one of the most important things you teach.

The Ladder Became a Maze

The old career pitch was simple. Work hard, get the qualification, climb the ladder. That said, that ladder is largely gone. In its place is a maze of uncertainty, where your students will move between being the expert in the room and the novice in the room many times over, often in the same year. As a result, a maze doesn’t reward certainty. Instead, it rewards a willingness to be wrong, notice it, and adjust. That’s confident humility, and most of our assessment systems were never built to teach it.

The Overconfidence Loop

Here’s the part that should concern every teacher using AI tools in their planning or handing them to students. For example, a 2026 study published in Science found that leading AI chatbots affirmed users’ actions roughly 49% more often than a human would, especially in situations involving poor judgement or outright error. As a result, participants who received that affirmation became more convinced they were right and less willing to reconsider, after a single exchange.

Confident humility isn’t the absence of confidence. It’s confidence that keeps asking what would change its mind. Generative AI, left unchecked, quietly removes the second half of that sentence.

That’s the overconfidence loop. A student floats a half-formed idea, the AI validates it, the student’s confidence grows, and no friction ever arrives to test whether the idea deserved that confidence. The right AI prompts can push back against this, but only if a student already knows to ask for pushback. In other words, that’s a taught skill, not a default setting.

What Your Assessment System Already Taught Them

This isn’t only an AI problem. For instance, Kruger and Dunning’s classic 1999 study found that the people least skilled at a task were often the most confident in their performance on it, simply because the skills needed to do a task well are the same skills needed to recognise you’ve done it badly. Standardised assessment has spent decades rewarding a quick, certain answer over a reconsidered one. Combine that history with an AI that never disagrees, and, as a result, you get students who mistake speed and certainty for understanding.

Four Moves That Build Confident Humility

None of this requires a new unit or a new app. Instead, it requires building a habit of checked confidence into work you’re already doing.

🧭 Four Moves to Try This Term

Calibrate confidence before the answer. Before a quiz, task, or discussion, have students rate their understanding: know it, think I know it, unsure, don’t know it yet. Compare that rating against actual performance afterwards. The gap is the lesson, not the score.
Build revision into the process, not just the marking. Use a simple stem: “I used to think ___, but now I think ___.” Ask peer feedback to focus on one question: what could this person do differently, not what did they do wrong.
Teach disconfirmation seeking. Ask students to argue against their own claim: what’s the strongest evidence against what you just wrote? A short pre-mortem, imagining the project has failed and working out why, does the same job for group work.
Protect the student’s own questions. Higher-order questioning usually means the questions a teacher asks. Confident humility needs room for the questions students ask, before you’ve told them whether their thinking is right.

Where This Meets Productive Struggle

If this sounds familiar, it should. Disconfirmation seeking is close cousin to the productive struggle we’ve written about before, the discomfort of not immediately knowing that has to happen before real learning can. That said, the difference here is the source of the false certainty. Productive struggle protects against a student giving up too early. Confident humility protects against a student, or an AI tool, deciding too early that they’re finished.

AI will tell your students they’re right. Your classroom is one of the few places left that consistently doesn’t, and that’s not a bug in teaching. It’s the point.

None of these four moves ask you to distrust AI outright, or to ban it from your room. Instead, they ask you to make sure a student’s confidence was earned somewhere other than a chatbot’s agreement. That’s a habit worth building well before your students are relying on AI for every first draft of their thinking, because by then, the habit is much harder to install.

Try this: Pick one class this week and run the confidence calibration move before their next quiz or task. Ask them to rate their understanding first, then compare it to how they actually go. Tell us in the comments what the gap looked like, and whether it surprised your students as much as it usually surprises teachers.

References

Cheng, M., et al. (2026). Sycophantic AI decreases prosocial intentions and promotes dependence. Science.

Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one’s own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 77(6), 1121–1134.

Spencer, J. (2026). Why confident humility is a critical skill in a generative age. Spencer Education.

3 Feuerstein Moves That Make Growth Mindset Real

A growth mindset poster won’t change how students handle failure on their own. Here are three Feuerstein mediation moves that actually build that habit.

Pedagogy · Mindset

3 Feuerstein Moves That Make Growth Mindset Real

A growth mindset poster tells students effort works. It never teaches them how.

⏱ 6 min read · By Matt Faulkner · Pedagogy / Mindset

A teacher leaning in with a warm smile to talk through a primary school student's work, a moment of growth mindset mediation in action.

You have put the growth mindset poster up. “Not yet,” not “I can’t.” You have said it in front of the class, probably more than once. And you have still watched a capable student stare at a hard maths problem, write nothing, and tell you it’s “too hard” the second it stops being easy.

