Stability and change in ethnic identification in Australia: An aggregate level analysis

A cohort approach is used to detect changes in ethnic identification between the 1986 and 2001 Australian censuses. The aim is to observe whether particular ethnic groups were more or less likely to state the same ancestry in 2001 as in 1986. Age-sex specific survival ratios are applied to the ancestry groups in three 15-year age cohorts in 1986 to estimate the number that would have survived to 2001, adjusting for emigration and underenumeration. Some ethnic groups appear to demonstrate remarkable consistency in their ancestry response in the two censuses. Others show an increase in size between the two censuses that could be attributed to differences in the format of the ancestry question and the guidelines and examples provided on the census form. The coding of only two ancestries appears to have contributed to a decrease in the size of some groups.