The disjunction effect in the prisoner’s dilemma game shows that humans tend to cooperate more under uncertain condition (U) than under the two complementary known conditions—one being competitive (D) and the other being cooperative (C)—a well-known violation of the classical decision principle. Our study explores the potential role of prosociality in the disjunction effect. We measured prosocial trait via the SVO Slider Measure, and prosocial bias via the drift diffusion model (DDM). By using the SVO Slider Measure (for prosocial trait) and the DDM starting-point bias parameter (for prosocial bias), we found that the variation in prosocial bias between uncertain and certain conditions substantially contributes to the disjunction effect. At the aggregate level, prosocial bias significantly decreased from U to D (competitive) but did not differ between U and C (cooperative). At the individual level, participants showed heterogeneous bias changes across prosocial-trait groups: intermediate participants had the largest bias shifts. This heterogeneity underlies the observed inverted U-shaped relationship between prosocial trait and effect size of the disjunction effect. Our study fills a critical gap by clarifying how prosocial inclination influences the disjunction effect.
Xin et al. (Fri,) studied this question.