Structural Equation Modeling of the Effects of Family, Preschool, and Stunting on the Cognitive Development of School Children

A recent study based on a sample of 1,580 children from five adjacent geographical locations in KwaZulu-Natal, South Africa, was carried out to examine the association of nutrition, family influence, preschool education, and disadvantages in geographical location with the cognitive development of school children. Data were collected on the children from 2009 to 2011 for this developmental study and included cognitive scores and information on the health and nutrition of the children. The current study analyzed the association of demographic variables (geographical location (site)), child variables (sex, preschool education and socioeconomic status), parental level of education (maternal and paternal), child's health (HIV status and hemoglobin level) and anthropometric measures of nutritional status (height-for-age) with children's cognitive outcomes. The hypothesis is that the nutritional status of children is a pathway through which the indirect effects of the variables of interest exert influence on their cognitive outcomes. Factor analysis based on principal components was used to create a variable based on the cognitive measures, correlations were used to examine the bivariate association between the variables of interest in the preliminary analysis and a path analysis was constructed, which was used for the disaggregation of the direct and indirect effects of the predictors for each cognitive test in a structural equation model. The results revealed that nutritional status directly predicts cognitive test scores and is a path through which other variables indirectly influence children's cognitive outcome and development.