2018 Census External Data Quality Panel: Assessment of variables

Stats NZ has adopted a rating scale to assess the quality of each variable measured in the 2018 Census of Population and Dwellings. The 2018 Census External Data Quality Assurance Panel (EDQP, the panel) has summarised its findings on the variables it has assessed by using this scale as well.

The scale has five options – Very High, High, Moderate, Poor, Very Poor – and there are three metrics that Stats NZ has adopted to calibrate this scale. A summary of these metrics is provided in Appendix 1; further detail can be found in Data quality assurance for 2018 Census.

The quality assessments by Stats NZ are a welcome initiative. Such measures enable users to assess the fitness for use of the statistical estimates and models that Stats NZ produce, and to have those assessments independently challenged. As a result, the richness of the information obtained in each population census can be used more critically in public policy, social and economic research and analysis, and community development.

When assessing the quality of data for several of the variables covered in this report, the panel has found a single category on the rating scale was not appropriate. For example, the data may be ‘moderate’ in terms of its quality at certain levels of aggregation and for certain groups in the population, but ‘poor’ at other levels of aggregation and for other groups. In some of the assessments in this report the panel’s consensus judgement is that a fair assessment spans two categories. 

The panel assessed key variables such as age, sex, usually resident population count, Māori descent electoral, and ethnicity (levels 1 and 2) in the Initial report of the 2018 Census External Data Quality Panel, which was published to coincide with the 23 September first release by Stats NZ of 2018 Census data. 

The panel has assessed an additional 31 variables in this report. These variables, along with the ones assessed in the initial report, and a small number that will be commented on in the panel’s final report to be published in early 2020, are listed in Tables 1.1 and 1.2. The shaded areas in these tables highlight the variables covered in this report. 

Download the full report below.

ISBN: 978-1-98-858376-1