Alternatively--or rather additionally--the 2006 totals may well reflect the ongoing entry of more than two ancestries by many census respondents, regardless of the new instruction.
With only two ancestries being recorded it is difficult to know which of these ancestries is more likely to have been lost where both and a third ancestry (such as a tick-box option) were stated.
The second matter is that, quite apart from any "squeezing out" or loss of ancestries, many Maori in Australia appear quite content not to enter "Maori" as an ancestry response at all.
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.
Among the ancestry groups shown in Table 4, the second generation of Eastern European ancestries such as Polish and Hungarian shows the highest proportion (among those intermarried) with spouses who are of Australian or Anglo-Celtic ancestries.
The second generation of the three Asian ancestries shown is less likely to partner with people of Other European ancestries compared with the second generation of the two Middle Eastern ancestries shown.
While more educated men and women of some ancestries have higher intermarriage rates, there is no difference by education among men and women of other ancestry groups and in a few ancestry groups, men and women of lower education have higher intermarriage rates than do those who are better educated.
People stating Australian, English, Southern European and Middle Eastern ancestries show an increase in intermarriage rates with educational attainment.
No relation between education and intermarriage is observed for men of the second generation of Lebanese, Turkish or Macedonian ancestry; but the more educated among the women of these ancestries do have higher rates of intermarriage than do the less educated.
The low proportions of men and women with spouses of a different ancestry among the first generation of Middle Eastern, Asian and some of the Southern European ancestries partly reflect the migration of family units from these regions.
The increase is quite striking for some ancestries.
About 40 to 50 per cent of the first generation with Eastern European ancestries had spouses of a different ancestry.
Are the second and third generations of non-English-speaking ancestries who have married outside their ethnic group intermarrying with the Anglo-Celtic Australian majority or with people of similar ethnicities?
A comparison of these ancestry groups from Tables 2 and 3 shows that the majority of the second or third generation who had intermarried had spouses who were of Anglo-Celtic or Australian ancestries.
Figure 1 summarises the intermarriage patterns by generation of Australians of non-English-speaking ancestries in terms of broad regional groups with Australians of English-speaking ancestries.