2001 Census data quality
For the 2001 Census of Population and Dwellings, Statistics New Zealand put in place a Quality Management Strategy (QMS) to ensure data from the census was fit for use. The QMS identified key actions Statistics New Zealand was to undertake during all phases of the 2001 Census cycle to ensure data quality.
At the processing phase of the census, certain quality assurance procedures were implemented to ensure that the data was recorded accurately and that any inconsistencies were minimised as the data was coded. These processes did not attempt to change the respondents' answers. The objective was to minimise introduced error in the data.
This strategy means that some inconsistencies may remain in the census data. There may, for example, be slight inconsistencies in table totals for some variables. These inconsistencies do not affect the fitness for use of data.
Further information on data quality in the 2001 Census can be found in chapter 7 of the Introduction to the Census.
Data quality of income questions
The income question is not as well answered as other questions in the census for a variety of reasons. These include: social aversion (including a perceived connection to IRD); embarrassment due to low income; proxy form fillers not knowing incomes (this particularly affects people in hospitals and retirement homes); pooling of household incomes (family businesses, shared investments), individuals not handling the 'books' or not knowing the gross amount received from a benefit or pension. This has resulted in a higher non-response rate than for other questions. Groups particularly affected are the elderly, the young (who usually give less detail), those in receipt of a benefit or pension and certain ethnic groups.
Changes to the form of the ethnicity questions used in the 1996 and 2001 Census of Population and Dwellings have resulted in some data that is not consistent between 1991 and 1996 or between 1996 and 2001. This applies particularly to the 'European' ethnic groups, including the 'New Zealand European' ethnic group, but also to the 'Māori' ethnic group. Data between 1991 and 2001 may not be affected.
To measure real effects of population change, comparisons should be made between 1991 and 2001 Census data.
Changes in the ethnicity question
The ethnicity question in the 1996 Census had a different format from that used in 1991 and 2001. In 1996, there was an answer box for 'Other European' with additional drop down answer boxes for 'English', 'Dutch', 'Australian', 'Scottish', 'Irish', 'other'. These were not used in 1991 and 2001.
Furthermore, the first two answer boxes for the question were in a different order from those of 1991 and 2001. 'NZ Māori' was listed first and 'NZ European or Pakeha' was listed second in 1996. The 1991 and 2001 questions also used the words 'New Zealand European' rather then 'NZ European or Pakeha'. The 2001 question used the word 'Māori' rather than 'NZ Māori'.
In all three years, people could provide more than one response to the question, for example Māori and Samoan. The wording of the 1996 question made it more explicit that respondents could tick more than one box for the ethnicity question.
Copies of the layout of the questions are contained in the more detailed technical paper referred to later in this note.
Impact on data
Some data from the 1996 ethnicity question is not consistent with data from 1991 and 2001. This is due in part to the effect of the different question format rather than an effect of real change.
- There have been increased counts for the additional 'Other European' categories from 1996. The increase was restricted to the 1996 data, with 1991 and 2001 being at much lower levels.
- There was also a drop in 1996 in the level for some categories such as 'European not further defined'. This again is a result of people in 1996 ticking the additionally specific 'other European' categories but in 1991 or 2001 just writing 'British' or 'European'.
- The count for the 'NZ European or Pakeha' category decreased in 1996 compared with that of 1991 and 2001. This may have been because some people in 1996 saw the additional 'other European' categories as better describing their ethnicity and so answered these rather than the 'NZ European or Pakeha' category. There may also be some impact on 1996 data from this category being second on the list rather than first.
- The count of 'Māori' ethnicity increased between 1991 and 1996 but there has been little change between 1996 and 2001. There may be some effect from the category being listed first in 1996 but second in the other years. Furthermore, additional categories may have increased multiple response, and there may be some effect from people answering on the basis of ancestry.
- There were more people who answered several categories in 1996 in comparison with 1991 and 2001. This has affected data comparisons for those who want to analyse single ethnicity response categories, including the sole Māori ethnic group.
Cautions for users
As a result of these question and data changes, it is recommended that users note:
- Data for the European ethnic groups in the 1996 Census is not consistent with that of 1991 or 2001. This includes data for the 'New Zealand European' subgroup, as well as other European subgroups such as 'Irish', 'Welsh', 'Dutch', and 'Italian'.
- Data for the 'Māori' ethnic group in the 1996 Census may not be consistent with that of 1991 or 2001.
- Data for the 'Asian' ethnic groups for the 1996 Census is less affected by the question changes.
- Data for the 'Pacific peoples' ethnic groups for the 1996 Census is less affected by the question changes.
- Data for single and combination ethnic groups for the 1996 Census is not consistent with that of 1991 or 2001.
- Data for prioritised ethnic groups for the 1996 Census is not consistent with that of 1991 or 2001.
Statistics New Zealand will not revise the 1996 Census data to impose consistency with the 1991 or 2001 ethnic data.
A technical paper with further details and analysis of the data is available.
General terms and definitions
General definitions can be found in Definitions and Questionnaires.