In 2003 the Australian Bureau of Statistics attempted to calculate the scale of the lost ancestries by sampling some 367,000 census forms.
This perhaps raises the question as to whether all those who had named at least two other ancestries as well as Maori would have regarded themselves as being of Maori ethnicity, or whether the 40,303 who gave sole Maori ancestry would have been the most likely to have identified as such.
Doubtless as a response to the problem of lost ancestries, in 2006 the question and guide introduced an instruction to respondents to enter two ancestries only.
By the third generation, two-thirds of men and women of these ancestries have partnered outside their ethnic group.
Men and women of Western European ancestries have similar rates of intermarriage.
Alternatively, if intermarriage is predominantly with persons who claim Australian or Anglo-Celtic ancestries it implies a higher degree of social integration into Australian society, which is composed predominantly of persons of English-speaking background.
Unlike the GSS, the census does not ask a respondent with multiple ancestries to identify the ancestry group with which the person is most closely associated.
The first two ancestries reported are coded in the order reported.
See Lieberson and Waters (1988) for further discussions on pooling data and alternative approaches in handling mixed ancestries.
A census guide handed out with the census form suggested that people should answer the question with the ancestry or ancestries that they most closely identified with, and that they could count their ancestry as far back as their great grandparents.
Most were families where both partners reported Australian ancestry or the same Anglo-Celtic or European ancestry as their sole ancestry or the first one coded of their multiple ancestries.