Stats NZ has a new website.

For new releases go to

www.stats.govt.nz

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Proposed topics for 2018 Census

We have grouped content into broad topics to help show the possible changes to existing content and new content being considered. The topics are:

  • Population structure
  • Location
  • Second address / residence
  • Ethnicity, culture, and identity
  • Education and training
  • Work
  • Income
  • Families and households
  • Housing
  • Telecommunications
  • Transport
  • Health

Each topic discussion covers the value of the data, current and emerging information needs, and issues we have identified. We then recommend whether the content should be included, excluded, or changed for the 2018 Census.

See 2013 Census forms and guide notes to find out what we included in the 2013 Census individual and dwelling forms.

Population structure

Overview of population structure content being considered for the 2018 Census

Existing content to remain the same

  • Age(1)
  • Name(1)
  • Number of occupants on census night(1)
  • Absentees

Recommended new content or improvements to existing content

  • Sex(1)
  • Partnership status in current relationship

More information required to recommend inclusion

  • Number of children born alive
  • Legally registered relationship status

Note: 
1. Content that is required to be collected under the Statistics Act 1975.

The value of this data

The priority for the census is to get the population count right. Beyond this, understanding and monitoring changes in population structure is important for an informed society. All policies and social services are influenced by population change.

Population structure is one of the census topics most frequently used by census data customers. Customers include central government agencies, local authorities, private organisations and businesses, researchers, communities, and the public.

Existing content to remain the same

Age, name, number of occupants, absentees

Age is regularly used to help analyse other census variables. For example, to investigate population ageing and to provide the base population for fertility, mortality, morbidity, suicide, and accident and crime statistics.

Name, number of occupants, and absentees (people who were temporarily away from home on census night) are used for internal collection and processing purposes. These pieces of information help us ensure we have an accurate count of the New Zealand population and its families and households. We use name and number of occupants on census night to check that the individual forms completed in a dwelling match the people listed on the dwelling form. We use information on absentees to help derive household composition, family type, and other family and household variables.

Recommended improvements to existing content

Sex

Sex is the distinction between males and females based on the biological differences in sexual characteristics (from current Statistical Standard for sex). The sex standard is used in all Statistics NZ household surveys.

Sex is a fundamental demographic characteristic used in social and population analysis. It is used to analyse most social statistics such as employment, health, and education. Collecting data on sex is a legal requirement under the Statistics Act 1975.

While including sex in the census is not in question, there is potentially a need for changes to the categories used. Currently there are two categories: male and female. However, some people are born biologically intersex and some people make transitions. Although this group is small, currently they cannot represent their biological sex in this question.

In the lead-up to the 2013 Census, a ‘two ticks for sex’ online campaign was launched. It urged respondents to express their dissatisfaction with the sex question by ticking both the ‘male’ and ‘female’ boxes on their individual census form. The campaign achieved a high profile in the lesbian, gay, bisexual, and transgender (LGBT) online communities, and attracted some mainstream media coverage.

We recognise the need to further explore and test the implications of collecting information on those who are biologically intersex. Concerns have been raised around how this would affect the quality and comparability of the sex data from the census. Defining what is meant by biologically intersex may be difficult on a self-completed form with finite space. Another potential issue is that including an intersex category may elicit false responses from some respondents.

To date, no any other country has included an intersex option in their national census.

Sex is distinct from sexual orientation and gender identity – which we discuss in the ethnicity, culture, and identity section.

Partnership status in current relationship

We currently produce data on legal marital status and social relationship status from the census. We need to determine whether it is still relevant to produce data on legal marital status, and consider improving partnership status categories.

The move away from legal marriage is continuing, and the distinctions between legally registered relationships and other partnerships are becoming more blurred. For example, under the Relationships (Statutory Reference) Act 2005, people living in de facto relationships for more than 12 months are legally treated the same as couples who are married or in a civil union. Interest in people’s social relationship status over their legal relationship status appears to be increasing.

One issue with the data produced on social relationship and legal marital status is the poor quality of the civil unions data. The 2006 and 2013 Census data on civil unions was rated unfit for use and was not output. The census figures were significantly higher than those from civil union registrations data. We believe this was caused by respondents misinterpreting what a civil union is.

There may be less interest in the civil unions data now than previously. Since same-sex marriages were legalised in August 2013, the number of civil unions registered has dropped dramatically and some civil unions have been transferred to marriage. Interest from our customers seems to have shifted to distinguishing between same-sex and opposite-sex relationships rather than between marriage and civil unions.

If information on civil unions is collected in 2018, one way to improve data quality may be to include a note on the census form that refers respondents to an explanation of civil unions in the guide notes.

Please tell us what data you need about different forms of relationship status. 

More information required to recommend inclusion

Legally registered relationship status

We’re examining whether data on people’s legally registered relationship status still needs to be produced from the census.

While demand for information on legal marital status appears to be decreasing, removing the legal marital status question in the census would mean we would no longer be able to produce marital rates, divorce rates, and nuptial and ex-nuptial birth rates.

We want to know if our customers still need us to produce these rates.

Number of children born alive

Number of children born alive is a cyclic variable (normally asked every second census).

While birth registrations (administrative data) are the key source of information on fertility, census information on the number of children born alive helps us understand changes in family size, patterns and timing of childbearing, and patterns of childlessness. It is our only source of data on childlessness and overseas births to women now living in New Zealand. It permits detailed socio-economic and socio-cultural studies of fertility.

We are considering including the question on number of children born alive in the census on a cyclic basis, as the data typically shows less change over time than other census variables. Including this in the last two censuses (2013 and 2006) and 10 years before that (1996) allows us to explore if the trends are variable or consistently changing. In other words, do the trends vary enough and at different rates to warrant inclusion for 2018, or is inclusion every 10 years sufficient?

Declining family size and increased childlessness are factors to consider when deciding the need to include number of children born alive in the 2018 Census. Census data indicates childbearing and childlessness patterns are changing, but in a relatively consistent direction and pace over the past 20 years (since 1996). Given the constant direction and pace of change since 1996, the need for inclusion in 2018 is questionable.

Beyond national changes, previous census consultation submissions stated that fertility is a continually important component of population change nationally, at detailed geographic levels, and for particular groups, because of:

  • New Zealand’s changing ethnic composition
  • large population flows in and out of New Zealand
  • an ageing population.

All the major ethnic group populations had corresponding increases in childlessness. Conversely, the average number of children born (completed family size) of women decreased in all the major ethnic group populations.

As well as considering the consistency of change in the data over time, we need to consider the sensitivity of this information. During the 2013 Census, there was a petition about the sensitivity of this question because it excluded stillborn babies. In recognition of its sensitive nature, respondents have the option of objecting to answering the question.

However, respondents who find this question sensitive may still prefer not to respond at all. This may be contributing to the relatively high non-response rate for this question.

Proposed recommendations for content on population structure

  • Collecting name, age, and number of occupants on census night is required by law under the Statistics Act 1975 so these must be included in the census. No changes are recommended for name, age, and number of occupants on census night.
  • Collecting data on sex is required by law under the Statistics Act 1975 and will be collected in the 2018 Census. We will do more testing and research to explore the possibility of collecting information on those who are biologically intersex.
  • We recommend absentees be collected with no changes in the 2018 Census.
  • We recommend that information on people’s social status with respect to relationships be collected in the 2018 Census. We welcome feedback on which categories are required for the data produced on relationship status, and whether we still need to collect and produce information that identifies people’s legal marital status.
  • Further evidence would need to be supplied by experts to support including ‘number of children born alive’ in the 2018 Census. [check edits for 'by experts']

Join the online discussion forum for ‘population structure’ content 

Location

Overview of location content being considered for the 2018 Census

Existing content to remain the same

  • Address on census night(1)
  • Address of dwelling(1)
  • Usual residence
  • Usual residence five years ago

Recommended new content or improvements to existing content

  • Usual residence one year ago(2)

More information required to recommend inclusion

  • Years at usual residence

Notes:
1. Content that is required to be collected under the Statistics Act 1975.
2. Content that was not collected in the 2013 Census.

The value of this data

The census is New Zealand's most valuable source of information about the size and characteristics of New Zealand's population at a subnational level. The location information collected in the census, combined with other topics, enables us to better understand the differences between our communities, and to target resources where they are needed.

As the census is the key source of information at low-level geographies, location information is used extensively to formulate, monitor, and evaluate central and local government planning and policy. Local authorities, local government, and health authorities have specific requirements for census-based populations at low levels of geography for use in planning and development.

Location information is also critical for understanding population movement within New Zealand (internal migration), which is a key component and driver for changes in population estimates and projections. Population estimates and projections are critical to the timely provision of infrastructure and services.

Existing content to remain the same

Address on census night, address of dwelling, usual residence, usual residence five years ago

Dwelling address and census night address are statutory requirements of the New Zealand census and allow population counts of both residents and visitors present on census night.

Usual residence is the address where a person considers themselves to usually live. It provides a count of all people who usually live in, and were present in New Zealand on census night. It is the only means of distinguishing between those usually resident in New Zealand and overseas visitors, and between those usually living in an area or just there on census night.

Usual residence five years ago is used to:

  • record population movements within New Zealand (internal migration)
  • produce the linked (or longitudinal) census dataset
  • estimate the inter-censal components of population change.

Recommended new content

Usual residence one year ago

Usual residence one year ago was last included in the 1981 Census and is proposed for inclusion in the 2018 Census.

New Zealand has high external (international) and internal migration rates by international standards. Therefore, we need accurate estimates of internal migration to produce accurate population estimates. Population estimates are the key census output and underpin many key government decisions, so they must be accurate.

The census is currently the only official data source on internal migration as captured through ‘usual residence five years ago’ and ‘years at usual residence’. However, to calculate annual subnational population estimates we require an estimate of movements over the last year. This information is currently not available from the census or other official sources. It is not possible to obtain an estimate from the data on usual residence five years ago because there is no sensible way to derive annual movements from five-year movements. This is a major limitation on the accuracy of the current subnational population estimates. For internal migration estimates, the value comes from combining the two measures.

Usual residence one year ago gives information on the characteristics of internal migrants, and assesses the propensity of people to change residence (internal migration). To calculate annual subnational population estimates we require an estimate of movements over the last year. We do not currently have this. This is a major limitation on the accuracy of our current subnational population estimates.

On the other hand, usual residence five years ago smooths random fluctuations and gives a robust spatial pattern over a longer period.

We are also investigating ways to increase the use of administrative data to produce subnational population estimates. However, a major limitation is not having a source of information on annual movements to calibrate the annual movements from administrative data against. As outlined above, we can’t use data on usual residence five years ago for this. The ability to calibrate administrative data against annual transitions from census data on usual residence one year ago would also be of substantial benefit for work on future censuses.

Internationally, the UK Office for National Statistics (ONS) includes usual residence one year ago but excludes usual residence five years ago (they carry out a census every 10 years). The Australian Bureau of Statistics (ABS) includes usual residence both one and five years ago.

We already collect up to four addresses in the census. Adding the proposed new content on usual residence one year ago, second address/residence (see the topic discussion on this), and educational institute (see the transport topic discussion) would mean we would collect up to seven addresses on the 2018 Census form if no address content is dropped. We may be able to reduce the burden on respondents by designing the questionnaire to reduce the amount of address information they need to provide.

