Multi-Level Logit Model for Infant Mortality in Nigeria

Abstract

An accurate understanding of the variance dynamics of infant mortality in Nigeria is as crucial, as accounting for the influence of predictor variables. This study thus accommodates the hierarchical structure due to the mother’s age. Data from the 2013 and 2018 Nigeria Demographic Surveys were used in this study. A total of 38459 complete records make up the study sample. Given that the outcome variable is binary, a Multilevel logistic regression is used in this research with the mother’s age category as the higher-level variable. Multilevel modeling enables us to appropriately account for factors contributing to infant mortality, providing a better understanding than the traditional logistic regression model. The study found out geopolitical zone/Region, number of children ever born, and Toilet type (Exposed toilets) were significant contributors to the prevalence of infant mortality in the study. Particularly, the SouthEast zone was identified as the most affected area, with the highest odds of infant mortality as compared to its other counterpart. Similarly, homes with toilets that are exposed are more likely to experience a higher rate of infant mortality than those using flush toilets. The study findings also revealed that increased childbirth implodes infant mortalities within the population. Thus, policy efforts should be geared towards mitigating the current trend of infant mortality in Nigeria.