The Effects of Food Social Security: A Randomized Experiment on Higher Education Students

The intervention involves a combination of financial support and educational activities aimed at fostering sustainable eating habits among students while addressing food insecurity. This multi-faceted intervention aims to address both immediate and long-term challenges associated with food insecurity while promoting knowledge and practices that align with sustainability goals.

1) Food Vouchers.
Participants in the treatment group will receive a monthly food voucher worth €125 for four months (January to April 2025). These vouchers can be redeemed in a curated list of partner stores that sell organic and/or local products and/or environmentally-friendly items and are mainly located in the 5th, 13th, and 14th districts of Paris. The selection of partner stores and eligible items will be finalized through a participatory process involving the students during the workshops. At this stage (on 11 March 2025), 15 partner stores have agreed to participate in the program, and this list of stores is not final. The vouchers aim to alleviate financial constraints and encourage the purchase of healthy and sustainable food products.

2) Workshops on Sustainable Eating
In November and December 2024 and January 2025, students in the treatment group will participate in a series of workshops designed to raise awareness about sustainable food practices. These workshops will include sessions on (i) the principles of sustainable and healthy eating, (ii) strategies for identifying and purchasing sustainable food products, and (iii) discussions to co-design the criteria for voucher usage, ensuring relevance and practicality for participants. The workshops will also emphasize collaborative decision-making, allowing students to propose and agree on criteria for store selection. By involving participants in shaping the intervention, the program seeks to increase engagement and empowerment.

3) Community Engagement Activities
To foster social interactions and build a sense of community, treated students will also be encouraged to attend regular group activities between the end of January and April/May 2025, running in parallel with the distribution of food vouchers. These events are designed to provide a supportive environment where participants can share experiences, discuss challenges related to food choices, and collaborate on solutions. Potential activities include cooking sessions that emphasize sustainable recipes, hiking in the forest and going to the cinema or the theater. By participating in these activities, participants are expected to build social connections, form peer support networks, enhancing both their knowledge and motivation to adopt sustainable eating practices.

2024-11-03

2025-04-30

Dietary habits (food access, food knowledge, nutritional quality, embedded GHG emissions footprint, share of organic consumption)
Health and well-being (general health, loneliness, depression, life satisfaction)

We build indices by combining several items using the methodology from Anderson (2008).

Food access is measured with an index based on two questions of the INCA 3 survey.
Food knowledge is measured with an index based on one question on the GHG footprint of dishes and four questions on nutritional facts, also asked in baseline.
Nutritional quality and embedded GHG emissions footprint are measured using a food frequency questionnaire in endline. The food groups will be those required to compute the comprehensive Diet Quality Index (cDQI). The share of organic consumption (in value) will be measured by analyzing the receipts associated to food purchases during one week in March and one in April.

Loneliness, depression and life satisfaction will be measured in endline using the same items as in the baseline questionnaire. Depression is an index built from several items.
General health is also a self-reported measure, based on a single item introduced in the endline questionnaire.

More details are available in the attached document.

Academic achievement (academic ranking, year completion)
Environmental attitudes and behaviour (support for environmental food policies, non-food-related pro-environmental behaviour)
Social relations (social support, trust)
Civic and political engagement (political self-efficacy, political participation, social engagement)

Academic achievement will be observed in administrative data after the end of the program. All the other secondary outcomes are measured in the endline questionnaire.
Except for non-food-related pro-environmental behaviour,, the corresponding questions were already asked in the baseline questionnaire. Except for the methodology from Anderson (2008).


More detail are available in the attached document.

There is one control group and one treatment group. The treatment group benefits from the intervention described above. The control group will receive 125€ in May 2025 to compensate their participation to the quesitonnaires and teh weeks of receipt collection.

Not available

The randomization was done in office by a computer. We stratified the randomization based on 1) self-reported gender 2) whether students live in the area of the shops where the vouchers can be used 3) self-reported food expenditures at baseline and 4) self-reported number of relatives with whom you can discuss intimate matters.

Individual (= students)

No

No cluster.

195 students observed in baseline, probably between 180 and 185 in endline.

99 students in the treatment group, 91 students in the control group in baseline.

We assume 85 students in the control group and 95 in the treatment group in endline, anticipating attrition. A two-sided two-sample t-test with 5% error has a power of 90% to detect an effect of 0.5 of a standard deviation in the outcome. A one-sided two-sample t-test with 5% error has a power of 80% to detect an effect of 0.38 of a standard deviation in the outcome. This power may be higher when introducing stratum fixed effects in endline. It may be even higher for the outcomes that were also measured in baseline when introducing individual-fixed effects.