Evaluation of geographical variation in live-birth registrations using Bayesian method

Abstract

Registration of births is very important and crucial in human’s life. This study examines the variation of live-birth registrations among the local governments in Ondo State. The data used was a secondary data collected from National Population Commission for the period of four (4) years (2012 − 2015). Spatial analysis was based on Poisson regression model and estimation was through Bayesian approach. Deviance information criterion was used to showcase the goodness of fit of the visualize variations and the 95% credible intervals show the variation thereby showcasing the significant clustering of each of the variations among the local governments. The result shows that Akure and Ondo are highly significant with live-birth registrations because of their consistency in year-in year-out registrations. It has been practically established that registrations among Akoko North East, Ilaje Ese odo, Okitipupa, Odigbo, Irele, Ile-oluji/Oke-igbo, Ifedore, Idanre, Owo, Ose, Akoko South and Akoko North West local governments are not consistent and are significantly low. The findings can guide the registrars on where to focus on for live-birth registrations.

Keywords:

• Live-birth registration
• deviance Information Criterion
• age-group
• clustering
• bayesian Model

Acknowledgments

A special thanks to National Population Commission for granting us access to the data analyzed in this study.