Detecting Accurate Emotions in Faces

Author

Emma Greenwood, Utah State University

Date of Award

5-2022

Degree Type

Thesis

Degree Name

Departmental Honors

Department

Psychology

Abstract

Anger race bias is the tendency to misidentify expressions of emotion, specifically anger, in Black or racially ambiguous faces that are fearful or neutral (Hutchings & Haddock, 2008). Anger is often associated with aggression (Murphy et al., 2005). Therefore, the inaccurate perception of anger and threat may lead to an inappropriate response and could increase the likelihood that a police officer will shoot at a suspect (Correll et al., 2007). From 2015 to 2020, police officers shot and killed over 100 unarmed Black males (Washington Post, 2020). This study examined if anger race bias could be reduced through emotion identification training. Faces from the Chicago Face Database were used to train participants on the emotions of neutrality, fear, and anger. Participants identified emotions on a series of Black, White, LatinX, and Asian faces on pre-and post-test measures. Two weeks following the post-test, participants were invited to complete a follow-up test to determine their retention of the training. We found that the experimental group accurately identified more facial expressions on average than the control group. These preliminary results demonstrate feasibility in developing emotion recognition trainings to decrease anger race bias with the potential to be helpful in police settings.

Recommended Citation

Greenwood, Emma, "Detecting Accurate Emotions in Faces" (2022). Undergraduate Honors Capstone Projects. 926.
https://digitalcommons.usu.edu/honors/926