Assessing the quality of partnership status attributes from admin data describes the method for deriving the two census attributes 'Legally registered relationship status' and 'Partnership status in current relationship' from admin data, and assesses the quality of the resulting admin attributes by comparing them with 2018 Census data.
Summary
This paper is one of a series of investigations by Stats NZ's census transformation research programme aimed at identifying and exploring the potential for admin data sources to provide census-type information.
One important reason for having a census is to provide information about the characteristics of a population, including information about relationships, for example, to understand family structure within New Zealand.
This investigation explores the potential for admin data to provide census-type information on the two partnership status attributes Legally registered relationship status (LRRS) and Partnership status in current relationship (PSCR).
Note that, whereas census groups the LRRS and PSCR attributes under 'relationship status', we use the term 'partnership status attributes' in the title and throughout this paper to avoid confusion with a future paper on familial relationships.
Note that this investigation took place in 2021, prior to both the third iteration of the admin population census (APC) and the 2023 Census. Methodology from this research was incorporated into both. Since 2023 Census data was not available at the time of this investigation, we used the 2018 Census for comparative purposes and the 2013 Census as a data source for deriving the two admin partnership status attributes.
The main findings of this investigation are:
- Overall, the two partnership status attributes derived from admin data sources currently provide census-type information reasonably well.
- The scope of admin PSCR is broader than census which limits partnerships to those between individuals living in a household. This opens opportunities for researchers to study this portion of the population that was previously unavailable.
- Overall coverage of the admin target population is 81 percent for LRRS and 88 percent for PSCR. For both attributes, coverage is lower for young adults, the Asian ethnic group, and individuals born overseas.
- Overall, 95 percent of admin LRRS values and 81 percent of admin PSCR values are consistent with the 2018 Census. LRRS consistency is high across all sub-populations investigated. PSCR consistency is lower for young adults, the Māori and Pacific peoples ethnic groups, and those of Māori descent.
- For both attributes, data from the 2013 Census is crucial to obtain sufficient coverage of the target population. This impact is largest for individuals aged under 25 years or over 50 years, those born overseas, the Asian ethnic group, and the categories 'never married', 'separated', and 'de facto'.
- Admin data sources have been identified that could improve the quality of both attributes, once incorporated into the Integrated Data Infrastructure (IDI).
- Over time, historic census will have a diminishing contribution to the admin partnership attributes. There is the potential for the new data sources to provide additional partnership information. In addition, the number of partnerships missing from New Zealand registers because they have not been digitised will reduce over time.
ISBN 978-1-991307-51-4 (online)
