A Proportional Odds Model with Complex Sampling Design to Identify Key Determinants of Malnutrition of Children under Five Years in Rwanda

Abstract

The main objective of this study is to identify the key determinants of malnutrition of children under five years in Rwanda. The Rwanda demographic health survey (2010) data was used as application. The anthropometric indicator for underweight (weight-for-age) was considered and categorized as severely undernourished when z-score <-3.0, moderately undernourished when -3.0 ? z-score <-2.0 and nourished when z-score ?-2.0. The score test and Brant test were used to test the proportional odds assumption and it was satisfied. As the data was collected using multistage sampling, this study extended the classical proportional odds model to include the complex sampling design. This research revealed that birth order, mother's education, gender of the child, birth weight of the child, marital status of the mother, body mass index, Anemia, multiple birth and whether or not the child had fever before the survey were found to be determinants of malnutrition of children under five years in Rwanda. The influence of these actors can be used to develop the strategies of reducing child malnutrition in Rwanda. When a complex design sample is used to draw a sample from finite population, the sample design should be incorporated in the analysis of the survey data in order to make statistically valid inferences for the finite population.

DOI: 10.5901/mjss.2014.v5n23p1642