From Sender to Receiver: Investigating Children’s Roles in Food Communication Through a Randomized Controlled Trial

The intervention is based on the delivery of a food-related message addressing either health or sustainability.
The delivered message consists of a short text that is read aloud, followed by the day's school canteen menu.
The short texts were developed based on focus groups with 4th-grade pupils, during which food practices, health, and sustainability were discussed.

2026-01-12

2026-01-30

The primary outcome is the consumption of fruit and vegetables by participating pupils at the school canteen. We will use the variation of quantities at the beginning of the meal (before the child eats) and at the end of the meal (when the child clears her tray).

One label with each participating child's anonymous number will be affixed to the tray for the day. At the beginning of the meal, a research assistant will launch an app using AI that allows taking a photo of each participating child's tray, before they eat. A photo of the tray will be taken again at the end of the meal. The data will then be transcribed into an Excel file, indicating the quantity variation of each food item on the tray before and after the meal. A dedicated AI app will be used to estimate quantity variations.

We will also examine variations in the quantities of other food items consumed and in food waste.
The acceptability of vegetarian meals, when available, as reflected in children’s consumption, will also be assessed.

Those quantities will be measured using the same method as the primary outcome.

This RCT aims to investigate the mechanisms through which "food messages" are transmitted among children in school settings, and how these messages influence children's dietary behaviors. The study focuses on 4th-grade pupils across 10 primary schools.

Not available

Two schools were randomized into each treatment group, ensuring that no two schools within the same group were located in the same district. Furthermore, each group included one school from a low socioeconomic background and one from a high socioeconomic background. Randomization was performed in the office using a computer-based procedure.

Randomization was done at the school level.

Yes

10 schools

556 pupils aged between 8 and 10 years old.

The target is 100 pupils in each treatment and control group.

From the literature, we can assume that, prior to the intervention, approximately 50% of children in our context will consume at least one portion of fruit and vegetables in the school canteen (one out of two children). For the power analysis, we consider the following parameters: a Type I error rate α = 0.05, a power of 1 - β = 0.8, and a moderate effect size of 0.3. Consumption will be recorded twice a week for three weeks in each school, resulting in six observations per child during each phase. Data collection will follow a timeline based on four phases: baseline (before the intervention), during the intervention, post-intervention, and three months after the intervention — leading to a total of (maximum) 6 × 4 = 24 observations per child. Expecting a 0.4 variation rate, with the above-mentioned parameters, 93 pupils are needed in each group. Having multiple observations per child helps to mitigate potential absences during data collection. Hence, we plan for a sample size of 100 children per group, which allows for potential attrition over time, resulting in a total sample of 500 pupils.