Decomposing Implicit Bias

Abstract

In their article, “Implicit Bias 61⁄4 Bias on Implicit Measures,” Gawronski, Ledgerwood, & Eastwick (this issue) characterize implicit bias and then discuss how the implicit association test (IAT) (Greenwald, McGhee, & Schwartz, 1998) fails to meet the requirements of a test for implicit bias. One of the central arguments is that people can predict their behavior on the IAT, indicating awareness of their implicit bias. However, part of the definition of implicit bias is that it occurs outside of awareness. In my commentary I discuss the thorny issue of awareness and suggest a more pragmatic definition of implicit bias that may help resolve the discord. I also discuss computational-modeling and process-tracing tools that allow us to decompose decisions in ways that can identify the mechanisms underlying behavioral biases. Together, hopefully these approaches will yield better insight into the nature of implicit bias.