Interpretive Nutrition Labels, Parent-Child Interactions, and Food Choices in Indonesia

The aim of this project is to investigate if interpretive nutrition labels may potentially food consumption of parents and children and that, in turn, may potentially inform current nutrition policies in Indonesia. First, we apply a “traffic lights” food labelling approach in the Indonesian context to examine its potential effect on the nutritional content of the grocery basket for family consumption. Second, we build on previous research studying the effect of “emoji” labels on children’s snack choices by extending the analysis from the lab experiment to a close-to-the-real-life food environment. Third, we assess the spillover effects of “traffic lights” labelling on parental ability and willingness to “correct” children’s choices. Fourth, we explore the nutrition knowledge and motivations behind children’s and parents’ decisions and choices related to food consumption.

2024-08-13

2024-09-30

1. Number, value and proportion of healthy classified products purchased by parents in treatment (“traffic lights” labels) and control groups.
2. Number, value and proportion of healthy classified snacks chosen by children in treatment (“traffic lights labels) and control groups.
3. Parental approval of children snack choices across four parent-child experimental conditions (binary variable where 1=approval).
4. Acceptance of parental (dis)approval by children (binary variable where 1=acceptance) across four parent-child experimental conditions.

1. Parental and children’s attitudes towards the “traffic lights” labels (attention, understanding, perception) among parents and children exposed to the “traffic lights” labels.
2. Children’s motivations of choosing the snack (perceived tastiness and healthiness of various snacks).

The study uses a randomized control design at the individual level based on the parents of 7-10 years old children from public primary schools. Subjects of our study are the parents and their children that agreed to participate in the laboratory experiment. The schools are sampled at random from the list of public primary schools in the area.

The parents are invited to a room where they are briefed about the procedure, task, the lottery with a chance to win a weekly supply of food for family consumption, and informed about a post-experimental survey. Participants are told that they are given 15 minutes and a fixed (virtual) budget to buy food and drink items in an online supermarket using their mobile phones. The online experimental supermarket offers a range of fresh fruits and vegetables, processed and non-processed foods, and healthy and sugary drinks. Following the completion of the experimental task, the participants fill out a post-experimental survey questionnaire. At the end of the session, the parents are invited to another room to pick up their children.

We invite children to participate in a quiz and bring them one-by-one to a separate room making a stop on the way to pick up a snack. In that room the children are waiting for their parents, fill a quiz, and their height and weight are measured. The parents are invited to pick up their children and every parent is asked to tell the child whether he or she is allowed to consume the snack.

Not available

Randomization of parents and children is done by a computer

Individual

No

20 schools (we plan to add more schools if participation in the randomly selected 20 schools is lower than expected)

800 parents and 800 children

800 dyads parent-child

The minimum required sample size of 800 observations (parent-child dyads) is calculated assuming 80 percent power and a 5 percent level of significance. We expect the minimum detectable treatment effect of 9.7 percentage points. The value share of processed/prepared foods accounts for 36 percent of per capita monthly food expenditures in urban Indonesia (National Statistics Bureau of Indonesia). If the share of unhealthy foods (labeled “red” in treatment “Traffic lights”) in the experimental shopping basket of adults at a baseline will be 40 (rounded) percent, we are powered to detect a positive treatment effect (“traffic lights”) decreasing the share to 30 percent within our setting given that we look at repeat exposure to labels in the online supermarket.