Combining individual- and population-level data to develop a Bayesian parity-specific fertility projection model
Combining individual- and population-level data to develop a Bayesian parity-specific fertility projection model
Fertility projections are vital to anticipate demand for maternity and childcare services, among other uses. Models typically use aggregate population-level data alone, ignoring the richness of individual-level data. We hence develop a Bayesian parity-specific projection model combining such data sources. We apply our method to England and Wales, using individual-level data from Understanding Society. Fitting generalized additive models gives smooth projections across age, cohort and time since last birth. We also incorporate prior beliefs about the relative importance of the data sources. Our approach generates plausible forecasts by individual-level variables including educational qualification, despite their absence in the population-level data.
Bayesian methods, Combining data sources, Fertility forecasting, Generalized additive models, Parity
Ellison, Joanne
d1560ac9-2c6c-49e8-b5c4-aa2258624e97
Berrington, Ann
bd0fc093-310d-4236-8126-ca0c7eb9ddde
Dodd, Erengul
b3faed76-f22b-4928-a922-7f0b8439030d
Forster, Jonathan
e3c534ad-fa69-42f5-b67b-11617bc84879
3 November 2023
Ellison, Joanne
d1560ac9-2c6c-49e8-b5c4-aa2258624e97
Berrington, Ann
bd0fc093-310d-4236-8126-ca0c7eb9ddde
Dodd, Erengul
b3faed76-f22b-4928-a922-7f0b8439030d
Forster, Jonathan
e3c534ad-fa69-42f5-b67b-11617bc84879
Ellison, Joanne, Berrington, Ann, Dodd, Erengul and Forster, Jonathan
(2023)
Combining individual- and population-level data to develop a Bayesian parity-specific fertility projection model.
Journal of the Royal Statistical Society: Series C (Applied Statistics), [qlad095].
(doi:10.1093/jrsssc/qlad095).
Abstract
Fertility projections are vital to anticipate demand for maternity and childcare services, among other uses. Models typically use aggregate population-level data alone, ignoring the richness of individual-level data. We hence develop a Bayesian parity-specific projection model combining such data sources. We apply our method to England and Wales, using individual-level data from Understanding Society. Fitting generalized additive models gives smooth projections across age, cohort and time since last birth. We also incorporate prior beliefs about the relative importance of the data sources. Our approach generates plausible forecasts by individual-level variables including educational qualification, despite their absence in the population-level data.
Text
Main text
- Accepted Manuscript
Text
qlad095
- Version of Record
Text
Supporting information
- Other
More information
Accepted/In Press date: 31 August 2023
Published date: 3 November 2023
Keywords:
Bayesian methods, Combining data sources, Fertility forecasting, Generalized additive models, Parity
Identifiers
Local EPrints ID: 482402
URI: http://eprints.soton.ac.uk/id/eprint/482402
ISSN: 0035-9254
PURE UUID: ca2d44bd-3d90-423d-82af-8336d9f5acde
Catalogue record
Date deposited: 02 Oct 2023 16:39
Last modified: 14 Dec 2024 03:03
Export record
Altmetrics
Contributors
Author:
Jonathan Forster
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics