摘要：Human gaze is a subtle cue to deliver information and helps impression formation in social interactions. People automatically follow the gaze direction of others and shift their attention accordingly, as well as determine the trustworthiness of others based on the predictable validity of their gaze behavior, yet it remains unclear how this works at the collective level. Therefore, the current study is the first to explore the incidental learning of trust from a group’s gaze behavior. To simulate different patterns of perceiving collective information in real life, two ways of presenting group member gazes were used in the object categorization task, the simultaneous way in Experiment 1 and the sequential way in Experiment 3, and a sampling strategy was ruled out in Experiment 2. Converging findings in experiments demonstrated a typical gaze-cueing effect, and more importantly, the Predictive-valid group obtained more trust compared to the Predictive-invalid group. To enrich and expand the applicability of the incidental trust learning effect from gazes, the current study provides supportive evidence at the collective level, confirming that humans have an efficient capability to process gaze information of groups.
