AI in Schools
Mark five pieces and plan the whole-class feedback move
Mark five pieces and plan the whole-class feedback move
Tuesday of Term 3 Week 2, and the first proper marking pile is landing on your desk this weekend. The default move is to sit at the kitchen table on Sunday afternoon, work through twenty-seven pieces, write a comment per piece, and hand them back on Wednesday. That evening costs about four hours and gives students twenty-seven comments they will read for six seconds each before checking the grade at the top. The move that pays better is to mark five deliberately chosen pieces first, spot the patterns, plan the whole-class response for tomorrow's lesson, and only then write the individualised comments off a shorter, tighter feedback vocabulary. The five pieces take twenty minutes. The patterns take five. The tomorrow-morning re-teach is what students will actually remember.
The prompt below is the middle step. Mark the five yourself, paste them in, get back the pattern and the plan.
The prompt
You are a senior [SUBJECT] teacher in an Australian [PRIMARY / SECONDARY / TERTIARY] school. You have marked this task, or its close relative, enough times to know that the point of marking is not the twenty-seven comments at the bottom of twenty-seven pieces. The point is what you notice across the pile, and what you do about it in the next lesson before students have forgotten what they wrote. You mark a small deliberate sample first, spot the patterns, teach against them, and only then write the comments.
Today I have marked five student pieces from a class of [CLASS SIZE] on the same task. I have chosen them deliberately across the bands: one strong, two mid, one weak, one interesting outlier. I need you to help me turn the sample into the whole-class feedback move for tomorrow, and into shorter individualised comments I can write off a pattern grid rather than from scratch.
CONTEXT
- Year level: [YEAR LEVEL]
- Subject and unit: [e.g. Year 9 English, unit on persuasive writing]
- Curriculum frame: ACARA v9.0 [PASTE the specific content descriptor or achievement standard], or the state study design line
- Class composition in one line: [e.g. "27 students, mixed ability, three EAL/D at intermediate, two on ILPs for written output, two extension"]
- The task brief students saw: [PASTE the brief, including conditions and word or time limit]
- The rubric I marked against: [PASTE the full rubric, every criterion and every band. Do not paraphrase.]
- The one thing this task was really assessing, in one sentence: [e.g. "whether students can hold a thesis across three paragraphs of evidence, not whether they can list four persuasive techniques"]
- Next lesson is in [HOURS OR DAYS] and lasts [MINUTES] minutes.
THE FIVE PIECES (paste each in full with my mark and any comments I wrote as I went)
- PIECE 1 (I marked this as [BAND / MARK]): [paste the student work]
My margin notes: [any moves you wrote as you marked]
- PIECE 2 (I marked this as [BAND / MARK]): [paste]
My margin notes: [any moves]
- PIECE 3 (I marked this as [BAND / MARK]): [paste]
My margin notes: [any moves]
- PIECE 4 (I marked this as [BAND / MARK]): [paste]
My margin notes: [any moves]
- PIECE 5 (I marked this as [BAND / MARK]): [paste]
My margin notes: [any moves]
PRODUCE SIX SECTIONS, IN THIS EXACT ORDER. NO PADDING. NO FLATTERING PRE-AMBLE ("great sample", "rich data"). START WITH SECTION 1.
SECTION 1. THE PATTERN GRID
A table with three columns and four to six rows. One row per pattern you can see across the five pieces.
- COLUMN 1. THE PATTERN, in plain teacher language. Not "students demonstrated weaknesses in analytical writing", but "students name the technique and then move on, instead of unpacking what the technique does to the reader". A pattern is a repeated move (or a repeated missing move), not a single student's mistake.
- COLUMN 2. WHICH PIECES SHOW IT. List by piece number. If it appears in three of the five, that is a class-level pattern. If it appears in one, it is a single-student note, and it does not belong on this grid; put it in Section 5 instead.
- COLUMN 3. THE RUBRIC CRITERION IT SITS UNDER. Name the criterion and the band it is dragging the piece to.
The grid is the whole point of this prompt. Cut ruthlessly. Four patterns are better than eight; the class cannot re-teach eight things.
SECTION 2. THE MODAL ERROR
One paragraph naming the ONE pattern from the grid that, if fixed, would move the most pieces up one band. Say why, in reference to the rubric. If the honest answer is that two patterns are tied for modal, name both and say why you cannot separate them without seeing the other twenty-two pieces.
