People perception and stereotype-based responding: task context matters.

Whether group impact social perception is a topic of renewed theoretical and empirical interest. In particular, it remains unclear when and how the composition of a group influences a core component of social cognition-stereotype-based responding. Accordingly, exploring this issue, here we investigated the extent to which different task requirements moderate the stereotype-related products of people perception. Following the presentation of same-sex groups that varied in facial typicality (i.e., high or low femininity/masculinity), participants had to report either the gender-related status of target words (i.e., a group-irrelevant gender-classification task) or whether the items were stereotypic or counter-stereotypic with respect to the preceding groups (i.e., a group-relevant stereotype-status task). Critically, facial typicality only impacted performance in the stereotype-status task. A further computational analysis (i.e., Diffusion Model) traced this effect to the combined operation of stimulus processing and response biases during decision-making. Specifically, evidence accumulation was faster when targets followed groups that were high (vs. low) in typicality and these arrays also triggered a stronger bias toward stereotypic (vs. counter-stereotypic) responses. Collectively, these findings elucidate when and how group variability influences people perception.