Using data from the Household Disability Survey and the 2023 Census

This paperprovides an overview of the disability information collected in the 2023 Household Disability Survey and the 2023 Census. It explains when to use each data source and gives guidance and best practice information for data users. It also includes examples of how the two data sources can be used together to draw insights about Aotearoa New Zealand’s disabled population.

Contents

1……Accessing the data.

1……2023 Census disability data.

2……Household Disability Survey.

3……How to use disability data.

5……2023 Census detailed activity limitations data.

5……Using both data sources to draw insights about the disabled population.

5……Comparing outcomes for disabled and non-disabled populations.

6……Comparing outcomes for subgroups of disabled populations.

6……Comparing outcomes between areas.

7……Learning more about a subgroup.

7……Comparing disability prevalence.

Accessing the data

Aotearoa Data Explorer has tables with disability information from the 2023 Census.

Disability statistics: 2023 has the results from the 2023 Household Disability Survey.

2023 Census disability data

The 2023 Census collected information about disability (referred to as ‘Activity limitations’) using the Washington Group Short Set on Functioning (WGSS). This set of questions asks about difficulty experienced with the following activities:

• seeing (even if wearing glasses)
• hearing (even if using a hearing aid)
• walking or climbing stairs
• remembering or concentrating
• washing all over or dressing
• communicating in usual language (for example, understanding or being understood by others).

A person was considered disabled in the 2023 Census if they had a lot of difficulty with or could not do at least one of these activities. In the census, having a lot of difficulty with or being unable to do an activity is referred to as having an ‘activity limitation’. A person can have more than one activity limitation.

Children aged under five years were not asked this set of questions and are therefore not captured in census disability statistics. 

The Washington Group questions have been extensively tested internationally. The activities covered by the 6-question WGSS are those that were found by the Washington Group to most often limit an individual and result in participation restrictions. Because the WGSS is relatively short, it can be used in surveys that have limited space for new questions which otherwise would not produce data about disabled people.

A limitation of the WGSS is that it does not identify all disabled people within a population. The purpose of the WGSS is to identify enough disabled people to enable comparisons between disabled and non-disabled people for key socio-economic outcomes. There are also high proportions of missing data for areas and subpopulations with low census response rates. Therefore, any counts of disabled people from the census (such as the total number of disabled people in New Zealand, or the number of disabled people who are employed or enrolled in education) will be undercounts. For this reason, census data should not be used as an official measure of disability prevalence. This also applies to other Stats NZ surveys that use the WGSS or Washington Group Short Set – Enhanced.

Household Disability Survey

The Household Disability Survey (HDS) is New Zealand’s most comprehensive source of disability data and only official source of disability prevalence estimates. HDS estimates are representative of people of all ages living in private dwellings or residential and community care facilities.

The HDS is used to produce estimates of the size of New Zealand’s disabled population, including estimates by age group, gender, and ethnicity. It also provides information on the outcomes and experiences of disabled people.

Because the HDS is used to estimate disability prevalence, it aimed to identify all disabled people within its sample. To do this, the HDS used a much longer set of questions than the 2023 Census. The HDS question set was based on the Washington Group Extended Set on Functioning.

The 2023 HDS asked about the same six activities covered by the census, but with a broader set of questions to capture more people with those difficulties. Additionally, the HDS asked about areas not captured by the census. For adults, these areas were:

• flexibility
• dexterity
• learning
• intellectual function
• mental health
• socialising
• pain
• fatigue.

For children aged 5–14 years, these were:

• flexibility
• dexterity
• learning
• intellectual function
• mental health
• accepting change to routine
• controlling behaviour
• making friends.

Household Disability Survey 2023 – findings, definitions, and design summary has more information about how disabled people were identified in the 2023 HDS.

HDS estimates are accompanied by sampling error. The HDS is unable to produce reliable estimates about very small groups of disabled people or to produce highly specific information (for example, the highest qualification for Māori disabled people living in a particular region).

How to use disability data

There are different ways that census and HDS data can be used together to learn more about the disabled population. Data users should be aware of how the disabled population has been identified and of any related limitations when drawing insights from disability data.

