Why it matters
The ALiF self-assessment questionnaire (ALiF-SAQ) is the main individual-level instrument in the ecosystem. Validation tells adopters how far they can trust its profiles, and it is also the main test of the Standard’s own structure: whether the five components and sixteen sub-competencies show up in real data.
Phase 1 (active)
Phase 1 follows the Development Protocol:
- Content validity: an expert panel of 5 to 10 rates items for relevance and clarity.
- Cognitive testing: about 10 people, two per role, think aloud as they respond.
- Pilot: 20 to 30 respondents across all five roles.
- Exploratory factor analysis: an initial look at whether the expected structure appears.
Phase 2 (planned)
- Confirmatory factor analysis with at least 300 respondents.
- Measurement invariance: whether the questionnaire works the same way across roles and demographic groups.
- Criterion validity: whether SAQ scores relate to actual AI tool use outcomes.
- Item expansion for sub-competencies currently measured by only one or two items.
The Standard Roadmap also plans multi-site validation and cross-cultural adaptation, including language versions.
Questions the Standard expects to face
The Standard names likely findings in advance. Some sub-competency pairs may correlate highly (EVA.1 and EVA.2, PRA.1 and PRA.3, ETH.1 and ETH.2) without that necessarily meaning they should merge. DAT.4 may overlap with Practical Application. The boundary between Technical Understanding and Data Literacy is the most likely point of ambiguity. Results feed into the Standard through its refinement clause and the Versioning and Review Policy.
Known limits
Self-report measures perceived competence, not performance, and can be inflated, especially at lower levels. The evidence so far comes from health care and higher education in one country. To protect validation, item wording is not published on this site.