A Data Base Of Iwi

Report for The Waitangi Tribunal (May 1995)

keywords: Maori; Statistics

The Data Base

Every individual living in New Zealand on the night of 5 March 1991, filled in a census form, which included a question about iwi membership.[1] Some 511,278 respondents (15.2 percent of the national total) gave a positive answer, of some form, to this question.[2] In addition 165,913 households (14.1 percent of the national total) were classified as Maori dwelling households with an iwi identification.

This computer program provides socioeconomic information about those respondents, and their households grouped together by iwi. (It does not hold unit data records.) The following is a description of the contents of the program, and the means of accession to the data. It should be emphasized that the access protocols are not simple, because as explained below, there are some restrictions on access. In addition there are numerous caveats which should be taken into consideration when obtaining and interpreting the data.

Some understanding of the data involved in the data base, may be obtained by examining Iwi Population and Dwellings (Department of Statistics,[3] 1993), hereafter called the SNZ Iwi Report (SNZIR). The report also includes copies of the census questionnaires and definitions. However this program holds the data in a much more flexible way.[4]

The Census and the Maori

The 1991 population census asked the following three questions in relation to each respondent’s Maori status.

Question 7: Which ethnic group do you belong to?
Possible answers included “New Zealand Maori” and “New Zealand European”. A respondent could tick more than one box if that was appropriate.

Question 8: Have you any New Zealand Maori ancestry?
Possible answers were “No”, “Don’t Know”, and “Yes”.

If the response was “Yes”, the respondents were asked

Question 9(a): What is the main iwi (tribe) you belong to? (please state one only)
In the case of being unable to state a tribe the respondent could answer “Don’t know” and “Don’t belong to any iwi (tribe)”.


Question 9(b): What other iwi (tribes) do you have strong ties with? (Please state no more than two iwi.)

This was the first recent census in which a question about iwi was asked.[5] As SNZIR reports there were some difficulties with question (p.12). An obvious one is that many Maori claim allegiance to belong to more than three iwi. For some iwi it was not possible to locate them geographically, and for others respondents gave a confederation, rather than iwi, name. While these issues should be addressed in the 1996 census, they warn that the figures should be used with caution. For instance SNZIR notes “the figures obtained for some iwi differ quite markedly from the numbers that are registered as beneficiaries of the iwi”.

The responses involve self categorization. Sometimes respondents miscategorize or misunderstand the question. sometimes they may recognize subtle distinctions which the user needs to be aware of. For instance, 111,330 who said they were of Maori descent said their ethnicity was solely European, including 39,204 who were nonetheless able to identify a iwi (or hapu) affiliation.[6] Note that in principle a person who was not of Maori descent could claim to be of Maori ethnicity, as in the case of a pakeha brought up in a Maori family.

As well as recording the ancestry of each respondent, the census assigns some households an iwi status according to the following rule. If the “occupier” is of Maori ancestry then the household is assigned according to the occupier’s iwi.[7] If the occupier is not of Maori ancestry, but the spouse is, then the household is assigned according to the spouse’s iwi. It is to be noted that if the occupier and the spouse are both of Maori ancestry, then the household is assigned only according to the occupier’s ancestry, ignoring the spouse iwi. One implication is that a household in a particular rohe will be assigned according to the occupier’s ancestry, even though the occupier is not tangata whenua, and the household has an active life with the local iwi through the spouse descent.

Data Accuracy

Note that not only there is an inherent inaccuracy arising for data derived form self response, but that Statistics New Zealand has a practice of rounding personal data to three in order to enhance confidentiality. For large numbers this rounding has little effect, but for small numbers the error may be significant. (For instance a cell containing one observation may be reported as containing three – or zero.)

More generally, even without the rounding, where small numbers are involved – as they are for many iwi – conclusions can be more subject to normal statistical error, and should be presented with caution.

Data Accessability

The Statistics Act 1975 has a number of provision which protect the confidentiality of the information in the Population Census (and other data Statistics New Zealand collects). The practices of Statistics New Zealand are systematically designed to effect those requirements. One can say with confidence that no individual or household would have information disclosed about them from their population and dwelling census records, except to a degree of aggregation which does not infringe their privacy.

While this resolution may be satisfactory for those without tribal origins, there is a complication in regard to tribes, since it could be argued that a tribe has some of the rights to protection of privacy to which an individual is entitled.[8] This is not the place to resolve such issues, and we have proceeded on the following basis.

First we observe that much of the data is already in the public domain via, for instance, in SNZIR. This data bases however makes the data very much more accessible. Thus one cannot simply say that since some of the data is in the public domain, then this data base should be also there. Instead the following rules have been set down for access:

1. Any iwi (or group of iwi) has access to the data base in regard to their relevant iwi. authorization of this access will come from the Director of Research of the Waitangi Tribunal, on request from an authorized representative of the iwi.