That gap is not a sign growth mindset is wrong. Carol Dweck’s research on how students respond to difficulty is some of the most replicated work in educational psychology. The gap is between naming the belief and building the habit that makes the belief true. A poster can tell a student what to think about effort. It cannot show them what to actually do the next time a task gets hard.

That is where Reuven Feuerstein’s work becomes useful, and where most classroom mindset language quietly stops. Growth mindset says ability can grow. Feuerstein spent a career building the actual mechanism for how.

What Dweck’s Growth Mindset Research Actually Showed

Dweck’s best-known finding did not come from asking students what they believed. It came from watching what happened after they were praised. In one influential study, students who were praised for being smart chose easier follow-up tasks and performed worse under pressure. Students praised for their effort and strategy chose harder tasks and kept improving. The praise itself changed the students’ behaviour, not just their mood.

The mechanism matters. A fixed mindset treats every task as a test of a trait you either have or don’t. A growth mindset treats a task as a process you can get better at. Students with a fixed mindset protect their self-image by avoiding anything that might expose them. Students with a growth mindset lean into difficulty, because struggle is evidence of learning, not evidence of a limit.

None of that is in dispute. What is missing from most classroom applications is the part Dweck herself was clear about: mindset shifts through experience, not slogans. Telling a student “you can grow your ability” without ever showing them how to work through a hard problem is asking them to trust a claim they have no evidence for.

The Missing Mechanism: Mediated Learning

Feuerstein’s core claim, Structural Cognitive Modifiability, is close to Dweck’s growth mindset research in spirit: the structure of a person’s thinking can change at any age, given the right conditions. Where Feuerstein goes further is in specifying what those conditions actually are.

He called it a Mediated Learning Experience. Instead of a child encountering a hard problem alone and forming their own conclusions about what it means, a mediator, meaning you, steps between the child and the task. You do not just present the problem. You shape how the child takes it in: what to notice, why it matters, and how this moment connects to something bigger than itself.

A growth mindset poster tells students to believe effort works. Mediation is what actually teaches them how.

This is not a call to run a formal Instrumental Enrichment programme. Most teachers will never need to. It is a call to notice that “not yet” is a belief statement, and beliefs change through repeated, structured experience, not through repetition of the phrase.

The Three Moves That Make Growth Mindset Real

Feuerstein identified several criteria for a mediated interaction, but three do almost all of the work in an ordinary classroom moment. Used together, in the sixty seconds after a student gets stuck, they turn “not yet,” the core growth mindset phrase, from a slogan into an experience the student has actually lived through.

🧠 Three Moves of Mediation

Intentionality and reciprocity. Name what you are doing and why, out loud, then wait for the student to engage back. “I’m going to show you how I check my working, watch this part” is mediation. Silently correcting the page is not, even if the content taught is identical.
Transcendence. Push the moment past the single task in front of you. Once a student solves one problem, ask where else the same strategy would work. A student who only ever hears “well done” learns that this problem is finished. A student asked “where else could you use that” learns the strategy travels.
Mediation of meaning. Say why the struggle matters, and say it like you mean it. “This is the exact skill you’ll need for the assessment next week, and I’ve watched you do harder things than this” carries more weight than “keep trying.” Students recalibrate their effort based on what an adult they trust tells them the moment is for.

None of these three moves take long. What they require is deliberately slowing down the sixty seconds after a student gets stuck, instead of rushing to either rescue them or move on. That pause is the actual intervention. Everything else is context for why the pause matters.

Where This Overlaps With What You Are Already Doing

If you have read our post on designing tasks for productive struggle, this will sound familiar, because it is the missing half of that argument. Task design gets students into a moment of genuine difficulty. Mediation is what you do once they are there. A well-designed task without mediation still leaves plenty of students stuck and unsupported. Mediation without a task worth struggling over has nothing to attach itself to. The two need each other.

The same logic connects to higher-order questioning. A mediated response to a stuck student is not “here’s the answer” and it is not “keep trying.” It is a question that hands the thinking back: what have you tried so far, what does this remind you of, what would happen if you started from the other end. The questioning moves you already use for stretching thinking are the same moves that make mediation work in the moment a student hits a wall.

Building It In, Not Bolting It On

The three moves work best when they are not left to whatever you remember to say in the moment. When you are planning backwards from the outcome you want, build in where a genuine struggle point sits in the lesson, and plan your language for it the same way you plan your success criteria. Decide in advance: what will I say to name the strategy, what will I ask to push it further, and what will I say about why this moment matters.

That small shift, from reactive encouragement to planned mediation, is the difference between a classroom where “not yet” is something students hear and a classroom where it is something they have actually experienced enough times to believe.

