Quality Statement
Household crowding index
en-NZ
Household crowding is based on the Canadian National Occupancy Standard (CNOS) and provides a derived index of whether a household is crowded or not crowded, based on the number of bedrooms needed for the typology of the household and the number of bedrooms in the dwelling.
Households in occupied private dwellings
‘Subject population’ means the people, families, households, or dwellings that the variable applies to.
Household crowding uses a 2-level hierarchical classification shown in the tables below. Level 1 of the classification is the household crowding index and level 2 of the classification indicates the number of bedrooms needed or spare.
Household crowding index (Canadian National Occupancy Standard) V1.0.0 - level 1 of 2
| Code | Category |
|---|---|
| 01 | Crowded |
| 02 | Not crowded |
| 09 | Not stated |
Household crowding index (Canadian National Occupancy Standard) V1.0.0 - level 2 of 2
| Parent Category | Code | Category |
|---|---|---|
| Crowded | 1 | Two or more bedrooms needed (severely crowded) |
| 2 | One bedroom needed (crowded) | |
| Not crowded | 3 | No bedrooms needed and none spare |
| 4 | One bedroom spare | |
| 5 | Two or more bedrooms spare | |
| Not stated | 9 | Not stated |
‘Two or more bedrooms needed (severely crowded)’ and ‘One bedroom needed (crowded)’ are ‘Crowded’ households. ‘No bedrooms needed and none spare’, ‘One bedroom spare’, and ‘Two or more bedrooms spare’ are included in ‘Not crowded’ households.
Households coded to the ‘Not stated’ residual category were not attributed a household crowding value for methodological or data source reasons which are covered below.
The 2023 Census classification for household crowding is consistent with that used for the 2018 Census.
Standards and classifications has more information on what classifications are, how they are reviewed, where they are stored, and how to provide feedback on them.
Crowding data is used for a number of purposes, by Stats NZ and outside Stats NZ, including:
- as an indicator of deprivation, used on its own or combined as in the NZDep Index of Socioeconomic Deprivation
- for use in health research, particularly as a risk factor for infectious disease and for analysis of links between housing and health outcomes (along with housing quality indicators such as dampness and mould)
- analysis of housing situations for different demographic groups.
Household crowding is determined by comparing the number of bedrooms a household needs, and the number of bedrooms available in the dwelling. The number of bedrooms needed is assessed based on the demographic composition of the household, and is calculated using the following rules:
- there should be no more than two persons per bedroom
- children aged less than five years may reasonably share a bedroom, irrespective of their gender (male, female, another)
- children aged 5 to 17 years may reasonably share a bedroom only with the same gender (male only with another male, female only with female, and people of another gender only with someone else of another gender)
- single adults 18 years and over and any additional children require a separate bedroom
- household members 18 years or over should have a separate bedroom, as should parents or couples.
These rules have been adapted for gender-based data on household members based on the new standard. Prior to 2023, CNOS was calculated using sex as gender was not available.
Household crowding is calculated using a population consisting of people living in households where:
- the individuals included in the household are the New Zealand usual residents living in them. This includes individuals in the census usually resident population count, absentees, and usual residents who are overseas on census night.
- the household is a private dwelling.
People in non-private dwellings and people who could not be placed in a dwelling are excluded.
See 2023 Census data user guide for more information on subject populations.
A household is coded to ‘Not stated’ if number of bedrooms data or usual residence address data used in the derivation was sourced by statistical imputation. Other imputed data is still included (age, gender, and individual’s role in family nucleus).
Editing, data sources, and imputation in the 2023 Census describes the alternative data sources used in 2023 Census including historical census responses, admin data, statistical imputation.
A small number of records were set to ‘Not stated’ if the household was a private dwelling in a retirement village and the minimum age of the people in the household was 65 years or more, and the number of bedrooms needed was 9 or more. This is based on the assumption that retirement villages would not have private dwellings with nine or more occupants without their own bedroom.
