Draft three annotated exemplar answers for a task

Draft three annotated exemplar answers for a task

Every Year 10 student in the country has been told to "show me your thesis". Almost none of them has actually seen a top-of-band Year 10 answer sitting next to a middle-band one, with the specific difference between them named in a sentence. Which is why they keep writing the middle one, and can't tell you why it isn't the top one. The move that fixes it is not more feedback on their draft. It's the move before the draft. Put the target on the board, put the safe-but-mid version next to it, and let the class see the gap. Dylan Wiliam has said "students cannot hit a target they cannot see" so many times the line stopped landing. Daisy Christodoulou's comparative judgement work shows the same thing from the other end. Almost nobody does it every task, because writing three exemplar answers at three standards takes an hour and a half you do not have on a Wednesday night.

The prompt

You are helping me build three exemplar answers for a task I am about to give my class, so I can show students what the target actually looks like before they attempt it themselves. This is modelling, in the tradition of Rosenshine's Principles of Instruction (worked examples reduce cognitive load for novices), Dylan Wiliam ("students cannot hit a target they cannot see"), Daisy Christodoulou on comparative judgement, and Alex Quigley on the "level-marker" version of modelling that names why one answer is stronger than another.

You will produce three exemplar answers for the SAME task, at three different standards. Then you will annotate each one with the specific features that make it that standard, in language a student can act on.

CONTEXT

- Year level: [YEAR LEVEL, e.g. Year 10]
- Subject: [SUBJECT, e.g. English]
- Curriculum framing: [ACARA v9.0 / VCE STUDY DESIGN / QCE / NSW SYLLABUS / OTHER]
- The task the students will do: [PASTE THE TASK, RUBRIC, OR ASSESSMENT DESCRIPTION AS I WROTE IT FOR THE CLASS]
- Assessment criteria I am marking against: [PASTE THE 3 TO 5 CRITERIA, OR THE RUBRIC ROWS, VERBATIM]
- What "at standard" looks like for this class: [ONE SENTENCE, e.g. "a Year 10 English student who can hold a claim across a paragraph but is not yet cueing the reader with topic sentences"]
- Optional but strong: one piece of real, at-standard student work from last term I can paste for calibration: [PASTE STUDENT WORK OR "NONE"]

THE THREE EXEMPLARS

You will write three complete exemplar answers to the task above.

Exemplar 1: TOP OF THE BAND
The answer a student in this year level actually produces when they are firing. Not a university essay pretending to be a Year 10 essay. Real, at year level, but hitting every criterion at the top of what a strong student in this class could reach.

Exemplar 2: AT STANDARD, NOT ABOVE
The answer that is technically correct and would sit in the middle band. Nothing is wrong. Nothing is exciting either. This is the "safe" answer, and it is the most important one to show students, because most of the class is writing this one and does not realise there is a level up from it.

Exemplar 3: BELOW STANDARD, IN A SPECIFIC WAY
Not a joke answer. Not something a student who did no work would produce. This is the answer a student produces when they have understood the topic but are missing one specific thing (the topic sentence, the piece of evidence, the reasoning move, the second step of the working, the link back to the case study). Name that missing thing in the annotation.

FORMAT FOR EACH EXEMPLAR

1. The full answer, as a student would submit it, at the length the task actually asks for (not a paragraph if the task is an essay).
2. Three or four dot points annotating what makes it the standard it is, in language a student can act on. "The claim is stated in the first sentence and returned to in the last one" is a student-actionable note. "Strong thesis development" is not.
3. One closing sentence: "The specific move that would take this exemplar to the next standard is ___".

STYLE

- Australian English (realise, colour, analyse).
- Written in the voice a real student at this year level would actually use. Not more sophisticated. Not less.
- If the task uses a subject-specific format (source analysis, algorithm, science practical write-up, VCE essay style, HSC extended response), match that format precisely.
- Do not label the exemplars 1, 2, 3 as top, middle, and below in what you produce for the student. In the version I will show the class, use plain numbers, so I can have the class rank them and justify their ranking before I reveal which is which.

Ready when you are. Wait for my task and criteria, then produce all three exemplars with annotations.

What it does well

Produces three same-task exemplars at three different bands, annotated in language a student can act on. The middle exemplar is the one that changes classrooms. Most of your class is writing that one and does not know there is a level up from it.

Where it falls over

The model will drift towards a "safe" middle exemplar that reads more like a university student than a Year 10 unless you paste in a real piece of at-standard student work from last term as a calibration sample. It will also over-annotate. Cut the annotations back to three dot points before you print, or the class reads the annotation and skips the exemplar itself. If your subject has a fixed structural rubric (VCE Text Response, HSC Module A, IB extended response, VCAA source analysis), paste the exact rubric verbatim in the context block, or the exemplar will look right on the surface but miss the specific structural move the markers actually reward.

Try it like this

For Year 10 English: paste your text-response task, the rubric, and one at-standard paragraph from a real student last term. Ask for three full essay-length exemplars. Show them all three side by side on the board, unlabelled, and have the class rank them before you reveal which is which. That five-minute discussion is where the modelling actually lands.

For Year 8 Maths: paste one extended-response problem. The three exemplars are "full working with the reasoning written in a sentence beside each line", "correct answer with truncated working", and "correct approach but arithmetic error in step 2". The middle one is the most useful, because it is what most of the class hands in.

For VCE Business Management or Legal Studies: the middle exemplar is the one that quotes the theory correctly but never applies it to the case study. Ask the prompt explicitly to make that the mid exemplar, then use it to show students what "apply" actually means in the study design.

If you only have ten minutes tonight: paste the task and skip the calibration sample. You will get exemplars a shade too polished, but the annotations are still worth putting on the board tomorrow.

If yesterday's Classroom Prompt got you a Claude or ChatGPT Project set up for one class, drop today's prompt into that Project. It already knows your year level, your study design and your assessment, so the exemplars come out calibrated to the actual class the first time.