Stepwise Geoadditive Regression Modelling of Levels and Trends of Fertility in Nigeria: Guiding Tools Towards Attaining MDGs

Abstract

Relative to other sub-Saharan African countries, transition from high to low fertility in Nigeria appears to be one of the least and slowing. Considering the adverse effects of this on national development couple with a desire to achieve a sustainable development which in turn can lead to the attainment of the Millennium Development Goals (MDGs), Nigerian government has targeted a marked reduction in total fertility rate through recent population policy which has made it inevitable to study determinates of fertility in Nigeria. In this Chapter, we applied stepwise regression approach for regression models with structure additive predictors; a modelling technique that simultaneously performs model selection and estimation in a complex situation where modelling involves components such as trend, fixed, random, nonlinear, interaction and spatial effects with inferences based on penalised likelihood, to analyse individual level fertility in Nigeria. This approach enables us to explored possible trend, geographical variability and determinants of fertility in Nigeria using the 1999–2008 Nigeria Demographic Health Survey datasets. The results show that women’s level of education and husbands’ desire for fewer children are significant driving forces of fertility in Nigeria, and that age at marriage and during of marriage are have nonlinear effect on individual level fertility. The spatial effects show evidence of geographical variation in children ever born to women in Nigeria. The results provide insights for policy makers in designing strategies that will contribute towards attaining the MGDs.