Summary
Highlights
Implicit bias can significantly alter decisions and behavior. Evidence shows it predicts which resumes get callbacks, influences negative evaluations of ambiguous actions by African Americans, affects perceptions of confident women in hiring, shapes how we read facial expressions, and increases the likelihood of mistakenly shooting an unarmed African American in video game simulations.
A 2009 meta-analysis involving nearly 15,000 subjects found a correlation of R = 0.236 between implicit bias, as measured by the Implicit Association Test, and discriminatory behavior. This correlation was a better predictor than explicit bias measures. Another meta-analysis found an even lower R value of 0.15.
Social scientists consider an R value of 0.236 a low correlation or a small effect. However, comparing this to medical examples like smoking and lung cancer (R = 0.08) or low-level lead exposure and reduced childhood IQ (R = 0.12) shows that even small statistical correlations can represent significant real-world risks and warrant concern, especially when considering equity for all.