Second address / residence

We are considering including second address/residence in the 2018 Census. This could potentially cover a range of information needs, some of which may be relevant to those interested in the location topic.

See the topic discussion on second address / residence for more information 

More information required to recommend inclusion

Years at usual residence

Years at usual residence is the number of completed years up to census night that a person has lived at their usual residence.

To date, the data collected on years at usual residence has been used to help produce internal migration estimates, which are a key component of population estimates and projections. However, with the suggested inclusion of usual residence one year ago and usual residence five years ago, having data on years at usual residence becomes lower priority for internal migration purposes.

One potential use of the data on years at usual residence is to indicate the social composition of an area by breaking it down into new and longer-term residents. However, the value and need for this information is unclear at present.

If we continue to collect information on years at usual residence, we could reduce the respondent burden associated with asking multiple address questions. Respondents who had lived at their usual residence for more than five years could be routed past the questions on usual residence one year ago and usual residence five years ago.

Please tell us how you use information on years at usual residence so we can make an informed decision its inclusion or exclusion for the 2018 Census.

Proposed recommendations for content on location

  • Collecting information on census night address, dwelling address, and usual residence is required by law under the Statistics Act 1975 so this information must be collected in the census. No changes are recommended for census night address, dwelling address, and usual residence.
  • We recommend that usual residence five years ago be included with no changes in the 2018 Census.
  • We recommend that usual residence one year ago is included in the 2018 Census.
  • Given the recommendation to include usual residence one year ago, we invite feedback on whether we still need to collect information on years at usual residence. We need information on how customers use this information so we can make an informed decision on its inclusion or exclusion for the 2018 Census.

Join the online discussion forum for ‘location’ content 

Second address / residence

Overview of second address / residence content being considered for the 2018 Census

Recommended new content or improvements to existing content

  • Second address / residence(1)

Note:
1. Content that was not collected in the 2013 Census.

The value of this data

We are considering collecting information on second addresses or residences in the 2018 Census. The most appropriate name for this new topic depends on the exact type and scope of the information to be collected, which is yet to be determined. The underlying concept for this topic is of another address at which a person stays some of the time. This does not include properties that a person owns but rents out.

Potentially, this data could be valuable for planning services and allocating resources at the local level, for research and policy concerning families and households, and for improving understanding of New Zealand’s housing stock. It could also be of value to Statistics NZ to improve operational processes and ensure data quality.

Recommended new content

Second address / residence

Many people in New Zealand move between different addresses or have more than one residence. Examples of this are:

  • children whose week is split between separate parental homes
  • students whose term-time address differs from their home address
  • people whose working week address differs from their home address
  • people with seasonal lifestyles who spend part of their time in New Zealand and part of their time overseas.

These complex living arrangements highlight limitations with the usual residence information currently collected in the census. This question can prove difficult to answer for those who move between different addresses.

Several information needs relating to second addresses or residences have been identified. These are diverse and multi-dimensional, ranging across different topic areas. The information requirements identified are a need to understand:

  • regional population fluctuations
  • family connections across households
  • New Zealand’s housing stock
  • second home ownership rates.

Collecting information on second addresses or residences could also provide operational gains for Statistics NZ and help to ensure the quality of the population count in the census.

The relative priority and value of these different information requirements is not clear at present and needs to be established. Potential pieces of information to be collected could include the location of a second address or residence and the purpose of that address or residence.

Regional population fluctuations

In regions where the population fluctuates significantly at different times of the year (eg holiday home areas), demand on services and resources such as medical services and transport also fluctuates. Information on second addresses or residences would improve understanding of these fluctuations and be helpful for planning services and allocating resources in these areas.

Family connections across households

An acknowledged limitation of the census data on families and households is that it relates to people living together in one dwelling (ie people who are members of the same household), and does not provide information on family connections and support that extend across households.

Some children whose parents live in separate households spend time living with each parent. These children are not identified in the census data. They may be classified as being in a one-parent family even if they spend equal amounts of time with each parent. This is contrary to the common understanding of what a one-parent family is and has implications for analysis of different family types. Accurate information about children in shared care is required.

Another situation the census does not capture is couples who live apart, or only live together some of the time and have separate homes. This is known as living-apart-together.

Collecting information on second addresses or residences could provide useful insights into movements of people between different households and dwellings. This could be used together with existing family and household data to gain a better understanding of families and households.

Housing stock and second home ownership rates

Customers who use housing data have said that we are not capturing sufficient information about New Zealand’s housing stock because we don’t collect information about ownership or use of other dwellings such as second homes and holiday homes. These dwellings may be unoccupied at the time of the census. The information currently collected on unoccupied dwellings is limited to their location and whether they were empty or the residents were away.

Collecting information about second addresses or residences could provide information on second home or holiday home ownership and improve our understanding of the country’s housing stock. However, it would not provide information on ownership of other dwellings in general, such as properties that are rented out. This is a different topic, and would require collecting a different type of information.

Census data quality and operational gains for Statistics NZ

The primary purpose of the census is to provide an accurate count of the population. To do this, people must only be counted once. This is becoming more difficult with more people moving between different dwellings. The shift to a mail-out and online-first approach for collection also makes it more challenging to check that the correct number of people has been counted at each dwelling.

Collecting information on second addresses or residences could help us make sure people are only counted once, because we could identify duplicate returns. These are cases where people return a census form from two locations – eg two forms may be received for children of separated parents.

Another benefit of collecting information on second addresses or residences is that it could help explain discrepancies between census and administrative data at low geographic levels and inform development work for future censuses.

Internationally, the benefits of collecting this type of information in the census have been recognised. The Office for National Statistics (ONS) in England and Wales collected information on second address and the purpose of that address in their 2011 Census. They used a definition of 30 days or more a year to identify second addresses. Statistics Canada identified second address as something of interest but did not include it in their 2011 Census, nor does it appear they are considering it for their 2016 Census. The Australian Bureau of Statistics (ABS) found support for collecting information on second address in their content submissions process and is currently considering including it in their 2016 Census.

If this consultation shows strong support for including second address or residence, we’ll need to determine exactly which types of information to collect, and the definitions that would be applied. As with all new census content, we would need to consider its suitability for inclusion in a self-completed form, and test to ensure that good-quality data could be produced.

It may not be possible for the census to meet all information needs identified in this topic area, because the amount of information we can collect on any one topic is limited. Therefore, we need to establish the relative priority of the different information requirements identified.

We also need to think about respondent burden when considering including second address/residence in the 2018 Census. We already collect up to four addresses per respondent in the census. Including second address/residence, as well as usual residence one year ago and educational institution would increase this to up to seven addresses per respondent if no existing address content is dropped. We may be able to reduce the burden on respondents by designing the questionnaire to reduce the amount of address information they need to provide.

Proposed recommendations for content on second address/residence

  • We recommend that second address or residence be considered for inclusion in the 2018 Census. We invite feedback is invited on information requirements relating to second addresses or residences, and which types of information would be of the greatest value and highest priority to collect.

Join the discussion on second address / residence 

Ethnicity, culture, and identity

Overview of ethnicity, culture, and identity content being considered for the 2018 Census

Existing content to remain the same

  • Ethnicity(1)
  • Māori descent
  • Birthplace
  • Languages spoken
  • Years since arrival in New Zealand
  • Religious affiliation

Recommended new content or improvements to existing content

  • Iwi affiliation

Content not recommended for inclusion

  • Generational attachment(2)
  • Citizenship(2)
  • Sexual orientation(2)
  • Gender identity(2)

Notes:
1. Content that is required to be collected under the Statistics Act 1975.
2. Content that was not collected in the 2013 Census.

The value of this data

The variables included in the ethnicity, culture, and identity topic provide vast and distinct value to a range of users. The value of each of these variables is discussed below.

Existing content to remain the same

Ethnicity

Ethnicity is the principal measure of cultural identity in New Zealand, and is used across official statistics. The census provides a baseline count for all official ethnicity statistics and is widely used with other demographic characteristics to inform research and policy development. Collecting ethnicity information in the census is required under the Statistics Act 1975.

We initiated a Review of the Official Ethnicity Statistical Standard in 2008 to understand the increase in written responses of 'New Zealander', from 2.4 percent in the 2001 Census to 11.1 percent in the 2006 Census. The review was used in conjunction with census cognitive testing and a research programme to inform decisions about collecting ethnicity data in the 2011/13 Census.

This review found an almost universal view amongst customers that the format of the census ethnicity question should remain unchanged. They emphasised the importance of consistency in all official statistics and comparability of the ethnicity data over time. This also reflected concerns based on past experiences which show that even a minor change in questionnaire format can have unintended but significant effects on responses and subsequent statistical outputs.

Following this review, the Government Statistician decided there would be no change to the ethnicity question for the 2013 Census. We also decided that ‘New Zealander’ responses received on the 2013 Census form would continue to be reported under the ‘other ethnicity’ category. Responses of ‘New Zealander’ have fallen since 2006, to 1.6 percent in the 2013 Census.

Given that ethnicity underwent a significant review before the last census, and that no other significant issues have been identified, we recommend that the information collected on ethnicity should not change for the 2018 Census.

Birthplace

Data on birthplace is used to develop, monitor, and evaluate settlement programmes for immigrants, and analyse their socio-economic status and demographic characteristics. Government agencies and researchers widely use it, and it is of considerable interest to immigrant groups themselves. It is particularly important given increases in immigration to New Zealand in recent years.

Years since arrival in New Zealand

The data on years since arrival in New Zealand for people who were born overseas supplements the birthplace data. Years since arrival in New Zealand has several uses, such as:

  • distinguishing recent immigrants from those who have lived in New Zealand for longer periods
  • monitoring changes in the characteristics of the immigrant population over time
  • evaluating and developing immigration policies and programmes
  • developing, monitoring, and evaluating settlement programmes for immigrants.

Languages spoken

Data on languages spoken is used to monitor knowledge of languages other than English among different groups. It can help with:

  • measuring the level of language change and the impact of government initiatives, such as Government’s Māori language revitalisation programme
  • understanding the health of the Māori language
  • determining what languages to provide services in
  • developing policies and planning services for the deaf community.

Religious affiliation

Data on religious affiliation is used:

  • by researchers and religious organisations to trace the changes in values and belief systems in New Zealand society
  • to assess the need for various types of religion-related or religion-sponsored services, including those of churches, mosques, temples, and religious schools; and for the care of the aged and other disadvantaged groups
  • by churches to assess their level of support
  • as an explanatory variable for studies on subjects such as marriage formation and dissolution, educational attainment, and income
  • by Māori, Pacific peoples, and other ethnic groups for whom this information is an important aspect of their culture.

Although many uses have been identified, religious affiliation is frequently identified as a variable of decreasing relevance, most recently in consultations with key census users for the 2013 Census. There are several reasons why people question the ongoing relevance and inclusion of religious affiliation in the census.

Firstly, census data on religious affiliation is not widely used by government agencies or deemed highly important for policy development, evaluation, or monitoring. The census is not legislated to collect the information and it is not closely connected to the core reasons for conducting the census.

Secondly, New Zealand is becoming increasingly secular, with just under half of the population either stating they have no religion or that they object to answering the question. The percentage of people reporting no religion has steadily increased for the previous three censuses (29.6 percent, 34.6 percent, and 41.9 percent, respectively). However, those objecting to answering the question have decreased since in the same period (6.9 percent, 6.5 percent, and 4.4 percent). It could be argued that with more and more people not identifying with a religion, its relevance and importance to New Zealand society decreases further.