2. Any researcher can have access to any iwi data, with the agreement of the relevant iwi, as in rule 1.

3. In regard to a bona fide researcher wishing to have access to data from most, or many, iwi, where it is impracticable to obtain agreement from all those involved, the Director of Research will authorize access where:
The research approach is cross-sectional and does not depend upon identifying the precise characteristics of any specific iwi, or group of iwi; and where
Any report of the research will not identify characteristics of the iwi from the statistical base, except after agreement from the iwi involved.

4. The data by Territorial Local Authority is available on request from the Director of Research.

5. These rules apply to all researchers, including those acting on behalf of the Crown.

These rules should be sufficient to cover all the likely reasonable requests for data. They require no more of the researcher than normal research ethics demand: that the research should be sensitive to and respectful of the dignity and the mana of any subject, or in this case collective subject, that is being considered.

The Structure of the Data Base

The data base might be thought to be a large array or matrix, organized on a spreadsheet. In the case of this data base there are six separate arrays, as follows.

Main Iwi
1. MIWIPER: Individuals by their main iwi, by their personal characteristics.
2. MIWIHOU: Households by their main iwi, by household characteristics.

Any Iwi
3. AIWIPER: Individuals by any of their iwi, by their personal characteristics.
4. AIWIHOU: Households by any of their iwi, by household characteristics.

Territorial Local Authority
5. TLAPER: Individuals of Maori ancestry by their territorial local authority, by their personal characteristics.
6. TLAHOU: Households of Maori ancestry by their territorial local authority, by household characteristics.

In the case of the first four arrays, the horizontal rows represent the data for each of the 65 iwi and the 6 other categories (such as “don’t know my iwi”, plus the Maori and National totals. Information for individuals are by gender, so each iwi has two rows, one for males and one for females.

In the case of the last two arrays, the horizontal rows represent the 74 territorial local authorities (district and city councils) plus a small residual (including address not specified).

In the case of all six arrays the vertical rows represent the characteristics of those who reported iwi affiliation. The available information is reported in the next section.

Spreadsheets 1 and 2 allow the rows from the main affiliation to be aggregated together. Thus it is possible to add the people (or households) reporting their main affiliation with Ngati Iwi to those who report their main affiliation as Ngai Maori. This is not meaningful to do for the all affiliation (including main and subsidiary) since every member who claimed their main iwi to be Ngati Iwi, say, might also have put down Ngai Maori as a subsidiary affiliation, and vice versa. Adding together the “all” affiliations would involve double counting.

Spreadsheets 5 and 6 allow the rows from the main territorial local authority to be aggregated together. Thus ir is possible to add together all the people who live in the South Island (i.e. South Island TLAs).

The Available Characteristics

The characteristics available in the data base (i.e. in each column) are those asked in the Census plus in some cases summary variables (such as the average income). The details of the data are summarized in a manual prepared by Statistics New Zealand, and which reflect standard Census definitions and conventions (which are extractable from other reports such as SNZIR.) The broad characteristics they cover as follows:

Individual Characteristics
Regional Council of Usual Residence (not for Territorial Local Authority)
Age (there is also an average)
Ethnic Group
Religious Denomination
Marital Status
Total Income (there is also an average)
Type of Voluntary Work
Highest School Qualification
Tertiary Qualification
Employment Status and Work Status
Industry (Major Division)

Household Characteristics
Dwelling Type
Dwelling Tenure
Rent per Week
Household Composition

Note the data is all that which the census asks. Some may not be of direct use for Maori purposes. For instance there are very few Maori born outside New Zealand.

Using the Data Base

The data base is currently held as a spreadsheet in the Waitangi Tribunal computer system, in a Microsoft Excel (for Windows) format. Access to the data base is by approval of the Director of Research of the Waitangi Tribunal.

The following instructions assume that the user can use Microsoft Excel. An example is given in italics.

Having got Excel operating, open up the required file which will be one of:

1. MIWIPER: Individuals by their main iwi, by their personal characteristics.
2. MIWIHOU: Households by their main iwi, by household characteristics.
3. AIWIPER: Individuals by any of their iwi, by their personal characteristics.
4. AIWIHOU: Households by any of their iwi, by household characteristics.
5. TLAPER: Individuals of Maori ancestry by their territorial local authority, by their personal characteristics.
6. TLAHOU: Households of Maori ancestry by their territorial local authority, by household characteristics.

The file should open up on spreadsheet “SELECT”. If not, go to Spreadsheet “SELECT”. (This spreadsheet is “protected” – that is the user cannot directly change it – except for B5 and the A column below A5.)

In B5 enter the collective name for the group of Iwi being considered. e.g. “Nga Kohore-Iwi”.