It is important to remember that the census disabled population will be an underrepresentation of the true disabled population due to the shorter question set. However, detailed information for smaller populations or subnational areas from the HDS is often not reliable due to sample size limitations. Census data can be used in these instances to compare the outcomes of one group of disabled people with another group of disabled or non-disabled people. This is the intended use of the WGSS in the census.

The following table provides a summary of what each data source can and cannot be used for. Examples of how to use the data sources together are provided in Using both data sources to draw insights about the disabled population.

Summary of uses for Household Disability Survey and 2023 Census disability data

The census disability rate can be used to compare the likelihood of disability between groups and identify trends about the disabled population. For example, ‘group X is more likely to be disabled than group Y’ or ‘the likelihood of being disabled increases as X increases’. However, it should not be used to make statements about disability prevalence (for example, ‘X percent of group Y are disabled’).

Disability rates from the census data should be used with extreme care to avoid misinterpretation as official disability prevalence rates. It is best practice to:

• avoid reporting counts of disabled people from census data
• be clear that any disability rates from census data are not official disability prevalence rates
• use prevalence rates from HDS data if available.

See Comparing disability prevalence for an example.

2023 Census detailed activity limitations data

The 2023 Census activity limitations data includes breakdowns by the six activities captured in the WGSS. This allows data users to disaggregate the census disabled population by type of activity limitation. For example, the median income of adults with a walking activity limitation.

Like the census disabled population, people with each activity limitation will be an underrepresentation of the actual population. These should not be reported as official prevalence rates of each activity limitation. For example, people with a hearing activity limitation should not be reported as representative of the entire deaf population.

All guidance and best practice provided in this report also applies when using detailed activity limitations data from the census.

Using both data sources to draw insights about the disabled population

The following are examples of how HDS and census data can be used together to provide insights about the disabled population. Specific values, for example the smoking rate of disabled adults, may differ between the HDS and census because the disabled populations are not the same. Where reliable information is available in the HDS, this should always be used over the census. Information about the census disabled population should be considered as an indication of trends, outcomes, or differences between groups.

Comparing outcomes for disabled and non-disabled populations

This is an example of how a comparison of outcomes for disabled and non-disabled people from the HDS can be supplemented with more detail from the 2023 Census. Smoking rates by gender, age, and ethnicity from the HDS would be very low quality and unfit for most purposes.

The HDS estimated that the smoking rate for disabled adults was 11 percent, compared with 7 percent for non-disabled adults. The smoking rate for disabled adult females was 9 percent compared with 5 percent for non-disabled adult females.

The 2023 Census data can be used to make comparisons between disabled and non-disabled people for more specific demographics. In the 2023 Census, across all life-cycle age groups (15–29, 30–64, and 65 years and over), disabled females had higher smoking rates than non-disabled females. The difference in smoking rates between disabled and non-disabled females was largest for the 30–64-year age group, with a difference of 8.8 percentage points (15.8 percent compared with 7 percent).

When looking at the age and ethnic group of females, females aged 65 years and over who belong to the Pacific peoples ethnic group were the only age and ethnic group where the smoking rate was lower for disabled females compared with non-disabled females (5.2 percent compared with 7.4 percent).

Comparing outcomes for subgroups of disabled populations

This is an example of how an outcome for disabled people from the HDS can be supplemented with a comparison of disabled subgroups using 2023 Census data. Dampness data from the HDS, broken down by age groups and income bands, would be very low quality and unfit for most purposes.

The HDS estimated that 1 in 5 disabled people (20 percent) lived in a home that was sometimes damp, and an additional 4 percent lived in a home that was always damp.

The 2023 Census data can be used to break this down for more specific disabled populations. According to the census, disabled people with a total household income of $20,000 or less were most likely to live in damp homes – 22.9 percent lived in a home that is sometimes damp, and 9.0 percent lived in a home that is always damp.

Disabled people with a total household income of $20,001–$30,000 were the least likely to live in damp homes – 17.8 percent lived in a home that is sometimes damp, and 4.8 percent lived in a home that is always damp. Compared with the overall disabled population, a larger proportion of disabled people with a total household income of $20,001–$30,000 were aged 65 years or older. Generally, older people are more likely to live in owner-occupied homes, which are less likely to be damp.