However, demand for this data continues from religious organisations, academic researchers, Māori and Pacific communities, and the media. Interest in religious affiliation has been increasing with the growth in the cultural diversity of New Zealand, which has been accompanied by a shift away from mainstream Christian affiliations. Those who identify as having no religion also constitute their own demographic, and religious affiliation varies across different age, ethnic, and other population groups.

Without census data on religion, any changes in this significant cultural area would be difficult to monitor.

Māori descent

The New Zealand census has produced information about Māori descent in various forms since the mid-1800s. Māori descent information is used in conjunction with Māori electoral registrations to calculate the Māori electoral population, which is subsequently used to determine the number of Māori electorates. Māori descent information is also used for projections of the Māori descent population. At the 2013 Census, Māori descent and ethnicity data were used to establish the sample for Te Kupenga, a post-censal survey of Māori well-being.

Māori descent is used as a routing question for iwi as they are formed through ancestry (whakapapa). Respondents who indicate they have Māori descent are asked to proceed to the iwi question and give the names of their iwi.

Two areas relating to Māori descent could be considered for 2018 to improve data quality:

  • trying to reduce the non-response rate
  • reviewing how non-response to the Māori descent question is dealt with, given its role as a routing question for iwi affiliation. The Māori descent question has traditionally had a high non-response rate (9.9 percent in 2013).

With an online-first approach for 2018, and the expectation that more people will complete their forms online than on paper, there is potential to reduce non-response. We could direct people who answer iwi affiliation without answering Māori descent back to the Māori descent question, with a note that this question must be answered.

Although we have collected all iwi responses since 2001, we define the subject population as those who said 'Yes' to the Māori descent question. We don’t include the iwi responses of those who did not answer the Māori descent question, or said 'No' or 'Don't know', in the output data.

Improving the non-response rate for Māori descent would reduce the number of people not counted in the iwi affiliation data. An investigation of the Māori descent question is underway to improve collection and responses. Additionally, a point often raised is that a person who answers the iwi affiliation question without responding to the Māori descent question could be assumed to have Māori descent.

Any change to how we collect these variables are collected and processed would need to be carefully considered as it would affect the comparability of the data over time.

Improvements to existing content

Iwi affiliation

Iwi affiliation data is used to:

  • monitor the performance of Treaty of Waitangi obligations by the Crown and iwi
  • help the allocation of resources and funds to iwi
  • help with Waitangi Tribunal decisions on, for example, land ownership and fishing rights
  • help in iwi planning and social and economic development
  • help central, regional, and local government agencies to plan and provide services to iwi in areas such as housing, health, social welfare, and special assistance programmes
  • help local government in administrating the Resource Management Act 1991.

The census is the only source of comprehensive iwi information. Iwi listed as part of the iwi classification can get comprehensive information regarding their people, at highly detailed levels. Iwi use the classification to formalise their iwi status, as it is used as the source of official statistics on iwi population numbers for Treaty claims and public policy. Therefore being on the classification is significant for iwi, particularly where there are long-held disagreements over identity and self-determination.

Academic research has identified that there is a degree of mismatch between census data on iwi affiliation and iwi registers (Walling, Small-Rodriguez, & Kukutai, 2009). It has been suggested that the information collected in the census should include registration status instead of relying upon self-identification to address this.

The 2009 iwi classification update working group noted several issues with the iwi classification review process, and with the guidelines for inclusion that are part of the iwi statistical standard (Iwi classification). The working group accept that some groups are excluded from the classification, and some are incorrectly classified. Considerable time has been invested responding to these complaints, including several requests under the Official Information Act 1982, but the underlying issues with the iwi statistical standard and classification, and Statistics NZ being placed in a role of arbitrating who is and is not an iwi, have not been resolved.

The 2009 working group recommended that the statistical standard and classification be reviewed before the 2018 Census. This project is currently in progress and is expected to be completed in time to inform the collection of iwi affiliation data in the 2018 Census.

Content not recommended for inclusion

Citizenship/residence status

Interest in information on citizenship was identified during consultation for the 2006 Census, and to a lesser extent, 2011 content consultation. The types of organisations that have expressed a strong interest in this data include central government agencies, community groups, private organisations, and researchers. Examples include the New Zealand Immigration Service and the University of Waikato.

Outside of the census, information on people’s visa status is available as they cross the border, as is information on visa approvals and applications. However, there are difficulties in measuring citizenship. For example, some people not born in New Zealand have citizenship of right (eg people born in the Cook Islands). Others have lived in New Zealand for many decades and have never taken – or needed to take – citizenship (eg people under assisted passage provisions in the 1950s).

A significant population of people are here on other residence provisions. People who are permanent residents in particular are equivalent to New Zealand citizens for almost all administrative purposes. Related concepts such as birthplace, religious affiliation, and ethnicity are collected in the census and may be used as alternative measures. The inability of a population census to disaggregate the residence status of people born overseas is a current limitation.

The argument can be made that the census is not the best way to collect the detail to distinguish between different citizenship, residency, and visa statuses. Multiple questions would be required to meet all information needs, and these would be difficult and overly technical for respondents. In a self-completed questionnaire, the questions needed to accurately collect the information would likely cause significant confusion and burden for respondents. A respondent’s visa or citizenship status may even deter them from filling in the census.

Generational attachment

Generational attachment refers to information on the origin of respondents’ parents for the purpose of understanding long-term migrant outcomes.

The settlement of migrants and their children, and how this affects labour markets and social services, is becoming a wider policy consideration both in New Zealand and internationally. Length of residence in New Zealand is associated with settlement outcomes (eg labour markets and social services), and is currently collected. Capturing generational attachment would require further development.

Key policy agencies have identified, over several censuses, a strong interest in data to help track migrant outcomes, particularly among the children of migrants. Ethnicity or country of birth are often used as a proxy to understand migrant outcomes, but this is not entirely satisfactory.

Information on generational attachment is available from the General Social Survey, but the level of disaggregation that is possible is less than what we could obtain through the census. A new question, probably about the birthplace of respondents' parents, would directly identify first-generation migrants. A further question would be required to identify second- and third-generation migrants, and due to complexity is unlikely to be possible for 2018.

Gender identity

Information on gender identity has been identified as an area of interest. There is an emerging need and desire to recognise gender equality, although the potential uses of census data on this are not yet clearly established.

Gender identity is a person’s internal, deeply felt sense of being male or female (or something other or in between). A person’s gender identity may or may not correspond with their sex (see Human Rights Commission definitions).

Currently there is no standardised approach for collecting and storing gender identity data in New Zealand. Development is well underway on statistical standards and classifications on gender identity. While gender identity is not currently collected in any Statistics NZ source, several organisations collect the information in some form, including the Ministry of Health, Department of Corrections, and the New Zealand Transport Agency.

A 2014 report from the Human Rights Commission (HRC) highlighted the need for visibility as a means of promoting equality and decreasing the likelihood of discrimination. HRC argued that insufficient data limits policymakers’ ability to quantify the issues affecting gender identity minority groups and to develop ways to adequately address the health and social needs of these populations.

Whether a standalone question would work on a self-completed form such as the census is not well understood and has not yet been tested. People may confuse or not understand the difference between ‘sex’ and ‘gender’ – which are conceptually different, and not interchangeable.

Sexual orientation

Information on sexual orientation is an area of interest in New Zealand and internationally. The gay, lesbian, and bisexual communities are becoming more visible as societies become less discriminatory about non-heterosexual relationships.

Sexual orientation is not currently collected in the census, or any other Statistics NZ survey. In 2008 we produced a discussion paper as part of a review of cultural identity statistics (Statistics NZ, 2008). A statistical standard for sexual orientation has not yet been developed.

We received submissions requesting information on sexual orientation be collected for the 2001, 2006, and 2013 Censuses, and the General Social Survey (GSS). These submissions identified a desire to be recognised as well as wider information needs. Independent research we commissioned supports a need for information on sexual orientation, with growing evidence that members of sexual minority populations are disadvantaged across a range of social well-being, health, and economic indicators (Pega, Gray, & Veale, 2010).

If a statistical standard for sexual orientation is developed, the census may not be the best way to meet customer requirements for this information. Sexual orientation is multi-faceted in its definition, including elements of behaviour, desire, and identity. It is conceptually difficult to define and measure, particularly given the self-completed nature and limited space of census forms.

We must also consider data quality and intrusiveness. Many respondents may not answer the question, and although census data may produce indicative figures, these may not be accurate enough to be analysed against other variables or published. Our focus group research identified quality issues and concerns about the accuracy of information people are comfortable disclosing.

To date, we are not aware of any other country that has collected information on sexual orientation in their census. Similar concerns about quality and complexity have been raised. Statistics Canada (2011) carried out extensive testing, which indicated that sexual orientation is best suited to inclusion in surveys that have a particular context (eg health surveys). The Ministry of Health collects data on sexual orientation in the NZ Health Survey.

Proposed recommendations for content on ethnicity, culture, and identity

  • We recommend that ethnicity, birthplace, years since arrival in New Zealand, languages spoken, and religious affiliation be included with no changes in the 2018 Census.
  • We recommend that Māori descent be included in the 2018 Census with no changes to the type of information collected, but that options for improving non-response be investigated.
  • We recommend that iwi affiliation be included in the 2018 Census. However, the iwi affiliation question, iwi statistical standard – including guidelines for inclusion in the iwi classification – and the iwi classification itself may change depending on the outcomes of a review of the statistical standard for iwi. This review is in the initial stages, but is planned to be completed in time for developing the 2018 Census content.
  • We recommend that citizenship, generational attachment, gender identity, and sexual orientation not be included in the 2018 Census.

Join the online discussion for ‘ethnicity, culture, and identity’ content 

Education and training

Overview of education and training content being considered for the 2018 Census

Existing content to remain the same

  • Highest secondary school qualification

Recommended new content or improvements to existing content

  • Post-school qualification
  • Study participation

The value of this data

Information on qualifications is widely used by central government agencies including the Ministry of Education (MoE), the Ministry of Business, Innovation and Employment (MBIE), the Ministry of Youth Development, the Ministry of Social Development, and the New Zealand Council for Educational Research.

Data on education and training is required to monitor current levels of participation and achievement of qualifications and long-term changes over time, both for the general population and for particular groups and regions. This is important for:

  • making evidence-based educational policy
  • allocating and targeting resources at disadvantaged groups
  • the provision of training services
  • monitoring the skill level of the workforce.

Long-term data continuity and international comparability are important.

The most important measure for users is the highest level of qualifications held by individuals. This is a Tier 1 statistic. Tier 1 statistics are the official statistics regarded the most important – those which are essential to understanding how well New Zealand is performing. We derive the census data on highest qualification using highest secondary school qualification and post-school qualification.

Specific uses of data on education and training include:

  • monitoring the impact of educational reform policies on skill levels
  • examining the extent to which people use their qualifications in employment, eg the link between level and field of study and occupation
  • examining the relationship between the level of educational investment and the return on investment as reflected in occupation and income
  • identifying potential skill gaps in the labour market and planning education and training programmes
  • studying efficiency and equity issues in education policy
  • preparing occupational priority lists for recruiting skilled migrants
  • determining decile rankings for schools receiving government funding
  • determining teacher supply projections.

