Files
Abstract
While Mali has experienced agricultural growth in recent years, child malnutrition has remained alarmingly high. This thesis examines the contributions that including indices can make in modeling child nutritional status. All indices are derived from three index categories: childcare, sanitation, and feeding. Four indices are constructed through two methods: factor analysis and ad hoc. The models with indices are compared to a preexisting model in order to identify immediate and underlying determinants of height-for-age (HAZ) Z-scores (an indicator of long-term health and nutritional status) in rural Mali and examine the relative impact that changes in these factors will have on average HAZ scores for rural children. Though the original model has a larger number of observations and will, ceteris paribus, have a larger R-squared value, the higher R2 does not necessarily signify that the model is an accurate reflection of the underlying relationships.
Explanatory power of the individual indices varies. Coefficients of four of the indices are not significant in the regression. Four of the indices are significant and deal with care and feeding. None of the sanitation indices are significant.