Trust learning in the repeated trust game: A meta-analytic study

Abstract

Trust involves making oneself vulnerable by relying on the expectation that others will reciprocate and act in a trustworthy manner, leading to mutual benefit. In behavioural economics and psychology, the Trust Game (TG) is a widely used paradigm to measure trust. The repeated TG is a modified version of the TG in which participants encounter the same partner(s) multiple times, allowing for reputation and trust learning. The aim of the present meta-analysis was to identify features of the repeated TG, participant characteristics, and manipulations of partner trustworthiness that affect trust learning. This is the first meta-analytic study to specifically assess trust learning in the repeated TG and included 404 effect sizes from over 8000 participants from 68 studies. Our findings indicate that the partners' behavioural trustworthiness, in the form of their reciprocation rate, is by far the most influential factor in participant trust learning (β = 3.0). Furthermore, the results reveal that manipulating prior information about partners can have an effect on the amount of learning, but only for manipulations of trustworthiness/morality. Notably, in ingroup–outgroup studies, participants learn from their partners' trustworthiness and it is not affected by their partners' group membership.