Although information on education and training is collected elsewhere (eg in the Household Economic Survey (HES), the Household Labour Force Survey (HLFS), and through administrative data), the census is a valuable source because it provides information on small geographic areas and for small population groups. Low-level geographic information is also available from administrative sources such as the data collected by the Ministry of Education.

The census data is currently the only source of information on field of study. Field of study is not collected in the HLFS, GSS, or HES.

We are in the early stages of investigating whether administrative data could meet the need for information on education and training. This is part of our development work for future censuses. Early indications are that there is potential for administrative data to meet this information need in the medium to long term. Ongoing findings from these investigations will help to inform us about the best way to collect and process information on education.

We need to better understand how well the information needs of Māori are being met. For example, whether information on skills apart from formal qualifications would be of value.

Existing content to remain the same

Highest secondary school qualification

School achievement is of high interest to central government, which has set a public service target to boost skills and employment by increasing educational achievement for New Zealanders. The current target is for 85 percent of 18 year-olds to achieve Level 2 or equivalent qualifications by 2017. This is to ensure New Zealanders are meeting their potential through education, to boost skill levels in the New Zealand workforce, and to improve the overall productivity of New Zealand.

Although these targets can be monitored using administrative data, census data has the advantage of indicating which population groups to target. This is because of the unique ability to cross-tabulate qualifications data from the census with other variables collected in the census.

As well as being an output variable in its own right, highest secondary school qualification is used along with post-school qualification to derive highest qualification.

We need confirmation from key customers such as the MoE on how this data is used – including whether census data is being used to support and monitor education targets.

Recommended improvements to existing content

Post-school qualification

The information collected on post-school qualifications is the name of the qualification, including the level if applicable, and the subject.

We collected this using write-in boxes, recognising the great diversity of qualifications. However, it was difficult to code, and errors affected the quality of the data produced on post-school qualification and highest qualification.

We need to consider alternative ways to collect this information, such as a shift to a tick-box approach, or using administrative data. A tick-box approach is used in the Household Labour Force Survey (HLFS), Household Economic Survey (HES), and General Social Survey (GSS), so changing to this approach in the census could improve data comparability across surveys.

Currently, subject is used in conjunction with the name of the qualification to determine the qualification level. However, if information on post-school qualifications was collected using a tick-box question, this would no longer be necessary from a processing point of view. We therefore need to better understand how people use the field of study data.

Using tick-boxes to collect qualifications information has known limitations. Respondents usually know their qualification name, but they may not know the level.

Internationally, the Office for National Statistics (ONS) in England and Wales uses a tick-box approach, but the Australian Bureau of Statistics (ABS) uses a write-in box.

Early investigations indicate there is potential to use administrative data to improve the quality of the data on qualifications gained in New Zealand. This work has also shown that there is value in identifying which qualifications were gained overseas. This would help inform our understanding of whether administrative data on qualifications can replace census data on qualifications in the future.

Study participation

Some changes to the type of information collected on study participation may be needed.

Study participation measures those aged 15 years and over attending, studying, or enrolled at school or anywhere else. The categories used currently for those who are studying are: full-time study (20 hours or more a week), and part-time study (fewer than 20 hours a week).

Study participation can be cross-tabulated with a wide variety of other census variables – eg income, age, and ethnic group – to understand the education participation rates of different groups. These cross-tabulations can provide valuable information, such as information about mature students living at their parents’ homes. Information on study participation is important for Māori and iwi who are concerned about their youth participating in study at all levels. If we collect information on travel to education in the 2018 Census, which is being considered, we may need to ask about study participation as a lead-in for travel questions.

Consultation with the MoE, one of the main users of education data from the census, indicated that census information about study participation is not useful to them since we do not collect information on the level of study (eg tertiary education). However, if we include information on travel to education, the level of study could potentially be derived from that (eg from the educational institution name or address).

Study participation information is available through alternative data sources, including administrative records and household surveys (eg HLFS). However, the scope for those participating in study varies between each source. The definition and scope of the information collected on study participation within the census needs to be considered.

The census uses definitions of full-time and part-time study based on the number of hours of study each week, whereas Statistics NZ surveys such as the HLFS ask questions to understand formal versus informal study.

Administrative records provide good information on study participation, but currently do not distinguish between part-time and full-time study.

With an increase in the types of courses available, users’ needs regarding the scope of study participation data may be changing. For example, it may be useful to include industry training. Currently it is not clear whether we are capturing this type of information in the census.

We should consider whether information needs can be met through other available sources. If we ask questions on travel to education in the 2018 Census (see transport section), we would likely still have to include study participation to route or validate answers.

Proposed recommendations for content on education and training

  • We recommend that highest secondary school qualification be included with no changes in the 2018 Census.
  • We recommend that post-school qualifications be included in the 2018 Census and that options for changing the way we collect the data be investigated further. These changes could include using a tick-box question and identifying whether the qualification was gained overseas.
  • We recommend that study participation be included in the 2018 Census. We welcome your feedback on your current and anticipated uses of this data, including anticipated uses if information on travel to education is collected. We also want to know which types of information would best meet your needs (eg full-time versus part-time study, or formal versus informal study, and the inclusion or exclusion of industry training).

Join the online discussion forum for ‘education and training’ content 

Work

Overview of work content being considered for the 2018 Census

Existing content to remain the same

  • Hours worked per week
  • Industry
  • Work and labour force status
  • Occupation
  • Sector of ownership
  • Status in employment
  • Workplace address

Content not recommended for inclusion

  • Veteran population(1)
  • Unpaid activities

Note:
1. Content that was not collected in the 2013 Census.

The value of this data

Statistics NZ produces a number of employment and unemployment measures, each with its own strengths and weaknesses. The Household Labour Force Survey (HLFS) provides the official measure of employment and unemployment in New Zealand, giving a regular, timely, and comprehensive portrayal of New Zealand's labour force. The collection of data on work in the census complements the HLFS. Because it covers the whole population, it allows more useful regional and demographic breakdowns.

Information is collected by the census on several areas relating to paid work to understand the labour market position of small population groups and small geographic areas. Information is also collected on unpaid activities.

Paid work information from the census is used extensively by a variety of organisations, from central and local government to community groups, academic researchers, and businesses. The data can be used to analyse:

  • occupation and industry composition at a fine level of detail
  • the size and characteristics of the labour force
  • the links between income, qualifications, and labour market outcomes
  • changes over time.

Labour force status is also used as a broad indicator of socio-economic status.

Data on workplace address is also used in combination with other variables to produce information on transport flows, which is highly valuable for transport planning.

Examples where census work data gives a significant input into a variety of products include:

  • Production of employment forecasts by councils.
  • Long-term fiscal projections produced at least every four years produced by Treasury. A significant input to Treasury’s long-term fiscal model is the demographic and labour force projections arising from census-based material.
  • Job vacancy monitoring and assistance by the Ministry of Business, Innovation and Employment (MBIE). Census data is used to allow research into factors affecting supply and demand for selected occupations at a regional level; monitor outcomes for migrants and their settlement patterns; develop regional labour market reporting; monitor trans-Tasman skills flows; and analyse housing supply and demand, including future projections.
  • Reweighting the labour cost index (LCI) by Statistics NZ. This index provides a measure of wage inflation and is used in wage negotiations, contract escalation clauses, economic research, and policymaking.

Information collected in the census on unpaid activities covers activities performed in the four weeks before the census date, without payment, for people living either in the same household, or outside. This ranges from housework to unpaid care and voluntary work. This information provides indicative data on the nature of unpaid activities undertaken, as well as the characteristics and location of those involved.

Users of data in this area include central and local government, and researchers. A key user of unpaid activities data is the Ministry for Women, to understand the difference in unpaid work between men and women. Other interest in this data appears to be around the voluntary work and care-giving components of this data.

However, unpaid work data is seen as currently being of relatively low value in terms of uses and outputs, mostly because no time component is collected. This is discussed in more detail below.

Existing content to remain the same

Paid work information

Collecting information about paid work is a vital function of the census. All the existing variables are valued and used. We expect the content to remain the same for 2018 Census.

While we do not expect to change the content of paid work information, we are considering making relevance improvements about the people who complete this information.

The census collects information from all usual residents. A usual resident is defined as anyone living in or who intends to stay in New Zealand for a period of at least 12 months.

But the census does not collect paid work information on people on temporary work visas who are not classed as usual residents. For some industries and geographical regions this group can be a significant portion of the workforce and population. Not capturing this group can lead to data on particular industries and regions being distorted. There is a possibility that some of this population is captured in the census if they don’t follow the guide notes saying they shouldn’t fill in the entire form.

A lot of temporary work is seasonal in nature, so the census is unlikely to be the best way to capture this section of the labour force. We know it may be difficult to obtain a high response rate from this population group because they are mobile. Another potential issue is that a significant number of this identified group may not be here on official temporary work visas and will likely not complete the census accurately.

Despite this, the need to have indicative data about workers who are not usual residents is high. We will investigate the potential to track this population group, especially through using administrative data sources.

Content not recommended for inclusion

Veteran population

Whether information on the veteran population should be collected in the census has been raised previously. Information on veterans was last included in the 1971 Census.

We received submissions during the consultation periods for the 2006 and 2011 Censuses to collect information about the number and characteristics of the veteran population in New Zealand. Veterans’ Affairs is responsible for providing services to people who have seen active service in war, including those in a peacekeeping role. The issue identified at the time was that the total number of people – and therefore the total liability – was not known to the government.

The need for this data was considered insufficient to warrant inclusion in the census, and this was not considered cost effective or the best way to satisfy the information need.

Options for providing this data have already been assessed. In 2010 Statistics NZ and Veterans’ Affairs recommended that modelling existing data could provide an estimate of the overall size and age distribution of the veteran population that was sufficiently accurate for costing and planning purposes. This data was supplied to Veterans’ Affairs to meet these planning needs.

Therefore this information does not need to be collected in the census.

Unpaid activities

We are considering whether the census is the best place to gather data on unpaid activities due to the limitations of collecting this information in the census.

The questions about unpaid activities allow people to record the work they do for their own households, relatives, friends, neighbours, and the community in general. This information provides indicative data on the nature of unpaid activities, as well as the characteristics and location of people involved.

Information on unpaid activities has been collected in the census because of the value to economic statistics, particularly in estimating the extent to which unpaid work underpins total economic activity (gross domestic product). Information about unpaid activities also has value in developing social indicators such as social and cultural capital (unpaid work as a measure of interconnections between individuals).

Key stakeholders such as the Ministry of Social Development have indicated that data on the caring and voluntary aspects of unpaid work may be increasing in importance. Voluntary work and unpaid care work provides significant savings for the national economy. Data on this is valuable, especially with an ageing population.

However, the census data on unpaid activities has limited value because it does not indicate the hours spent on various activities. Non-response is also relatively high (9.5 percent in 2013).

Looking internationally, information on unpaid work was collected in the Australian Bureau of Statistics’ (ABS) 2011 Census, along with the hours worked. However, it was not included in the Office for National Statistics (ONS) 2011 Census, or Statistics Canada’s 2010 Census.

Other Statistics NZ surveys collect information on unpaid activities. These data sources may meet the need for this information better than the census.

