RESUMO
Being able to identify trustworthy strangers is a critical social skill. However, whether such impressions are accurate is debatable. Critically, the field currently lacks a quantitative summary of the evidence. To address this gap, we conducted two meta-analyses. We tested whether there is a correlation between perceived and actual trustworthiness across faces, and whether perceivers show above-chance accuracy at assessing trustworthiness. Both meta-analyses revealed significant, modest accuracy (face level, r = .14; perceiver level, r = .27). Perceiver-level effects depended on domain, with aggressiveness and sexual unfaithfulness having stronger effects than agreeableness, criminality, financial reciprocity, and honesty. We also applied research weaving to map the literature, revealing potential biases, including a preponderance of Western studies, a lack of "cross-talk" between research groups, and clarity issues. Overall, this modest accuracy is unlikely to be of practical utility. Moreover, we strongly urge the field to improve reporting standards and generalizability of the results.