Try this in your next tricky lesson: pick one moment where you know a few students will get stuck, and plan your intentionality, transcendence, and meaning language for it ahead of time. Tell us in the comments how it changed the sixty seconds after they hit the wall.

References

Dweck, C. S. (2006). Mindset: The new psychology of success. Random House.

Feuerstein, R., Feuerstein, R. S., & Falik, L. H. (2010). Beyond smarter: Mediated learning and the brain’s capacity for change. Teachers College Press.

Feuerstein, R., Rand, Y., & Hoffman, M. B. (1979). The dynamic assessment of retarded performers: The learning potential assessment device, theory, instruments, and techniques. University Park Press.

Mueller, C. M., & Dweck, C. S. (1998). Praise for intelligence can undermine children’s motivation and performance. Journal of Personality and Social Psychology, 75(1), 33–52. pubmed.ncbi.nlm.nih.gov

Structural Learning. (2024). Feuerstein’s mediated learning: A guide for teachers. structural-learning.com

4 Proven Steps to Fix Your Student Choice Problem

Pedagogy · Student Agency

4 Proven Steps to Fix Your Student Choice Problem

Student choice doesn’t fail because students can’t handle freedom. It fails because we hand it all over at once.

⏱ 6 min read · By Matt Faulkner · Pedagogy / Student Agency

Students working on different self-chosen projects as part of scaffolded student choice.

Student choice sounds great, right up until it goes wrong. You have seen it happen: Genius Hour, passion projects, “pick your own topic” Fridays. The first year you try it, half the room lights up and the other half sits there for 40 minutes doing nothing, or quietly finishing maths homework because at least that has instructions. By week three you have quietly stopped mentioning it.

The instinct after that is to conclude students weren’t ready, or that choice doesn’t work with this particular class. That conclusion is almost always wrong. The problem is rarely the students. It is the size of the jump.

Handing a class complete autonomy in one step is the same mistake as handing them a cold-open essay question with no modelling. We would never do that with content. We do it constantly with choice, then act surprised when it collapses.

What Actually Goes Wrong With Wide-Open Student Choice

Educator John Spencer has written honestly about his own “epic fail” running Genius Hour early in his teaching career. He gave students full ownership on day one: pick a topic, pick a question, go. Within weeks it had collapsed. High achievers defaulted to homework because at least homework had a clear finish line. Everyone else drifted into conversation. His own diagnosis was blunt: “This was too much choice and ownership without any actual guidance. I had failed to provide a vision for these projects or a road map.”

That is worth sitting with, because it names the actual failure point. It was not that students lacked interests worth pursuing. It was that nobody had built the scaffolding that lets students act on an interest productively: no checkpoints, no criteria, no model of what a finished piece of work looks like. Freedom without structure does not feel like empowerment to most students. It feels like being dropped somewhere with no map.

Borrow the Model You Already Use for Everything Else

Here is the useful reframe: you already know how to scaffold this. It is the same logic behind gradually releasing responsibility for higher-order questioning — you model a skill, practise it together, then hand it over once students can carry it alone. Student choice is not a different category of teaching decision. It is the same gradual release model, applied to ownership instead of content.

The “I do, we do, you do” structure works for choice because it treats independence as a skill, not a switch. Students do not arrive able to self-select a topic, plan their own research, and manage their own timeline. They build that capacity in stages, the same way they build capacity to answer a why-question instead of a what-question.

🪜 The 4-Level Choice Menu

Level 1 — Embedded Choice (mostly teacher-directed). Everyone works on the same task, but chooses how to complete a sub-part of it. Same maths problem, choice of solving it with a diagram, an equation, or manipulatives.
Level 2 — Simple Choice Menu (somewhat student-directed). One shared learning objective, several assignment options. Students pick their own novel, then choose from a set of book-report formats you provide.
Level 3 — Advanced Choice Menu (mostly student-directed). Students select which learning target to focus on and find their own resources, then choose how to present what they have learned.
Level 4 — Independent Project (entirely student-directed). Students choose their own groups, set their own questions, source their own research, and design their own product. This is Genius Hour done properly — after the earlier levels have built the muscle for it.

The point of the ladder is not that Level 4 is the goal for every task, every time. Some units genuinely call for Level 1. The point is that you cannot skip from “no choice” to Level 4 in one lesson and expect it to hold.

Two Skills Students Need Before They Are Ready for More

Spencer draws a distinction worth borrowing directly into your planning for student choice: self-starters and self-managers are two separate skills, and students can be strong in one without the other.

A self-starter takes initiative without waiting for permission — driven by “why not?” rather than “what do I do?” A self-manager is the one who sticks to a deadline once the excitement has worn off, who pushes through the messy middle of a project instead of abandoning it for something easier. This overlaps directly with what we covered in designing for productive struggle: a student who has never been asked to sit inside difficulty on a teacher-set task is unlikely to manage that difficulty alone, on a task they chose themselves.

