Fear acquisition and liking of out-group and in-group members: Learning bias or attention?

•We report results from human differential fear conditioning Out+, Out−, In+, In− with in-group and out-group faces.•Our measures includes physiological variables (skin conductance, pupil dilation), explicit evaluative ratings and a test for implicit liking (APT).•Across all measures our results reveal a better discrimination of out-group faces where the Out+ was an accentuated signal for harm but the Out− face also was an effective signal for safety from harm.•We provide computer simulation of the Rescorla-Wagner theory to illustrate that the pattern of results is not in accord with the hypothesis that out-group members in general are more readily associated with an aversive events.•We propose an alternative attentional explanation that rests on the assumption of a positive in-group bias.