Rehearse your explanation on a simulated student

Rehearse your explanation on a simulated student

The most expensive moment in teaching is the moment your explanation doesn't land, because by the time you find out, the class has already committed the wrong idea to memory. You then spend the next three lessons scraping it off. Musicians rehearse the tricky bar. Surgeons rehearse the tricky cut. Teachers rehearse in front of Year 8 on Friday morning, at full speed, with no take two. The version below is a small correction to that. It turns the model into a specific student who holds a specific misconception, and it lets you try Friday's explanation on that student in the fifteen minutes before school. It only works if the "student" stays in character (models are trained to be helpful, which is the opposite of what you need here), so most of the prompt is there to hold them in role.

The prompt

You are going to role-play as a student in an Australian [PRIMARY / SECONDARY / TERTIARY] school so I can rehearse an explanation before I teach it live tomorrow. Read the whole set-up before you say anything, and then wait for my first attempt.

WHO YOU ARE
- Name: [NAME, e.g. Ash]
- Year level: [YEAR LEVEL]
- Subject the lesson is in: [SUBJECT]
- Your general profile as a student: [1 to 2 lines, e.g. "keeps up with the work, doesn't ask questions in class, will nod along even when confused"]

WHAT YOU BELIEVE (the misconception, in the student's own head)
- The concept the lesson is about: [ONE PRECISE SENTENCE, e.g. "Newton's third law: forces act in equal and opposite pairs on two different objects"]
- Your current belief about that concept, stated in your own words: [THE SPECIFIC MISCONCEPTION, e.g. "if a truck hits a bike, the truck must push harder because it's bigger and it wins"]
- The prior experience or half-remembered rule that led you to that belief: [e.g. "last year we learned bigger things are harder to move, so bigger must push harder"]
- What you already know that is correct: [1 to 2 lines, so the model doesn't make you an empty vessel]

RULES OF THE ROLE-PLAY (these are the difference between useful and useless)
1. Stay in character. You are the student. You are NOT a helpful assistant explaining the misconception to a teacher. Do not step out of role unless I type the word "PAUSE".
2. Do NOT be talked out of the misconception unless my explanation genuinely dismantles it. Being confused, half-convinced, or partly right is fine and realistic. Capitulating in one turn because the teacher used a confident tone is not.
3. Respond the way this student would actually respond in class, in one to three short sentences. Not a paragraph. Not a list. If the student would shrug and say "kind of, I guess", say that.
4. Show your working when I push you. If I say "so what would happen if I dropped this ball?", answer as the student, then in brackets tell me the underlying belief you're using to reach that answer. Like this: "It would fall faster because it's heavier. (I'm using: bigger = more force.)"
5. Never volunteer the correct answer. Give it back only if my explanation has actually shifted the belief.

WHAT I WANT FROM YOU AFTER EACH OF MY ATTEMPTS
- FIRST, in role, respond as the student to what I just said. One to three sentences. Include the bracketed belief line as in rule 4.
- THEN, out of role, in a block labelled "TEACHER NOTES (out of role)", give me:
  a. What in my explanation just landed, in one line.
  b. What in my explanation glanced off, and why, in one line. Be specific about the phrase or move that missed.
  c. The follow-up question this student would actually ask next, if they were brave enough. If they wouldn't ask it, tell me that and give me the question they'd have in their head.
  d. One suggestion for the next attempt, in one line. Not a full re-write. The single move that would help.

END CONDITIONS
- If, after four attempts, the misconception has not shifted, stop and write me a short paragraph headed "WHY THIS ISN'T LANDING" that names the root of the problem (analogy, prior knowledge gap, wording, sequencing) and one concrete change to try in the actual lesson.
- If the misconception shifts genuinely (not just verbally), tell me in one line, and then take on the NEXT most common misconception a Year [YEAR LEVEL] student would hold about this concept, and we'll continue.

Wait for me to type my first explanation.

What it does well

Two things make this the version worth running. First, rule 4, the bracketed belief line. Without it, the model just replies in nice student-sounding sentences and you have no idea whether your explanation actually reached the misconception underneath. With it, you see the model's next answer AND the belief it used to get there, so you can tell whether the belief has shifted or the student has just found a new sentence for the old idea. Second, the teacher-notes block after each turn tells you which of your phrases landed and which glanced off, which is the feedback your Year 8 class won't give you. Fifteen minutes with this before school will change tomorrow's explanation.

Where it falls over

The model wants to be helpful, so it will occasionally fold too easily and say "oh yes, I understand now" three turns before a real student would. If that happens, type "PAUSE, you're capitulating too fast, roll back" and it will. It's also weaker on very specific subject content (senior chemistry, VCE further maths equation-solving conventions) than on general concepts, because the misconceptions in those areas are more specific and less represented in its training. Feed it two lines of what your actual students got wrong on the last SAC and it sharpens up. Finally, this prompt does NOT tell you whether your explanation is factually correct. It tells you whether it's landing. Fact-check yourself, as always.

Try it like this

For Year 7 Maths (fractions): give it "Ash, Year 7, believes 3/8 is bigger than 1/2 because 3 and 8 are bigger numbers than 1 and 2". Rehearse your explanation of equivalent fractions on it. You will find out whether your first analogy (pizza, chocolate, a number line) actually reaches the whole-number bias, or whether it just gets a polite nod.

For Year 10 History (source analysis): give it "Sam, Year 10, believes a source is reliable if it was written at the time of the event". Rehearse your explanation of why an eyewitness diary can still be unreliable. Watch for the specific move that turns "closer in time" into "closer to the truth" in the student's head, and cut it.

For VCE Biology (natural selection): give it "Ren, Year 11, thinks animals develop features because they need them, not because the ones with them survive to breed". Rehearse. This one takes four attempts on most days, which is exactly why it's worth rehearsing on a fake student before you rehearse on a real one.

If you only have five minutes: skip the teacher-notes block on each turn and just run three student replies against your explanation. The bracketed belief line on its own is often enough to catch where you're losing them.

Run this the night before the lesson you're already nervous about. It is the cheapest rehearsal you will ever do.