Not lie detection but stereotypes: Response priming reveals a gender bias in facial trustworthiness evaluations

Using an innovative variant of the response-priming paradigm, ten Brinke et al. (2014) seemingly found evidence for unconscious lie detection. In the present study, however, we conducted a reanalysis of this experiment showing that the data can be explained better by the assumption of a gender bias in facial trustworthiness evaluations than by the assumption of unconscious lie detection abilities. To corroborate the hypothesis that female faces are subjectively perceived as more trustworthy than male faces, we conducted two response-priming experiments: Participants (Ns = 39 and 61) had to categorize target words as truth-related or lie-related. These words were preceded by male and female face primes explicitly rated as trustworthy or untrustworthy by an independent sample. First, categorization times showed that trustworthy-appearing faces biased responses towards “truth” and untrustworthy-appearing faces biased responses towards “lie”. Second, within the group of trustworthy-appearing faces, there was a gender effect with female faces biasing responses towards “truth” and male faces biasing responses towards “lie”. These results suggest that (a) trustworthiness evaluations are biased by face gender and that (b) trustworthiness evaluations can be measured with the response-priming paradigm.