Choice is not a lesson you switch on. It is a skill you build, one level at a time, the same way you build every other skill in the room.

Before moving a class up a level, it is worth asking which of the two skills is missing. A class full of self-starters who abandon projects halfway needs more structure around follow-through — checkpoints, exit tickets, a visible progress tracker. A compliant class that never volunteers an idea needs more low-stakes practice generating options, before you ask them to generate an entire project.

Building This Into Your Planning, Not Bolting It On

This only works if it is a planning decision, not an improvisation on the day. When you are planning backwards from the outcome you actually want, choice level is one of the variables you set deliberately, alongside the learning intention and the success criteria — not something you default into because open tasks feel more engaging on paper.

In practice, that means deciding up front: which level does this unit warrant, given what this particular group has already shown you they can manage? What is the checkpoint that tells you they are ready to move up a level next term? What does the rubric or success criteria look like at this level, so “choice” does not quietly become “no expectations”?

None of this removes the appeal of student agency. It just stops treating agency as something you either grant or withhold, and starts treating it as something you teach.

Where does your class currently sit on this ladder? Tell us in the comments — and if a choice-based lesson has ever collapsed on you, we would love to hear what you changed the second time around.

References and Further Reading

Fisher, D., & Frey, N. (2014). Better learning through structured teaching: A framework for the gradual release of responsibility (2nd ed.). ASCD.

Pearson, P. D., & Gallagher, M. C. (1983). The instruction of reading comprehension. Contemporary Educational Psychology, 8(3), 317–344.

Spencer, J. (2024, August; updated 2025, April). Phasing in student choice with the gradual release of responsibility model. Spencer Education. spencereducation.com

4 Proven Steps to Fade Support Across Lessons

Pedagogy · STEAM

4 Proven Steps to Fade Support Across Lessons

Fading isn’t just a single-lesson skill — you fade support across lessons too.

⏱ 6 min read · By Matt Faulkner · Pedagogy / STEAM

Part 2 of this series looked at fading inside a single lesson. It covered the signals that tell you when to pull back support moment to moment. Even so, some fading decisions don’t happen in a moment. They happen between lessons, and they’re easy to miss — this post is about how you fade support across lessons.

Once you’re teaching a sequence rather than one lesson, the question changes. It’s no longer just “should I help now?” Instead, it becomes “is this class ready for the next stage?” Or does it need another lesson at this level first? That’s how you fade support across lessons, not just within one.

Here’s what that looks like across four lessons of a real Science sequence.

Why You Fade Support Across Lessons, Not Just One

The gradual release of responsibility model — I do, we do, you do — was built for a single lesson. In practice, most worthwhile learning takes longer than one lesson to land. From there, the same three phases need to stretch across several.

Even so, the mechanics don’t change. You’re still watching for the same signals from Part 2. They show up as questions turning exploratory, errors turning careless instead of conceptual, and students finishing before you do. Put simply, the unit of decision changes. For example, instead of asking whether to fade in five minutes, you’re asking whether to fade in the next lesson.

This isn’t just intuition. It echoes Rosenshine’s principles of instruction, which found that reviewing previous learning before releasing new independence improves retention. That’s a bigger, slower-moving decision. It deserves its own checkpoint, planned in before the sequence starts.

A Four-Lesson Fade in Science

For example, take a Year 4 forces sequence — how push and pull forces change an object’s motion. Four lessons, each one releasing more responsibility than the last.

Lesson 1 (I do): You run a full guided friction test aloud — including the parts you’d get wrong first. In addition, students watch and record, not design.

Lesson 2 (we do): Second, the class runs a similar investigation together. At the same time, you still frame the method, but students choose the materials to test and predict the results.

Lesson 3 (you do, supported): Third, small groups design their own fair test using a partially worked checklist. The variables are named; the steps aren’t. You circulate, watching for the same signals from Part 2.

Lesson 4 (you do, independent): Finally, groups plan and run an investigation on a new but related question. No checklist. No template. Just the habits built across the first three lessons.

As a result, this is what it looks like to fade support across lessons. Each lesson releases one more piece of the method. By lesson four, nothing about the process is being handed to them anymore.

🪜 4 Signs Your Class Is Ready for the Next Lesson

Exit tickets show fluency, not just participation. Students can explain the “why,” not just complete the steps.
Independent work has few careless errors. Mistakes are about precision, not the underlying idea.
Questions shift toward planning, not procedure. “What if we tried…” replaces “What do I do next?”
Peer helpers start emerging on their own. Students you didn’t assign are already checking each other’s work.

