Some Potential Problems with Implicit Bias Interventions Qua Indirect Prejudice Interventions

Abstract Implicit bias interventions, qua attempts to decrease implicit bias measures, are often indirect prejudice interventions. As such, they face a number of potential problems. In particular, implicit bias interventions are more susceptible to generate unforeseen side-effects, are more prone to overcompensation, and less clearly decrease prejudicial behaviour, than at least some direct prejudice interventions. These particular problems are not widely recognized (even though other problems with these interventions are well known, e.g. those raised by Oswald et al. in J Personal Soc Psychol 105:171–192, 2013; Lai et al. in J Exp Psychol Gen 145(8):1001–1016, 2016; Forscher et al. in J Personal Soc Psychol 117(3):522–559, 2019). In this paper I describe the three problems in detail, and explain why they result from indirectness. I thereby hope to show that the problems beset an entire approach to deprejudicing, and point towards deeper problems than unfortunate empirical results found for particular implicit measures. The problems do not point towards finding new indirect measures and interventions to replace old ones, but away entirely from indirect measures, toward direct measures and interventions.