“…People have the visual mechanism to extract summary statistical information from visual scenes, including emotions, diversity, and hierarchy from faces (e.g., Haberman & Whitney, 2009;Phillips, Slepian, & Hughes, 2018). For gender categorization, people can rapidly perceive the sex ratio of a mixed-sex display, and this ratio further affects judgments of threat (Alt, Goodale, Lick, & Johnson, 2017) and social attitudes (Goodale, Alt, Lick, & Johnson, 2018), as well as perceiver's sense of belonging (Goodale et al, 2018). For race categorization, people can perceive the average race (e.g., Jung, Bülthoff, & Armann, 2017) and estimate the majority race (Thornton, Srismith, Oxner, & Hayward, 2019) from arrays of faces, and perceive difference in average emotions between two racial subgroups in a mixed-race display (Lamer, Sweeny, Dyer, & Weisbuch, 2018), implying that they encode the racial identity of the constituent faces.…”