Before you decide which population data to use...
Define the subject population
Is it historical, current or projected? Does it include all people who live in New Zealand or just those who live in a specific area, or some subset or combination of either of these groups, or only those who were counted in a survey or census?
Information in these links discusses which data to use for historical, current and future (projected) populations.
Know what the data will be used for
It is important to match the subject population to the planning need. For instance, when analysing local social needs, the resident population may be the appropriate subject population. But for other issues such as civil defence planning, it may be necessary to consider visitors to the area as well as residents.
Other considerations include where people live, where they lived in the past and what time period is being considered. A policy may relate to a location or source of people who usually live overseas, to historical population change or to future population prospects.
Ensure the data is appropriate for your purpose
To know this, you need to know how the data was collected, compiled and disseminated.
How data is collected and processed affects information in a number of ways. Each data source has its own features to be considered. This may entail looking at the source questionnaires and understanding collation issues.
For some data sources, not all data may be available, or it may not be available in a convenient form. In some cases, part of the data will be available in one form but the remainder needs to be accessed in a different form. This is an issue particularly with administrative and historical data.
Examples of when data may be misleading
Using the fertility rate for a rural area
Subnational fertility rates may be misleading, for example, when contrasting the fertility rate for a rural area with other areas. This is because a birth recorded for a mother living in a rural area contributes to the fertility rate for that area, even if she moved to that area after becoming pregnant.
Estimating crime rates in the inner city
Daytime and transient populations of the inner city are very much larger than the resident population, and the perpetrators of crime may live somewhere else. Therefore neither resident population estimates nor usually resident census data for that area are the relevant population data to be used in the calculations.
In this particular example, one possible subject population to use would be people who usually work in the area together with people who live in the area but are not employed. However, this only gives the population at risk of being affected by crime. To derive rates of offending requires research into the dynamics of criminal activity.
Using 'people born overseas' as your population
Care is needed when using a sub-group such as 'people born overseas' as your population. This is because for any census or other data source, a proportion of people born overseas do not supply information about how long they had lived in New Zealand. And some people do not give their country of birth.
The best overseas-born population to use as a base is people who live in New Zealand and said they were born overseas. For more detailed analysis of migration patterns from particular birthplaces, the population should be restricted further to those who gave their country of birth and gave the year they arrived in New Zealand.
If you assumed that those who did not say they were born in New Zealand were born overseas, you would get the wrong population because it would include many people who were born in New Zealand but failed to answer the question.
When appropriate data is not available
When suitable data is not available, you will need to make assumptions and identify these in the results along with an analysis of the uncertainties. In the following example, prior to 1996 appropriate data to derive mortality by ethnicity is not available because of issues about its validity and non-comparability. These issues would affect use of mortality rates derived from base populations measured in different ways.
Before 1996, death registrations recorded ethnicity only when the deceased were of Māori or Pacific ethnicities. If the deceased were of other ethnicities, their ethnicities were not recorded. Therefore, there is no measure of mortality by ethnicity for people other than Māori and Pacific peoples before 1996.
In addition, it cannot be assumed that all Māori or Pacific deaths were recorded correctly - and the level of non-response is unknown. As a result, there is no valid measure of non-Māori and/or non-Pacific mortality before 1996.
Back to What is population?
Back to Frequently asked questions
Back to Population counts
Back to Population estimates
Back to Population projections