|
Accompanying the report are interactive graphs providing data sources for key 2018 Census variables.
- Read the final report of the 2018 Census External Data Quality Panel
- View the graphs at 2018 Census External Data Quality Panel: Data sources for key 2018 Census individual variables
Background to the report
The panel was convened by the Government Statistician in August 2018 to provide an independent, external review of the quality of 2018 Census data and to provide recommendations to the Government Statistician around improvements to census data quality. The eight-member panel includes experts on census methods, statistics, Māori data, demography, and equity.
It was the Government Statistician’s intention that the panel’s reports would be released publicly and unedited, as a matter of transparency, so all New Zealanders could see both the quality of the variables and the composition of the data.
Copies of the previous reports can be accessed below:
The panel’s key findings
Below is a summary of the panel’s findings in their own words.
- The use of government administrative records has improved the overall coverage of key variables such as age, sex, ethnicity, and place.
- Changes in methods for the 2018 Census have introduced breaks in the time series compared to earlier censuses. Measurement of change should be undertaken with caution.
- The metric 1 quality measure at level 4 of the ethnicity classification (e.g., French, Tongan, Indonesian) ranges from very high to very poor.
- The metric 1 quality measure at level 4 of the language classification (e.g., English, te reo Māori, Fijian) is estimated to range from high to poor. Data on te reo Māori (poor) is not of sufficient quality to undertake time series analysis.
- Stats NZ initially rated the 29 family and household variables as being of very poor quality. Stats NZ is continuing to work on understanding and improving the quality of census data, and users should check with them on quality ratings before using these variables.
- There were 24 small areas (SA2s; 1 percent of the total) with populations of 300 or more where less than 60 percent of data came from census forms. Most low response SA2s have high Māori and/or Pacific populations.
- Caution should be exercised when using data rated as being of poor or very poor quality. The panel encourages Stats NZ to control access to data rated as being of poor quality overall.
- The panel have made 24 recommendations for Stats NZ to consider, several relating to preparations for the census in 2023.
- Regaining the trust and support of iwi and other Māori organisations, as well as key users of census data, will be critical for a successful 2023 Census.
Our response
Stats NZ would like to thank the panel for the thorough and dedicated work they have done to assess 2018 Census data quality. We acknowledge the impact that a change in census methodology has had, both on our customers and wider census data beneficiaries.
We agree with the report’s position around the importance of helping our customers understand what has gone into building each variable in the 2018 Census dataset and what that means for customers, including any constraints they may face.
Stats NZ agrees with most of the panel’s recommendations and progress has already been made on five of them, with a further 16 scheduled for review and consideration in the coming months.
We know our customers will view and analyse our data in ways we haven’t, so may find differences or variations in quality that we have not seen. We encourage anyone who has this experience to get in touch with us so we can broaden our assessments and make sure other customers can learn more about how census data can be appropriately used.
We continue to be committed to helping our key customers, and all New Zealanders, understand both the value and the limitations of 2018 Census data.
More information
If you’d like to know more about 2018 Census data quality, you can find a range of reports and methods papers, as well as a breakdown of all our quality ratings on the Data quality for 2018 Census webpage.
|
