School gardens have become a widely used approach to influence children's food knowledge, preferences and choices in low- and high-income countries alike. However, evidence indicates that such programs are more effective at influencing food knowledge and preferences than actual food choices. Such finding may occur because school gardens insufficiently influence the food behavior of parents and because healthy food items are not always available in children's homes. We tested this hypothesis using a one-year cluster randomized controlled trial in Nepal with 15 treatment and 15 control schools and a matched sample of 779 schoolchildren (aged 8-12) and their caregivers. Data were collected before and after the intervention during the 2018-2019 school year. In addition, children's food consumption was monitored using a monthly food logbook. Average treatment effects were quantified with a double-difference estimator. For caregivers, the intervention led to a 26% increase in their food and nutrition knowledge (p < 0.001), a 5% increase in their agricultural knowledge (p = 0.022), a 10% increase in their liking for vegetables (p < 0.001), and a 15% increase in home garden productivity (p = 0.073). For children, the intervention had no discernible effect on food and nutrition knowledge (p = 0.666) but led to a 6% increase in their liking for vegetables (p = 0.070), healthy food practices (p < 0.001), and vegetable consumption (October-December +15%; p = 0.084; January-March +26%; p = 0.017; April-June +26%; p = 0.088). The results therefore indicate both schools and parents matter for nudging children toward healthier food choices.
Keywords: Healthy eating; Impact evaluation; Nepal; Nutrition-sensitive agriculture; Randomized control trial; School garden.
© 2020 The Authors.