Learn Better With AI

Methods · Assessment & diagnosis · Short guide

H11

Comparative exemplar ranking

Rank three anonymised exemplars of differing quality and justify the order, building the evaluative criteria you need before you write.

When to use this
When checking yourself
Learning impact
Reliable gain
Learning evidence
Credible with caveats
Where you are with the topic
Checking readiness
How often to use it
Worth repeating often
What AI is doingEvaluate your work

When to use it

  • You can recite the rubric and still cannot order three real samples.
  • You are about to draft and want criteria talk that is grounded in examples.
  • Peer review feels vague because you have never practised comparative judgment.

What to ask AI

Give the brief and rubric. Ask for three anonymised exemplars of different quality with order hidden.

Rank them and justify pairwise. Then ask where your criteria drifted versus the intended ranking.

Finish by listing the criteria you will use on your own draft.

How it helps

Rubrics stay abstract until you apply them to real work.

Ranking exemplars forces criteria talk: why B beats C on evidence, why A still wins on structure.

That judgment transfers to self-marking and peer review.

You must rank before seeing any official order; otherwise you are recognising, not judging.

How to run it

  1. Get three exemplars on the same brief with a quality spread. Hide grades.
  2. Rank 1-2-3 with written reasons per pair.
  3. Map your reasons onto rubric dimensions.
  4. Reveal intended order if available; calibrate.
  5. List criteria you will use on your draft.

Example exchange

You

Here's the brief and rubric. Can I have three anonymised exemplars at different quality levels? Don't tell me which is which. I'll rank them.

AI

A, B, and C below, same prompt, different quality. Rank them now.

You

C, then A, then B. C has a clear claim and weighs evidence; A mostly describes; B is polished but empty.

AI

Intended order was C, then B, then A. You correctly punished polish without argument, but you undervalued B's paragraph structure. Carry “claim, weigh evidence, and structure” into your outline.

Copyable prompt

Brief: [ ]. Rubric: [ ].
Give me three anonymised exemplars (A/B/C) of different quality. Do NOT reveal
order. I will rank and justify. Then compare to your intended ranking and show
where my criteria drifted.

The Tell

You can recite the rubric and cannot order three real samples.

That is why self-assessment has been failing, criteria were never practised on work.

Ranking by length or polish only also fails: force pairwise reasons on the rubric dimensions.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Ranking anonymised exemplars draws on comparative-judgment ideas: relative quality decisions can be reliable when criteria are applied. It is discrimination practice more than a grading substitute.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI-supplied exemplars are speculative. Prefer anonymised real samples from your course when you can get them.

Related methods