How competencies, items and scores are governed.
The assessment system uses a defined competency framework, versioned content and recorded review. Score bands and comparative interpretations remain provisional until appropriate pilot data and expert standard-setting support them.
What is measured
Eight reporting domains cover ten connected areas at stage-appropriate difficulty: analytical and critical thinking; problem-solving and decision-making; AI literacy; digital and technological readiness; awareness of global trends; cultural intelligence and adaptability; communication and collaboration; research and information evaluation; leadership potential; and professional judgement.
The framework evaluates transferable competencies through defined tasks. It does not certify subject mastery, diagnose personality, predict individual success or replace academic records, qualifications and contextual professional judgement.
How an item becomes active
- 01Competency blueprint
- 02Structured draft
- 03Automated schema and duplicate checks
- 04Subject-matter review
- 05Cultural, bias and accessibility review
- 06Pilot item
- 07Psychometric analysis
- 08Approved item bank
AI has a limited role
AI can draft structured candidates for the review queue and help identify possible duplication. Generated content never moves directly into a live form. Model, prompt and reviewer records are versioned so the origin of an item can be audited.
GlobalPass+ prohibits emotion recognition, personality inference, facial-expression scoring, voice-based personality scoring and AI-only integrity or selection decisions.
Scoring and comparability
The provisional GlobalPass Index scale runs from 300 to 900, with competency subscores from 0 to 100. Candidates are compared only within the appropriate assessment stage. The score model is intended to support a common interpretation across regions, without region-specific curves. Comparative claims will be published only when appropriate reference data supports them.
Initial readiness bands are provisional. Final cut scores require expert standard-setting, representative pilot data, reliability analysis and documented review of subgroup performance. A confidence range will communicate measurement uncertainty rather than presenting a score as perfectly precise.
Versioning and traceability
Once a completed attempt uses an assessment, test form, scoring model or norm table, that version becomes immutable. Corrections create a new version instead of altering historical evidence.
Every certificate is traceable to the assessment version, exact form, delivered item versions, responses, scoring model, norm table, integrity outcome, template and signing key.
Validation publication
Before paid public launch, GlobalPass+ will publish the cohort, reliability, completion-time distribution, item-quality criteria, subgroup review method, limitations and the evidence supporting each score interpretation. Pilot findings will be labelled as pilot findings, not independent validation.
External review or advisory participation will be named only after appointment, consent and conflict-of-interest disclosure. No endorsement is implied while those appointments are in progress.