When the Sequence Doesn’t Go to Plan

Sometimes a class isn’t ready to move on schedule. Even so, that’s useful information, not a failure. If lesson two’s exit tickets show shaky fluency, repeat lesson two’s shape with a new context before releasing further.

That said, some classes are ready to compress two stages into one. That’s especially true if the content connects to something they’ve already practised. As a result, the four-lesson shape is a starting plan, not a fixed script.

Beyond that, the decision to fade support across lessons stays visible and evidence-based. You’re not guessing whether to move on — you’re checking the signals against what the last lesson actually showed.

Planning the Whole Arc Before Lesson One

This is where backward design earns its keep again. If you know the independent task you want by lesson four, you can plan backward from there. Each earlier lesson needs to release one more piece to make that possible.

First, build your checkpoints into the plan itself. What evidence from lesson one tells you lesson two can begin? What would tell you to slow down instead? Answering those questions before teaching starts means the mid-sequence decision is already half made.

In other words, this is how you fade support across lessons. You design it, then confirm it against the evidence as you go.

A sequence doesn’t fade support by accident — it’s designed to, one checkpoint at a time.

💬 Over to You — How many lessons does your unit take to move from “I do” to independent? Tell us below — the shape of other teachers’ sequences is always worth comparing against your own.

References

Pearson, P. D., & Gallagher, M. C. (1983). The instruction of reading comprehension. Contemporary Educational Psychology, 8(3), 317–344.

Rosenshine, B. (2012). Principles of instruction: Research-based strategies that all teachers should know. American Educator, 36(1), 12–19, 39.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.

5 Proven Signs It’s Time to Fade the Scaffold

Pedagogy · New Teachers

5 Proven Signs It’s Time to Fade the Scaffold

Fading isn’t a feeling you talk yourself into — it’s a decision you can plan for.

⏱ 6 min read · By Matt Faulkner · Pedagogy / New Teachers

A student finishes the worked example, tries the next one alone, and stalls. You’re three metres away, marking book in hand. The moment asks a genuine question: do you step in, or fade the scaffold and wait ten more seconds?

Most of us know the gradual release of responsibility model well: I do, we do, you do. What it doesn’t tell you is when, inside that “you do” moment, to actually let go. That judgement call happens dozens of times a lesson. In practice, it’s rarely taught explicitly.

This is where Vygotsky’s zone of proximal development (ZPD) earns its keep. It isn’t another framework to memorise. By contrast, it’s the diagnostic tool. It tells you whether a student is ready to work without you, or still needs the scaffold.

GRR Sets the Stage, ZPD Tells You When to Move

The gradual release of responsibility model comes from Pearson and Gallagher’s 1983 research on reading comprehension instruction. They found that comprehension improved when teachers deliberately shifted cognitive responsibility from themselves to students across a sequence. In other words, that sequence — I do, we do, you do — is the structure.

Fading is what happens inside that structure. It’s the technique of reducing support in increments: fewer prompts, shorter sentence stems, more open questions. As a result, a student moves from needing you to not needing you in stages, not all at once. In other words, GRR tells you the phases exist; fading is how you move through them.

That’s where ZPD tells you when to fade the scaffold. The zone of proximal development is the gap between what a student can do alone. It’s also what they can do with support close by. For example, fade too early and you’ve asked them to work outside that zone, unsupported. Fade too late, and they stay inside it long after they were ready to leave. That has its own cost, too.

The Two Ways Calibration Goes Wrong

We’ve covered the most common GRR misfire before. It’s an “I do” phase so thorough that “you do” becomes copying, not struggle. Put simply, that’s fading too slowly. The scaffold stays up well past the point a student could carry their own weight. As a result, the task stops asking anything of them.

By contrast, the opposite failure gets less attention. Fade too fast, and you haven’t created independence — you’ve created guesswork. That happens when the worked example, sentence stem, or check-in disappears before “we do” has landed. As a result, what looks like “productive struggle” up front is often just confusion with better posture.

Even so, both are calibration errors. Neither is solved by scrapping GRR — both come down to knowing when to fade the scaffold, not whether to.

🪜 5 Signals It’s Time to Fade the Scaffold

Their questions change shape. “How do I start?” becomes “What if I tried it this way?” — that shift from procedural to exploratory is a reliable ZPD marker.
Errors move from conceptual to careless. Getting the method right but rushing the last step means the concept has landed; only precision is missing.
They finish the “we do” example before you do. Hands go up, pencils move — students are already a step ahead of your modelling.
They start correcting each other. Peer correction only happens once a student trusts their own grasp enough to check someone else’s.
The room goes quiet in a different way. Not confused-quiet — working-quiet. That’s a felt cue, and it’s worth trusting alongside the others.