In Column B from B6 down are the iwi in alphabetical order,[9] or the Territorial Local Authority (TLA) in (roughly) north to south order. For individual information each iwi or council has two entries, one male and one female, which are shown in column C.

In column A, assign next to the selected iwi (or TLA) a “1”, for each gender where appropriate.[10]
e.g. in A138, A139, A140, A141, A144, A145, A148, A149 (which combine “Dont know my Iwi”, “Dont belong to any Iwi”, “Other/Not elsewhere specified”,[11] “Not Specified”).[12]

At this stage you may want to check A3, to see how many iwi are selected. In this case “4”.

Change to the “TABLES” spreadsheet (which is fully protected)

If satisfied, the tables can now be printed. (There is a print icon on the second row, and a print procedure in “File”.)

The “MAIN” spreadsheet (which is also fully protected) is the working spreadsheet, where the numbers for tables and the charts are calculated. Normally it can be ignored.

Each of the other spreadsheets is a chart, the contents of which are indicated by the spreadsheet name. Clicking to that chart, followed by a print instruction, will give a printout of the chart.

When closing the workbook do not save. This will make it easier for the next user, who will not have to clear the previous choices.


The data base presents a powerful means to enhance our knowledge and understanding of Maori social and economic characteristics by iwi and location. However the data is the data base is subject to various limitations, and must be interpreted with care. In addition it is the mana and dignity of those who are being investigated must be respected.

1. Throughout this report the term “iwi” will be used to mean “tribe”. It is nowhere used to mean “people” or “nation”.
2. Including 152,896 who were unable for one reason or another, to specify a main, or usually any, iwi.
3. In 1993 the Department of Statistics changed its name to Statistics New Zealand. Its Maori name is Te Tari Tatua.
4. Two other useful 1991 census publications by Statistics New Zealand are National Summary (1992) which gives data for the aggregate population as a whole, and New Zealand Maori Population and Dwellings (1992), which gives figures for the Maori in aggregate, but not by individual iwi.
5. In 1987 a more limited iwi question was asked in the “Attitudes and Values” Survey for the Royal Commission on Social Policy. The April Report, Vol I, pp.397-700. The question was asked earlier in the censuses from 1874 to 1901, see J. Lowe
6. On the other hand of the 146,118 who could not identify a iwi or hapu affiliation 42,009 said they were of solely Maori ethnicity.
7. The “occupier”, a technical term for census purposes, is defined as the “person who is in charge of the dwelling on Census night”. It may be the owner, the person in whose name the dwelling is rented, or some other responsible person. Where a private dwelling is owned or rented jointly by two or more persons only one of these is “the occupier”.
Little is known about how a married couple decide which of then is the “occupier”, although there is tendency for the husband to take that role. For flats and the like, the issue is even murkier. Thus any use of the household data, based on the occupier’s characteristics should be done so with caution. As explained further this applies especially to the iwi affiliation of a household.
8. While data about them is not collected in the Population and Dwellings Census, business enterprises (and other corporate bodies) have similar confidentiality protection over any data collected about them by Statistics New Zealand. Thus the iwi claiming confidentiality is not unique, in terms of a corporate entity.
The distinction might be made that while the business enterprise data is collected directly from the institution itself, the iwi data is collected from the members of the iwi. Thus the parallel is not exact. Nevertheless the point remains that the recognition of rights of confidentiality of a corporate body already exists.
9. The data base follows the convention that “te”, “ngati”, and “ngai” are not part of the iwi name for alphabetical purposes. Hence the first identified iwi is “Te Aitanga a Mahaki”, and the last specified iwi is “Ngati Whatua”.
10. Note that after the specified iwi are “other” grouped in to eight regions, plus 6 categories of unspecified iwi, the largest of which is “dont know my iwi”.
11. It is not entirely clear whether this group should be included. However the numbers are so small the decision is not vital.
12. In the case of the “all iwi”, in which the respondents identify up to three iwi of affiliation, only one iwi should be identified. The spreadsheet has been booby-trapped so that if more than one iwi is identified, it will immediately become apparent in the tables. (These spreadsheets are of the form AIWI*.XLW.)

Appendix: EXAMPLE

In the above illustration considered those of Maori ancestry who do not have, or know, an iwi affiliation. This avoids the protocol problems discussed in Chapter 1. Because this is primarily an exercise to show use, only salient features of the Nga Kohore-Iwi will be discussed here.

By way of background, the Nga Kohore-Iwi are likely to consist of a mixture of two main groups. Those who have become most absorbed into Pakeha society and those who have become marginalized from all society, are each likely to have lost their iwi affiliations. It may be that further investigation in each of the four categories of Nga Kohore-Iwi will show different balances between the groups.

