Skin-Tone Bias and Bias Self-Correction Processes
Loading...
Date
2021-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Evidence suggests that, much like racial biases, the consequences of feature-based biases (e.g.,
skin-tone and Afrocentric biases) are prevalent and detrimental. This study examines people's
ability to detect and correct for skin-tone bias within the context of hiring decisions. Participants
(N = 233) were randomly assigned to rate a lighter- or darker-skinned Black target on how well
they fit a job position. Participants also rated whether features of the target influenced their
hiring decisions and were randomly assigned to receive either a general (referring to the target's
appearance) or a specific (indirectly referring to the target's skin-tone) warning to not let features
of the target influence their judgment before rating the target again. I predicted that (1) initial
ratings would be more favorable for the lighter- vs. darker-skinned candidate, (2) warnings about
bias could lead to different changes in evaluations for lighter- vs. darker-skinned candidates,
though it was not clear whether more specific warnings about bias might be necessary to induce
them, and (3) measures of theories of bias (i.e., perceived or expected biases) would predict how
people attempted to correct for possible biases of target features. Results suggest some instances
of skin-tone bias, such as participants initial ratings showing a greater likelihood of offering the
job to the lighter-skinned participant. However, this tendency was not consistent across different
hiring-related ratings. Also, mostly inconsistent with expectations, participants' ratings shifted
positively after correction instruction, regardless of skin-tone or type of bias warning. This study
may have implications for the limitations of bias correction research regarding skin-tone bias to
be considered in future studies.
Description
Keywords
colorism, skin-tone bias, bias correction, attitudes, flexible correction model