How to Fade the Scaffold, Move by Move

Fading rarely works as an all-or-nothing switch. In practice, it’s a series of small withdrawals, each testing whether a student can hold the next level of responsibility.

For example, a worked example becomes a partially worked example — the first two steps shown, the rest left blank. Beyond that, sentence stems shrink from full frames to a single connective word. At the same time, questions move from closed (“What’s the next step?”) to open (“What would you try next?”). Each version removes one layer of support while the task itself stays constant.

As a result, you’re never asking a student to jump from fully supported to fully independent in one move. You’re testing the edge of their ZPD in small enough steps. That way, you can pull back if the gap turns out to be wider than expected.

Planning the Fade, Not Just Reacting to It

In-the-moment judgement matters, but it’s easier to trust when it’s backed by planning. If you’ve already set up your lesson using our backward design approach, the same thinking applies to fading. For this reason, decide your checkpoints before the lesson starts, not during it.

First, mark two or three points before the lesson where you’ll deliberately step back and watch rather than help. For example, good checkpoints include after the first worked example, halfway through practice, or at a new question type. Second, at each checkpoint, look for the signals above rather than the clock. That way, the decision to fade isn’t a feeling you’re second-guessing mid-lesson — it’s a checkpoint you already built in.

This matters even more once lessons stop being single, standalone events. From there, planning shifts from one lesson to a sequence built around learner variability. When it does, the question of when to fade the scaffold travels with it too. Part 3 of this series answers that next.

Fading isn’t a feeling you talk yourself into — it’s a decision you can plan for, checkpoint by checkpoint.

💬 Over to You — Which signal do you already watch for without realising it? Which one are you going to add this week? Tell us below — the best classroom cues on this site always come from teachers who’ve already tested them.

References

Pearson, P. D., & Gallagher, M. C. (1983). The instruction of reading comprehension. Contemporary Educational Psychology, 8(3), 317–344.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.

5 Easy Ways to Plan for Every Learner With UDL

Differentiation isn’t something you bolt on after the lesson is built. It’s a decision you make before you write a single activity.

⏱ 6 min read · By Matt Faulkner · Pedagogy / Planning

Picture the lesson plan taking shape on your screen. Content descriptor pasted in. Learning intention written. Main activity mapped out for the class. Then, right at the end, you add a line: “adjustment for Sarah,” or “extension for the early finishers.”

It feels like good practice. In practice, it’s also the wrong order. To illustrate, you built the lesson for an imagined average student, then patched it for everyone who didn’t fit. The patch usually works, for a while. As a result, you solve the same problem lesson after lesson — instead of solving it once at the planning stage.

Universal Design for Learning (UDL) flips that order. Instead of building one pathway and adjusting it afterwards, you plan multiple ways in from the very first decision. Beyond that, it sits neatly alongside the backward design framework we’ve covered before. Same starting point — one extra lens over every choice that follows.

Why “Adjustments Later” Never Quite Works

The trouble with retrofitting isn’t the intent. It’s the timing. Every adjustment gets made under pressure, at the end of a planning session. In practice, that’s when your energy for creative problem-solving is lowest and the deadline is closest. Consequently, the graphic organiser gets left out — or the extension task becomes “read on quietly.”

At the same time, there’s a quieter cost. When adjustments arrive last, they read as exceptions to the “real” lesson. In other words, they’re not part of the plan — just a patch on it. Even so, students notice that framing, whether anyone says it aloud or not.

For example, we’ve covered low-prep differentiation moves elsewhere — the kind you reach for once a task is already built. UDL asks an earlier question instead. What if the task already worked for the widest range of learners, before any adjustment was needed?

What UDL Actually Asks You to Plan For

UDL, developed by CAST, organises planning around three questions rather than one. First, multiple means of engagement. Why would this student care — and does that reason hold for more than one kind of student? Second, multiple means of representation. Is the content available through more than one channel — text, image, spoken explanation? Third, multiple means of action and expression. Can students show what they know in more than one way, while still being judged against the same success criterion?

Put simply, none of those questions ask you to lower the learning intention. Instead, they ask you to widen the road leading to it.

UDL is not about lowering the bar. It’s about building more doors into the same room.

Where UDL Fits Inside Backward Design

If you already plan backwards from the learning intention, UDL doesn’t replace that process. In practice, it sharpens it. At Stage 1, ask whether the intention can be reached — and cared about — through more than one entry point. At Stage 2, ask whether your formative evidence has to look identical for every student. Even so, a spoken explanation can count as evidence too — just as validly as a written one. At Stage 3, build the variety in before you teach the lesson, not after you notice it’s needed.

