Poole, J., Davies, R. and Huxley, K. (2026) Validating measures of Welsh language use. Cardiff: Administrative Data Research Wales (ADR Wales).

The Census is used as the key source of data for tracking progress towards the Cymraeg 2050 target of reaching one million Welsh speakers by 2050. Different data sources, however, show very different levels of Welsh language skills. This Data Insight demonstrates the contribution that administrative data can make in terms of validating alternative measures of Welsh language skills. Analysis reveals that the recording of fluency within school records provides a good predictor of the decisions of pupils to subsequently study through the medium of Welsh at university. This measure aligns closely with students’ own assessments of their Welsh language fluency. The Census measure of Welsh language skills is, by comparison, a poor predictor of subsequent choices made to study through the medium of Welsh.