Abstract
In a brief glance, observers are able to evaluate gist characteristics from crowds of faces, such as the average emotional tenor or the average family resemblance. Prior research suggests that such high-level ensemble percepts rely on configural and viewpoint-invariant information. However, it is also possible that feature-based analysis could be sufficient to yield successful ensemble percepts in many situations. To confirm that ensemble percepts can be extracted holistically, we asked observers to report the average emotional valence in crowds of Mooney faces. Mooney faces are two-tone shadow-defined images that cannot be recognized in a part- or feature-based manner. To recognize a part or feature of a Mooney face, one must first recognize the image as a face by processing it holistically. In Experiment 1, we asked participants to report the average emotional valence of 6-face crowds viewed for 1 second. Participants successfully extracted the average emotional tenor of the crowd by effectively integrating 5 out of the 6 faces presented. In Experiment 2, we asked participants to report the average emotional valence of Mooney face crowds presented in a rapid sequence (10 Hz, 50% duty cycle). Participants were able to successfully report the average emotional valence of the sequentially-presented crowds, integrating up to 6 faces. In Experiment 3, we confirmed holistic processing by interleaving upright and inverted displays of Mooney face crowds. As expected, participants’ ensemble percepts of emotional valence were negatively impacted by inversion. Taken together, these experiments are strong evidence that ensemble perception can operate selectively on holistic representations of faces.