Previous research on the relation between oxytocin and trustworthiness evaluations has yielded inconsistent results. The current study reports an experiment using artificial faces which allows manipulating the dimension of trustworthiness without changing factors like emotions or face symmetry. We investigate whether (1) oxytocin increases the average trustworthiness evaluation of faces (level effect), and/or whether (2) oxytocin improves the discriminatory ability of trustworthiness perception so that people become more accurate in distinguishing faces that vary along a gradient of trustworthiness. In a double blind oxytocin/placebo experiment (N=106) participants conducted two judgement tasks. First they evaluated the trustworthiness of a series of pictures of artificially generated faces, neutral in the trustworthiness dimension. Next they compared neutral faces with artificially generated faces that were manipulated to vary in trustworthiness. The results indicate that oxytocin (relative to a placebo) does not affect the evaluation of trustworthiness in the first task. However, in the second task, misclassification of untrustworthy faces as trustworthy occurred significantly less in the oxytocin group. Furthermore, oxytocin improved the discriminatory ability of untrustworthy, but not trustworthy faces. We conclude that oxytocin does not increase trustworthiness judgments on average, but that it helps people to more accurately recognize an untrustworthy face.
