Process and Effect Evaluation of the Project "Aan Tafel in 1,2,3 Euro"

"Aan Tafel in 1 2 3 euro" is a project of retailer Colruyt in cooperation with social organizations in Belgium. The project/intervention was launched in 2016 and has been implemented throughout Belgium since. In 2021 about 8000 families have already been subscribed, and the numbers are still growing. Families with a lower SES (vulnerable families) are informed by a social organization (OCMW, CAW, kind \& gezin, etc.) and can voluntarily register for the project. When a family is registered, they receive a recipe booklet every two weeks at home. The booklet consists of six easy, child-friendly recipes with accompanying shopping lists. Recipes are targeted at three people and include fish, meat and veggie dishes. Each recipe is guaranteed to cost no more than one, two or three euros per serving. The participants are registered via the Xtra-card (loyalty card of Colruyt), so that the discount can be offered anonymously and discretely. Since the intervention is an already existing one, we made an overview of the intervention components, behavior change techniques and practical applications, based on the Intervention Mapping Theory (Kok et al., 2015) and the Behavior Change Wheel (Abraham \& Michie, 2008). Based on this overview, we determined goals and outcome measures for our research project. In the qualitative study (WP1), participants of the intervention, as well as representatives of social organizations will be recruited to participate in an interview and/or focus group interview. Interviews will be conducted until data saturation is reached. In the interventional study (WP3), participants of the project (intervention condition) will be compared to participants in a control condition who do not participate in the project. A total sample size of 200 participants is calculated, based on an a priori power analysis (two groups, two measurements, α=0.05; power=0.80, ES=0.25). Participants in both conditions will complete a survey in a pre-measurement and a post-measurement, three months after the start of the intervention. The second effect study (WP2) has a longitudinal pre-post design, and is based on secondary purchase data from 2018 until 2022. The intervention group (i.e., participants of the intervention) will be compared to a control group (i.e., not subscribed in the intervention), by using general customer data. A multilevel growth model will be used to look at each individual's different purchase moments, with data before and during the intervention. The model represents three levels: 1) point of time/purchase, 2) customer, 3) store/location.