Diverse FACES – A database of facial expressions in young, middle-aged, and older individuals from underrepresented racial populations

This dataset complements the 2010 FACES database created by Natalie Ebner and her colleagues at The Max Planck Institute [1] . This collection contains 432 validated expressions from 36 Black/African American and Hispanic/Latinx individuals expressing 6 emotions: Happiness, Sadness, Anger, Disgust, Fear, and Neutral.This database's goal is to make facial expression stimuli represent the racial and ethnic diversity of the U.S. population. This dataset could combat what other researchers have called the own-race bias and lead to more accurate results for non-White research participants.To create the photos, we replicated the FACES protocol, recruiting Black/African American and Hispanic/Latinx community members to pose as face models (models) in the lab’s amateur photo studio. We selected the two images displaying the best representations of each emotion and randomly sorted them into A and B groups. These images were then placed into a Qualtrics survey for online validation.Validation participants (raters) were pooled from online survey panels and selected based on age (25-85), race (Hispanic/Latine, Black/African American, White/Caucasian), and gender (male/female). Raters were randomly placed into Group A or B and rated 36 images according to five criteria: emotion, age, race, representativeness of race, and trustworthiness. For each criterion, these ratings were averaged for each image.