Every lesson has a moment it hinges on: the point where the second half only makes sense if students understood the first. The hinge question — a single, carefully designed diagnostic question placed at exactly that point — is arguably the highest-leverage move in formative assessment. It takes two minutes, involves every student at once, and its answers tell you, unambiguously, whether to proceed, pivot, or reteach. This guide explains what makes a hinge question work, how to write diagnostic questions whose wrong answers reveal specific misconceptions, and how to act on the results without losing your lesson’s momentum.
What a hinge question is — and what it isn’t
A hinge question is a multiple-choice or short-response check given to the whole class at a lesson’s decision point, designed so that each possible answer maps to a specific next step. Dylan Wiliam, who popularized the term, set demanding criteria: all students must be able to answer within about two minutes; the teacher must be able to interpret the responses within about thirty seconds; and — the crucial part — you must be able to tell, from the pattern of answers alone, exactly what to do next.
That rules out most questions teachers casually ask. “Does everyone understand?” is not a hinge question (no interpretable evidence). A recall question with one obviously right answer is not either — it can’t distinguish the student who understands deeply from the one who memorized a phrase. Even a good quiz question fails as a hinge question if its wrong answers don’t tell you why students missed it. The hinge question is engineered, not improvised: you design the distractors first, by listing the misconceptions you expect, then write the question that will expose them.
A contrast makes the design principle vivid. Weak version, mid-way through a lesson on evaporation: “What is evaporation?” — four options, three absurd, one memorized definition; the whole class picks correctly and you learn nothing. Hinge version: “A wet towel dries faster on a windy day than a still day. Why?” — with distractors blaming the wind for “blowing the water off” as liquid, crediting it with cooling the towel, or correctly identifying that moving air carries away water vapor and speeds evaporation. Now each wrong answer corresponds to a real, nameable misunderstanding you can address in the next three minutes. Same topic, same two minutes — completely different information.
what makes a good diagnostic question
So what makes a good diagnostic question? Six qualities separate diagnostic questions from ordinary quiz items:
- It targets one concept or skill — the hinge point of the day’s lesson, not a review grab-bag.
- Every wrong answer is a diagnosis. Each distractor corresponds to a known misconception, so a cluster of B’s tells you precisely which wrong idea is circulating and how widely.
- It cannot be answered correctly by guessing or surface cues — no longest-answer giveaway, no grammatical clues, no “all of the above.”
- It resists the right-answer-wrong-reason problem. The strongest formats ask for an answer and a reason, or pair two linked items, so lucky guesses expose themselves.
- It is fast — readable and answerable in under two minutes by a student who understands, including your slower readers.
- The class result is scannable — letter cards, whiteboards, or an instant poll result you can read at a glance, not thirty papers to sort.
A classroom example: mid-way through a fractions lesson, the hinge question asks which is larger, 3/5 or 5/8, with options designed so that “5/8, because 8 is bigger” reveals the whole-number misconception, while the correct choice requires genuine comparison. One glance at the room’s answers tells the teacher whether the next activity will land — or collapse.
Writing hinge questions step by step
Start from the lesson, not the question bank. Step 1: find the hinge. Identify the idea today’s lesson cannot proceed without — usually the first genuinely new concept, about a third of the way in. Step 2: list the misconceptions. Write down the two or three wrong ways students typically think about this idea; your past exit tickets and error patterns are the best source. Teachers who keep the daily one-line records described in our guide to formative assessment in practice find this step gets faster every unit.
Step 3: draft the stem so it presents a small decision or prediction, not a definition recall. Step 4: write distractors from your misconception list, one per wrong idea, plus the correct answer, matched in length and style. Step 5: plan the branch. Decide in advance what happens at roughly 80%+ correct (proceed, with a stretch task for the ready), at a split (deploy the prepared alternative explanation or peer-teaching round), and below half (stop and reteach with a different representation). Writing the branch into your plan is what turns the question from a survey into a steering wheel. Step 6: pilot and revise. After class, note which distractor attracted nobody (dead weight — replace it) and which attracted the students you expected to be secure (a better diagnostic than you planned). Hinge questions improve with reuse; a personal bank of proven ones is a career asset.