For disabled people with a total household income of $20,001–$30,000, younger people were more likely to live in a damp home than older people. For those 65 years and over, only 11.8 percent lived in a home that is sometimes damp, and 2.5 percent lived in a home that is always damp. For those aged 15–29 years, 36.1 percent lived in a home that is sometimes damp, and 10.8 percent lived in a home that is always damp.

Comparing outcomes between areas

This is an example of how outcomes for disabled and non-disabled people from the HDS can be supplemented with comparisons between areas, with more detailed breakdowns, using 2023 Census data. The HDS can only produce reliable subnational estimates for regional council, health district, and district health board areas, and is limited in what it can produce for these areas.

The HDS estimated 58 percent of disabled adults had an annual personal income of $30,000 or less (compared with 33 percent of non-disabled adults) and 13 percent had a personal income of $70,001 or more (compared with 32 percent of non-disabled adults).

The 2023 Census data can be used to make comparisons across areas. Across territorial authorities (TAs), disabled adults living in the Queenstown-Lakes district had the highest median income ($36,600), while disabled adults living in the Buller district had the lowest ($22,800).

These results are different from the TAs with the highest and lowest median income for non-disabled adults. Non-disabled adults in Wellington city had the highest median income ($59,400), and those in Kawerau district had the lowest ($29,500). For disabled adults in the Queenstown-Lakes district, those aged 45–49 years had the highest median income ($67,000) across five-year age groups.

Learning more about a subgroup

This is an example of how a finding from the HDS can be supplemented with further detail from the 2023 Census to better understand a subgroup of disabled people. Information at this level of detail from the HDS would be very low quality and unfit for most purposes.

The HDS estimated that 56 percent of disabled adults (aged 15–64 years) had a paid job. It also estimated that the most common industries for disabled adults to work in were retail trade, and accommodation and food services.

The 2023 Census data can be used to learn more about this subgroup of disabled people. While ‘retail trade’ and ‘accommodation and food services’ are grouped together in the HDS to increase sample size, the census is able to look at these two industries separately.

In the 2023 Census, for disabled adults who worked in the retail trade industry:

• 44.5 percent were sales workers
• the average number of hours worked per week was 30.9.

For disabled adults in the accommodation and food services industry:

• 25.8 percent were community and personal services workers
• the average number of hours worked per week was 28.8.

Comparing disability prevalence

This is an example of how an estimate of disability prevalence for a population from the HDS can be supplemented with more detailed information using 2023 Census data. Care is required when reporting information like this from the census to ensure it is not misrepresented as an official disability rate.

The LGBTIQ+ population includes people aged 15 years and over who are lesbian, gay, bisexual, transgender, intersex, or who identify with other minority genders or sexual identities, based on 2023 Census data. The LGBTIQ+ (or rainbow) indicator used in the HDS and 2023 Census can be presented for the total LGBTIQ+ population or for the more specific subpopulations of:

• another gender
• transgender (which includes people of another gender)
• sexual minorities (homosexual, bisexual, or sexual identity not elsewhere classified)
• people who know they were born with a variation of sex characteristics, also known as intersex (limited availability due to data quality).

Because the LGBTIQ+ population is a small proportion of the total population, information at the more detailed level from the HDS would be low quality and unfit for most purposes.

The HDS estimated that 31 percent of the LGBTIQ+ population were disabled.

The 2023 Census data can be used to compare the more detailed subgroups of the LGBTIQ+ population. Note that these subgroups are not mutually exclusive. People who identified as another gender were most likely to be disabled compared with the other subgroups of the LGBTIQ+ population. People who knew they were born with a variation of sex characteristics were the least likely to be disabled of these subgroups – although still more likely than the non-LGBTIQ+ population.

For adults that identified as a sexual minority, those who identified as belonging to the Māori ethnic group were most likely to be disabled compared with other level 1 ethnic groups. Those who identified with the Asian ethnic group were least likely to be disabled.

ISBN 978-1-991307-63-7

Enquiries

Katy Auberson
04 931 4600
info@stats.govt.nz