Build hinge questions that catch the misconception

Build hinge questions that catch the misconception

The thirty seconds after you finish explaining a concept is the most expensive moment in a lesson. Move on when they don't have it and the rest of the unit sags. Reteach when they do have it and you burn a period on cotton wool. The move that closes the gap is a well-designed hinge question, four options, distractors that each mean something specific, decision made in two minutes. Dylan Wiliam has been saying this for twenty years, Harry Fletcher-Wood has written the practical playbook, and every teacher who has run a proper hinge knows the room changes when you do. The reason most of us don't do it every lesson is the same reason we don't do a lot of good things: writing six of them takes forty-five minutes on a Sunday. It doesn't anymore.

The prompt

You are helping me build a set of six diagnostic questions for a hinge point in tomorrow's [YEAR LEVEL] [SUBJECT] lesson in an Australian school. I will use them at the moment I finish explaining the concept, to decide inside two minutes whether to move on or reteach.

You are working in the tradition of Dylan Wiliam's "Embedded Formative Assessment" and Harry Fletcher-Wood's hinge-question work. That means:
- Each question is answerable in under 60 seconds and readable from the back of the room on one slide.
- A student who really understands the concept gets it right. A student who holds the specific misconception below gets it wrong in a specific, informative way.
- Wrong answers are NEVER random. Every distractor must map to a plausible misconception a real student would hold. If you cannot name the misconception a distractor represents, cut the distractor and rewrite it.
- The set as a whole must let me decide, on the spot, whether to reteach or move on.

CONTEXT
- Year level: [YEAR LEVEL, e.g. Year 9]
- Subject and unit: [SUBJECT AND UNIT, e.g. Science, forces and motion]
- The specific concept the hinge is on, in one precise sentence: [e.g. "Newton's third law: forces act in equal and opposite pairs on two different objects"]
- What most of them get wrong first time: [KNOWN MISCONCEPTION, e.g. "students think the reaction force acts on the same object as the action force, or that the bigger object exerts the bigger force"]
- What they should already know before this lesson: [PRIOR KNOWLEDGE, one or two lines]

FORMAT each of the six questions like this:

Question N of 6
  Stem: [plain, direct question. Do not use "which of the following".]
  A: [option]
  B: [option]
  C: [option]
  D: [option]
  Correct answer: [LETTER]
  What each distractor tells me about the student who picked it:
    A: [named misconception]
    B: [named misconception]
    C: [named misconception]
    D: [named misconception]
  If more than a third of the class picks [DISTRACTOR], my next move is: [ONE sentence, a specific reteach, worked example or analogy, not "revisit the concept"]

Then finish with one line labelled "Hinge decision rule": "If fewer than [X]% of the class gets questions [Y] and [Z] right, do NOT move on. Reteach using [SPECIFIC MOVE]."

Rules for the whole set:
- Australian English throughout. No em-dashes, no en-dashes.
- Two of the six must be application questions (a short scenario), not just recall.
- If the concept is numerical, at least one distractor must come from a specific procedural error, not a random wrong number.
- No padding. Six tight questions.

What it does well

Forces the model into the actual craft of question writing, which is the distractors. Every wrong option has to name the misconception it represents, so you end up with questions that diagnose thinking, not just check recall. The "next move" line for each distractor is the part that makes the lesson survive: you're not scrambling for a reteach at 11.14am, you've already got one.

Where it falls over

The model will invent plausible-sounding misconceptions that aren't real for your cohort. You have to sense-check the distractors against what you have actually heard your students say in the last fortnight. On genuinely tight VCE metalanguage or Year 12 methods content, its default distractors are still a bit crude. Feed it three sentences of what your kids wrote in the last quiz and the distractors sharpen up sharply. Its "specific reteach" suggestions will read as generic, because it does not know your class. Rewrite that line yourself in ten seconds.

Try it like this

For Year 8 Maths (equivalent fractions): concept is "two fractions with different numerators and denominators can represent the same value". Feed the known misconception "students think 3/8 is bigger than 1/2 because 3 and 8 are bigger numbers". You'll get diagnostics that pull the whole-number bias into daylight.

For VCE English Language (metalanguage on register): concept is "distinguish between formal and informal register on the basis of lexical choice, syntax and mode". Paste three lines from the article you're studying into the CONTEXT block. The distractors get much more useful when the model has real examples to work from.

For Year 7 HASS (primary vs secondary sources): concept is "identify a source as primary based on its origin, not on its age or format". Common misconception "if it's old, it's primary; if it's a photo, it's primary". Ask for two scenario questions pulled from the unit you're teaching.

Run the six in Period 3 and watch whether the decision rule actually fires. That is the whole point of the move.