  • The Time Use Survey includes hours in unpaid activities.
  • The New Zealand General Social Survey collects data on the number of people doing formal unpaid work.
  • A module on voluntary work is also planned to be incorporated into the Household Labour Force Survey.

Comparing census data with data from these surveys suggests reporting errors in the census exaggerate the gaps in participation between different population groups. This also raises the question of whether the census is the most suitable place to collect this data.

Proposed recommendations for content on work

  • We recommend that information on paid work (work and labour force status, occupation, hours worked in employment per week, industry, status in employment, sector of ownership, and workplace address) be collected with no changes in the 2018 Census.
  • We recommend that information on paid work by people in New Zealand on temporary work visas not be collected in the 2018 Census.
  • We recommend that information on the veteran population not be collected in the 2018 Census.
  • We recommend that unpaid activities not be included in the 2018 Census.

Join the online discussion forum for ‘work’ content 

Income

Overview of income content being considered for the 2018 Census

Recommended new content or improvements to existing content

  • Total personal income
  • Sources of personal income

The value of this data

Statistics NZ produces several income measures, which are fit for different purposes.

The census data on income covers the population as a whole, enabling detailed regional and demographic breakdowns. It provides long time series data – it was first introduced in 1926. Census income data can be cross-tabulated with other census variables and is useful as an indicator of the spread of income levels across the population.

On the other hand, surveys such as the Household Economic Survey (HES) are best for providing national annual measures of household and personal income in New Zealand and are the source of Tier 1 statistics on income. (Tier 1 statistics are the most important official statistics, essential to understanding how well New Zealand is performing.)

Key users of census income data include central government agencies, local authorities, private organisations, and researchers. Examples of how household income (derived by combining personal income for all household members) informs important policies and adds value to decision-making include:

  • The Ministry of Education’s (MoE) decile funding system for schools and early childhood education. A wide range of funding is affected by this system, including targeted Funding for Educational Achievement (TFEA), Special Education Grant (SEG), and Priority Teacher Supply Allowances (PTSA).
  • The Ministry of Social Development’s (MSD) Social Report and a range of needs assessment exercises.
  • The New Zealand Deprivation Index – household income and information on benefits (from sources of income data) are required for creating this index, which is a measure of New Zealand’s most vulnerable people. This in turn is widely used for a range of research and policy work, targeting services, and spending aimed at helping New Zealand’s most vulnerable people.
  • Housing affordability – income data is useful for analysing trends such as the income levels of different household types. This informs housing affordability analysis.
  • Statistics NZ links income (and other census variables) to the data from post-censal surveys (the disability survey and Te Kupenga) to reduce respondent burden and increase the value of the survey data.

Recommended improvements to existing content

Sources of personal income

We are considering changes to the level of detail collected for sources of income. The information collected currently has detailed categories for different income sources, including separate categories for different types of benefits.

Collecting information on sources of income was introduced to focus respondents on providing accurate total income figures. While it is likely to fulfil this purpose, it is not clear whether the current level of detail collected is necessary for this purpose. It also appears that respondents may have difficulty providing this detailed information. Interviewers for household surveys such as the HES have found that many respondents are able to report whether they receive a benefit, but have difficulty reporting which type of benefit they receive. This is also likely to occur in the census.

We need to know whether our customers require information on different types of benefits. Census data on benefits is used in creating the New Zealand Deprivation Index, but this only requires data on whether a benefit is means-tested, not what type of benefit it is.

From a research perspective, a limitation of this data is the inability to determine how much income was received from each source or at which period of the year. It would not be appropriate to collect that type of information in the census due to its self-completed nature, finite length, and because it would increase respondent burden.

Total personal income

While the value of census income data for policy, planning, and research is clear, quality is an issue for this data. We assessed the 2013 Census data on income as moderate quality. We need to investigate ways of improving its quality.

We are also considering changes to the income bands, to maintain the relevance and usefulness of this data.

Non-response to the personal income question has been consistently high (9.7 percent in 2013, 10.2 percent in 2006). This is also an issue for interviewer-administered household surveys. We think this non-response is caused by a lack of understanding of the question, difficulty recalling income, and reluctance to answer due to the sensitivity of the subject.

In addition, Australian Bureau of Statistics (ABS) testing indicated that some people may not answer because they think it does not apply to them (eg those not in the labour force, pensioners, and self-funded retirees).

Non-response is higher among groups of particular policy interest – including people of Māori, Pacific, or Asian ethnicity, older people, and those receiving benefits.

As personal income is used to derive family and household income, non-response to personal income affects the quality of family and household income data. It can result in income levels being unknown for a significant proportion of families and households.

Some changes to the information collected may help reduce non-response and improve data quality. One way of improving data quality may be to impute for non-response using administrative data. We are considering this for the 2018 Census. Reducing the amount of detail collected on income sources by collapsing the benefit categories may also help to improve response rates for this question.

The income information collected currently is annual income from all sources, but some people (particularly beneficiaries) may only know their weekly or fortnightly income rather than their annual income. This may be contributing to non-response. To help address this, in 2013 we provided a table in the guide notes to help respondents work out their annual income.

It does not appear that collecting weekly income instead of annual income would help improve response rates or data quality. Work done for the New Zealand Income Survey (NZIS) showed that respondents were more likely to know their hourly wages or annual salary than their weekly wages or weekly salary amount, and weekly income would be difficult for self-employed people to provide. Shifting to weekly income could also lead to undercounting of total income as people would be led to think about their weekly income from wages or salary only, rather than their income from all sources.

As income levels generally tend to increase over time, we are considering changing the income bands to keep the data relevant. This may mean increasing the top category and potentially the number (or size) of bands. However, we will need balance making changes to improve relevance and maintaining the comparability of the data over time. We welcome your views on which income bands would be most useful.

Previously, data users have noted that broad income bands are not ideal for complex analysis and have suggested that exact amounts would be more useful. We have not implemented this because it could further increase non-response by making the question more sensitive, and because respondents would not know the exact amount. This is supported by evidence from the NZIS, which showed that asking for the specific amount led to a drop in response rates.

Proposed recommendations for content on income

  • We recommend that sources of income be included in the 2018 Census.
  • We also recommend that, except for New Zealand Superannuation, the benefit categories be collapsed into one category unless a clear need for information on different benefit types is identified.
  • We recommend that total personal income be included in the 2018 Census, and that the income bands be reviewed. We also recommend investigating ways of improving the quality of this data.

We invite comments on changes to the income bands and the possible use of administrative data to impute income where no response was given.

Join the online discussion forum for ‘income’ content 

Families and households

Overview of families and households content being considered for the 2018 Census

Existing content to remain the same

The existing range of derived family and household information, for example:

  • Family type
  • Household composition
  • Extended families
  • Child dependency status
  • Dependent young person indicator
  • Grandparents in a parental role

Recommended new content or improvements to existing content

  • Stepfamilies(1)

Note:
1. Content that was not collected in the 2013 Census.

The value of this data

There is an enduring need for data on families and households. Family and household statistics are used extensively for formulating social policy, planning, monitoring programmes, and research. Census data on families is also used in creating the New Zealand Deprivation Index and Statistics NZ uses it to derive household and family projections. Family and household data from the census is used by central government agencies, local authorities, private organisations, and researchers.

The census is a valuable source of information on families and households. The need for family and household information at low geographic levels and for particular population groups cannot be met by sample surveys (eg the General Social Survey) due to their small sample sizes. Initial investigations of the ability of administrative sources to provide family and household data have showed that this is limited at present, but may improve in the future.

Consultation with key census customers in 2013 indicated that information on families and households had become more important over the previous five years, and that the importance of this information for local government had increased.

Existing content to remain the same

We don’t ask direct questions about family type, household composition, and other aspects of families and households on the census forms. The family and household data produced from the census is derived from information collected about people’s relationship to a reference person, their living arrangements, usual residence, and census night address.

The information produced includes family type, household composition, extended families, and child dependency status (whether a child is dependent or not). 2013 Census data dictionary shows the full range of family and household data available from the census.

We are investigating whether we can collect more information on families and households in the 2018 Census (see the ‘recommended new content’ section below). At this stage we do not expect to change the existing range of information.

The legislative change regarding same-sex marriages means we need to change how we collect information on people’s relationships and living arrangements. This will not affect the data produced on families and households.

Families and households is a complex topic. Because the census is self-completed, and covers a broad range of topics, the complexity of the information we can collect is limited. We can’t expect to capture all aspects of families and households.

The concept of ‘family’

Family data from the census is based on the nuclear family concept (ie parents and children living in the same household). Each family nucleus is classified as couple only, couple with children, or one parent with children. This is the standard way of defining and classifying families used for Statistics NZ surveys (and available for use in other surveys as well). It also aligns with the definition and classification used internationally.

However, a limitation of using this nuclear family concept is that it may not accurately reflect some family structures, such as the nature of whānau for Māori.

If we review the family concept that census data is based on, it would have to be done as part of a wider review of standards and classifications for families and households. Such a review has been done previously and the concept of family used for these standards and classifications was not changed.

Although the census family data may not reflect the nature of whanau for Māori, the existing census data on household composition and extended families may go some way toward providing information that reflects the nature of whanau and gives a fuller picture of family structures. The household composition data indicates whether households contain multiple family nuclei or other people in addition to a family, and whether these families and other people are related to each other. The extended families data provides some information about wider family groupings, which is the number of generations present within an extended family in a household.

At this stage we do not recommend changing the concept the family data is based on for the 2018 Census. However, other changes we are considering may expand the range of information on families available from the census (see the ‘recommended new content’ section below).

Child dependency status

There is strong interest in distinguishing dependent children from other children in family data. Some customers have pointed out that data including children of any age (eg a 50-year-old son living with his 75-year-old mother) does not meet their needs. However, it is possible to get information on families with dependent children from the existing data. Although some family data from the census includes children of any age, data with breakdowns by child dependency status is also available, and tables that focus specifically on families with dependent children are available.

Dependent young people and grandparents in a parental role

Two information gaps identified previously were dependent young people and grandparents in a parental role – both of considerable policy interest. Information on both of these is available from the 2013 Census.

The dependent young person indicator developed for the 2013 Census provides information on children in family nuclei who are aged 18 to 24 years and are not employed full-time.

There is also interest in information on students living with their parents. This information need can be met by cross-tabulating study participation with the dependent young person indicator.

The census data on grandparents in a parental role covers situations where grandparents live with grandchildren aged less than 15 years whose parents are not present. It excludes situations where the parents are present in the household. Collection of this more detailed information would be better suited to a specialist survey.

People living in non-private dwellings

Although no family and household information is collected for people living in non-private dwellings (eg boarding houses), it is possible to obtain data on their personal characteristics (eg age, partnership status in current relationship). This can provide some information that is relevant to families and households, such as the number of children living in certain types of non-private dwellings.

Longitudinal data on families and households

The need for longitudinal information on how families and households change over time can now be met by the new longitudinal census data. As previously, time series data can also be used to examine general changes over time.

Recommended new content

Stepfamilies

Stepfamilies are those containing children who are the biological or adopted child of one partner in a couple but not the other partner. Currently these families are not identified in census data. All types of couples with children are included in the family type category ‘couple with children’.

Stepfamilies have been identified as a family type of interest to users of family statistics, and one of the main gaps in information about families and households. Currently there are no national estimates of the number of stepfamilies in New Zealand. Research shows stepfamilies face additional challenges, but services are limited because they are an unknown population. There is keen interest in obtaining information on the prevalence and well-being of stepfamilies.

