Editorial: Race query confuses Census data

Opinion

No question in the quinquennial census causes more problems than that which asks us to identify our ethnicity. At the last census nearly 87,000 people ignored all the stated options andwrote in "New Zealander" or "Kiwi", which is not an ethnicity, it is a nationality. Statistics New Zealand assumed these people were what it called New Zealand European had included them in that category when it reported the census results. This time it will bow to their wish. "New Zealanders" will be reported as a distinct group.

This is going to reduce the value of the figures for social research. Nobody will know the actual ethnicity of any of those who identify themselves by their nationality. In one sense that does not matter because social research has long been content to regard ethnicity as a matter of personal choice rather than objective definition. The figures will tell us how many people regard themselves as Maori, European, Chinese and so on, and how many do not distinguish their ethnicity from their nationality.

But if all of the latter are of European extraction, as Statistics NZ previously assumed, their absence from that category will cause the Pakeha component of the population to be under-recorded in the census result. And those who use the figures cannot make the assumptions Statistics NZ has previously made. Anyone determined to write New Zealander into the ethnicity box needs to be aware that his or her ethnic group will be diminished.

This is not the only confusion introduced to the ethnicity question for the 2006 census. Statistics NZ has also changed the way in intends to count those who give two or more ethnic identities. Previously the department arbitrary gave priority to one identity. So that if a person ticked the boxes for Maori and Pacific Island, for example, they were recorded as Maori only. If they were Pacific Island and European they were recorded as Pacific Island. Maori had the highest priority, European the lowest.

The affect has been to inflate the Maori and Polynesian population and deflate the Pakeha proportion. When the distortion was magnified by population growth projections it produced wildly exaggerated predictions, particularly of Maori population growth. The department has now changed the way it will count dual ethnicity but the previous demographic projections have been much quoted and may take some time to dispel.

This time the statisticians will make no assumptions about people's priorities but simply record all ethnicities stated. That brings new problems. The ethnic components of the population will add up to more than 100 per cent and will be practically meaningless as a proportionate concept, which is the main use made of ethnic measures. Social policy needs to know what proportion of the population an ethnic group represents if policy makers are to know how its faring against the figures for all.

But how are statisticians to provide that information without asking people of dual or multiple ethnicity to nominate the one they primarily identify with. In many cases that would be an invidious choice to demand of them. And in any case, social policy wants to know more than the number who identify themselves as Maori or other. Self identification may be enough for deciding representation in Parliament and the like, but not in treating social problems if some of those who suffer most for their ethnicity deny their ethnicity to themselves.

There is no more sensitive subject in the census and Statistics NZ has still not got it right. This time we are going to find so little precision in the ethnic make up of the population that we might begin to pay less attention to it. That might be no bad thing.