Mechanisms of Coordination during Foraging in Cooperative and Competitive Interactions

Successful coordination in dynamic environments often requires individuals to integrate their own preferences with predictions about others' behavior. In this study, we tested 148 participants in a Social Foraging Task, where pairs selected between resource patches to maximize shared outcomes. Behavioral clustering revealed four distinct coordination profiles: Economic, Sticky, Avoiding, and Mixed. Each group showed characteristic decision patterns and levels of coordination success. To uncover the cognitive strategies behind these profiles, we applied hierarchical Bayesian inference using twelve computational models. These models systematically varied in how participants evaluated options for themselves and how they estimated their partner's behavior, incorporating features such as sensitivity to visible resources, patch preferences, tendency to repeat their own choices, outcome-oriented updating, and leader-follower dynamics. Different clusters were best explained by different combinations of these features. Participants in the Economic and Mixed clusters were best captured by models that combined sensitivity to patch values and stickiness in their own choices, while also assuming their partner used similar decision processes. Sticky participants also used island resources and repetition themselves, but assumed their partner followed a simpler, repeat strategy. Participants in the Avoiding cluster relied mostly on repetition themselves but modeled the partner as someone using complex information, including resource values and outcomes. This suggests that even when participants limited their own engagement with certain patches, they still attended closely to their partner's behavior. Coordination success was linked to similarity in internal value functions between partners. This was especially true in the Economic cluster, where alignment of value representations supported consistent coordination. In contrast, participants in the Mixed cluster showed more variability in strategy and poorer alignment. Together, these findings reveal that people coordinate using diverse mechanisms, ranging from shared valuation strategies to asymmetric social tracking, depending on their behavioral tendencies and mental models of others.Keywords: Social Coordination, Foraging, Social Inference, Mentalizing, Social Decision-Making, Cooperation, Competition.