Esther Goldberg¹, Cambria Revsine², Wilma Bainbridge²; ¹University of California, Los Angeles, ²University of Chicago

Celebrities are known around the world because of their prevalence in various media platforms and advertisements. For that reason, it is common for celebrity faces to be used as stimuli in visual memory research since fame can represent a shared form of familiarity. However, it is possible that celebrity faces as a stimulus set include perceptual biases that may impact memory performance. For example, celebrities could fundamentally differ in the intrinsic memorability of their faces, separate from their repeated cultural exposure. In this study, we assess the memorability of famous people independent of their level of familiarity by testing memory for celebrity faces cross-culturally. We selected participants and stimuli from different cultural backgrounds; Indian participants (n = 266) viewed Brazilian celebrity and non-celebrity faces, while Brazilian participants (n = 236) viewed Indian celebrity and non-celebrity faces during a continuous recognition task. Across both cultures, celebrity faces were remembered significantly less accurately than non-celebrity faces, contradicting the intuition that celebrities have inherently more memorable faces. A separate attribute rating study revealed that, beyond celebrity status, only perceived confidence predicted memorability, where more confident-looking faces were remembered less often. Other attributes, including attractiveness and distinctiveness, were significantly different between celebrities and non-celebrities, but were not predictive of memorability. These findings suggest that celebrities and non-celebrities differ significantly in their intrinsic memorability, and that memory effects that often have been attributed solely to familiarity may instead come from underlying differences in the memorability of the stimuli. This work challenges the assumption that celebrity faces can serve as neutral “familiar” stimuli in cognitive research and emphasizes the need to control for intrinsic memorability when selecting stimuli for memory studies.