An analysis of fertility differentials in Liberia and Ghana using multilevel models
An analysis of fertility differentials in Liberia and Ghana using multilevel models
This thesis investigates differentials in the levels of fertility, nuptiality and contraceptive use in Liberia and Ghana, using data from the recent Demographic and Health Surveys in these countries. Of particular interest is the effect of the community in which a woman lives on her current and past fertility, her marital status and her use of contraception. This interest stems from the fact that, although the community in which a woman lives is integral to anthropological explanations of fertility, statistical models of fertility have rarely included an assessment of community effects.
The method of analysis used is multilevel modelling. This involves fitting variables measured at the woman level, variables measured at the community level and also includes the use of random effects to assess the extent to which community effects have not been captured by the fixed explanatory variables. Multilevel log-linear models are used in the analyses of fertility and multilevel logistic models are used in the analyses of nuptiality and contraceptive use.
This thesis demonstrates not only that there is significant variation between communities in both Liberia and Ghana for number of births 0-4 years before survey, children ever born, marital status and use of contraception but also that in each case significant community effects are found even after controlling for woman's age, education, religion and ethnicity.
Parr, Nicholas John
60ff080f-e2d1-4c52-babc-9d14499a4b85
April 1992
Parr, Nicholas John
60ff080f-e2d1-4c52-babc-9d14499a4b85
Diamond, Ian
b4e9ea54-fced-4314-9286-727256504de9
Hinde, Andrew
0691a8ab-dcdb-4694-93b4-40d5e71f672d
Cooper, Philip
2d586394-4f46-4d60-a477-7649f48d2ff8
Parr, Nicholas John
(1992)
An analysis of fertility differentials in Liberia and Ghana using multilevel models.
University of Southampton, Department of Social Statistics, Doctoral Thesis, 496pp.
Record type:
Thesis
(Doctoral)
Abstract
This thesis investigates differentials in the levels of fertility, nuptiality and contraceptive use in Liberia and Ghana, using data from the recent Demographic and Health Surveys in these countries. Of particular interest is the effect of the community in which a woman lives on her current and past fertility, her marital status and her use of contraception. This interest stems from the fact that, although the community in which a woman lives is integral to anthropological explanations of fertility, statistical models of fertility have rarely included an assessment of community effects.
The method of analysis used is multilevel modelling. This involves fitting variables measured at the woman level, variables measured at the community level and also includes the use of random effects to assess the extent to which community effects have not been captured by the fixed explanatory variables. Multilevel log-linear models are used in the analyses of fertility and multilevel logistic models are used in the analyses of nuptiality and contraceptive use.
This thesis demonstrates not only that there is significant variation between communities in both Liberia and Ghana for number of births 0-4 years before survey, children ever born, marital status and use of contraception but also that in each case significant community effects are found even after controlling for woman's age, education, religion and ethnicity.
Text
92090604.pdf
- Version of Record
More information
Published date: April 1992
Organisations:
University of Southampton
Identifiers
Local EPrints ID: 192425
URI: http://eprints.soton.ac.uk/id/eprint/192425
PURE UUID: ce615efa-f279-434a-b824-7bd1ea48fe09
Catalogue record
Date deposited: 11 Jul 2011 15:52
Last modified: 15 Mar 2024 02:45
Export record
Contributors
Author:
Nicholas John Parr
Thesis advisor:
Ian Diamond
Thesis advisor:
Philip Cooper
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