While one tends to use the tables for interpretation, in practice it has been found easier to look at the charts in the first instance.[1] Here we go through each characteristic briefly, highlighting the main differences.

In total there were 146,124 New Zealanders of Maori descent who were unable to identify an iwi affiliation. This was 28.6 percent of all Maori, and 4.3 percent of all New Zealanders. Note that over one quarter of those of Maori descent do not know their iwi.

Table 01 Regionally, there are relatively fewer Nga Kohore-Iwi than iwi affiliated Maori in the Northern part of the North Island, including the Hawkes Bay, but excluding Auckland city. Thus there are proportionally fewer of Maori descendants who do not know their iwi in these regions. Conversely in the South Island and the lower part of the North Islet (excluding Wellington) there are proportionally more Kohore amongst the Maori.

Table 02 In age terms, the Maori average almost ten years younger than the average New Zealanders, and the Nga Kohore-Iwi are about a 1½ younger than the Maori average. In terms of those over the age of 15, to which most of the subsequent tables (from Table 06) refer, the gap is a little smaller. All Maori are 7½ years younger than average than all New Zealanders, with the Nga Kohore-Iwi just over a further year younger on average.

Table 03 Not unexpectedly birthplace has proved to be an unrevealing statistic about the Maori, with over 98 percent born in New Zealand, in comparison to 83 percent of all New Zealanders. The Nga Kohore-Iwi are slightly more likely to have been born outside New Zealand than all Maori. If so it is in Australia and the Pacific Islands.

Table 04[2] In terms of their ethnicity, while 56.7 percent of Maori describe themselves as solely Maori, and another 20.2 percent as of part Maori ethnicity, only 28.8 percent of Nga Kohore-Iwi describe themselves as solely Maori and 19 percent mention part Maori ethnicity. Almost half describe themselves as solely European/Pakeha.

Table 05 Religion provides one of the most curious findings. The proportion of all Maori and all New Zealanders who associate themselves with some religious affiliation is much the same – just under 30 percent. However almost 40 percent of the Nga Kohore-Iwi have no religious affiliation. The proportion of the Nga Kohore-Iwi who associate themselves with a Maori oriented church, such as Ratana and Ringatu, is less than the all Maori 9.7 percent, Nevertheless some 4.0 percent do. Also the Nga Kohore-Iwi are more likely to be Presbyterians than other Maori.[3]

Table 06 In regard to marital status the Maori is more likely to be never married, and the non-Maori more likely to be married and divorced, perhaps because of their younger age structure. Nevertheless among both all Maori and the Nga Kohore-Iwi, separation (and divorce on an age adjusted basis) is higher with lower remarriage.

Table 07 The Nga Kohore-Iwi (over the age of 15) have higher average incomes than the Maori, $16460 against $15310, but markedly lower the national average of $18,990.

Table 08 While all Maori and all New Zealanders do about the same amount of voluntary work, the Nga Kohore-Iwi are likely to do a little less. The differences in the pattern between Maori and Pakeha probably reflect the multi-purpose marae based work subsumed under “domestic service”.

Tables 09 & 10 The Nga Kohore-Iwi are slightly better educated than all Maori, but less so than the national average.

Table 11 The Nga Kohore-Iwi are typically about midway between all Maori and all New Zealanders in their employment status.[4]

Table 12 There is not a lot of different between the Nga Kohore-Iwi and all Maori in their industries of employment, except the former tend to be more predominant in wholesaling and retailing, and the latter in community services. Surprisingly the Nga Kohere-Iwi ares lightly more predominant in the agricultural sector than all Maori.

Table 13 Again the occupational differences are not so great, although The Nga Kohore-Iwi seem to have a slightly better status than all Maori, but still not as good as the non-Maori.

In summary, the Nga Kohore-Iwi seem to be those Maori who are more successful, on material measures, than other Maori, but not as much as non-Maori. There are two major anomalies which suggest the story may be more complex than simply success (or greater intermarriage). The Nga Kohore-Iwi are younger than the Maori, although we might have expected them to have an age structure more like the non-Maori. And they are more detached from a religious affiliation than either all Maori or the average non-Maori.

It would be easy to jump to some judgements about the Nga Kohore-Iwi, but the implications are that their situation involves some subtle processes. There may well be a case for further investigation of them.

1. Incidentally, the entire set of charts and tables took less than ten minutes to produce, from entering up the Excel system to closing it down, and the last coming off a 4ppm laser printer. Admittedly the user in this case knew precisely what he was selecting, but the point is that obtaining this information from a spreadsheet is not onerous.
2. Here an in later tabulations the column heads are sometimes foreshortened to the point that SNZIR needs to be consulted.
3. Incidentally, the data confirms the anecdote of great diversity in church affiliation by iwi.
4. Note this table has two charts.