Identity and change 1991-2001

Bola Akinwale, Office for National Statistics, UK

The Longitudinal Study of England and Wales follows 1% of residents between censuses, showing the extent of stability and instability in ethnic group categories between the two censuses of 1991 and 2001 in which they has been asked. Instability may be caused by errors in responses, data capture and imputed missing items, and in changes of question, which have been partly quantified. The Longitudinal Study then helps to answer two sets of questions that researchers ask. First, which characteristics, life events, and environmental changes promote a change in ethnic group for an individual? Second, to what extent can tabular data using the two census ethnic group classifications be accurately compared using collapsed but reliably comparable categories? Do these conversions work better for some demographic and social subgroups than for others? Here we define and use measures of reliability and comparability, and quantify the conditions under which they are maximised.