Project: Census Evaluation - Social Science Research Centre - University of Canterbury - New Zealand

2001 Census Evaluation

Project by Christian Galbraith

Project sponsored and supervised by Statistics New Zealand

In advance of conducting the 5-yearly Census of Population and Dwellings.

Statistics New Zealand (SNZ) estimates of the number of people and dwellings in each mesh block throughout New Zealand on census night. Mesh blocks are the smallest geographic units used by SNZ -at present there are 39,300 with an average size of 110 people. Mesh blocks are aggregated to form sub districts, which are each assigned one census collector. Sub district estimates are used in planning census enumeration, including the distribution of census forms, the design of collector contracts and as a quality check when collecting census forms.

For the 2001 Census, estimates were produced for dwellings by combining 1996 Census counts with new building consents issued between 1996 and 2001. Increases in population were derived by applying the average number of people per dwelling in each mesh block at the last census to the new dwelling estimates. This method gave estimates for private dwellings where 92% of sub districts were within +/- 10% of actual counts and population estimates where 81% of sub districts were within +/- 10% of actual counts.

The aim of the review project was to explore ways in which estimate accuracy might be improved, or alternatively to demonstrate that current estimation methods are the best option available at present. Two new options were explored in detail. First, the use of small area population projections produced by the demography division to estimate the population in private dwellings. Second, the use of floor area information from building consent data in combination with Accommodation Survey data to estimate the population in non-private dwellings (NPD's) such as hotels, hospitals and motor camps. This evaluation has shown the use of demographic projections to produce similar results to present methods. However, because new NPD's were previously excluded from the estimates system, the addition of these is a key area where improvements might be made.

  • Social Science Research Centre
    University of Canterbury
    Private Bag 4800, Christchurch
    New Zealand
  • Phone +64 3 364 2987 ext 7898
  • Contacts Page
  • Follow us
    FacebookYoutubetwitterLinked In