Pairwise mean differences for trust conditions.

When deciding whether to reciprocate trust, people are typically strongly influenced by how much trust their interaction partner has originally shown them. If a partner has placed a lot of trust in you, there is a strong motivation to reciprocate, and indeed this factor often outweighs pro-self considerations to maximize one’s own financial payout. However, one important unanswered question in this regard is what people decide to do when this prior information is ambiguous; that is, when they do not know for sure exactly how trusting their partner has been. How then do people decide to reciprocate? This study utilizes a novel version of the Trust Game to directly address this question. Here, we develop, and validate, a computational model-based approach to quantify and categorize how participants assessed the trustworthiness of an unfamiliar partner when making reciprocity decisions. We find that participants spontaneously use their prior experience about the trustingness of game partners in general to inform their reciprocity decisions, even when they had the opportunity to strategically assume that their new, unfamiliar, partners were untrusting, and hence could have justified lower reciprocation rates.在决定是否回报信任时，人们通常会受到其互动对象最初展现出的信任程度的强烈影响。若对方对你抱有高度信任，个体便会产生强烈的回报动机，且该因素往往会超过追求自身收益最大化的利己考量。然而，该领域中仍有一个尚未解决的重要问题：当此类先验信息模糊不清时，即个体无法确切知晓对方的信任程度时，人们会做出何种决策？那么人们会如何做出信任回报决策？本研究采用一种全新版本的信任博弈（Trust Game），直接针对该问题展开研究。在此研究中，我们开发并验证了一种基于计算模型的方法，用于量化并分类参与者在做出信任回报决策时，对陌生互动对象可信度的评估方式。我们发现，即便参与者有机会策略性地假设其陌生的新互动对象缺乏信任倾向，从而为自己选择更低的信任回报比例找到合理依据，他们仍会自发地基于对博弈对象总体信任倾向的先验经验，来指导自身的信任回报决策。