15.1 Prejudice, Discrimination, and Stereotyping

The label “implicit bias” can have both positive and negative implications. Although the implicit bias label can be useful for explaining why our conscious attitudes may not align with our behaviors, the label can also lead people to shift focus toward unconscious bias and overlook overt, explicit prejudice and discrimination. When discrimination happens, people may argue, “I don’t discriminate; I just have implicit bias” as an excuse. It also leads people to believe that they do not need to take any action, as everyone has implicit biases.

Focusing too heavily on implicit bias can be an issue for some organizations because implicit bias is increasingly used to explain challenges that organizations face regarding diversity and inclusion. Thus, implicit bias training is used as a primary solution to address racial disparities in many organizations. However, implicit bias training has not been very successful. Lai et al. (2014) found that only eight out of 17 implicit bias interventions significantly reduced bias. Even among those that are effective, the effect fades away within 24 hours (Lai et al., 2016), and it is not associated with explicit attitudes or changes in behaviors (Forscher et al., 2019). Implicit bias training could increase the awareness of bias, but it does not seem to reduce the bias itself in the long term (Forscher et al., 2017). Thus, simply educating people about implicit bias may not be effective. In addition, inducing the guilt or fear commonly associated with implicit bias may backfire and lead people to be defensive (Lai et al., 2016). Instead of focusing on implicit bias to check a box, organizations need to focus more on the actual behavior changes. Lai et al. (2016) suggested that interventions may need to be repeated and more intensive to reduce biases. In addition, implicit bias interventions could be more effective for children than adults, so it may be better to apply the interventions earlier in development.

Although implicit bias interventions do not effectively reduce prejudice, there are other ways to reduce prejudice effectively. For example, altering intergroup boundaries using decategorization or recategorization has been shown to be effective on an intrapersonal level. Decategorization involves reducing group boundaries by considering outgroup members as individuals instead of as members of a group. For example, masking the gender of musicians can be effective in reducing gender discrimination in symphony orchestra hiring. Recategorization involves drawing outgroup members into the ingroup. For example, placing both ingroup and outgroup members into a situation that requires them to coordinate to achieve a common goal can be an effective way to reduce bias.

On an interpersonal level, using contact theory and fostering positive intergroup interactions can reduce bias. For example, Aronson and his colleagues (Aronson et al., 1978) developed a jigsaw classroom technique. The jigsaw technique creates a collaborative environment by having children of different racial backgrounds work together on standard course material. Each child is responsible for a subset of the learning material that they are required to teach to the other students in the group.

On a systemic level, one way to reduce prejudice is by using media interventions. For example, when students were exposed to posters that suggested diversity was encouraged on campus, they also had more inclusive behaviors than students who were not exposed to such posters (Murrar et al., 2020). In addition, reading books that feature multicultural characteristics, children with disabilities, and immigrant children can also reduce prejudice among children (e.g., Cameron & Rutland, 2006).

Video 15.1. “The problem with using The Implicit Association Test at your business” by CounterweightMedia.

Implicit bias measurements assess people’s automatically activated attitudes, and they have contributed greatly to our understanding of the link between attitude and behavior. The use of implicit bias measurements can also be useful in evaluating attitudes that can be influenced by social desirability. For example, some of the earliest implicit bias measurements were designed to measure racial attitudes. Fazio et al. (1995) found that those who had a strong motivation for controlling their prejudice rated a different ethnicity group positively using explicit measurements but rated the same group negatively using implicit measurements. However, the strengths of implicit bias measurements are not limited to addressing the issue of social desirability. Automatically activated attitudes can be referred to as “preconscious” because they often exist at a level below people’s conscious awareness. As a result, these attitudes are difficult to measure using self-report measurements, but they can be measured using implicit bias measurements (Sun et al., 2008).

One popular implicit bias measurement is the Implicit Association Test (IAT). The IAT directly measures how strongly two concepts are associated by measuring reaction time and accuracy of sorting stimuli (Nosek et al., 2007). It is based on the idea that when two concepts are strongly associated in one’s mind, the sorting tasks will be facilitated. The IAT test has demonstrated a sufficiently high test-retest reliability (e.g., Fazio & Olson, 2003). The IAT test also has good concurrent validity (Fazio & Olson, 2003).

The IAT also has some limitations. One criticism is that the IAT measures category-level associations rather than automatically activated attitudes toward specific stimuli. Fazio and Olson (2003) argued that the IAT measures the strengths between a category label (e.g., Black) and an evaluation (positive or negative) instead of automatically activated attitudes. This means that the IAT test may not measure the actual negative attitudes. In addition, the IAT test is susceptible to “extrapersonal association.” For example, Karpinski and Hilton (2001) found that people preferred apples over candy bars on an IAT test but not when they were tested using explicit measurements. Their responses likely reflected societal values that label apples as “good.” It is suggested that societal stereotypes can influence the results of the IAT, even if people do not endorse the bias (Goodall, 2011).

Figure 15.1. Implicit Association Test (IAT) sample screen stimuli [Image Description]

Image Descriptions

Figure 15.1 Image Description. The image is divided into four quadrants, labeled a, b, c, and d, each with a black background. Quadrant a: Features a grayscale photo of a person’s face centered between the headings “Black Patient” and “White Patient” at the top. Below the image, the words “or Bad” are under “Black Patient” and “or Good” under “White Patient.” Quadrant b: Displays only text with “Black Patient” and “White Patient” at the top. Below, aligned under “Black Patient” is “or Bad,” and under “White Patient” is “Pleasure.” Quadrant c: Similar to Quadrant a, it features the same grayscale photo of a person’s face between the same patient labels. Below, “or Bad” is under “White Patient” and “or Good” under “Black Patient.” Quadrant d: Contains text similar to Quadrant b, with the headings “White Patient” and “Black Patient” at the top. Below, “or Bad” is under “White Patient,” and “Pleasure” is below “Black Patient.” [Return to Figure 15.1]