This is the pattern the whole-class re-teach tomorrow is built around. Everything else waits.
SECTION 3. TOMORROW'S WHOLE-CLASS RE-TEACH
A time-stamped block for the first fifteen minutes of the next lesson, addressing the modal error. Structure:
- 0:00 to 3:00. THE NAMING. What the teacher says to the class about the pattern, without shaming individual pieces. Verbatim, in quotation marks. Do NOT project a student's actual writing on the board. Rewrite the error into a generic version the class can look at without recognising themselves in it.
- 3:00 to 8:00. THE MODEL. The teacher shows what the move looks like when it is done well. Include the exact text or worked example the teacher writes on the board, plus one line of teacher commentary per step.
- 8:00 to 13:00. THE PRACTICE. A short task where every student in the class practises the move, on paper or in book, on a fresh example related to the original task but not from the marking pile.
- 13:00 to 15:00. THE CHECK. The specific move that tells the teacher, before the end of the fifteen minutes, whether the re-teach has landed. Not "I check in with students". A specific check (thumbs up or down against a criterion, one line copied from a student, a hands-up count on a diagnostic question). Name the pass condition and the fallback if it does not pass.
The re-teach must be TIGHT. If it is more than fifteen minutes, the class remembers the re-teach and forgets the rest of tomorrow's lesson.
SECTION 4. THE FEEDBACK VOCABULARY (for the individualised comments)
For each row in the pattern grid, one short feedback comment I can paste into a student piece that shows this pattern, worded so it is actionable, not evaluative. Not "your analysis is underdeveloped" but "you named the technique in paragraph two. Add one sentence after each mention that says what the technique does to the reader, then check whether the paragraph now answers the question".
Format: PATTERN [1 to N]. Comment: [ONE OR TWO SENTENCES, action-oriented, in the register a Year [X] student can act on].
Four to six comments total, one per pattern. That is my feedback vocabulary for the pile. Every student's individualised comment will be a combination of one or two of these plus a single sentence about that specific piece. I do not want twenty-seven original comments; I want a vocabulary I can combine.
SECTION 5. THE FIVE INDIVIDUALISED COMMENTS
One comment per piece I pasted in. Each comment:
- Names one specific thing the piece did well (not "well done"; the specific move).
- Applies one or two comments from the feedback vocabulary in Section 4, keyed to the patterns THIS piece shows.
- Ends with the ONE thing this student should do on the next task to move up a band.
Between 40 and 70 words per comment. Not longer. If the comment runs longer, the student stops reading.
SECTION 6. THE HONEST NOTES
Two to four short bullets I need before I mark the other twenty-two pieces.
- Any pattern from the grid you are less sure of, given only five pieces to work from. Name it, and the piece or two I should look for it in when I mark the rest.
- Any single-student note from the five (the piece that showed a pattern nothing else in the sample showed) that is worth a longer individual comment or a quiet corridor conversation, not a whole-class re-teach.
- If the outlier piece I named is doing something the rubric does not reward but that I quietly value, say so. That is a moderation conversation with the KLA, not a feedback comment.
- If the rubric is the reason the sample is landing where it is (thin band descriptors, criteria that do not distinguish, a scale that collapses in the middle), say so plainly. Better to know now than at moderation in Week 5.
RULES
- Australian English throughout. Year 9, not 9th grade. Term 3, not Q3. Australian assessment vocabulary.
- No em-dashes and no en-dashes anywhere. Commas and full stops only.
- Do NOT project any student's actual work on the board in Section 3. Rewrite the modal error into a generic version.
- Do NOT include student names in any output. Use "Piece 1, Piece 2" throughout.
- Do NOT recommend I write a personalised comment starting with "Dear [name]", or a smiley face, or a sticker equivalent. This is secondary or upper primary work, and the comments are professional feedback.
- Do NOT assume I have a full lesson to give to the re-teach. Fifteen minutes is the ceiling.
- Do NOT tell me to "have students self-assess against the rubric" as the whole re-teach. If the class could self-assess, they would not have made the modal error.
- Do NOT finish with "let me know if you want me to mark the other twenty-two". I will mark them myself off the pattern grid.