Three design mistakes quietly ruin otherwise good hinge questions. The first is cue leakage — the correct option is longer, more technical, or grammatically parallel to the stem in a way test-wise students spot instantly. The second is double-barreling: asking two things at once (“which is bigger AND why”) so a wrong answer can’t be interpreted. The third is hinging on the wrong point — placing your best question on an interesting side idea rather than the concept the rest of the lesson stands on. A quick peer review fixes most of this: hand your draft to a colleague and ask only, “If a student picks B, what do they believe?” If they can’t answer, the distractor isn’t diagnostic yet.
Acting on the answers in real time
The response protocol matters as much as the question. First, collect from everyone simultaneously — raised whiteboards, ABCD cards, or a digital poll — so confident voices can’t anchor the room. Second, read the pattern before reacting: count roughly, don’t just feel the vibe. Third, execute the branch you planned, and tell students why: “Most of us chose B, which tells me the ‘bigger denominator’ idea is still hanging around — let’s kill it with a number line before we move on.” That narration teaches students their answers steer the class, which keeps response quality high.
If the split is close, the most efficient response is often structured peer discussion: pair a correct answerer with an incorrect one, give them ninety seconds to convince each other, then re-poll with the same question. The re-poll almost always shifts toward correct — and the students who changed their minds did the re-teaching themselves.
Collection method deserves a word of practical advice. Analog tools — whiteboards, letter cards, finger voting with eyes on you — cost nothing, never crash, and make answers visible to the whole room, which fuels discussion. Digital polls and interactive slides cost a login but give you a named record: you learn not just that 40% chose B, but which students did, which matters when the same names recur across a week and a quiet intervention is due. Many teachers run analog for daily hinges and digital when they want the record — before a unit’s midpoint, for instance, or when a PLC is comparing sections. This hinge-question family sits alongside fifty other classic techniques in the classroom assessment techniques tradition; hinge questions earn their special status because they are the only ones designed, from the start, around the decision.
how long should a diagnostic assessment last
Zooming out from a single question, teachers also ask how long should a diagnostic assessment last when they give one at the start of a unit or year. The practical answer: as short as it can be while still covering the prerequisite skills and misconceptions that will matter. For a unit pre-check, 10–20 minutes (roughly 8–15 focused items) is the sweet spot for most subjects and grades. Reading and math screeners run longer by design, but even those are typically 20–40 minutes, split across short sessions for young students.
Two rules protect diagnostic value. First, diagnose only what you will act on: every item should map to a decision (pre-teach, group, skip, watch). Items “for information” waste the goodwill students bring to low-stakes testing. Second, keep it early and light: a diagnostic given two weeks into a unit is an autopsy, and one heavy enough to need a full period steals the instruction it was meant to inform. Remember too that a diagnostic’s job ends at the map — the reteaching, grouping, and adjustment it triggers are the actual intervention, and they belong to your formative routines in the days that follow. For more on designing the questions themselves, Edutopia’s formative assessment resources include teacher-written examples across subjects.
Frequently asked questions
How many hinge questions should a lesson have?
One well-placed hinge question per lesson is the standard recommendation; a long lesson may support two. More than that fragments instruction and exhausts the format’s novelty.
Can I use questions from a textbook or question bank as hinge questions?
You can, but screen them hard: most bank items weren’t written with diagnostic distractors, so their wrong answers diagnose nothing. Rewriting the distractors from your own misconception list usually fixes them.
What if students just copy a neighbor’s answer?
Simultaneous reveal (all boards up on your count), answer-plus-reason formats, and the occasional re-poll after discussion all reduce copying. So does the low-stakes culture: students copy when answers feel risky.
Are hinge questions the same as exit tickets?
No. Hinge questions sit mid-lesson and steer it live; exit tickets sit at the end and steer tomorrow. They complement each other — many teachers’ exit tickets are simply yesterday’s hinge question, re-asked.
What percentage correct means “move on”?
A common working threshold is around 80%: above it, proceed and support the minority with a targeted check-in; between roughly 50–80%, use peer discussion and re-poll; below half, reteach with a different approach. Treat the numbers as a guide, not a law — a safety-critical prerequisite deserves a higher bar than a minor detail.
The bottom line
The hinge question is a small instrument with an outsized effect because it respects the two scarcest resources in teaching: time and information. Design questions whose wrong answers diagnose, place them where the lesson truly hinges, plan your branch before you ask, and act on what the room tells you. Do that consistently and mid-lesson becomes the moment you find out the truth about learning — while the truth can still change the lesson itself.