Fig. 5
The relationship between weight-for-age zscore by birth-year cohort and access to the midday meals program3. .3 Figures 5, 6, 7 Notes: Regressions are estimated using robust standard errors. The treatment variable is a binary variable equal to one if the child lives in a plantation which runs the midday meals interventions. Dots denote the point estimates for the treatment effect and horizontal lines denote the 95% confidence intervals. The left panel shows the results based on instrument-variable random effects (IV-RE) estimation, while the right panel shows the results using generalized least squares random-effects (GLS-RE). The vertical line represents the zero-reference line