Lorio Forensics

A Critique of Current Methods of Racial Bias Measurement

This article is the second in a 3-part series regarding racial bias measurement and law enforcement. The first can be found here; the second here.  


Self-report is the earliest and most well-established method for measuring racial bias. Unfortunately, self-report measures are highly susceptible to social desirability bias resulting in individuals distorting their results to appear less prejudiced (Quillian, 2006). Due to the issues that arise from researching prejudice using predominantly self-report, such measures are most beneficial when combined with other units of analysis. 

However, implicit methods present their own challenges when trying to evaluate prejudiced responding. Some researchers have used the term racial bias to describe the differences on physiological (e.g. amygdala activity, fMRI) and behavioral measures (e.g. IAT) whereas others have argued that the term bias is inappropriate. These researchers argue that implicit measures do not directly examine racial bias. Rather, they test the strength of an association learned through cultural stereotypes (e.g. White-Good or Black-Bad). What is actually being captured through the observed neurobiological, physiological, and behavioral responses is social knowledge of negative stereotypes associated with racial and ethnic minorities.  Knowledge of these negative stereotypes may increase amygdala activation, scores on the IAT, and other physiological and sociological measures (Eberhardt, 2005). 

A shared critique of explicit and implicit racial bias measurement is that both methods focus on individual differences. As a result, many of the studies in this area describe prejudice and related phenomenon but fall short of intervening in prejudiced behaviors (Jost et al., 2014). Another shortcoming is that explicit and implicit racial bias measures require individual administration. The cheapest forms of measurement (e.g. self-report) are largely inaccurate whereas more effective physiological measurement systems are too expensive or time consuming for practical use in widespread evaluations. 

The ultimate goal of measuring racial bias should be to identify it so that it can be addressed and eliminated from our institutions, policies, and systems. With the current measurement’s focus on individual differences, racist beliefs are centered as the target issue rather than the systemic and institutional nature of racism. Identifying racial prejudice on an individual level merely describes a part of a larger structural issue. 

The limited utility of existing racial bias measures for implementation of structural change raises the question – Is quantifying bias essential to reducing it? This is not to discount the valuable knowledge that racial bias measurement has contributed in terms of understanding how expressions of prejudice have changed over time. However, these contributions have the most utility on an individual level rather than a systemic level. The intent behind racial bias assessments is that once people become aware of their implicit biases, then they can exert cognitive control to mitigate them. The IAT has some evidence of predictive validity and is often used as a projection of behavioral intent.

However, a 2017 meta-analysis that included 492 studies found that reducing implicit bias did not result in changed behavior (Forscher et al., 2019). In a recent example, 27-year-old Derrick Sanderlin worked for years at the San Jose police department offering implicit bias training. In May 2020 Sanderlin was shot with rubber bullets by officers from that same police department while attempting to de-escalate tensions between civilians and police at a protest following the murder of George Floyd. Sanderlin suffered injuries that required emergency surgery and may prevent him from having children (Koran, 2020). 

Sanderlin’s example underscores that reducing implicit bias is insufficient. Diversity, equity, and inclusion, trainings as stand-alone interventions may also be insufficient. In a second meta-analyses of nearly one thousand intervention studies aimed at prejudice reduction (e.g. workplace diversity initiatives, sensitivity training, educational programs cognitive interventions) researchers found weak evidence that any type of intervention reduced biased behavior in real-world settings (Paluck & Green, 2009).

Even if an individual’s bias was able to be reduced through measurement and intervention, those changes would be short-lived without systemic intervention. Research on intergroup bias suggests that being in a prejudiced environment makes prejudice more socially acceptable and makes people more willing to express private prejudices (Effron & Knowles, 2015). Selecting the least biased individuals for civil roles would be ineffective because those same individuals would become more biased over time by being surrounded by and a participant in a biased system.

A New Way Forward: Systemic Intervention

During heightened times such as these, there is a natural urge for individuals and organizations to leap into action. Before engaging in training, stockpiling resources, and planning organizational overhauls, we must recognize a few simple truths. Institutional racism is an issue that is hundreds of years in the making. Simple solutions and quick fixes like attitudinal surveys and employee training are insufficient for the magnitude of the problem we hope to solve. 

Our impulse for measurement is in the right spirit but headed in the wrong direction. Rather than focus on the quantification of individual bias, attention must be trained on measures that capture inequitable outcomes, interventions that address them, and re-evaluation and iterating until equity is achieved. Such efforts should naturally include community accountability and collaboration.


Eberhardt, J. L. (2005). Imaging race. American Psychologist, 60(2), 181.

Effron, D. A., & Knowles, E. D. (2015). Entitativity and intergroup bias: How belonging to a cohesive group allows people to express their prejudices. Journal of Personality and Social Psychology, 108(2), 234.

Forscher, P. S., Lai, C. K., Axt, J. R., Ebersole, C. R., Herman, M., Devine, P. G., & Nosek, B. A. (2019). A meta-analysis of procedures to change implicit measures. Journal of Personality and Social Psychology, 117(3), 522.

Jost, J. T., Nam, H. H., Amodio, D. M., & Van Bavel, J. J. (2014). Political neuroscience: The beginning of a beautiful friendship. Political Psychology, 35, 3–42.

Koran, M. (2020). An activist gave police anti-bias training. Officers still brutalized him at a protest. The Guardian.

Paluck, E. L., & Green, D. P. (2009). Prejudice reduction: What works? A review and assessment of research and practice. Annual Review of Psychology, 60, 339–367.

Quillian, L. (2006). New approaches to understanding racial prejudice and discrimination. Annu. Rev. Sociol., 32, 299–328.

Share Post:

Stay Connected

Dr. Vinson Recently Featured on NPR

Dr. Vinson was recently featured on NPR. She’s featured in a new Ken Burns documentary on youth mental health entitled ‘Hiding in Plain Sight’. Listen