Statistical learning shapes face evaluation (data, code, and stimuli)

Past research has identified many configurations of facial features that predict specific trait inferences, and detailed computational models of such inferences have been recently specified. However, these configurations do not fully account for trait inferences from faces. In this work we propose a new direction in the study of inferences from faces. Any face can be positioned in a statistical distribution of faces extracted from the environment. We argue that understanding inferences from faces requires considering the statistical position of faces in this learned distribution. Four experiments show that the mere statistical position of faces imbues them with social meaning: As faces deviate from a learned central tendency, they are evaluated more negatively. Here we share data (raw .txt files, and processed .RData files), code for processing, analysis, generating figures, and the stimuli that were presented.