That’s the shift in one sentence. Instead of asking who will struggle with this activity after it’s built, ask which students the design already leaves out.

A Year 4 Worked Example

For example, here’s what that looks like against a real content descriptor, rather than in the abstract.

🎯 UDL in Action — Year 4 Persuasive Writing

Curriculum anchor: AC9E4LY06 — Plan, create and revise written and multimodal texts, including persuasive texts.
Learning intention: “We are learning to persuade a reader using reasons and evidence.”
Engagement: Students choose their topic from a short, locally relevant list — not one topic assigned to everyone.
Representation: The teacher models the structure through a think-aloud, a labelled example, and a simple graphic organiser.
Action and expression: Students meet the same success criterion by writing a paragraph, recording a spoken pitch, or building an annotated poster.

In other words, notice what stayed constant — the content descriptor, the learning intention, and the success criterion. Only the entry points changed.

This Isn’t Differentiation Downward

For this reason, it’s worth being explicit here, because the two ideas get confused constantly. Differentiating downward means changing what a student is expected to learn — a simpler text, a lower target, a smaller ask. By contrast, UDL keeps the expectation identical for everyone and changes how students reach it.

Beyond that, this distinction matters for AITSL Standard 1.5 — and for the students in front of you. Planning for variability is not the same as planning for less.

Finally, once a lesson is built this way, a second question follows naturally. How do you adjust support in the moment, as students actually work through the task? That’s where frameworks like the Gradual Release of Responsibility come in — and it’s exactly where this series goes next.

💬 Your Turn! Think about your last unit plan. Where did the “adjustments” get added last — and what would it look like to build them in first instead? Share your experience below.

References

Australian Curriculum, Assessment and Reporting Authority (ACARA). (2023). Australian Curriculum v9.0. https://v9.australiancurriculum.edu.au

Australian Institute for Teaching and School Leadership (AITSL). (2018). Australian Professional Standards for Teachers (Standard 1.5). https://www.aitsl.edu.au/teach/standards

CAST. (2024). Universal Design for Learning Guidelines (version 2.2). https://udlguidelines.cast.org

Novak, K. (2022). UDL now! A teacher’s guide to applying universal design for learning (2nd ed.). CAST Professional Publishing.

Rose, D. H., & Meyer, A. (2002). Teaching every student in the digital age: Universal design for learning. ASCD.

Wiggins, G., & McTighe, J. (2005). Understanding by design (2nd ed.). ASCD.

3 Easy Ways to Spot What Assessment Really Costs You

Assessment · Honest Trade-Offs

There is no assessment strategy that gives you time, consistency, and depth all at once. Every choice trades one for another — the trick is choosing on purpose.

⏱ 7 min read · By Matt Faulkner · Assessment / Honest Trade-Offs

Somewhere around the fifteenth persuasive writing piece on a Sunday night, most of us have had the same thought: could I just use a checklist instead? The rich rubric felt right in the planning meeting. At 9pm, with thirteen scripts still in the pile, it feels like a decision someone else made on your behalf.

That moment is worth sitting with, because it’s not really about tiredness. It’s about a trade-off nobody named out loud when the assessment was designed. In Part 2 of this series, we looked at matching assessment strategies to record-keeping systems. This final part asks the harder question: what does each strategy quietly cost you, and is that cost one you’d choose if someone made you say it out loud?

Three costs show up again and again in the research: time, consistency, and depth of insight. No strategy escapes all three. The honest move isn’t finding the mythical assessment that avoids every trade-off — it’s knowing which one you’re carrying, for which task, and why.

Trade-Off One: Time

Let’s start with the cost every teacher already feels in their body. Australian Education Union data puts full-time teachers at an average of 49.3 hours a week, with marking and planning dominating the hours beyond the timetable. A 2022 AEU South Australian survey found teachers were doing 30 hours a week beyond face-to-face teaching alone. Richer assessment — detailed rubrics, portfolios, oral check-ins — produces better evidence of learning. It also produces more hours you don’t have.

This isn’t an argument against depth. It’s an argument for honesty about the exchange rate. A five-minute exit ticket and a forty-minute portfolio conference both tell you something real about a student. They do not cost the same, and pretending otherwise is how burnout starts. For quick formative checks that protect your time without sacrificing useful data, the formative assessment approaches we’ve covered before are worth revisiting — they’re built for exactly this trade-off.

In practice, the fix isn’t cramming more hours into the week. It’s matching assessment weight to decision weight. A task that shapes tomorrow’s teaching deserves a fast, low-cost check. A task that reports a term’s learning to families deserves the time it takes. Treating every piece of work as if it needs the same depth is what makes the Sunday night pile unmanageable.

