Dialogue: How the census can misread us

By JOHN ROUGHAN

Now that the census forms are arriving, take a look at the ethnic question. It is much the same as last time except that Pakeha are no longer recognised. They want to call the likes of me "New Zealand European," whatever that is.

Never mind. There is a box beside "other" which looks a perfect size for those who want to write in something more indigenous. That is not what this is about.

When you look at the question asking which ethnic group you belong to, you will notice you are allowed to tick more than one box, naturally enough.

But when people feel they belong to more than one ethnic group, how do official statisticians work out the ethnic composition of New Zealand? Well, here's the thing.

First, those who tick Maori and any other box are classified Maori.

Next, those who tick Pacific Island and any other box except Maori are classified as Pacific Islanders.

Then, those who identify themselves as Chinese and anything else except Maori or Pacific Island are Chinese.

And so it goes on until the only people officially classified as European/Pakeha (as we were in the last census) are those who have no other ethnicity.

It rocked me to hear this explained by a Waikato University professor of demography, Ian Pool, at a conference last year. As he said, the classification system takes no account of the priority people themselves might accord to the components of their ethnic background.

And it has the effect, he pointed out, of inflating the Maori population and deflating the number of Pakeha at each stage of the count. It will deflate the Pacific Island tally, too, if more of them have married Maori than Pakeha.

After the 1996 census this counting method produced a national ethnic composition of 79.6 per cent European/Pakeha, 14.5 per cent Maori and 5.6 per cent Pacific Island. (Add the 2.2 per cent Chinese and 1.2 per cent Indian and the composition exceeds 100 per cent, so there was still some double counting somewhere.)

That often quoted 14.5 per cent figure for the Maori population consisted of about 524,000 people in the census. Nearly half of them - 250,000 - said they belonged to more than one ethnic group.

We do not know whether they regard themselves primarily as Maori because the census did not ask, and nor should it. People of mixed heritage should not have to choose one over the other.

What is disturbing is not so much the doubtful value of these statistics as the fact that they form the basis of a great deal of social science, public policy and decisions of social equity. And we are never warned to take them with a grain of salt.

Social scientists, who spend their working lives marshalling these figures into data sets that support their preconceptions, do not want to question them. And governments dare not.

Shortly after the census, an official advertising campaign for the Maori electoral option will get under way. Every time officials go to this trouble they end up being castigated, and castigating themselves, for their failure to reach much more than half the supposed number of potential Maori electors.

A great deal of time, effort, anguish and money is expended chasing what is probably a statistical phantom.

The Maori electorate is further inflated by a statutory requirement to add in the result of another question you will find in the census, which asks whether you are of Maori descent.

There are usually more people who declare Maori descent than there are who say they belong to the Maori ethnic group. In the last census 56,343 people with Maori ancestry did not identify ethnically as Maori.

Presumably they are included in the potential Maori electorate because they are indeed potentially Maori. At each census more of them appear to shift into the Maori ethnic category. But even then, they probably retain divided loyalties.

Think about what this all means. If you have wondered at the astonishing rise in official counts of the Maori population over the past decade or so, wonder no more.

And the implications for allocation to iwi of resources returned under the Treaty of Waitangi do not bear thinking about.

Tribal numbers are determined by asking those who declare Maori descent to specify the iwi in the census. Yet some of those people do not even give their ethnicity as Maori.

When it comes to closing the gaps, well, the "gaps" are statistical, based entirely on comparisons of census counts of Maori and the population overall.

One of the few social scientists to question the statistical distinctions was a Labour Department analyst, Simon Chapple, who rocked the Wellington establishment last year with a paper challenging the ethnic thrust of policies to close the gaps.

It pointed out that variations of wealth among Maori (widely defined) are much as they are among non-Maori even if the Maori average lags slightly behind the rest.

That was a useful reminder that the vast majority of Maori are not unemployed, not poor, not in prison, not in such bad health or uneducated. For so long we have been hearing they are "over-represented" in those circumstances that we have lost a sense of proportion.

Dr Rangi Walker has described them so often as a "brown underclass" and "bottom of the heap" that they themselves have come to believe it.

But when Mr Chapple concentrated on those who identified solely as Maori in the census, he found gaps. And if their number is half the supposed Maori population, the proportion registered as unemployed, in prison, in poor health and so on, is higher.

The gaps, in other words, are worse.