HCD.15:4.5 - Match evidence and reassessment to the receiving use

Specify what will be observed and which inference it may support. Separate the person’s first attempt from feedback-assisted correction; familiar performance from unfamiliar transfer; immediate from delayed performance; normal support from changed or withdrawn support; participation from contribution; and individual evidence from cohort or community reports.

Choose the lightest evidence that can change the current decision. A local practice adjustment may need one fresh discriminating attempt. A safety- or qualification-bearing reliance may need representative cases, critical-error coverage, delayed or perturbed-support observations, qualified criteria, and independent checking.

State the reassessment rule before collecting more data: which observation retains, adapts, narrows, reopens, or stops the arrangement; which rival explanation survives; and what remains unknown. A favourable observation under one configuration does not establish general transfer, long retention, or causal effect.