Multivariate Analysis of Fertility: an Application of the Generalized Poisson Regression Model

Abstract

Total fertility rate (TFR) is a standard measure commonly used to estimate fertility levels and trends. However, TFR is a period measure and does not offer reasons for observed variations in fertility rates and trends. Using data from the 2017 Ghana Maternal Health Survey (GMHS), this study examines current fertility levels and trends of ever-married women and employs a generalized Poisson regression (GPR) model to analyze the determinants of fertility. Findings suggest that the current fertility level of 5.4 for ever-married women is high. The results also reveal that age at first marriage, educational level, household wealth, area of residence, use of contraceptives; and ownership and use of a bank account are significant determinants of total fertility (p<0.05) and are thus factors that affect the fertility levels of women. The study concludes that the GPR analysis provides a clearer picture of the nature and determinants of fertility compared to the standard TFR analysis.

Keywords

• total fertility
• generalized poisson regression model
• determinants
• sub-saharan africa