Sample surveys (eg the General Social Survey) can provide information on the well-being of stepfamilies but are not considered suitable for measuring their prevalence (ie not suitable for providing an accurate count). This is because it is not possible to include sufficient numbers of relatively small population groups such as stepfamilies in sample surveys.

The census is considered the best way to obtain information on the frequency and distribution of stepfamilies, so we need to consider whether we can do this.

Looking internationally, information on stepfamilies has been collected in the Canadian census and in the UK census.

One way to collect information on stepfamilies from the census would be to add this into the existing questions about relationships and living arrangements within households. However, this would require many changes to the processing and derivation of family and household data, which is already complicated. A stand-alone question may be a better way to collect it.

Collecting information on stepfamilies in the census would be subject to the availability of space on the form and to testing showing that data of suitable quality could be produced. It would also be important to make sure that collecting this extra information does not adversely affect the quality of the existing information about families and households.

Second address / residence

See the second address / residence section for a discussion on the possibility of including this as a new topic in the 2018 Census. Including this topic may help to improve the range of family and household information available from the census. For example, it could provide an indicator of children in shared care who move between separate parental homes.

Proposed recommendations for content on families and households

  • We recommend that the information from which family and household variables are derived is included in the 2018 Census and that the current range of family and household information is produced from the 2018 Census.
  • We recommend that options for providing information on stepfamilies from the 2018 Census be explored.

Join the online discussion forum for ‘families and households’ content 

Housing

Overview of housing content being considered for the 2018 Census

Existing content to remain the same

  • Dwelling counts (occupied, unoccupied, under construction)
  • Number of bedrooms
  • Number of rooms(1)
  • Weekly rent paid by households

Recommended new content or improvements to existing content

  • Occupied dwelling type
  • Tenure of household(1)
  • Tenure holder
  • Sector of landlord

More information required to recommend inclusion

  • Types of unoccupied dwellings (empty, residents away)
  • Housing quality(2)
  • Mortgage payment amount(2)
  • Fuel types used to heat dwelling

Notes:
1. Content that is required to be collected under the Statistics Act 1975.
2. Content that was not collected in the 2013 Census.

The value of this data

Housing statistics are essential for formulating and evaluating housing policy and legislation and monitoring housing trends, including habitability, suitability, tenure security, and affordability. Census data on housing is used together with other census data to create the New Zealand Deprivation Index. Users of housing statistics include central government agencies, local authorities, private organisations, housing providers, policy analysts, and researchers.

The national significance of information on housing and the importance of time series information make the census an appropriate source of this information. Another major reason the census is an appropriate source of housing data is that this data is required for small areas and population groups (eg for monitoring trends in dwelling density, crowding, and home ownership).

Housing is a key priority area for government. Consultation with key census customers in 2013 indicated the need for information on housing had increased over the previous five years. There is increased interest in housing demand, home ownership, and in tracking the trend to higher density that is occurring in some parts of New Zealand.

Initial investigations suggest the ability of administrative data to provide housing information for the whole population is limited. However, some housing information could be produced from administrative sources – eg information on housing that is rented from private landlords, or from Housing New Zealand or a community agency.

Existing content to remain the same

Dwelling counts – occupied, unoccupied, and under construction

Counting dwellings is a core part of the census, as it is a census of dwellings as well as the population. Dwelling counts are a Tier 1 statistic. Tier 1 statistics are the official statistics considered essential for understanding how well New Zealand is performing. The census is the only source of a complete count of dwellings in New Zealand. It provides an overview of New Zealand’s housing stock at the national and local levels.

No changes are proposed for the counts of dwellings that are occupied, unoccupied, or under construction.

However, we are considering whether we should continue collecting information on different types of unoccupied dwellings. See the ‘more information required to recommend inclusion’ section below.

Number of bedrooms

The number of bedrooms data collected for private occupied dwellings indicates the number of rooms or sleepouts furnished as bedrooms. It includes caravans that households use as a bedroom.

Number of bedrooms data is valuable for determining whether households are living in a crowded situation according to the Canadian National Occupancy Standard (CNOS). This standard is considered to be the most appropriate way of measuring crowding in New Zealand. The census data on number of bedrooms is also useful as an indicator of the size of private dwellings and for examining how the cost of renting relates to the number of bedrooms a dwelling has.

Number of bedrooms is an important variable that is relatively easy for respondents to understand.

No changes are currently proposed for number of bedrooms.

Number of rooms

The number of rooms data collected for private occupied dwellings indicates the number of habitable living spaces including lounges, kitchens, dining rooms, bedrooms, studies, and games rooms. Open-plan areas are counted as if there were walls between the different areas. It excludes bathrooms, laundries, halls, and garages.

Collecting information on the number of rooms is a statutory requirement under the Statistics Act 1975, so it has to be included in the census.

Number of rooms provides an indicator of the size of private dwellings and can also be of value for measuring household crowding according to the people per room measure that is used internationally. This crowding measure is not widely used in New Zealand, because CNOS is regarded as a better measure in the New Zealand context.

The data on number of rooms does not appear to be used as frequently as the data on number of bedrooms.

This is a more difficult question for respondents to answer than number of bedrooms, because it requires them to work out which rooms to include, and because room layouts have become more complex. This can affect the quality of the data.

No changes are currently proposed for number of rooms.

Weekly rent paid by household

The census data on weekly rent paid by households provides valuable information on housing affordability, which is a key policy concern. Collecting this information in the census means that low-level geographic data is available and a range of cross-tabulations with other census variables can be done, such as sector of landlord and household income.

No new information needs regarding rent paid have been identified at this point, so no changes to this data are currently proposed.

Recommended improvements to existing content

Occupied dwelling type

Some changes to the occupied dwelling type data could be useful to improve its relevance and value. The new approach to collection for 2018 also has implications for the collection of this information.

The data on occupied dwelling type classifies occupied dwellings according to whether they are private or non-private. Private dwellings (eg houses, units, or apartments) are further classified according to whether they are separate or joined, and the number of storeys. Non-private dwellings are further classified according to their type or function (eg hotels, hospitals, boarding houses). This data is derived from the dwelling form questions asking respondents for a description of the dwelling and the number of storeys, and from the field-use box that collectors complete.

The changes to collection for 2018 mean that collectors will no longer be visiting each dwelling and determining whether it is private or non-private. Given this, it would be useful to know how important the non-private dwelling data is to customers (eg which particular types of non-private dwellings are of interest and what cross-tabulations are required).

Occupied private dwellings that are joined to others are classified according to whether the building they are in has one storey, two or three storeys, or four or more storeys. This means that apartments in high-rises are in the same category as those in four- or five-storey buildings. Better measurement of apartments in high-rise buildings is needed. It has been suggested that the top category be raised to 10 or more storeys.

Homelessness

There is interest in obtaining more information on homelessness from the census. Currently no count is produced directly from census data, but some relevant information is collected, including data on night shelters, boarding houses, motor camps, and improvised dwellings (eg garages, sheds, cars). Estimates of the number of people in this situation have been produced by Amore, Viggers, Baker, and Howden-Chapman (2013) by applying criteria to the census data and combining this information with administrative data.

Strategies for improving the information on homelessness that is available directly from census data include better identification of boarding houses and accommodation for the homeless through the use of administrative data. Another possibility is to have a ‘no usual address’ tick box on the form.

The Australian Bureau of Statistics has developed a methodology for counting the homeless in their census and a ‘flag’ to identify shelters for homeless people.

Retirement villages

An emerging area of interest is private dwellings in retirement villages. Currently, these are not identifiable in census data because there is no separate category or indicator for these dwellings. Private dwellings in retirement village complexes are classified in the same way as private dwellings elsewhere – ie, according to whether they are separate or joined, and the number of storeys.

It is not clear if there is a strong need to collect this information in the census. Alternative data sources may be available. If there is a strong need to collect this information in the census, possible options include a stand-alone question or collection via a ‘licence to occupy’ category in the tenure holder data.

Dwelling age and accessible housing

Other areas of interest are the age of private dwellings and accessible housing. It seems unlikely that good quality data on the age of private dwellings could be collected in the census because people may not know this information, particularly if they are renting. Administrative data may provide more accurate information on the age of private dwellings.

Information about accessible housing (special building features and modifications to housing to improve access) was collected in the 2013 New Zealand Disability Survey.

Tenure of household

We are considering changes to tenure of household questions, due to concern about data quality and gaps in the information.

The tenure of household data from the census indicates whether households living in private dwellings own their home (with or without a mortgage), have it in a family trust (with or without a mortgage), rent it, or occupy it rent-free. This data relates to the usual living situation of households. To contain a household, a private dwelling must have at least one usual resident.

A perceived information gap is that no data is collected on ownership of dwellings that are not a person’s or household’s usual residence, and the location and use made of these dwellings. This information would be helpful for understanding changes in home ownership rates and assessing under-use of dwellings. These dwellings could include properties that are rented out, holiday homes, dwellings being renovated or repaired, and dwellings that a household stays at part of the time but which are not their usual residence. Some of these dwellings may be unoccupied at census time or may be occupied but have no usual residents.

If included, the second address / residence topic has the potential to provide some of this information, but it would not cover ownership of properties that are rented out. The data from the second address / residence topic may be at the individual level rather than the household level, depending on how it is collected.

There is also concern that the family trust question on the dwelling form does not capture all of these situations correctly. Addressing this issue may require changes to questionnaire design, or to the information provided to help respondents answer this question, such as the family trust definition in the guide notes.

No information was provided in the ownership question on how respondents with a licence to occupy (eg people living in self-care units in retirement villages) should answer. Information directing these respondents to answer ‘yes’ was provided in the guide notes only. How these respondents answered this question, and where they are included in the data produced, may have depended on whether they read the guide notes.

Answers of ‘no’ to the mortgage questions do not necessarily indicate no mortgage exists – only that no payments were being made. The guide notes direct respondents on short-term mortgage repayment holidays to answer yes, but whether such respondents gave ‘yes’ answers may depend on whether they read the guide notes.

A disadvantage of making changes to tenure of household would be that it would reduce the comparability of the data with previous data, making it less reliable for measuring changes in home ownership over time, which is of particularly high interest.

There need to be clear benefits from making changes to this data, and we need to balance the benefits of making changes with the need for data continuity.

Tenure holder

Some changes to the information collected on tenure holder (home ownership at the individual level) may be useful to improve the value of this information as well as how respondents feel about providing it.

This data indicates whether people aged 15 years and over own or partly own the home they live in, with or without a mortgage. This information is collected on the individual form. The data produced has two categories:

  • own or partly own usual residence
  • do not own usual residence.

Respondents whose home is in a family trust are included in the ‘own’ category.

An advantage of collecting data on home ownership at the individual level is that it allows analysis of the personal characteristics (eg age) of those who own or do not own their home. Another advantage is that it covers people whose home was unoccupied at census time, whereas the tenure of household data (which comes from the dwelling form) excludes cases where a private dwelling was unoccupied at census time.

Two issues with this data are the treatment of family trusts and licence to occupy. Information in the question directs respondents whose home was in a family trust to mark yes. However, legally they do not own their home, so they may have felt uncomfortable about answering yes.