Trade-Off Two: Consistency

Here’s the uncomfortable one. Even careful, experienced markers are not as consistent as they believe themselves to be. A 2023 experimental study found that a student’s grade in one subject was significantly influenced by how that same student had performed in an entirely unrelated subject — the classic halo effect, and it held even after researchers controlled for gender and status (Schmidt et al., 2023). This isn’t a character flaw in individual teachers. It’s a known feature of how human judgement works under load, and it’s exactly what assessment for, as, and of learning tries to guard against when it separates the purpose of a judgement from the moment it’s made.

There’s a useful nuance here, though. Experienced teachers are better than novices at reading both the obvious “surface” signals — hands up, eyes down — and the harder “deep” signals, like a quiet student who might be disengaged or might just be thinking. That skill develops. It doesn’t develop by accident, and it doesn’t fully protect against bias on its own — structure still matters more than good intentions.

🎯 Moderation Moves That Actually Fit a School Week

Check the extremes first. Re-read your highest and lowest grades before anything else. That’s where halo effects and fatigue errors cluster.
Mark one question across all scripts, not one script across all questions. This keeps your internal standard consistent instead of drifting as you tire.
Run a five-minute calibration before you start. Grade one sample together with a colleague before marking solo. It costs almost nothing and catches drift before it happens, not after.
Blind-mark when the stakes are high. Covering the name is a small step that removes a documented source of bias for borderline or subjective work.

This is also the territory AITSL Standard 5.3 points to directly — making consistent and comparable judgements isn’t a bonus skill, it’s a named professional standard. Worth remembering next time moderation feels like an add-on rather than the job itself.

Trade-Off Three: Depth of Data

The third cost is the one we talk about least, possibly because it’s the hardest to see from inside a busy term: what a grade actually tells you about a student’s understanding.

Recent research has raised real doubt about that link. Comparisons between teacher-awarded grades and external, course-aligned tests have found meaningful mismatches — evidence that a grade can look confident and still not accurately describe what a student understands. That gap widens when assessment rewards compliance — submitting on time, following the format, filling every box — over genuine thinking. AJ Juliani’s phrase for this, “doing school,” captures it well: students becoming skilled at producing what a task asks for without the underlying understanding ever being tested.

A grade that’s easy to collect and a grade that’s true to what a student understands are not always the same grade. Depth is the trade-off we notice last, because it’s the one that doesn’t show up until much later.

The fix isn’t abandoning efficient assessment formats. It’s building in occasional deeper checks — a follow-up question, a brief verbal explanation, a second look at borderline work — before a grade goes into the record permanently. Even one extra “tell me how you got there” question can surface whether a correct answer reflects understanding or a lucky guess. This connects directly to performance of worth: tasks that ask students to apply learning somewhere real are far harder to complete through compliance alone.

Choosing Your Trade-Off on Purpose

No single assessment strategy solves time, consistency, and depth simultaneously. A quick multiple-choice check is fast and reasonably consistent, but shallow. A rich portfolio conference is deep and genuinely revealing, but slow. A shared rubric with clear criteria protects consistency, but only if you also protect the time to moderate it properly.

That’s not a flaw in your planning. It’s the actual shape of the problem. The teachers who manage assessment well aren’t the ones who’ve found a loophole — they’re the ones who’ve stopped expecting one, and instead ask a simple question before choosing a strategy: for this task, which cost can I afford, and which one matters most?

🪜 A Quick Decision Check Before You Design a Task

Low stakes, shapes tomorrow’s teaching? Choose speed. A fast formative check beats a slow, perfect one you never get to use.
Reported to families or used for placement? Choose consistency. Build in moderation time even if it means assessing less often.
Central to whether real understanding happened? Choose depth. Add one verbal or applied check before the grade is locked in.

That’s the honest version of assessment design. Not a search for the perfect tool, but a series of deliberate choices about which cost you’re willing to carry, task by task, term by term.

💬 Your Turn! Which of these three trade-offs do you find hardest to accept — time, consistency, or depth? Tell us what you’ve changed (or given up trying to fix) in the comments below.

References

Australian Education Union. (2023). The heavy hours: Work intensification for teachers and school leaders. aeufederal.org.au

Australian Institute for Teaching and School Leadership (AITSL). (2018). Australian Professional Standards for Teachers — Standard 5.3: Make consistent and comparable judgements.

Juliani, A. J. (2026). The game of school is out of control. ajjuliani.beehiiv.com.

NSW Department of Education. (2024). Assessment practice — consistent teacher judgement. education.nsw.gov.au

Schmidt, F. T. C., Kaiser, A., & Retelsdorf, J. (2023). Halo effects in grading: An experimental approach. Educational Psychology, 43(2–3), 246–262.