Implicit bias is usually measured with the “Implicit Association Test” (https://implicit.harvard.edu). In this test, people make speeded responses to a set of stimuli over repeated trials presented on a computer screen. The test measures the extent to which people implicitly associate certain groups with certain constructs. Large-sample studies suggest that most Americans indeed have implicit race biases, i.e., they implicitly associate ethnic minorities with negative terms and Whites with positive terms. Other measures of people’s implicit associations are the “Affect Misattribution Procedure” and the “Evaluative Priming Task.”

The Implicit Association Test has been shown to have a rather low “test-retest reliability.” The same individual might have a score indicating high implicit bias at the beginning week but a score indicating no implicit bias at the end of the week. Such unreliability makes the test unsuitable for measuring individual differences related to prejudice. People’s Implicit Association Test scores are heavily influenced by information that they have been exposed to recently, primarily in the hour prior to taking the test.

Implicit bias, as measured with the Implicit Association Test, is uncorrelated with behavior. In recent meta-analyses, Oswald et al. (2013; 2015) and Carlsson and Agerström (2016) found no relation, but Kurdi et al. (2019) found a correlation of r = .14. Gawronski (2019) discusses why it doesn’t make sense to expect that people’s scores on implicit association measures should be correlated with all intergroup behaviors.

Furthermore, correlation does not imply causality. There is not a single study showing that people’s implicit associations have a causal impact on behavior. In a meta-analysis involving 492 studies and 87,418 participants, Forscher et al. (2019) showed that procedures to change implicit bias have no effect on behavior. There is also not a single study showing that a training-induced shift in implicit bias leads to a subsequent shift in behavior (Lai et al., 2013; Kulik & Roberson, 2008). In addition, procedures designed to change people’s implicit associations have effects that do not last longer than a day (Lai et al., 2016).

Raising awareness about implicit biases has negative side effects. References made to implicit (versus explicit) bias make inequality between groups seem more inevitable (Pietri, Hennes, Dovidio, Brescoll, Bailey, Moss-Racusin, & Handelsman, 2018). Attributing discrimination to implicit bias causes people to believe perpetrators of discrimination as less responsible and less worthy of punishment (Cameron, Payne, & Knobe, 2010; Daumeyer Onyeador, Brown, & Richeson, 2018). Completing the Implicit Association Test reduces positive intergroup interaction behavior, possibly because raising awareness about implicit bias enhances caution and inhibition, reduces self-efficacy, or primes categorical thinking (Vorauer, 2012).

In most settings, there are no data about the number of individuals who treat/judge members of marginalized groups more negatively than members of non-marginalized groups. Although members of marginalized groups are undeniably the target of acts of discrimination and exclusion far too frequently (see here), it is unclear whether these acts are committed by a numerical majority or a numerical minority of individuals. In most settings, there is no empirical evidence for claims such as “Most of us engage in microaggressions” and “Because of implicit bias, most employees in this company engage in discriminatory behavior.”

Recent evidence has shown that people’s Implicit Association Test scores are correlated with behavior at the aggregate level. For example, counties in which the average resident has a relatively high Implicit Association Test score are also the counties with the most police shootings of ethnic minorities. It appears, then, that implicit bias should be thought of as an effect rather than a cause. Based on their explicit attitudes, people self-select themselves into different social/geographical environments (Motyl, Prims & Iver, 2019) and expose themselves to different media. As a result, they are more likely to be exposed to certain types of information in the hour prior to taking the Implicit Association Test, which then influences their score on the test.

Many journalists have written about the doubtful utility of implicit bias as a means to reduce discrimination and promote inclusion. Here are some examples:

• Goldhill (2017). The world is relying on a flawed psychological test to fight racism (https://qz.com/1144504/the-world-is-relying-on-a-flawed-psychological-test-to-fight-racism/)
• Jussim (2017). Mandatory Implicit Bias Training is a bad idea (https://www.psychologytoday.com/us/blog/rabble-rouser/201712/mandatory-implicit-bias-training-is-bad-idea)
• Singal (2017). Psychology’s favorite tool for measuring racism isn’t up to the job: Almost two decades after its introduction, the implicit association test has failed to deliver on its lofty promises (http://nymag.com/scienceofus/2017/01/psychologys-racism-measuring-tool-isnt-up-to-the-job.html)
• Bartlett (2017). Can we really measure implicit bias? Maybe not. (http://www.chronicle.com/article/Can-We-Really-Measure-Implicit/238807)

 

Return to “How to Promote Inclusion in 750 Words” here