Experimental Design
Our experimental design consists of four different treatments:
Baseline: Children participating in the experiment are presented with five different food trays. Each tray includes five different food items of similar nutritional value, selected by a nutritionist. Subjects have to pick four food items from any of the trays. They have the possibility of choosing items from the same or from different trays, and are able to select more than one item from the same tray and more than one unit of the same item.
Grades Treatment (GT, hereafter): Similar to Baseline, with the difference that students see labels with a “grade” associated with each of the five trays before making their choice. The grades corresponding to each tray depend on their nutritional content, and are analogous to those used to mark children’s academic schoolwork (in Spanish schools grades range from 0 to 10, so the five trays have assigned marks of 0, 2.5, 5, 7.5, and 10, respectively). All the items included in a given tray have the same grade assigned. These grades are the only pieces of information that children allocated to this treatment receive.
Parents Treatment (PT, hereafter): Similar to GT, with the difference that parents receive information about the average mark (linked to the nutritional composition) of their child. Before selecting the items, children are aware of the fact that their parents would receive a report about their “performance”. Note that parents do not receive exact information about children choices, but just the average grade obtained.
Nutritionist Treatment (NT, hereafter): Similar to Baseline, with the difference that the first day of the experiment and before children make their decisions, a nutritionist give a short talk explaining the benefits of eating in a healthy way. Note that the nutritionist give a general talk for all the participants and he avoids leading children to pick any particular item from the trays.
Each day of the intervention in each school, experimental subjects belonging to the same class are gathered in a room (Room A). Participants are not given any information about the experiment. Each student is then – independently and sequentially - asked to move to another room (Room B) in the company of one of the experimenters. In this second room, the subject is walked through a short orientation session - the details of which depend on the subject’s treatment assignment. Next, the student pick the four food items they prefer. Finally, the subject is taken to a third room (Room C), where she joins other classmates who already completed the task. This procedure ensures that individuals’ decisions are not directly influenced by their classmates (who remain either in Room A or C at the time the participant is choosing her lunch in Room B).
Subjects’ dietary choices and grades are recorded by a member of the research team, who is present during the decision process. Note that in order to minimize the interaction with the participants, children make their decisions alone. The bags used to keep the food are transparent so the researcher can see what children pick through it, without being directly involved during the decision.
A staff member of each school is present in Room C where children gather together after making their choices. This member controls that no food is wasted.
To analyze the dynamics and long-term effects of the incentives, we collect data twice a week over three weeks, for a total of 6 observations per subject. In the case of students assigned to PT, information about grades is also sent to their parents. To that end, and in order to keep the same conditions across treatments, prior to the experiment, the parents of all the participants (regardless the treatment) receive the same explanations/clarifications and are requested to provide their contact details before the beginning of the experiment. Parents in PT receive the information only once at the end of each week, which implies that each time they receive the information from two different dates.
After the end of the experiment children are administered a questionnaire asking them to identify their closest friends in class. We also conduct post-experiment questionnaires with children, teachers and parents in order to obtain information about participants’ socioeconomic variables and self-control indicators. Regarding the socioeconomic background of the student, we borrow a questionnaire administered by the Spanish Government. Among the set of variables, we can find: i) the highest educational level achieved by the parents, ii) parents’ jobs, iii) number of books in the household, iv) proxy variables for the economic level of the household, v) parents’ involvement in the school-related activities (homework) of their child. For the self-control indicators, following Tsukayama, Duckworkth and Kim (2013), we include questions related to interpersonal impulsivity and also to schoolwork impulsivity. Finally, in the questionnaire administered to the teachers, we gather information regarding students’ average performance in class, average attendance to the school, and an external measurement of self-control.
Finally, four months after the end of the first intervention, all subjects participate in a second round of the Baseline. The aim of this is to study whether the effect of the different incentives remains after some time and once incentives had been removed.