What it does well
Turns the twenty-seven-comment evening into a twenty-minute mark, five-minute pattern-spot, and fifteen-minute Wednesday re-teach. The pattern grid in Section 1 is the block that pays for the whole prompt, because most teachers write individualised feedback without ever naming the pattern the class shares, and the class then repeats the same modal error on the next task. Section 2's insistence on ONE modal error is what stops the re-teach becoming an eight-point lecture nobody remembers by the end of Wednesday. Section 3's time-stamped fifteen-minute re-teach with a specific pass condition at 13:00 is the discipline most re-teaches skip; the "we spent the lesson going over the essay" that quietly wasted forty minutes is what happens when you do not build the check in from the start. Section 4's feedback vocabulary is the move that changes what marking costs you: once you have four short action-oriented comments keyed to the patterns, the individualised comment in Section 5 is a combination of two plus one sentence, not an original composition. That is the difference between four hours on Sunday and ninety minutes.
Where it falls over
Five pieces is a sample, not the pile. If your five are unrepresentative (you picked one strong, two mid, one weak, one outlier as instructed, but the class actually has six students at the top who all made a different error you did not sample), the pattern grid will miss it and the Wednesday re-teach will land wrong for those six. Section 6's honest notes are where you catch this, but the harder fix is to mark the other twenty-two with the pattern grid in hand, marking against the grid rather than from scratch, and to add a row to the grid if a fifth pattern emerges. Second failure: the prompt cannot see your rubric well enough to notice a rubric problem that a marker with three years of teaching this task would notice. If the model is confidently placing a piece in the wrong band by your judgement, believe yourself, not the model. Third failure: for subjects where the assessment is a performance rather than a written artefact (music, drama, PE practical, VET workplace observation), five pieces is not the sample shape. Do not force this prompt onto assessments it is not designed for; the prompt for performance-based marking is a different one. Fourth failure: the individualised comments in Section 5 are competent but generic. If a student in your class needs a comment that names their specific piece in a way only you would know how to name (the callback to something they wrote in Term 2, the reference to a private conversation you had after Week 1), write that sentence yourself. The prompt gives you 90% of the comment. The last 10% is you.
Try it like this
- For a Year 9 English persuasive-writing pile: the modal error in Term 3 is almost always "names the technique, does not unpack the effect". Section 3's re-teach becomes a five-minute worked paragraph on the board where the teacher unpacks one technique in three moves, then a five-minute practice where every student unpacks a different technique on a fresh short text. The pass condition at 13:00 is a hands-up count on "did you write more than one sentence after your quotation".
- For a Year 10 Maths pile on quadratics: the modal error is often "extracted the common factor but did not check whether the remaining expression factorises further", not "cannot factorise". The re-teach in Section 3 becomes the two-step check written on the board with three quick examples, plus a three-minute practice on a fresh set of five. The feedback vocabulary in Section 4 has one comment for the students who did not extract the common factor at all and a different one for the students who extracted it but stopped a step early.
- For a Year 7 Science practical report: the modal error is often "wrote the method in the past tense from memory, not in the passive imperative", which is a genre convention students have not seen enough of yet. The re-teach becomes a two-minute naming of the convention, a three-minute annotation on the board of a sample method paragraph, and a five-minute rewrite of one paragraph from the student's own report. Section 4's comment for this pattern is a one-liner students can act on for the next practical.
- For a VCE analytical response: the modal error at the start of the year is usually the thesis-in-paragraph-two problem. The re-teach is the "sentence one is the answer to the question" move, modelled on the board with three variations, then practised in two minutes. The feedback vocabulary lets you write a six-word comment for every student in the pile who buried the thesis, rather than writing a fresh comment for each.
- If you only have thirty minutes tonight: mark three pieces instead of five (strong, mid, weak), run the prompt for Sections 1, 2, and 3 only. The whole-class re-teach for tomorrow is the block that saves the next lesson. The individualised comments can wait until Sunday, and they will write faster off Wednesday's re-teach than they would have from a cold start on Sunday afternoon.
The marking pile you sit down to on Sunday is the pile that decides whether the class moves up a band or repeats the same errors on the Week 5 summative. Marking five deliberately and teaching against the pattern on Wednesday buys back the evening, and gives the class the one thing twenty-seven original comments never did. The next attempt is better.