The purpose of this update is to ensure customers are well informed about and prepared for the second release of 2018 Census data on 16 December 2019.
This update includes:
- Second 2018 Census data release and supporting products
- Second external data quality panel report
- Update on census data variables
- Census Transformation research on admin data uses
- Additional technical methods papers
- Upcoming supporting activities.
On 16 December 2019, we published our second release of data and supporting products for the 2018 Census. This release comprised:
- small areas dataset (SA1 dataset) 2018 Census data broken into small geographical areas – this data set contains 59 selected output variables at the highest level [level 1] of the classification for our small and local geographic areas, wards, district health boards (DHBs), territorial authorities, local boards, regional councils, and at a national level. These have been released in both Excel and CSV formats
- three further 2018 Census data tables in NZ.Stat.
A further four variables (hours worked, unpaid activities, travel to work by usual residence, and workplace address) from the small areas dataset will be released in the first quarter of 2020.
The tables being released into NZ.Stat were selected based on enquiries we received and most used tables from the 2013 Census data. The tables are:
- Ethnic group (detailed single & combination) by age and sex, 2013 and 2018 Censuses
- Ethnic group (detailed total response – level 3) by age and sex, 2006, 2013, and 2018 Censuses
- Māori descent indicator by age group and sex, 2006, 2013, and 2018 Censuses.
Village of 100
We also released an addition to our popular New Zealand as a village of 100 people: Our population infographic. The infographic displays key information about New Zealand from the 2018 Census by imagining all of New Zealand as a village with 100 people living in it. The additional infographic, New Zealand as a village of 100 people: Education and employment, looks at employment and education data. Together, these two infographics provide a complete picture of New Zealand at a national level.
2018 Census microdata
In the coming weeks, the 2018 Census microdata will also be added to the Data Lab. Microdata is unit record data that can only be accessed in controlled circumstances. Microdata access is only provided to approved users. For more information, see Apply to use microdata for research.
Proposed electorate boundaries
On 20 November 2019, the Representation Commission released proposed new electorate boundaries for public consultation. The release of updated electoral boundaries is a key step in our democratic process and is heavily reliant on up-to-date census data.
We released very small areas (meshblock) electoral population data on Datafinder and maps of the proposed boundaries on our Geographic Boundary Viewer at the same time to support the proposed boundaries release.
We also published a news story about the data and copies of the maps behind the proposed electorates. See Proposed electorate boundaries released for more information.
Upcoming 2020 data and product releases
In November 2019, we updated the Stats NZ release calendar to include indicative dates for 2018 Census data releases through to June 2020. These dates may be subject to change. See our Release calendar for updates.
Open data API availability
Release of the open data API has been deferred until 2020.
Place summaries are coming!
The next significant release of 2018 Census data will be place summaries, expected in late February 2020. These summaries will enable you to find information and quick stats about the area you live in and how your area compares with the next highest geography level.
The place summaries release will cover seven topics across local geographic areas (SA2), territorial authorities, local boards, regional councils, and at a national level. The finalised topics and subtopics are grouped and named as follows:
|Population and dwellings||
|Ethnicity, culture, and identity||
Cigarette smoking behaviour
|Work, income, and unpaid activities||
|Education and training||
Travel to work
The associated data tables will follow the release. These will be available in .xls format.
Initial Report of the 2018 Census External Data Quality Panel was published on 23 September 2019 and focused on the release of the 2018 Census data.
There are two further panel reports on 2018 Census data:
- 2018 Census External Data Quality Panel: Assessment of variables (released 16 December 2019)
- Final report (due to be released in early 2020).
The final report will cover families and households, ethnicity to level 4, small areas dataset (SA1) with associated graphs, and other subjects.
External data quality panel provides more information about the panel and their terms of reference.
Census data quality
We take our role as stewards of New Zealanders’ data seriously. We know that incorporating checks and balances around data quality is fundamental to ensuring the data we gather and hold is managed and used appropriately. That’s why we will be proactively integrating data quality assurance into future censuses.
Data quality for 2018 Census is our new, easy-to-access webpage hub for all our data quality information.
As highlighted in previous updates, the majority of key variables from the 2018 Census are of very high, high, or moderate quality, however, some variables are of poor or very poor quality.
We are now able to provide more information about the work that has been taking place to support the three very poor key variables:
Families and households
We have completed a complex and robust assessment on the family and household data. We are using findings from this assessment to re-evaluate the suite of data at a detailed level, identifying which groupings of variables have higher concentrations of data quality issues and which are less affected. It is likely that we will be applying a range of more specific quality ratings as a result. We are currently working to understand what this means for outputs and further engagement.
Absentees has been considered as part of the family and household assessment. We will make a decision about quality and any future output plans once the rest of the family and household outcomes have been announced.
