Methods · Metacognition & self-regulation
E7
Exam post-mortem
Get more out of a returned paper than a feeling about the mark, by turning it into a specific list of what to change.
When to use it
- A paper just came back and you are tempted to absorb the mark emotionally and put it in a drawer.
- You want a study plan built from classified losses, not from “revise harder.”
- Your last post-mortems keep naming the same concentration and nothing changed.
What to ask AI
After you classify losses yourself, ask where they concentrate and which single fix would recover the most marks.
Compare: “Before the exam I'd have said my weakest topics were X. Actual losses were in A. What does that mismatch tell me?”
Ask which losses were gambles on material you knew you had not covered, those need a different response from tried-and-failed.
How it helps
A returned exam is the highest-quality feedback you will get: real performance, real conditions, material you prepared. Almost nobody uses it.
Stop looking at the score and start classifying the losses. “62%” is not actionable. “Eleven marks to command-word failures, nine to one misconception, four to timing” is a study plan.
Six different problems need six different remedies. Responding to all of them with “revise harder” fixes one of them.
Your reconstruction of what went wrong often disagrees with the marker. That disagreement is information about self-knowledge as well as content.
How to run it
- Wait a day or two if you need emotional distance, then work without staring at the total score.
- Before reading marker comments, reconstruct what you think went wrong on each loss from memory of sitting it.
- Classify every lost mark into exactly one category: didn't know; wrong answer type; misconception; ran out of time; misread; couldn't start.
- Ask where losses concentrate and which one fix recovers the most marks.
- Compare predicted weak topics with actual losses. Separately name which losses were deliberate coverage gambles.
- Convert the top concentration into one concrete practice change before the next assessment.
Example exchange
I classified the marks I lost on the mock. Can you tell me where they're concentrated, not the overall percentage?
Paste the categories with mark weights. I'll look for concentration and refuse a vague 'gaps everywhere' answer.
Wrong command-word answer type: 11. One misconception on rates: 9. Timing unfinished: 4. Misread: 2. Didn't know: 3.
Command-word habit is the largest recoverable block. Before we touch the rates misconception, what would fixing that habit look like in practice this week?
Three past questions where I label the command word first and only write what it asks, timed.
Do that. Rates stay a separate workstream. Don't lump it into 'revise chemistry.'
Copyable prompt
Here's my classified list of lost marks. Where are they concentrated? Is this
one problem or several? If I could only fix one thing before the next exam,
which would recover the most marks?
CLASSIFICATION:
[paste lost marks with categories and weights]The Tell
If your last three post-mortems show the same concentration and you have not changed practice. The analysis is not the intervention.
Classification without a concrete fix is an autopsy.
Principle evidence
Strength of the underlying learning idea, not a claim about AI products.
Strong as practice: turning summative papers into formative data by error type is classic assessment-for-learning applied after the fact. Scores become diagnostic only when losses are classified and acted on.
AI delivery evidence
Whether an AI tutor delivers this method well is a separate question.
AI can help spot concentration and suggest remedies, but error classification is only as honest as your categories. Treat AI typing of causes as provisional. No claim here that automated exam diagnosis is validated.