This report presented key results from New Zealand’s second post-enumeration survey, which was held during 21 March–3 April 2001 following the 2001 Census. Like the first survey in 1996, its expressed objective was to measure the completeness of census coverage. "For the 2001 Census, there was an intensive public information campaign. Advertising on television, radio and in the press was the key to raising awareness of the census. The main messages covered: when and how to take part in the census, why it is important and how the information is used, guarantees of confidentiality, and the availability of bilingual (English/Māori) forms. The campaign was aimed at the population as a whole, but Statistics New Zealand had also identified youth, Māori and Pacific peoples as target groups" (Statistics New Zealand, 2001).
Overall, census coverage was high. The 2001 Census missed about 85,000 or just 2.2 percent of the New Zealand residents in the country on census night. This compares with a net undercount of 1.6 percent at the 1996 Census. However, comparisons between 1996 and 2001 should be made with caution. In both the 1996 and 2001 Post-enumeration Surveys the sample size was small, the sample error was relatively significant, and the effect of non-sampling errors is unknown. In 2001 the estimate of net undercount ranged between 1.9 and 2.5 percent, and in 1996 between 1.4 and 1.8 percent. Thus the lower limit for the 2001 estimate was quite close to the upper limit for the 1996 estimate. Statistics New Zealand's experience with other household surveys suggests that the general public is becoming less willing to cooperate with information gathering. Other national statistical agencies have had a similar experience.
As Redfern (2001) points out, the net undercount in a census "is a function of many factors, including demographic, social and housing structures, public attitudes to the census, the census methods employed and the effectiveness of the instrument used to measure undercount".
All population sub-groups (with minor exceptions) returned higher net undercount rates in 2001 than in 1996. Significantly though, the results were broadly consistent between 1996 and 2001 and with overseas findings. Females were better enumerated than males. Older New Zealanders were better enumerated than children or young mobile adults. The majority European ethnic group was better enumerated than either the Māori or Pacific peoples.
It is important to note that a post-enumeration survey is one aspect of examining the quality of census output and processes. As such, the 2001 PES results will assist Statistics New Zealand in developing the enumeration strategy for the 2006 Census. A key element of this will need to be an improvement in the enumeration of the Māori and Pacific peoples, who now make up 20 percent (or 1 in 5) of all New Zealand residents.
4.1 Adjusting the base population
The PES results are integral to the derivation of reliable national and subnational post-censal population estimates and demographic projections. Following a careful appraisal of the 1996 PES results, the Government Statistician decided to adjust the population base for net undercount for the purpose of deriving post-censal population estimates (national and subnational) as well as demographic projections (for background information and details on methodology, see Statistics New Zealand, 1998b). This was an integral part of a broader change in estimation methodology introduced for post-1996 demographic series. It also included an adjustment of the base population for New Zealand residents temporarily overseas on census night, and a shift from the traditional 'de facto' concept to a 'resident' population concept (Statistics New Zealand, 1996). Census counts were not adjusted as such.
Following the approach in 1996, Statistics New Zealand intends to use an adjusted base population for deriving the post-2001 population estimates and projections. The base population will be adjusted for the census undercount, as estimated by the 2001 Post-enumeration Survey, and for the estimated number of New Zealand residents temporarily away overseas at census. Once again, the ‘adjustment methodology’ is perceived as a restricted statistical initiative. The 2001 Census figures will not be adjusted as such. Post-censal demographic series will use as the base the estimated resident population at 30 June 2001.
The assumption implicit in this approach is that the adjusted base yields a more realistic estimate of the people who, for example, normally reside in an area, pay rates, etc, and therefore should be regarded as the target population for planning, administrative and decision-making purposes.