Measure competence without creating arbitrariness
Tests and assessment
Tests may form part of a public selection process, but they are neither a mandatory ritual nor a licence to measure irrelevant attributes. Each assessment instrument must have a demonstrable relationship to the function and known, contestable marking criteria.
Pre-Foundation · planned structure, not yet established
Today
There is no entry assessment under way. When a real selection uses an assessment, its format, criteria and appeal mechanisms must be published before it is administered.
In brief
Measure the real work
The assessment must evaluate knowledge, reasoning or performance materially connected to the function.
Verifiable criteria
Objective answer keys, rubrics and panel decisions must be documented according to the type of stage.
May assist objective components
Automated marking may be used where answers are verifiable, audit records exist and review is possible.
Public employment is high-impact
Automation cannot become the final arbiter of a significant decision without a competent authority, effective review and the ability to contest it.
How it works
Link competencies to the duties of the position.
Ensure accessibility, proportionate identity checks and integrity.
Apply previously disclosed criteria and record any changes.
Allow appeal, error correction and reprocessing where there is a systemic failure.
Limits and safeguards
- An assessment cannot test political loyalty, ideological personality or economic status.
- An automated model that materially influences public employment must comply with the high-impact-system regime.
- A platform provider cannot be the final authority on the legality of its own marking process.
What this means for you
If part of your assessment is marked automatically, that must be disclosed where material, with a genuine possibility of review. A system may count answers; it cannot turn an opaque score into a fictitious administrative decision.