Template
Turn each criterion into a performance sample
A criterion becomes trainable when it produces something you can inspect.
When it fits
- The assessment map contains broad labels such as architecture, communication, leadership, grammar or problem solving.
When to avoid it
- One sample is evidence about that sample, not proof of competence across the whole domain.
A template to use
Criterion: [criterion]. Representative task: [prompt]. Output: [answer/case/demo]. Success evidence: [observable features].
Why it matters
For each important criterion, define one representative performance sample: an answer, solved case, explanation, decision, demonstration or follow-up exchange. Choose a sample that requires the same kind of response the real assessment is likely to require, not only recognition of the topic.
An example
Instead of 'review SAP MDG concepts,' prepare a five-minute explanation of a real design decision, then answer two challenge questions without notes.
Check your result
Someone can look at the sample and judge whether the criterion is becoming usable.
Keep this limit in mind
- One sample is evidence about that sample, not proof of competence across the whole domain.
Connected ideas
Use beforeTake a cold baseline before allocating study hours
Evidence and sources
A meta-analysis found that retrieval practice can transfer beyond identical questions, but transfer varies substantially with response congruency, elaboration and the target task.
Retrieval practice should not be treated as automatic proof of broad competence.
Transfer of test-enhanced learning: Meta-analytic review and synthesis · Abstract
Metacognitive and self-regulated learning approaches explicitly separate planning, monitoring and evaluation rather than treating study time alone as progress.
The evidence base is largely educational; adult professional assessment contexts may differ.
Metacognition and self-regulation · Toolkit overview