Detecting and Punishing Unconscious Bias

Abstract

We present experimental results demonstrating how ideology shapes evaluations of technology aimed at detecting unconscious biases: (1) liberals supported use of the technology to detect unconscious racism but not unconscious anti-Americanism, whereas conservatives showed the reverse pattern, (2) liberals and conservatives opposed punishing individuals for unconscious bias but supported punishing organizations failing to use the technology to root out, respectively, racism or anti-Americanism, (3) concerns about researcher bias and false accusations mediated the effects of ideology on support for the technology, and (4) participants taking strong initial stands were likelier than moderates to reconsider their positions. Our findings demonstrate that there is substantial concern about penalizing unconscious bias at the individual level and that it will be difficult to generate broad support for regulation of unconscious bias at even the organizational level unless the technology is a reliable detector of unconscious biases that lead to frequent or serious antisocial behaviors. Disciplines Management Sciences and Quantitative MethodsThis journal article is available at ScholarlyCommons: https://repository.upenn.edu/mgmt_papers/233 (1) liberals supported use of the technology to detect unconscious racism but not unconscious anti-Americanism, whereas conservatives showed the reverse pattern, (2) liberals and conservatives opposed punishing individuals for unconscious bias but supported punishing organizations failing to use the technology to root out, respectively, racism or anti-Americanism, (3) concerns about researcher bias and false accusations mediated the effects of ideology on support for the technology, and (4) participants taking strong initial stands were likelier than moderates to reconsider their positions. Our findings demonstrate that there is substantial concern about penalizing unconscious bias at the individual level and that it will be difficult to generate broad support for regulation of unconscious bias at even the organizational level unless the technology is a reliable detector of unconscious biases that lead to frequent or serious antisocial behaviors.

Red States, Blue States, and Brain States

“…this observation highlights the importance of our framing results: approval is only likely to differ based on partisanship if neuroscience is introduced in a way that excites partisan differences (e.g., helping defendants to "get off easy" for their crimes). this finding is consistent with previous research demonstrating the nuanced reaction to advances in science based on individuals' political or cultural leanings (Kahan et al 2009;Mccright et al 2013;tetlock, Mitchell, and Anastasopoulos 2013). the story is not as simple as a universal negative reaction to neuroscientific advances from the political Right.…”

“…to address this gap, this article presents the first empirical investigation of American citizens' knowledge and approval of neuroscience-based legal reforms. Building on research on public approval of other emerging technologies (Lee, Scheufele, and Lewenstein 2005;cobb 2005;Kahan et al 2009;Stewart, Dickerson, and hotchkiss 2009;tetlock, Mitchell, and Anastasopoulos 2013), our study examines the basic contours of approval for neurolaw, as well as how this approval is affected by framing and partisanship. We assess the extent to which framing neuroscientific evidence as a technological advance to help criminal defendants, versus one to help prosecutors, affects approval of neurolaw among different political groups.…”

Toward a meaningful metric of implicit prejudice.

“…Based on this distribution, researchers have argued for the high prevalence of implicit racial bias in American society, with some arguing that modal distributions of IAT scores point to a need for reforms to legal codes (Levinson & Smith, 2012), judicial practices (e.g., Bennett, 2010; Saujani, 2003), and organizational policies (Ayres, 2001). Indeed, there is the prospect that organizations may in the future seek to use IATs as screening tools on the assumption that positive IAT scores are meaningful indicators of a propensity to discriminate (Tetlock & Mitchell, 2009; Tetlock, Mitchell, & Anastasopoulos, 2013). Our findings indicate the need for greater caution and for more applied research designed to determine the true meaning of the IAT metrics that are now widely employed in efforts to gauge implicit biases.…”

Perceived moral responsibility for attitude‐based discrimination

“…Whereas, for example, religious people may attribute special harmfulness to religious discrimination, liberals may overtake conservatives in attributing harmfulness to racial discrimination. Conservatism is related to greater implicit and explicit racial bias than liberalism (Nosek et al ., ), and liberals value egalitarianism (e.g., Tetlock, Mitchell, & Anastasopoulos, ), so conservatism may negatively predict perceived harmfulness of racial discrimination. In turn, those who perceive discrimination as less harmful may attribute lower obligation and moral responsibility to a discriminator.…”