Evaluation of Ghana School Feeding Program: A Causal Mediation Analysis

Abstract

Introduction: School feeding programs (SFPs) deliver significant nutritional intervention during school children's critical growth period. Its effects on learning appear to work by both increasing school attendance and earning efficiency while at school. As part of the evaluation of the Ghana School Feeding Program (GSFP), the findings did not suggest any impact on the learning outcome, hence this research attempted to determine the mechanism by which the school feeding program (or not) impacts learning. Methods: A randomized control cluster design was adapted to scale the GSFP across the 10 regions. Baseline data was collected in June 2013 and follow-up in March 2016. A total of 30 districts were randomized in both research rounds. A two-stage stratified random sampling design was used. Twenty-three and twenty-nine schools were randomized at the intervention stage: respectively, Home Grown School Feeding Pilot (HGSF+) and its control. The study district selection criteria were based on national poverty and food insecurity. Household inclusion criteria were limited to children aged 5 to 15 years. Results: There was no significant association between learning outcome and treatment, HGSF+ (95% CI -2,25 to 1,64, P-value = 0,758). Likewise, there was also no significant interaction between mediator(weight-for-age) and treatment (95% CI -0.015 to 0.024, Pvalue = 0.636). The indirect effect of HGSF+ on the learning outcome through the mediator was 0.999 (95% CI: 0.972-1.032), and this is statistically significant. Conclusion: There was no evidence that the enhanced school feeding program improves learning outcomes, findings of the study suggest the design, execution, and timing of the study were causes to the failed desired impact of the enhanced SFP on learning outcomes