Respondents with a licence to occupy (eg those in self-care units in retirement villages) are currently not identified in this data. We do not provide information in the question or guide notes on how they should answer, so we don’t know which category they include themselves in.

We’d like to know if the current tenure holder categories meet your needs, or if more information would be useful – such as information on family trusts or licence to occupy.

As discussed in the tenure of household section, the potential new second address / residence topic might provide some information on ownership of other dwellings that are not a person’s or household’s usual residence. This information may be at the individual rather than household level.

Sector of landlord

A change to the information collected on sector of landlord (who households rent from) is likely to be required because in the future it is expected that some social housing will be provided by community agencies, iwi, and hapū. The information collected currently does not cover these types of landlords.

The number of additional categories required needs to be determined (eg one additional category covering community agencies, iwi, and hapū, or one new category for community agencies and another for iwi and hapū.)

Factors to consider when determining this include customers’ needs and the ability of respondents to provide the level of detail customers would like.

More information required to recommend inclusion

Types of unoccupied dwellings – empty, residents away

Unoccupied dwellings are currently classified as ‘empty’ or ‘residents away’. We are considering whether we need to continue distinguishing between these two types of unoccupied dwellings.

Previously there have been difficulties in determining which category an unoccupied dwelling should be assigned to. For the 2018 Census there will be additional challenges in collecting this information, because most people will be contacted by mail rather than by a collector visiting each dwelling.

We’d like to know more about how customers use the data on these two types of unoccupied dwellings, and how important this distinction is.

Housing quality

There is interest in expanding the housing information collected in the census to include housing quality. Adding this would be subject to testing and the availability of space on the form.

The only housing quality information currently available from the census is whether private dwellings are improvised (eg garages, sheds, and cars), mobile, or in motor camps; and whether heating is used. Dwellings that lack basic amenities or are dilapidated (unfit for habitation) are not identified.

Collecting information on dwellings lacking basic amenities has been recommended by Amore et al (2013) as a way of improving the measurement of homelessness. This data would indicate whether dwellings are habitable or not. People using uninhabitable dwellings are included in the New Zealand definition of homelessness (Statistics NZ, 2009a).

If information on housing quality was collected in the census, we would first have to determine the information requirements. Census questions on this would need to be quite simple (eg whether the dwelling has a working kitchen, water supply, and sanitary facilities).

An alternative source of information on housing quality is the New Zealand General Social Survey (NZGSS). In the future, administrative data on the quality of rental housing may be available.

Mortgage payment amount

We are considering whether new information on mortgage payment amounts could be collected in the census. This would be subject to testing and having sufficient space on the form.

Currently the information collected on mortgages is limited to whether households make mortgage payments or not for the home they live in. Expanding this to include the mortgage payment amount would provide more information for assessing patterns in housing affordability, including information for small geographic areas.

The Australian Bureau of Statistics collects information on the mortgage payment amount in their census.

Information on mortgage payment amounts is collected in the Household Economic Survey (HES). The advantages of the HES are:

  • new data is available every year
  • disposable income can be calculated – which is the preferred measure for analysing affordability and is too complex to collect in the census.

However, a limitation of the HES is that it does not provide sufficient information on regions and population groups.

If information on the amount households pay for their mortgage was collected in the census, it may only be possible to ask the total amount paid, rather than more detailed information such as the principle paid and the interest paid. We need to find out how valuable the type of information that can be collected in the census would be.

Including mortgage payment amount in the census also depends on whether it could produce data of sufficient quality and whether there is a strong need for small-area or population-group data on this.

Fuel types used to heat dwelling

Information is required on customers’ needs regarding data on fuels and heating, and the value of collecting this type of information in the census.

The information currently collected in the census is all the fuel types (eg electricity, wood, mains gas) that are used to heat a private dwelling. Although this data is collected at the dwelling level, it is possible to cross-tabulate it with personal variables that are also collected in the census (eg age, ethnicity) to analyse the characteristics of people in dwellings where certain heating fuels are used.

There has been interest previously in collecting:

  • different information on heating fuels (eg the equipment used, main form of heating rather than all forms)
  • more detail (eg whether gas heaters are flued or unflued)
  • fuels used for any purpose, not just heating
  • other types of information (eg whether dwellings have a ventilation system).

Reasons for these proposed changes were that changes in technology make it more important to collect data on appliances and systems, and that this provides useful information on energy efficiency, fuel conservation, and health.

It is unlikely that the census could meet all these information needs. There may be alternative data sources that could provide the information required.

Proposed recommendations for content on housing

  • Collecting information on the number of dwellings is one of the main purposes of the census, so this must be collected. Feedback is welcome on whether it is important to continue to distinguish between unoccupied dwellings that are empty and unoccupied dwellings that have residents who were all away at census time.
  • Collecting information on number of rooms and tenure of household is required by law under the Statistics Act 1975 so this information must be collected in the census. No changes are recommended for number of rooms.
  • We recommend that tenure holder be included in the 2018 Census and that options for improving the information collected on tenure of household and tenure holder be explored, including aspects relating to family trusts and licences to occupy.
  • We recommend that occupied dwelling type is included in the 2018 Census, that the top category for storeys is raised, and that other changes to improve the relevance of this data be investigated further.
  • We recommend that sector of landlord be included in the 2018 Census with changes to include housing providers not included previously, such as community agencies, iwi, and hapū.
  • We recommend that weekly rent paid by household be included with no changes in the 2018 Census.
  • We recommend that number of bedrooms be included with no changes in the 2018 Census.
  • We invite feedback on information requirements relating to fuels and heating and the value of collecting this information in the census so we can make an informed decision about collecting this data in the 2018 Census.
  • We recommend that the possibility of collecting information on housing quality in the 2018 Census be explored further.
  • We invite feedback on whether there is a need to collect information on mortgage payment amounts in the census so we can make an informed decision on its inclusion or exclusion for the 2018 Census.

Join the online discussion forum for ‘housing’ content 

Telecommunications

Overview of telecommunications content being considered for the 2018 Census

Recommended new content or improvements to existing content

  • Access to telecommunications systems

The value of this data

Information on access to telecommunication systems is used for monitoring access to different forms of telecommunication systems. It indicates the ability of households to access services including health care and civil defence communication in an emergency.

Access to information and communications technology (ICT) has become an important government policy area as more essential services become primarily accessible through phone and Internet. There is a need to identify those who are on the wrong side of the digital divide.

One important use of this data is in calculating the Deprivation Index, in which Internet access at home is currently the highest weighted variable in determining deprivation levels. This data informs customers across local and national government in the areas of health, education, and housing. Lack of access to certain telecommunication systems is viewed as an indicator of isolation and lack of community participation.

As census data on telecommunications access is the best source for understanding access to telecommunications systems over small geographic areas, it is also used by telephone companies, market research companies, and researchers.

Recommended improvements to existing content

Access to telecommunication systems

Forms of telecommunications and how they are used have changed greatly over the last decade. These trends are expected to continue, so it is important that census data collected in this area remains relevant for the future, while retaining comparability.

Historically, household telephone access was the main indicator used by the New Zealand Deprivation Index in this area. It has since been replaced by household Internet access, informed by Census 2013 data.

Data on household access to a telephone is still important to collect. Although Internet access has become more prevalent, in many households a telephone is still the only form of telecommunication available. Landline telephone access is dropping as growing mobile phone and Internet access has lessened reliance on a home telephone. Census data shows a steady drop in landline telephone access – from 96 percent in 2001, to 92 percent in 2006, and 86 percent in 2013.

We considered for the 2006 Census whether it was more appropriate to gather information on access to cellphones at an individual level. The report on this issue concluded that the question on the dwelling form currently met most users’ needs, and that asking it at an individual level in the census would not add value without significantly increasing respondent burden. To gather useful data on usage of specific devices would require a series of detailed questions not suited for inclusion in the census.

While usage of information and communication technologies has become increasingly personal in recent years, the census is not viewed as the best way to gather information on these trends.

In-depth information on information and communication technology usage has previously been gathered through Statistics’s NZ Household Use of Information and Communication Technology Survey. This gives information on trends (eg the devices used and time spent using them) that the census cannot provide.

Currently the main data need of users of census data on telecommunications is access at the household level. The existing question fulfils this information need adequately.

Whether information on fax access should continue to be collected in the census is questionable, as faxes are becoming obsolete due to widespread use of the Internet. Submissions we received during content consultation for the 2011 Census showed support for collecting data on faxes, so we decided to retain it at that time. However, fax access has dropped significantly since then. Household fax access was reported as 26.0 percent in the 2006 Census, dropping to 14.6 percent in the 2013 Census. We expect this trend to continue.

Proposed recommendations for content on telecommunications

  • We recommend that access to telecommunication systems be included in the 2018 Census, but that information on fax access no longer be collected.
  • We recommend that access to telecommunication systems continues to be collected on the dwelling form.

Join the online discussion forum for ‘telecommunications’ content 

Transport

Overview of transport content being considered for the 2018 Census

Recommended new content or improvements to existing content

  • Main means of travel to work
  • Main means of travel to education(1) 
  • Educational institution address(1)
  • Number of motor vehicles

Note:
1. Content that was not collected in the 2013 Census.

The value of this data

Transport data is extensively used by local government and transport planners, particularly in large cities where congestion is a problem. It is also used by private organisations. Consultation with key census customers in 2013 indicated that information on transport had become more important for private organisations in the previous five years, and that the need for this information was expected to increase in the next five years.

Currently in the census, we collect information on the main means of travel to work and the number of motor vehicles available to households for private use.

Main means of travel is considered the more important of these two types of information. It provides valuable data on transport patterns and trends which can be broken down to low geographic levels and for particular population groups. How people travel in New Zealand – and how this is changing demographically, geographically, and over time – is one of the enduring information needs identified in the Transport Domain Plan jointly led by Statistics NZ and the Ministry of Transport. Travel to work data also indicates the number of people working from home.

Data on the number of motor vehicles each household has is used to:

  • plan transport services
  • assist in locating public transport services, commercial, and welfare facilities
  • study energy conservation
  • create the New Zealand Deprivation Index.

Users of data on the number of motor vehicles include local authorities, national and local transport planners, and health and social service providers.

The information on usual residence and workplace address collected in the census also has value in the area of transport. Together these pieces of information allow measurement of traffic flows between home and work. Workplace address data is also useful for assessing daytime populations of particular areas for town planning and civil defence purposes.

Alternative sources of transport data are available (eg the New Zealand Household Travel Survey) but do not provide the low-level geographic data (down to meshblock level) that transport planners need. The need for low-level geographic data can be met by the census.

Recommended new content

Main means of travel to education and educational institution address

We are considering whether information on travel to education could be collected in the 2018 Census. This would be subject to testing and having space on the form. The ABS has been considering collecting this information in their 2016 Census.

A common theme from consultation with customers is that expanding the transport information collected to include travel to education as well as work would be valuable. Travel to education contributes significantly to traffic during the morning peak period. Expanding the transport topic to include travel to education would provide more complete data on traffic flows and is of particular interest to local authorities. Some information on travel to education is available from the New Zealand Household Travel Survey, but this survey does not provide the low-level geographic information that some customers need.

This information would be collected on the individual form for people who travelled to education. We would need to determine which levels of education (preschool, primary, secondary, tertiary) would be covered, and whether the information would relate to usual means of travel or means of travel on census day. Some students study part-time or only attend lectures or classes on certain days of the week, so they may not be travelling to their place of study on census day. If the information collected was means of travel on census day, the transport modes used by these students would not be included in the data.

