Draft an exit ticket that actually diagnoses

Draft an exit ticket that actually diagnoses

Exit tickets got a bad name in Australian staffrooms because most of us have been handed one at a PL day, made to circle a smiley face labelled "I felt engaged today", and left wondering what the point was. The point is not the ticket. The point is the pile on your desk in the ninety seconds before the bell, sorted into "has it", "doesn't have it", and "has a specific misconception I need to address at 9am tomorrow". Dylan Wiliam has been arguing for that pile since the nineties, the William and Leahy playbook makes it concrete, and NSW's What Works Best guide sits neatly on top of it. Nobody runs it every lesson because writing three questions that actually diagnose (rather than three that just check compliance) takes longer than the five minutes at the end of the lesson you were going to run it in.

The prompt

You are helping me draft an exit ticket for tomorrow's [YEAR LEVEL] [SUBJECT] lesson in an Australian school. The exit ticket is a three-question slip every student completes in the last five minutes of the lesson and hands to me on the way out. Its job is not to make students "reflect" or "rate the lesson". Its job is to tell me, before I walk into tomorrow's lesson, who has it, who has a gap, and what the gap is.

Here is what the lesson covered:

[PASTE LESSON PLAN, LEARNING INTENTION, OR A THREE-SENTENCE DESCRIPTION OF WHAT YOU TAUGHT AND THE KEY IDEA YOU WANT STUDENTS TO WALK OUT WITH.]

Here is the class context I need you to hold:

[PASTE ANYTHING RELEVANT: recent assessment data, known misconceptions you have already heard in the room, which sub-group you are worried about, prior knowledge you can and cannot assume, the ACARA v9 or VCE study-design code the lesson maps to if you know it.]

Draft me three questions in this exact order and format. Do not add a fourth. Do not include a "how did you feel about today's lesson" question. Do not include a smiley or frowny face scale.

Question 1. Retrieval. A closed-answer question that a student who was paying attention today can answer in one sentence or one calculation. The purpose is to check the fact, not to trick them. If a student gets this wrong, they were not with me.

Question 2. Application. A short scenario or transfer question a student needs to have understood the concept (not just memorised the fact) to answer correctly. One or two sentences of student response, no more.

Question 3. Misconception probe. A question specifically designed so that the WRONG answers tell me which of the two or three most common misconceptions the student is holding. Give me the question, then underneath, list the misconception each likely wrong answer maps to. This part is for my eyes, not the student's, so I know at a glance which pile each ticket goes in.

After the three questions, give me:
1. The two or three wrong-answer patterns I should scan for as students walk out.
2. The exact opening move I should run in tomorrow's lesson if more than a third of the class trips on Question 3.

Australian context: use Australian spelling. Do not use American examples (no Fahrenheit, no US grade equivalents, no "Common Core"). Keep the reading level appropriate for the year level. No emojis. No motivational language. Every question should feel like real classroom work, not a Google Form.

If any of the class context I gave you is thin or missing, tell me what you would need to sharpen Question 2 or Question 3, then draft the best version you can with what I gave you.

What it does well

Forces the exit ticket into the shape that actually matters: one retrieval, one application, one misconception probe, with the wrong-answer patterns pre-flagged so you can triage the pile on the walk from the door to your desk. In ninety seconds you have a three-question slip that will genuinely tell you something about the class as they leave the room.

Where it falls over

It cannot design Question 2 or Question 3 well if you paste in a vague "we did fractions today". The prompt is only as sharp as the context you feed it, and the more specific the misconception you already suspect, the more useful the diagnostic will be. It also cannot mark the tickets for you. That is still your job on the walk back to the desk with the pile in your hand.

Try it like this

For Year 9 science: name the misconception around confusing correlation with causation, and ask the model to build Question 3 around it. You will get a diagnostic worth ten minutes of Monday's lesson.

For VCE English: run the prompt on your own draft SAC question first. If the misconception probe surfaces something you had not spotted, your marking rubric will be sharper by Friday.

If you only have three minutes: skip the class-context block, paste the learning intention, and take the first draft. It will not be as good, but it will beat the smiley-face version.

This one pairs neatly with Saturday's misconception re-teach prompt: one collects the diagnostic, the other builds the lesson that responds to it.