We are continuing to analyse iwi affiliation data. We are working with technicians from the Data Iwi Leaders Group (Data ILG) of the Iwi Chairs Forum to enrich understanding of the 2018 Census dataset, enable collective rigorous analysis of quality, build insights into the data’s value, and identify possible approaches to improving the quality of the data.
Customer update on data quality of 2018 Census provides more information, including on the quality ratings.
The Census Transformation programme published two research reports in early September 2019 that explore whether admin data sources could provide census-type information that would allow comparisons between 2018 and 2013 Census results. The two reports looked at the census topics of: number of children born, and birthplace and years since arrival in New Zealand.
Drivers behind the assessments
Data about the number of children born provides information for use in the context of fertility. An earlier Census Transformation indicative assessment, the first step to exploring admin sources, identified that this variable was unlikely to be fulfilled by admin data.
Country of birth and years since arrival in New Zealand were variables that we identified as having the potential to be produced from admin data. Since the metadata analysis only indicated possibility, we needed to assess the coverage and consistency of the estimates produced from admin data against the 2013 Census results to confirm our conclusions.
Outcome of the assessments
When looking at the number of children born, we found that:
- for women born in or after 1974, the admin sources show good potential for providing an ongoing series of number of children born in the census context
- the main limitation in using admin sources is a small but systematic undercount of the number of births. This is due to a lack of information on overseas births and missed mother-child links for New Zealand births, which will need to be taken into consideration when using admin data for this purpose.
Comparing 2013 Census and admin data for number of children born provides more information about this assessment.
When looking at birthplace and years since arrival in New Zealand, we found that:
- there are some gaps in the admin data
- overall, both variables had good coverage and consistency, especially for New Zealand-born people and for recent immigrants
- both variables are attributes that do not change over time, for example, a person’s date of birth, so when combined with historical census data as an additional source, they have good potential for replacing the census question.
Potential for admin data to provide country of birth and years since arrival in New Zealand information provides more information about this assessment.
To support our first release of data on 23 September 2019, we released a large suite of technical and non-technical methods papers. These were designed to help customers gain a fuller understanding of the new methods we used to create the 2018 Census dataset.
Four additional technical methods papers are being made available this month:
Data sources, editing, and imputation in the 2018 Census: Describes the 2018 Census approach for detecting data errors and for filling in gaps when the characteristics of people or dwellings have not been provided on census forms and reports on the quality and results of the approach.
Overview of international peer review of 2018 Census methodology: Summarises the reviews received from international experts in statistical methodology on the administrative enumeration approach Stats NZ used in the 2018 Census.
Electoral boundaries sensitivity analysis of 2018 Census data for electoral boundaries summarises the first of two sensitivity analyses and analyses the decisions that were made in constructing the 2018 Census file and how different choices of model parameters might have affected New Zealand electorates.
Population counts sensitivity analysis of 2018 Census data summarises the second of two sensitivity analyses and analyses the decisions that were made in constructing the 2018 Census file and looks more broadly at all of the choices made during admin enumeration.
Quick guide to the 2018 Census provides more information about what was different about the 2018 Census and key changes introduced as we prepared for the 2018 Census and made once collection was complete.
2018 Census has faced some difficult and unique challenges, and we are aware the approach taken has resulted in significant changes and impacts both to users of census data and on the quality of the data itself. It is more important than ever that we educate people about the 2018 Census data and the particular challenges it brings.
We aim to deliver a range of educational activities in the new year that will build census data capability across a range of audiences and complement specific one-on-one engagement. This currently includes:
Are you a data-capable user? Would you like to develop your understanding of the changes to the 2018 model, the methods applied to resolve the challenges faced by 2018 Census, the complexities of the census data, the data quality issues, and the impact of this on data use? Then this half-day training session may be for you.
Data quality training will be available from late January 2020 and will be advertised in both the Expert Data Users newsletter and the Census Advisory newsletter. Locations will be confirmed based on interest and availability.
For more information or to register to attend, email [email protected].
Coming to a town near you in March 2020!
Have you been wondering how your local area has changed since 2006 or how it compares with the territorial authority / local board or how your regional council compares nationally? Now you can.
To coincide with the place summaries product release in late February 2020, we will be touring around the country [35-40 locations] showcasing place summaries and providing local area data back to communities through a series of interactive sessions.
Are you reasonably data-capable but not sure how to make the most of census data to aid your organisation? Do you know what census data you can use or how to access it? We are here to help!
This short workshop will help to increase your understanding of census data, provide the opportunity for some hands-on interaction and answer your organisational-specific scenarios.
If you are from a professional or community organisation and are interested in the tour and data capability workshop, please contact [email protected].