Delight in Disorder: Inclusively Defining and Operationalizing Implicit Bias

Gawronski, Ledgerwood, and Eastwick (this issue) address a timely issue of both theoretical and practical importance in the burgeoning study of implicit bias. The authors “highlight conceptual and empirical problems with the widespread equation of implicit bias and bias on implicit measures” (p. 139). They are not the first to raise and grapple with a question closely related to deciphering the conceptual meaning of implicit bias and its relationship to measures of implicit bias, but they distinguish themselves with their mastery of diverse literatures, sophisticated analyses of core theoretical issues, and original insights. While maintaining a steady focus on their core question, the authors’ review and synthesis of the work that they cover makes this a valuable resource for various audiences. It provides a detailed, yet accessible introduction for those who are interested in but relatively unfamiliar with the topic, as well as a thought-provoking and well-argued contribution for those who have considerable expertise in the area and may already have well-formed perspectives on the questions posed and answers provided. Importantly, in an area in which heated debate has been common, Gawronski et al. navigate through complex issues with logic and data in an even-handed way. This is an impressive piece of scholarship. A common colloquial expression is, “If the shoe fits, wear it.” The article is particularly impressive in the way the authors examine the many ways that scholars have attempted to define implicit bias. They try on many shoes for the term “implicit,” as compared to “explicit.” Gawronski et al. (this issue) consider distinctions in process, such as in differences between “mental levels.” For instance, they discuss how implicit has been treated as reflecting associative processes “involving unqualified mental links between concepts”, whereas explicit processes are propositional “involving the perceived validity of specific relations” (p. 141). Alternative, procedural distinctions are also reviewed. These tend to be instrument-focused. For example, a measure would qualify as implicit to the extent to which the response is automatic—that is, unintentional and difficult to control. By contrast, an explicit measure would be one in which people respond in a deliberative, intentional, and selfreflective way. Indeed, the first author of this commentary falls into this procedural camp, describing implicit as activation that occurs unintentionally (Dovidio, Kawakami, & Beach, 2001), automatically (Dovidio, Hewstone, Glick, & Esses, 2010), and which can operate without people being aware of the “biased associations or of the role those associations play in guiding their judgment and action” (Greenwald, Dsagupta, et al., 2022, p. 8). However, Gawronski et al. (this issue) skillfully argue how and why none of these shoes fit. In the end, we resonate with Gawronski et al.’s critical conclusion that “despite 25 years of extensive research, the current labeling conventions are still based on conceptually ambiguous lists according to which a measure qualifies as implicit if researchers have described it as implicit in the past” (p. 142; see also Gawronski, De Houwer, & Sherman, 2020). While we agree with Gawronski, Ledgerwood, and Eastwick’s analysis of current problems in the way implicit bias is conceived and studied, where we diverge is in the proposed solution. We deliberately use the word “diverge” rather than “disagree,” because the perspective that brings us here is quite different. Different perspectives carry different assumptions and dictate different priorities. Gawronski et al. (this issue) focus on implicit bias as a behavioral phenomenon that can be distinguished from the bias that is assessed by implicit measures. They write that “bias can be defined as the effect of social category cues (e.g., cues related to race, gender, etc.) on behavioral responses” and “to classify a person’s behavioral response toward a target as an instance of IB [implicit bias], one has to demonstrate that (1) the behavioral response is influenced by social category cues and (2) the person is unaware of the effect of the relevant social category cues on their behavioral response” (p. 5). As clear and as elegant in its directness that this set of definitions is, we are not convinced that this shoe fits, either. At first glance, Gawronski et al.’s (this issue) definition may seem quite similar to the definition recently used by Greenwald, Dsagupta, et al. (2022, p. 8) as a bias that