Assessment capability is where problems concentrate

Judgement about evidence, not administration of a task.

Assessing is frequently treated as the administrative end of teaching: issue the task, collect the response, apply the marking guide. The competence actually required is judgement about whether evidence is sufficient, authentic, valid and current, applied to submissions that do not match the guide.

The failures cluster at the margins, where a submission is nearly adequate, where the evidence is unusual, or where a learner has done something reasonable that the task did not anticipate. Those are the decisions that separate assessors, and nothing in a marking guide addresses them.

Consistency between assessors is the observable symptom. Where two assessors reach different conclusions on comparable evidence, that is a calibration problem, and it is invisible unless somebody deliberately compares.

The remedy is time spent together on real samples, disagreeing. A department that meets to mark the same three submissions independently and then compare will find differences and will resolve most of them in that conversation, which is worth more than any written guidance.

It is also the activity most often skipped, because it costs the whole team an afternoon and produces no artefact anybody requires. Where it is treated as validation and scheduled, it happens; where it is treated as good practice, it does not.