Multilevel Analysis of Fertility Rate in Northern Nigeria

Abstract

In this study the modification of the existing multilevel regression models was carried to analyze fertility rates in Northern Nigeria. By modifying the slope parameter, we aimed to improve the model's accuracy and compare its results to the original model. Additionally, we assessed the impact of fixed and random effects on the intercept. The data collected from Nigerian Demographic and Health Survey (NDHS), 2018, is used to validate the new model. Our analysis revealed that the modified model provided more accurate estimates of fertility rates compared to the unmodified model. Also, the impact of Fixed and Random effects on the regions (North-East, North-West, and North-Central), and the independent variables (Religion, Number of wives, Highest educational levels, Sex of child) was determined, it was discovered that the intercept and Education have significant effects with p-values of .000, indicating strong influence. Religion is marginally significant with a p-value of .051, while Sex of child has no significant effect           (p = .798), the number of wives shows a highly significant effect (p = .001). The modified Multilevel Regression model is more robust than the unmodified Multilevel Regression model.