What is it about?
The paper presents two research approaches at CSO, Ireland to compile population estimates from administrative data (data in Government databases) when no Central Population Register exists. One research approach is based on simple counts of records while the second includes an estimation step based on Dual System Estimation (DSE) methods. The paper also provides an overview of the DSE toolkit used - DSE methods are restated and extended in a manner allowing them to be applied in a broader set of scenarios including a scenario to deal with overcoverage.
Featured Image
Why is it important?
There is significant demand around the world for annual census like population estimates. It is not practical to undertake a Census every year, as such, census modernisation efforts are turning to the use of administrative data. This is particularly challenging for those countries that do not have a Central Population Register. This paper considers the use of administrative data for population estimates in such circumstances. The paper also provides a summary of the PECADO toolkit, a DSE based toolkit, and how it is applied to compile population estimates. The PECADO toolkit is novel in that it restates the DSE assumptions such that the methods can be used in a much broader set of circumstances, including a scenario where overcoverage exists or even a scenario where an administrative data list can be used to evaluate undercoverage in the traditional census operation.
Perspectives
The use of administrative data (data in Government databases) has a significant role to play in modernising National Statistical Systems. The potential of these data sources is huge, particularly when these data sources can be linked with each other (and survey data). However, Government systems are designed for efficient and effective public administration and not necessarily for statistical purposes. With this in mind, it is imperative that NSIs continue to explore/develop/share new and innovative ways to ensure the statistical potential of these data sources can be realised. We hope this paper makes a valuable contribution and that the perceived value/potential of administrative data is not diminished with the increased focus on newer and more novel data sources (MNO data, web scraped data and other big data sources) in the world of official statistics.
John Dunne
Central Statistics Office (CSO), Ireland
Read the Original
This page is a summary of: To count or to estimate: A note on compiling population estimates from administrative data, Statistical Journal of the IAOS, December 2023, IOS Press,
DOI: 10.3233/sji-230067.
You can read the full text:
Contributors
The following have contributed to this page
