Conference Agenda

We study how social interactions shape meal choices using detailed canteen transaction data comprising roughly two million purchases over one academic year. Leveraging purchase timestamps, we infer the social networks of university students and examine how exposure to vegetarian peers influences individual consumption patterns. We find that individuals that dining with a vegetarian peer on a given day reduces meat consumption by about 4 percentage points, corresponding to roughly 9 percent of baseline levels. Together, our findings highlight how social contexts can shape sustainable consumption behaviors.

Knowledge gaps and biased beliefs concerning both climate change and climate policy represent major obstacles to the decarbonization process. Climate education may offer a scalable solution to address such obstacles. In the context of a national reform of the school curriculum in Italy, we implemented a nationwide field experiment, training thousands of secondary school teachers across thousands of schools using a staggered design. Our intervention, a comprehensive course on climate change and climate policy, goes beyond the light-touch interventions typical in the literature. Using extensive survey data, we examine how training affects teachers’ knowledge, beliefs, attitudes, behaviors, and policy preferences and, in turn, those of students. Our study highlights important initial knowledge gaps and biased beliefs about climate change

among teachers and students, and provides evidence that climate education can address them at scale. Following our intervention, teachers and students also reconsider their support for climate policies.

Interventions to reduce meat consumption are often short-lived, leaving their long-term effects unclear. We study the post-intervention effects of a one-month removal of meat meals from a university canteen menu. Using transaction data on over 270,000 purchases by more than 4,500 guests, we estimate difference-in-differences effects in an intent-to-treat framework. Treated guests were 4% less likely to choose meat in the two months after the intervention, with no change in visit frequency. Purchase patterns and survey evidence suggest these effects are driven mainly by taste discovery rather than habit formation or shifting social norms.

Despite the urgency of addressing climate change and other environmental challenges, individual, collective and policy action has remained limited. Changing relevant beliefs, values, and norms through education programs is one potentially high-impact lever for increasing pro-environmental attitudes and behaviors. Immersive education interventions have been shown to have deeper and longer-lasting effects than other behavioral interventions. However, these studies typically measure only the direct effects of the intervention on those exposed to the treatment. To the extent that there are spillovers to others in the household, the total effects of the intervention could be even larger than previously estimated. Here, we design a study to understand both the direct and indirect effects of an immersive education program. We ask whether parents or children are more effective targets of educational interventions when accounting for indirect spillover effects. While most accounts of intrahousehold socialization or intergenerational transmission of beliefs and norms focus on the influence that parents have on their children, climate change and environmental challenges are intergenerationally asymmetric: parents have more agency over household decisions yet children are the ones who face the brunt of unmitigated climate change. These asymmetries could make children more effective conduits for the transmission of environmental knowledge, attitudes and behaviors, as they have greater moral leverage. Using a randomized field experiment with 1,545 child-parent pairs in Patna, India, we analyze both direct program effects on participants and indirect spillover effects on non-participating family members. Our design has four distinct treatment arms: a control group, child-only participants, parent-only participants, and joint child-parent participants. This allows us to systematically evaluate the effectiveness of targeting different household members for an environmental education program when accounting also for the indirect effects of the intervention. We find strong direct effects of the intervention on both adults and children. We also find spillover effects on non-participating family members, but these effects vary by the type of behavior and the direction of influence. While both parents and children influence each other’s observable pro-environmental behaviors, we find that children uniquely shape their parents’ attitudes toward climate change and perceptions of environmental risks. Interestingly, joint participation by both parent and child does not yield stronger results than targeting either member individually, suggesting that selective targeting may may offer a more scalable and cost-effective approach to promoting sustainable household behavior.