The overall quality ratings of input variables are in the tables below. Links to the information by concept covering the input variable are included. Information by concepts provide metadata information on census data such as variable definitions, details on how quality was assessed and the quality of the variables concerned, and recommendations for the data’s use.
Input variables either contribute to the derivation of the household crowding value or are used to filter to the appropriate individuals in the household and its dwelling.
The quality of age, gender, and usual residence address was assessed on census usually resident population count which does not capture the absentee population, who are included in the household. A separate assessment covering the combination of both populations was conducted and found:
- a very high proportion of age, gender, and usual residence address data for absentee comes from high quality data sources (census forms and admin data)
- the inclusion of absentees alongside census usually resident population count individuals has minimal impact on counts across ages, genders and usual residence address meshblocks.
| Quality ratings of input variables | ||
|---|---|---|
| Variable | Overall quality rating | |
| Age | Very high | |
| Dwelling type | Moderate | |
| Gender | High | |
| Number of bedrooms | Very high | |
| Usual residence address | High | |
| Quality ratings of filter variables | ||
|---|---|---|
| Variable | Overall quality rating | |
| Absentees | Moderate | |
| Census usually resident population count | Very high | |
| Dwelling count | High | |
| Dwelling occupancy status | High | |
en-NZ
Household crowding is consistent with expectations across nearly all consistency checks, with some minor variation from expectations or benchmarks which makes sense due to real-world change, incorporation of other sources of data, or a change in how the variable has been collected.
More than 100,000 crowded households in New Zealand covers details on household crowding’s quality assessment and insights.
When using this data, users should be aware that:
- When calculating crowding data, the aim is to ensure accuracy at the household level. This is why households were coded to ‘Not stated’ when input data was imputed or in special cases such as unlikely situations in retirement villages.
- In the 2023 Census, more individuals who were away on census night (absentees) were able to be repatriated back to their usual residence, improving the capture of individuals in a household relative to 2018 Census.
- Users should note that Māori and Pacific Peoples typically have lower census response rates, and therefore 'Not stated' is slightly higher for these populations. Overall, ‘Not stated’ has reduced for Māori and Pacific Peoples since 2018.
- Households with Māori and Pacific Peoples may also have a greater proportion of input variables that are alternatively sourced. Additionally, Māori and Pacific Peoples are overrepresented in the population who could not be placed in a household. As a result, Māori and Pacific Peoples living in crowded households are likely to be undercounted. However, a higher proportion of people (including Māori and Pacific Peoples) were able to be placed in households than in 2018.
- In 2023 Census outputs, time series data for 2018 counts differ from those in 2018 Census outputs for the following reasons:
- The 2018 Census household dataset was updated after the release of the crowding data.
- 2018 published counts excluded some households which were not derived values. These households have been coded to ‘Not stated’ and included in 2023 Census output counts.
- The 2018 Census household dataset was updated after the release of the crowding data.
- Caution is advised when examining data for small geographical areas due to high volatility between census years as a result of population change and urban development. Users should be aware that there may be additional volatility in areas affected by Cyclone Gabrielle, particularly in Gisborne and Hawke's Bay.
- It is important to recognise that the changes to census methodology in 2018 and 2023 led to a better coverage of populations, particularly Māori and Pacific Peoples.
Comparisons to other data sources
Although surveys and sources produce household crowding data, data users are advised to familiarise themselves with the strengths and limitations of the sources before use, for example, while surveys such as the General Social Survey measure household crowding, the survey is only based on a sample of the population.
Finding the crowding index that works best for New Zealand: Applying different crowding indexes to Census of Population and Dwellings data for 1986–2006 contains details on the selection and assessment of CNOS being New Zealand’s household crowding measure.
Almost 1 in 9 people live in a crowded house contains details on 2018 household crowding data.
Contact our Information centre for further information about household crowding.
