HCD.14:4.4 - Pair the change with a test and stopping rule

Name the expected observable difference and the fresh task or other observation that could show it. Include a meaningful failure condition and an observation that would point to a rival cause.

The test should elicit the affected contribution without reproducing the cue the person is supposed to supply. Preserve domain criteria and compare first performance separately from feedback-assisted correction. Check protected contributions as well as the intended gain.

State acceptable learner and practitioner cost, resources, timing and any delayed observation actually needed for the claim. Do not add a retention visit merely to complete a sequence. When delay matters, HCD.13 supplies that comparison under its own resources and evidence conditions.

Specify what to retain, adapt, abandon or return after the observation. A local before/after comparison can guide the next reversible revision; an estimate of causal effectiveness needs a suitable causal design and evidence beyond this comparison.