HCD.22:4.2 - Make each required action consequential

For each learning target, identify a visible action and the result another action can use. A diagnostic participant might propose rival causes, interpret a measurement, question an assumption or choose a discriminating observation. A second participant must use that contribution in comparing the models, not merely acknowledge that it was spoken.

Allocate actions across preparation, interaction and later attempts. Rotating the role name alone does not rotate the action: a new “analyst” must perform analysis rather than read the previous analyst’s conclusion. A recorder can have a genuine learning target if they reconcile inconsistent claims or preserve the conditions of a conclusion. Copying notes does not establish model selection.

Choose a personal first attempt when the learner has enough preparation and the initial judgement matters for diagnosis, comparison or protection from premature copying. Preserve it before feedback so a later correction can be interpreted. When the learner first needs a demonstration, supply it, then select a subsequent task that exposes the intended action without the demonstration’s decisive cue.

At shared equipment, a non-operator can predict, identify a mismatch, choose a pause or measurement, and explain a consequence. Arrange rotation and later personal execution for actions that require control. Determine how the result is observed without making unnecessary written reporting displace the physical or interpersonal work.

If the purpose includes learning to review or explain, a supplied AI critique or explanation cannot count as the learner’s own performance. It may be a contrast or support. Specify what the person must judge, change or explain and what observation would distinguish that action.