Racial bias in face perception is sensitive to instructions but not introspection

Abstract

Faces with typically African features are perceived as darker than they really are. We investigated how early in processing the bias emerges, whether participants are aware of it, and whether it can be altered by explicit instructions. We presented pairs of faces sequentially, manipulated the luminance and morphological features of each, and asked participants which was lighter, and how confident they were in their responses. In Experiment 1, pre-response mouse cursor trajectories showed that morphology affected motor output just as early as luminance did. Furthermore, participants were not slower to respond or less confident when morphological cues drove them to give a response that conflicted with the actual luminance of the faces. However, Experiment 2 showed that participants could be instructed to reduce their reliance on morphology, even at early stages of processing.Note: All stimuli used, code to run the experiments reported, raw data, and analyses scripts and their outputs can be found at https://osf.io/brssn/?view_only=f458374cf57d4d54b231f9cac252ce95, to be made public on publication.

An Exploration of Relationships Between Perceptual and Cognitive Racial Biases

“…After these initial questions, participants were presented with two separate tasks: the Implicit Association Test and Race-Lightness Task (whose order in presentation was randomly assigned to participants). The RLT was a replication of a study conducted by Travers, Fairhurst, and DeRoy (2020) on racial bias in facial perception, which investigated the factors that contribute to the perception of African features being perceived as darker than they actually are (Travers et al, 2020). The second task presented was a replication of the Harvard Project's Implicit Association Test (IAT).…”

“…They would then be asked two questions: first, if the target face was "darker" or "lighter" than the first face; then, how confident they were that their answer was correct, marked by moving a slider along a scale of zero to 100 (the exact numbers were not shown) with the lower end labeled "not certain at all" and the higher end labeled "totally certain." The wording of the instructions and questions were very similar to that used by Fairhurst, Travers, and DeRoy (2020). After reading the instructions, participants were given 20 practice trials and told how accurate their responses were.…”

“…The faces presented to participants were taken directly from Travers, Fairhurst, and DeRoy (2020). There were 72 images of faces, but each one originated from one of eight unique faces that were computer generated using FaceGen (Singular Inversions, 2016).…”

“…The Race-Lightness Effect (Travers et al, 2020) is used to assess perceptual racial biases. This task focuses on evaluating the relationship between a face's morphological features (e.g., how "Africanized" a face is) and a subject's perceived lightness or darkness of that particular grayscale face.…”

Conclusion: Revising Intersectionality

“…5 It seems that instruction can impact this relationship to some extent. For example, telling people that their reactions are biased by their concept of race influences the way they categorise skin shades (Travers et al 2020). It means that through instruction regarding concepts, people can partly adapt their reliance on visual evidence in categorising similarity and difference.…”