Published online by Cambridge University Press: 09 December 2009
The England and Wales 2001 Census was the first to include ‘Mixed’ categories which have now been adopted across government. The four ‘cultural background’ options were highly prescriptive, specifying combinations of groups. This paper assesses how satisfactorily these analytical categories captured self-ascribed cultural affiliation based on the criteria of validity, reliability and utility of the data for public services. Finally, the paper asks whether we now need a census question on ethnic origin/ancestry in addition to – or instead of – ethnic group or whether multi-ticking or a focus on family origins might give more useful public policy data and better measure the population's ethnic diversity.