Quick Read
One successful performance matters.
But before we call a capability stable, it is useful to know whether the result can reappear.
The second attempt should not be a carbon copy that can be solved by memory alone, nor a completely different task that tests transfer instead.
The Repeatability Test asks whether approximately the same capability produces approximately the same result when the important conditions are reasonably equivalent.
The One-Sentence Answer
A capability becomes more trustworthy when it reappears across equivalent opportunities without depending on the exact memory of one item, one lucky guess or one unusually favourable moment.
Where Repeatability Sits in Curie
Retrieval Stability deliberately changes time. The Robustness Test deliberately changes harmless surface features. The Transfer Test changes context meaningfully.
Repeatability asks a simpler question first: if we create another fair opportunity to demonstrate essentially the same capability, does the capability show up again?
Equivalent Does Not Mean Identical
An identical question repeated immediately may measure memory of the answer rather than repeatability of the capability.
Equivalent conditions should preserve the intellectual demand while avoiding exact item recall.
| Keep reasonably stable | Allow modest variation |
|---|---|
| Underlying capability | Specific example or values |
| Support level | Surface wording |
| Expected difficulty | Order of information |
| Time allowance | Context decoration |
| Required independence | Exact question instance |
Diagnosis: One-Off Success and One-Off Failure Both Need Caution
A single success may be luck, recent prompting or unusually good fit. A single failure may be fatigue, misreading or an accidental slip.
Repeatability protects both directions. It prevents us from over-celebrating one performance and from over-diagnosing one mistake.
English
If a learner identifies tone accurately once, a repeatability check uses another passage of comparable demand without reproducing the exact wording. If the same interpretive control reappears, confidence rises.
For developmental English stages, continue through the English Tutor route.
Mathematics
If a learner independently chooses a method in one mixed question, a second equivalent problem can test whether the route choice was stable rather than accidental.
For developmental Mathematics stages, continue through the Mathematics Tutor route.
Science
If a learner explains one causal chain correctly, another equivalent system can check whether the same mechanism reasoning reappears without the original diagram or sentence structure.
For developmental Science stages, continue through the Science Tutor route.
Additional Mathematics
A-Math performance often contains long chains. One complete solution can be unusually smooth. A repeatability check asks whether route selection, algebraic control and verification reappear in a comparable problem without being directly copied.
Repeatability and Natural Variation
Repeatability does not require identical scores or identical speed.
Human performance varies. The relevant question is whether the underlying capability remains recognisable despite ordinary fluctuation.
Stable learning can vary in surface performance without disappearing at the structural level.
Repair: Repeat the Capability, Not the Exact Item
- Define the capability narrowly.
- Record the conditions of the first observation.
- Create another opportunity of comparable demand.
- Avoid exact item memory where possible.
- Observe whether the same intellectual operation reappears.
- If results differ, use Evidence Conflict rather than averaging them away.
- If they agree, add the result to Evidence Convergence.
Tutorial and Tutor
A Tutorial can create an equivalent second opportunity while remembering which supports were present the first time. The Tutor dashboard should report repeatability plainly: “observed once,” “reappeared,” or “still inconsistent.”
What Parents Can Ask
- Has this capability appeared more than once?
- Were the conditions reasonably comparable?
- Was the second task merely remembered?
- Is the variation ordinary, or does the underlying operation disappear?
- What changes when the result does not repeat?
Voyage and Darwin
Across the Voyage, repeated independent performance matters more because responsibility is gradually moving inward. The Darwin Series asks whether a new adaptation has genuinely taken hold; repeatability is one of the first ways to see whether the repair persists beyond one success.
Boundaries
- Do not confuse exact repetition with equivalent re-observation.
- Do not demand identical human performance.
- Do not call a capability unstable from one failed repeat without examining conditions.
- Do not keep repeating once the next educational decision is already justified.
- Keep repeatability distinct from robustness, transfer and delayed retrieval.
The Core Rule of the Repeatability Test
Create another fair opportunity for the same underlying capability to appear, preserve the important conditions, avoid exact-item memory, and increase confidence only when the intellectual operation genuinely reappears.
Frequently Asked Questions
Is repeatability the same as consistency?
It is related, but Curie uses repeatability more narrowly: can we observe essentially the same capability again under reasonably equivalent conditions?
How many repeats are needed?
There is no fixed number. The evidence should be proportionate to the importance and breadth of the claim.
Marie Curie Series · Evidence Convergence · Sampling Window · Retrieval Stability · Robustness Test · Tutor
eduKate Sengkang | Small groups of up to 3 | 1.5-hour tutorials | Properly Taught Kids Shine a Bright Light Into the Future.
