Assessment for Learning: Quick Checks That Actually Tell You Who’s Got It
Assessment of learning, and assessment for learning
Every teacher has had this moment. You ask the class, “Does everyone get it?” A sea of nods comes back. You move on. Three days later the books come in and it’s obvious a third of the room didn’t get it at all. Nobody lied to you. They didn’t know they didn’t know, or they didn’t want to be the one holding the class up. That gap between what you think has landed and what’s actually landed is the reason assessment for learning exists, and it’s worth being precise about what the term actually means.
Assessment of learning is a judgement: the end-of-topic test, the mock, the GCSE paper itself. It’s a snapshot of what a pupil can currently do, usually reported as a mark, usually arriving after the teaching is finished. It matters for tracking and reporting, but by the time you’re marking it, the teaching that produced those results is already done. You can act on what it tells you for next term. You can’t act on it for the class sitting in front of you today.
Assessment for learning is different in purpose, not necessarily in format. It happens during the teaching, and its only job is to change what happens next, in this lesson or the following one. Paul Black and Dylan Wiliam’s 1998 review for King’s College London, which pulled together evidence from close to 250 studies, is still the reference point most of this thinking traces back to. They found that improving formative assessment was one of the more reliable ways to raise attainment, and one of the few interventions that seemed to help lower-attaining pupils more than anyone else, narrowing the gap rather than widening it.
The practical test is simple: a quick check only counts as assessment for learning if you’re going to act on it before the lesson ends, or before the next one starts. Ask a question, get an answer, and carry on regardless of what that answer was, and you haven’t checked anything. You’ve just talked.
Why hands up gives you the wrong picture
Hands-up questioning feels like it’s checking understanding. Mostly it isn’t. It’s checking who is confident, and quick, enough to want to answer in front of everyone else, which is a much smaller and far more self-selecting group than “the class”.
Black and Wiliam noticed that in most classrooms only a handful of pupils ever answer, and everyone else learns fast that they can leave it to that handful. The quiet ones aren’t necessarily lost. Some know the answer and just don’t fancy the exposure. Some genuinely don’t know. Hands-up questioning lets both groups hide behind the same small set of volunteers, and you can’t tell which is which from the front of the room.
There’s a timing problem too. Watch what actually happens after you ask a question: three or four hands go up almost instantly, a couple more join after a second or two, and if you hold out to around four seconds, roughly half the class will have offered a response. Half. Most teachers, without meaning to, take the first hand or answer their own question within two or three seconds, which isn’t long enough for a pupil to think an answer through rather than just recall it. Black and Wiliam’s advice on this is blunt: target a question at a named individual, then wait at least three seconds before expecting anything back.
None of this means questioning is a bad idea. It means hands-up questioning on its own tells you about a subset of the room and very little about the rest. If you want to know what the whole class actually thinks, you need a method where everyone answers, not just the volunteers.
Hinge questions: the check that makes you decide something
A hinge question is Dylan Wiliam’s term for a question planted at a genuine turning point in a lesson, the moment where what happens next should depend on how pupils respond. The name is the whole idea. The lesson hinges on the answer. If understanding is solid, you carry on. If it isn’t, you loop back before building anything else on a shaky foundation.
What makes a hinge question actually work
A good hinge question has to be quick both ways. Pupils should be able to answer it in a minute or two, and you should be able to read the whole class’s response within about thirty seconds, which rules out anything needing extended writing. It needs to reach every pupil rather than a sample, so it wants a format where everybody commits to something visible at the same moment. And it has to be diagnostic, not just right-or-wrong. A hinge question that only tells you “62% correct” is far less useful than one that also tells you which misconception is doing the damage.
