基于迭代囚徒困境的“助人任务”开发与验证：评估求助者关系动态

Self-other evaluations play a crucial role in understanding psychiatric disorders and improving collaboration between service users and providers. However, experimental paradigms of such relational processes tend to have low reliability and ecological validity. Quantitative analyses can validate the task’s measurement properties while qualitative analyses can empower service users to provide novel insights about task development and interpretation. In this study, we use both quantitative and qualitative analyses to develop the “Helping Task” based on the iterated prisoner’s dilemma. In this task, help-seeking participants knowingly engaged in an asymmetric relationship with a computerized help-provider, in which both parties invested effort (points). Initial qualitative feedback from 17 individuals (15 ‘neurotypical’, 2 having lived experience of relational difficulties) was used to refine key task properties (like task instructions, interface labels, and general navigation). Subsequent thematic analyses revealed that these participants interpreted helping exchanges through expectations about reciprocity, moral obligations, and underlying mental intentions, mirroring descriptions of real helping relationships. Moreover, quantitative analyses of 30 online participants validated the task’s sensitivity to helper behaviour and the within-session reliability of mean trust, decision satisfaction, and intended effort. Finally, computational analyses revealed probabilistic states in which trust, decision satisfaction, and effort were jointly lower or higher, indicating that self-other evaluations were linked to willingness to invest in the relationship. These findings support the reliability and validity of the Helping Task and pave the way for computational studies of self-other evaluations in helping relationships.