Logistic regression has been used to model exclusive breastfeeding (EBF) as a response variable conditioning on some predictors. To determine the factors influencing the practice of EBF, various maternal and infant characteristics were compared between exclusively breastfed infants and their counterparts. Descriptive analysis was performed to determine the differentials of EBF by explanatory variables and to provide insight into the structure of data. A stepwise backward logistic regression model was applied to test significant variables. A p-value of less than 0.05 denoted significance of the factors, and was retained in the final model. Our results show that being married is inversely associated with EBF {less common among married women}. This inverse relationship appeared to be due to 2 selection factors in the sample and possibly due to the fact that most unmarried women were the teen ages {very few} who are under the care of their families. The results further show that richer women are more likely to practice EBF. That is the wealth index directly affects EBF prevalence. Classification table was used to examine the consistency of the model in classifying the observed and predicted values of EBF and its overall percentage of correct classification was 60.4%.

logistic regression, exclusive breastfeeding, infants, odds ratio.