Building the answer options
This is where the planning time actually goes, and it’s the part worth doing properly. Multiple choice suits hinge questions because it forces you to decide the plausible wrong answers in advance, rather than reacting to whatever pupils happen to say on the day. Each wrong option should represent a real misconception, ideally one you’ve genuinely seen pupils make, not a distractor invented to fill a gap. Teaching negative number multiplication, for example, you’d want options built roughly like this:
- the answer a pupil gets if they forget the sign rule entirely
- the answer they get if they apply the rule to only one of the two numbers
- the correct answer
- possibly a fourth option covering a common arithmetic slip that has nothing to do with signs at all, so you’re not accidentally teaching pupils to hunt for “the sign trick” option
When the whiteboards go up, a well-built set of distractors tells you not just who is stuck but roughly where and why, which is what actually shapes what you do in the next five minutes.
Mini-whiteboards: the fastest read you’ll get on a room
If hinge questions are the “what to ask”, mini-whiteboards are one of the best ways to run them. Every pupil writes an answer, holds it up on a signal, and for a couple of seconds you get a picture of the whole room’s thinking at once. No one opts out by staying quiet, and no one coasts on somebody else’s raised hand.
The reason whiteboards work is fairly mundane: because everyone has to write something and hold it up, participation sits close to 100%, in a way hands-up questioning never manages. That single design feature is most of the benefit. You’re not getting cleverer questions out of a whiteboard. You’re getting an honest sample instead of a self-selected one.
The routine matters as much as the technique. Most teachers who use whiteboards well settle on something like: write your answer, don’t show your neighbour, hold up on the count of three, no changing your answer once boards are up. Standing at the back of the room to scan, rather than the front, lets you sweep the whole set in one glance instead of just the first couple of rows. Reading twenty-eight boards held in the air genuinely is quicker than marking twenty-eight books, and you get the information while it’s still useful rather than three days later.
The Education Endowment Foundation’s 2018 evaluation of a professional development programme built around Wiliam and Siobhán Leahy’s formative assessment ideas, whiteboard-style checks included, followed roughly 25,000 pupils across 140 secondary schools. It found an average gain equivalent to two months’ additional progress on GCSE Attainment 8, rated as a high-security finding. It’s a modest number rather than a dramatic one, and the trial didn’t find a separate significant effect isolated to English or maths specifically, but two months from a low-cost change in classroom habit, tested at that scale, is a reasonable return for something that costs a stack of whiteboards and a bit of routine-building.
Exit tickets that don’t create a marking backlog
An exit ticket is whatever a pupil answers on the way out of the room, usually in the last three to five minutes, usually a slip of paper or a single question on the board. The idea is old and simple. Where it goes wrong is when teachers turn it into extra marking, which is the surest way to stop using it by half term.
Keep it to one question, and pick a format you can read in a glance rather than one you have to read carefully. A single multiple-choice question, a traffic-light self-rating next to one line of working, or one sentence starting “I’m still not sure about…” all take a pupil under two minutes to complete and take you under two minutes to skim across the whole class. The researcher Robert Marzano set out four broad types of exit prompt worth knowing: a question that checks a specific piece of understanding, one that asks pupils to explain their own thinking, one that asks them to evaluate the lesson itself, and one that invites reflection on what they’d still like to know. You don’t need all four every lesson. Pick whichever matches what you actually need to find out.
The design discipline is the same one that makes hinge questions work: decide the question before the lesson, not during the scramble of the last five minutes, so it tests what you actually taught rather than whatever’s quickest to write on the fly. If putting together fresh checking questions every week is the bit that eats into your planning time, a bank of ready-made quiz and check content, the kind the Learnaroo Hub library provides, is worth having as something to adapt rather than a blank page to start from.
Cold-calling, done as support rather than a trap
Cold-calling means asking a question and then naming the pupil who’ll answer it, rather than waiting for hands. Doug Lemov set it out as a named technique in Teach Like a Champion, and it’s since been picked up and promoted heavily in England by Dylan Wiliam, largely because it solves the same problem as mini-whiteboards from a different angle: it makes participation the default rather than the exception.
