Table 7

Treatment Effects of School Feeding on Child Cognitive Scores, Full Sample, and Heterogeneity by Child Gender, Household Poverty, and Geographical Areas

Digit SpanStandardized Progressive MatrixesComposite Cognitive Score
Panel A: All Children
School feeding0.1190.1290.143
[–0.010 – 0.249][0.018 – 0.240][0.019 – 0.267]
Unadjusted p-value0.070*0.024**0.024**
R-W p-value0.070*0.046**
Observations2,3052,3072,321
R-squared0.0500.0340.064
Mean treatment endline0.09920.1190.171
Mean control endline-0.0348-0.03260
GirlsBoysGirlsBoysGirlsBoys
Panel B: Gender
School feeding0.1900.06490.1160.1480.1750.124
[0.0372 – 0.342][–0.207 – 0.336][–0.0358 – 0.267][0.0191 – 0.276][0.0215 – 0.329][–0.0223 – 0.271]
Unadjusted p-value0.015**0.4040.1330.025**0.026**0.01*
R-W p-value0.032**0.4030.1220.042**
Observations1,0851,2201,0861,2211,0911,230
R-squared0.0570.0530.0430.0370.0710.068
Mean treatment endline0.09770.1010.06870.1650.1400.200
Mean control endline0.1080.0288−0.09420.0208−0.08280.0717
Below the Poverty LineNonpoorBelow the Poverty LineNonpoorBelow the Poverty LineNonpoor
Panel C: Household Poverty
School feeding0.2690.07230.2340.09310.2930.0946
[0.108 – 0.429][–0.172 – 0.316][0.100 – 0.367][–0.0427 – 0.229][0.147 – 0.440][–0.0517 – 0.241]
Unadjusted p-value0.001***0.3390.001***0.1770.0002***0.202
R-W p-value0.002***0.3420.002***0.309
Observations 540 1,765 537 1,770 542 1,779
R-squared0.0960.0420.0710.0300.1120.056
Mean treatment endline0.1270.09070.1380.1130.1960.164
Mean control endline−0.2070.0175−0.1610.00626−0.1800.0547
NorthSouthNorthSouthNorthSouth
Panel D: Place of Residence
School feeding0.253−0.0020.2120.0510.2720.025
[0.0798 – 0.427][–0.323 – 0.318][0.040 – 0.384][–0.0922 – 0.194][0.0994 – 0.445][–0.147 – 0.198]
Unadjusted p-value0.006***0.9810.017**0.4770.003***[0.772]
R-W p-value0.009***0.980.013**0.726
Observations1,0931,2121,0961,2111,0991,222
R-squared0.0430.0640.0280.0420.0550.079
Mean treatment endline0.1310.0660.1760.060.2260.114
Mean control endline−0.1090.024−0.0498−0.019−0.05160.040
  • Notes:

  • * p < 0.1,

  • ** p < 0.05,

  • *** p < 0.01.

  • Confidence intervals clustered at community levels in squared brackets. R-W p-values were adjusted for multiple testing using the Romano–Wolf (2005, 2016) step-down method with 2,000 iterations and standard errors clustered at community level. The table above presents intent-to-treat effects on each outcome for the full sample and stratified by child gender, household poverty, and place of residence. Models were estimated through OLS. For each outcome, the model controls for its baseline value, a dichotomous variable related to the randomized assignment to school feeding, and region dummies. Household poverty is a dichotomous indicator having the value of one if the household had baseline per capita consumption levels falling below the national consumption poverty line in 2013. Northern regions include Upper West, Upper East, and Northern region. Southern regions include Western, Central, Greater Accra, Volta, Eastern, Asanti, and Brong Ahafo.