A big concern about polarization is whether its effects can spill over beyond the political realm and affect our everyday interactions, compromising our capacity of working together and collaborating in situations in which both sides can obtain gains. To analyze if polarization affects cooperation between citizens, we conduct a behavioral experiment in Brazil, United States, and Spain. Participants were asked to cooperate asynchronously with another subject on a simple one-shot task: converting to capital letters either a salad recipe or a political text written by the other subject. The experiment manipulates whether the other subject is described as voting for the respondent’s most liked or disliked political party, as well as the rewards and benefits of cooperation. The results reveal that cooperation is significantly reduced (by over 20 percentage points) when the other subject is an out-partisan, compared to a control group, although it is not more likely with co-partisans. However, the negative effect of polarization weakens considerably when cooperation contributes to a public good (e.g., a donation to an NGO), and it is amplified by higher levels of affective polarization. Additionally, among those who do choose to cooperate, task quality improves when a reward is offered but declines markedly when the task involves political texts.

These findings suggest that rising levels of affective polarization may extend beyond the political sphere, with serious consequences for cooperation in heterogeneous societies. When even basic interactions are undermined by partisan animosity, the potential for fostering higher forms of cooperation between groups with divergent political preferences becomes severely limited.