The “Implicit Bias” Wording Is a Relic. Let’s Move On and Study Unconscious Social Categorization Effects

In their article, Gawronski, Ledgerwood, and Eastwick (this issue; hereafter, GLE) explain why “implicit bias” (defined as the unconscious effect of social category cues on behavioral responses) should not be confused with “bias on implicit measures.” We see much value in their clarification and agree with their bleak assessment of research on implicit tasks when they are said to measure “implicit bias” (hereafter “implicit measures of bias”), the most prominent of which is the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998). The article opens with a puzzling statement, though. GLE celebrate the educational value of “implicit measures of bias”: “implicit measures of bias deserve enormous credit for providing a tool for the widespread dissemination of the idea that people can be biased without being aware of it” (Gawronski et al., this issue, p. 139). However, while reading their article, it becomes quickly clear (1) that “implicit measures of bias” have little conceptual consistency, and (2) that critical assumptions underlying their use and interpretation are unsubstantiated (e.g., the assumption that these tasks tap into unconscious mental contents or hold a special relation to associative learning). GLE also note that social cognition research has barely started to study the unconsciousness of category-driven biases beyond responses entered on computer keyboards. It is an open secret that we do not clearly know how to interpret outcomes from “implicit measures of bias” (see, e.g., Fiedler, Messner & Bluemke, 2006). The managers of Project Implicit, the largest educational and researchoriented platform conventionally said to study “implicit biases” feature an honest disclaimer on the website of the platform: the designers of the task, their promoters, and their associated institutions “make no claim for the validity” of their suggested interpretations of IAT scores (https:// implicit.harvard.edu/implicit/takeatest.html). If we want to be honest about it, we do not know much which and when social behaviors are driven by an unconscious influence of social categories either. If social cognition research relied on tasks and study settings that are detached from “implicit biases” (as GLE define them), then this begs the question of how accurate and profitable the education around this notion has been. As a case in point, introductory psychology textbooks generally fail to accurately portray the most prominent “implicit measure of bias” (Bartels & Schoenrade, 2022). We suspect that extraacademic education does not fare better. In the present commentary, we speculate on how we got here, we discuss how bad it can get when scientists conflate science with mere opinions, and we propose ways forward. We argue that strong research on “implicit bias” can finally see the light if drastic changes are implemented in social cognition research, starting with radical terminological changes.