It would probably be necessary to ask about travel to education separately from travel to work, because travel to education would apply to a different segment of the population (children and students) than travel to work (employed people aged 15 years and over). Other reasons to collect the information separately are that the range of response categories required for travel to education may differ from those required for travel to work, and some people who travel to education and work may use different transport modes to get to these different destinations.

Along with collecting information about the mode of travel to education, information about the educational institution attended would also be required so that travel patterns between home and educational institutions could be analysed. If the location of the institution can be determined using an address register (or other source of address information), then we would only need to ask the institution name, rather than the full address. This would reduce respondent burden and processing costs. Whether it is feasible to collect information on travel to education may depend on whether an address register of suitable quality is available.

A bonus of collecting information on the educational institution is it might increase the value of the study participation data, because it may then be possible to obtain information on the level of study (see education and training section). In addition to the current information collected on study participation (which is for people aged 15 years and over), we might also need to ask about preschool and school attendance as lead-in questions for the means of travel to education question.

One issue with collecting information on travel to education may be that some students do not travel directly from home to the educational institution they attend (eg tertiary students who work part-time, and children in before-school care). This would affect analysis of travel patterns between home and educational institutions. However, this also applies to travel to work (ie many people may not travel directly from home to work, but via their children’s schools, place of study, or elsewhere), and perhaps does not outweigh the value of collecting data on travel to education.

Recommended improvements to existing content

Main means of travel to work

We are considering whether the data collected on travel to work should be changed to the usual means – rather than the means used on census day. Information about the means of travel to work on census day has been collected in censuses since 1991. Data on usual means of travel to work was collected in the 1971 to 1986 Censuses.

Part of the rationale for collecting data on the means of travel on census day has been that the census is a ‘snapshot’ at a point in time, so this was the appropriate type of data to collect, and was consistent with how the data was used. This type of data indicates traffic volumes on a particular day, although only for that part of the traffic which is due to people travelling to work. It shows the actual numbers of employed people aged 15 years and over whose main means of travel to work that day was driving, cycling, taking the bus, or walking.

However, if respondents filled in their census forms before census day, it is likely that they indicated their usual means of travel. This may not necessarily have been the same as their means of travel on census day. There is some evidence that many respondents fill in their census forms early. In 2013, 46 percent of online census forms had already been received by census day (Statistics NZ, 2014).

Data on the means of travel to work on census day does not provide any information on the transport modes used by those who did not go to work that day because:

  • they were sick
  • they were away on holiday, or
  • census day was not a normal work day for them (eg retail and hospitality workers who do not work on Tuesdays).

The number of employed people who did not travel to work on census day can be substantial (207,141 people in 2013).

Another implication of collecting information about travel on census day is that the data may be affected by the weather or particular events occurring on that day. For example, wet weather may result in fewer people cycling and walking, or if public transport in a certain region was not operating that day, this would change that region’s data and could increase the numbers of people driving to work.

Data on the usual means of travel to work would indicate the total numbers of people who usually use a certain transport mode for this purpose. In this type of data, people who did not go to work on census day would be included in the transport mode categories and there would not be a ‘did not go to work today’ category. An advantage of collecting data about the transport mode people usually use is that it would not be affected by the weather or particular events occurring on census day. A disadvantage of changing to usual means is that the data would be less comparable with previous census data, making it more difficult to measure changes over time.

We need to know which type of travel to work data is more valuable to customers – usual means of travel or means on census day.

Comparability of data with that collected overseas is another factor to take into account. Some other statistical agencies collect information on usual means of travel to work – the Office of National Statistics (ONS) in their 2011 Census, and Statistics Canada in the 2011 National Household Survey (which replaced their census long form). However, the Australian Bureau of Statistics (ABS) collected information on travel to work on census day in their 2011 Census.

There is interest in capturing combinations of transport modes. For example, people using park-and-ride facilities would have a combination of driving a private vehicle and taking the bus. However, this would increase the complexity of the information collected and may be better suited to a specialist travel survey.

Number of motor vehicles

We are considering whether the top category for number of motor vehicles should be raised. The current data has categories of none, one, two, and three or more. It has been suggested previously that the top category should be raised to five or more motor vehicles. Another possibility is to raise it to four or more motor vehicles.

Reasons given previously for wanting the top category to be raised were that households tend to have more motor vehicles than previously because of lifestyle changes, so a top category of three or more does not provide sufficient detail. Raising the top category would allow more precise measurement of car ownership, and the effects of this on road congestion and demand for road space. Knowing the actual number of vehicles is important when looking at parking provision in urban planning.

The top category was not raised for the 2013 Census because the number of households with more than three motor vehicles was considered too small to make it worthwhile. There has been little change in the data since 2006. In 2013, 16.1 percent (237,468 households) had three or more motor vehicles, compared with 15.9 percent (222,204 households) in 2006.

We need to determine whether users of this data are still interested in raising the top category, or whether other sources of data (eg administrative data) could be used instead.

Another issue raised previously is that this data relates to vehicles for private use, and excludes those that are for work only – yet uses of this data include measuring demand for road space and the effects of this on road congestion. The potential benefits of including work vehicles would need to be weighed up against the importance of having data that indicates how many motor vehicles households have unlimited access to for any purpose, and the loss of comparability over time. The travel to work data provides some information on the numbers of work vehicles being used.

Another suggested change for number of motor vehicles is to collect it at the individual level instead of the household level, and aggregate data to household level as required. This has merit because different members of the same household may not have equal access to the motor vehicles belonging to that household. However, it is not clear at this point if there is a strong need to collect data on individual motor vehicle access in the census. Although the current data is household level, it can be cross-tabulated with individual variables (eg age, ethnicity) to provide information on the characteristics of people living in households with access to a certain number of motor vehicles or no motor vehicles.

Users have also suggested previously that information on other forms of transport (eg bicycles, motorbikes, and scooters) could be collected. This was not implemented for 2013 due to time constraints and the minimal-change policy in place. However, there are other reasons to collect information on motor vehicles rather than other forms of transport. Other transport modes have limitations that motor vehicles do not have. For example, For example, motorbikes, scooters, and bicycles may not be suitable forms of transport for small children or elderly people. Bicycles that a household has may not necessarily be in use. Data on the forms of transport people use to get to work or education may be more valuable than data on the number of other forms of transport a household has. The New Zealand Household Travel Survey provides an alternative source of information about households who have bicycles.

There has also been interest in collecting additional information such as the frequency of use, main reason for use, and distance travelled. However, it would be difficult to obtain good quality data on these aspects of travel from a self-completed survey such as the census. Collecting detailed information is better suited to a specialist, interviewer-administered transport survey.

Proposed recommendations on content for transport

  • We recommend that main means of travel to work be included in the 2018 Census. We welcome feedback on which type of travel to work data is more relevant and useful to collect – means of travel on census day, or usual means of travel.
  • We recommend that number of motor vehicles be included in the 2018 Census. We welcome feedback on whether any changes are needed for this variable, such as raising the top category.
  • We recommend that main means of travel to education and educational institution address be considered for inclusion in the 2018 Census. Feedback is invited on whether usual means of travel or means on census day should be collected, and which levels of education should be covered

Join the online discussion forum for ‘transport’ content 

Health

Overview of health content being considered for the 2018 Census

Existing content to remain the same

  • Disability

More information required to recommend inclusion

  • Cigarette smoking behaviour

The value of this data

Census information on cigarette smoking behaviour is used to:

  • monitor changes in smoking prevalence amongst the population of New Zealand
  • understand the characteristics and profile of smokers, to better target at-risk groups in the community with future health education programmes, and evaluate the success of those programmes
  • examine the relationship between smoking and other socio-economic variables and how these change over time
  • help assess the impacts of government policy on smoking, for example price rises, law changes, and public health campaigns
  • validate the New Zealand Deprivation Index.

Census information on disability is used to select a sample for the post-censal disability survey. We don’t publish census information on disability.

Existing content to remain the same

Disability

The purpose of disability information collected in the census has been to support the selection of a sample for the Household Disability Survey. Using census data is the most cost-effective way of identifying disabled people for a disability-specific survey.

Census disability questions were redesigned in 2013 to more closely align with those in the short question set from the UN Washington Group on Disability Statistics. These changes were meant to improve the efficiency of the sample selection for the 2013 Household Disability Survey. But the changes did not significantly improve the identification of the disabled population, so we are reviewing the quality of the census questions. This review will also inform any decision to include a short set of disability questions in other social surveys.

Statistics NZ and the Office for Disability Issues are setting up a joint working group to consult on current and future needs for disability statistics. The group will include representatives from across the disability sector as well as relevant government agencies. It will contribute to the development of a monitoring framework within which the role of official survey collections will be clarified.

We are carrying out ongoing consultation on disability statistics across sector groups and with the Human Rights Commission.

More information required to recommend inclusion

Cigarette smoking behaviour

Cigarette smoking refers to the active smoking of one or more manufactured or hand-rolled tobacco cigarettes, from purchased or home-grown tobacco, per day, by a person aged 15 years and over.

Smoking was first included in the 1976 Census, and repeated in the 1981, 1996, 2006, and 2013 Censuses. Cigarette smoking behaviour was historically a cyclic variable, meaning it is reviewed for relevance and inclusion each census due to typically less change in the data over time.

This information has been collected in the last two censuses as there has been strong government interest in data on smoking due to its recognised high social and monetary costs. In March 2011 the government committed to Smoke-free 2025, a goal of less than 5 percent prevalence of smoking across all populations. Smoking levels declined from 20.7 percent in the 2006 Census to 15.1 percent in the 2013 Census for the population overall, and from 42.2 percent to 32.7 percent for Māori over the same time period.

Data on smoking is widely collected through other surveys, most notably the Ministry of Health’s Health Survey. Data on national smoking levels (daily) from the 2013 Census and the 2013/4 Health Survey is comparable. The key value of census data compared with other data is that census data can be broken down geographically and for particular demographic groups to a level of detail that is not possible from a sample survey.

Previous consultation with external users has indicated a strong need for data with a level of detail which has previously only been available from the census, in order to target programmes towards specific population groups with higher smoking rates. The current smoking level is considered to be a Tier 1 statistic (the most important official statistics, essential to understanding how well New Zealand is performing). This information need is currently listed as being met through the data from the Health Survey, which has recently been changed to run continuously instead of every five years.

Given that the Health Survey is now continuous, we are considering whether it is still necessary for the census to collect data on smoking behaviour. There is potential for data from the Health Survey to be collated over several years to produce information on smoking levels for particular population groups. However, this data – or any other data from sample surveys – would not give an accurate indication of smoking levels below a regional (District Health Board) level, and could not be broken down for demographic groups to the same level of detail as the census data on smoking.

We welcome feedback on current information needs on smoking behaviour, and whether we need to collect this information in the 2018 Census, or if other data on smoking behaviour meets information needs.

Proposed recommendations for content on health

  • We recommend that disability be included in the 2018 Census to support the selection of a survey sample only if a disability-specific household survey is to be carried out in 2018.
  • We welcome feedback on customers’ current needs for information on cigarette smoking behaviour, and whether this information need can now be met by alternative data sources rather than the census.

Join the discussion forum for ‘health’ content

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Top
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+