The case for it is really about inclusion, not about catching anyone out. With hands-up, the pupils who never volunteer get to opt out of thinking as well as opt out of answering, lesson after lesson, and nobody notices because the lesson keeps moving regardless of who’s actually engaged. Cold-calling removes that option. Everyone has to be ready to answer, which in practice means everyone has to actually engage with the question rather than wait to see if somebody else will.
Whether it feels supportive or feels like a trap comes down entirely to how you run it, and done badly it genuinely is unpleasant to sit through. A few habits make the difference. Give real thinking time before naming anyone, rather than firing a name at a pupil who hasn’t had a chance to consider the question yet. Treat “I’m not sure, but I think…” as a real answer worth building on, not a failure to move past quickly. Come back to the same pupil later, once they’ve heard a couple of other answers, so being called on isn’t a one-shot public test. And keep your own reaction genuinely even, the same tone for a right answer as a wrong one, because pupils clock instantly whether a wrong answer gets treated as useful information or as something to be embarrassed about. Teachers who introduce cold-calling often report that it’s unpopular with almost everyone for the first couple of weeks, including the keen hand-raisers who’ve lost their chance to show off and the quiet pupils who’ve lost the option of staying under the radar. Both groups tend to settle once it’s clear the questioning isn’t a gotcha, just how the class works.
The part most advice skips: what you actually do with what you find out
This is the bit that separates assessment for learning from assessment theatre. Running a hinge question, a whiteboard check, or an exit ticket, and then carrying on with the lesson plan regardless of what it showed, isn’t formative assessment. It’s just a different way of asking the class a question. The value sits in the decision you make immediately afterwards, and there are really only three honest options.
Reteach it now
Reteach now, in this lesson, if the check shows a majority of the class sharing the same misconception or gap. If two-thirds of the whiteboards show the same wrong answer, pushing on to the next part of the topic just builds new content on top of a foundation you already know is cracked. Stop, reteach the specific point differently from how you taught it the first time (repeating the same explanation more slowly rarely fixes a genuine misconception), and check again before moving on. This is close to what the EEF’s 2021 feedback guidance means by its second recommendation, that feedback should be timed appropriately and focused on moving learning forward rather than filed away for later.
Flag it for next lesson
Flag it for next lesson if it’s a minority who are shaky and the rest of the class is solid. Derailing the whole lesson for six pupils out of thirty wastes the other twenty-four pupils’ time, and probably won’t fix the six anyway, since a rushed whole-class reteach delivered under time pressure rarely lands better than the first attempt did. Note who struggled. A seating plan with a quick tick or cross next to each name takes seconds and is far more useful later than a vague memory of “a few people looked unsure.” Build a short retrieval task, a starter, or five minutes of small-group support into the next lesson or the one after.
Move on
Move on if the check shows genuine, near-universal security. This is the option teachers seem to trust least, possibly because it feels like you should always be doing more. But reteaching a concept that’s already landed for the whole class isn’t cautious, it’s a waste of lesson time that could go toward the next thing, and pupils notice when they’re being taught something they already understand. Trust the check. If it says they’ve got it, let them have it and move at the pace the evidence supports.
One more habit worth building in: don’t let a single check make the decision alone if you can help it. A hinge question that half the class gets wrong is a strong signal on its own. A hinge question where three specific pupils get it wrong, and those are the same three who struggled on a whiteboard check two lessons ago, is a stronger one still, because now you know it’s a pattern rather than a bad day. These techniques work best used together over weeks, not as one-off tricks dropped into a single lesson.
Keeping it sustainable
None of this needs new resources or a policy change. A stack of mini-whiteboards, a slip of paper for an exit ticket, and a class list to cold-call from are the whole toolkit. What it costs is planning time up front, mainly in writing the hinge question’s wrong answers properly, and a bit of nerve in the room, mainly with cold-calling in the first few weeks. The Education Endowment Foundation’s Teaching and Learning Toolkit puts the average impact of improving classroom feedback at around six months’ additional pupil progress, among the higher-impact, lower-cost strategies available to a classroom teacher. That’s a reasonable trade for something you can start using in your next lesson, not your next planning cycle.
