November 25, 2013

Address Bias at its Roots

By Dr. Ken Broda-Bahm: 


Since 1998, more than 4.5 million people have discovered they have more of a racial bias than they thought they had. They did this by self-administering an online tool called the Implicit Association Test. Developed by researchers at Yale University and the University of Washington, the test looks at implicit bias, or the ways unconscious attitudes lead to unjustified assumptions and perceptions about individuals of other races. While explicit racial bias is generally seen as wrong, these implicit preferences, measured by reaction times when looking at faces, are more deeply rooted. In addition to bias based on race, the researchers have also applied the same tool to document implicit preferences based on gender, ethnicity, sexual orientation, age, religion, disability, body weight, and other factors. Even without rising to the level of conscious awareness, these implicit biases still have the ability to Influence judgment and behavior in a variety of contexts, including determinations of criminal guilt and sentencing. 

A recent article in Court Review (Elek & Hannaford-Agor, 2013) looks at the state of the research on implicit racial bias. Written by two researchers with the National Center for State Courts, the article reviews the state of the research and also surveys court efforts so far to address it. One wrinkle is that sensitizing individuals by raising awareness of the bias can help reduce it, or it can bring on a defensive backlash that actually increases the bias. That is obviously a concern in the context of racial bias in the criminal justice system. My focus in this blog post, however, will be to look at what this discussion has to say about not just racial bias, but about biases generally. 

The Lessons of Racial Bias (And What They Say About All Bias)

There are four factors in the National Center for State Courts’ analysis of implicit racial bias that I believe carry an important message regarded bias generally. These core facts should inform litigators’ task of addressing bias in voir dire and in legal persuasion. 

1. Bias Affects Outcomes

A central conclusion is that bias matters. Elek and Hannaford-Agor point to recent research looking at 122 studies supporting the conclusion that implict bias is a useful and independent predictor of behavior and judgment (Greenwald et al., 2009). These behaviors and judgments extend into the courtroom as well. As Elek and Hannaford-Agor note, “Contemporary researchers have become more convinced that much of the disparity in legal outcomes for African-Americans compared to whites is likely due to implicit bias”  

If widespread cultural recognition of the unacceptability of racism has not erased implicit bias, then the chances are good that the broader family of other biases brought into the legal process aren’t going to be erased simply through a juror’s efforts to be rational and to follow the evidence. At the same time, there is no justification for going to the other end of the spectrum by concluding that bias is the only thing that affects outcomes. Even in studies on racial bias, the research still shows that there isn’t a simple and deterministic relationship. Even as bias puts a thumb on the scale, the NCSC review shows that “strength of evidence is generally the overwhelming predictor.”  

2. Bias Doesn’t Require Conscious Awareness

One premise of most voir dire, particularly when conducted by the court, is that biases are subject to our own awareness and control: Those who have a bias will know and admit to it, and once they promise to overcome that bias, they’ll do it. Of course, a basic knowledge of social science will tell you these assumptions are unfounded. One of the central contributions of the research on implicit bias is the conclusion that even people who do not want to be biased, and indeed believe that they’re not, will still carry implicit associations that influence how they evaluate others. The legal process appropriately encourages evaluators to base their decisions on reasons and not biases. But when the bias motivates a search for reasons, then we tend to believe that our ultimate evaluations are based on the reasons and not on the motivated search. As Elek and Hannaford-Agor note, “It is difficult for decision makers to realize when their decisions are influenced by race, ethnicity, gender, or other extraneous factors if other selective information can be used to support their decision.” 

3. But Sensitizing Helps

So what does the legal process, relying as it does on conscious awareness and control, do with unconscious and uncontrollable bias? This is where Elek and Hanneford-Agor point to a somewhat tricky path forward for the justice system. They point out that lawyers and judges are looking for ways to address implicit bias through instructions, juror orientation, and voir dire. In most cases, the intervention relies on increased awareness. “Personal awareness of one’s own potential for any type of cognitive bias,” the authors note, “is necessary before an individual is capable of engaging in efforts to correct for it.” 

But the kind of awareness matters. In a racial bias context, for example, the research shows that raising awareness with a message focused on an appreciation of diversity can at least temporarily help to reduce racial bias, while a message focused on color blindness can have the opposite effect of making participants even less aware of their biases. For that reason, the NCSC researchers recommend that any debiasing strategy be empirically tested before it is used. Approaches that might seem intuitive can actually fail. 

4. As Long as Sensitizing Doesn’t Lead to Backlash

Another reason to test beforehand is to guard against a potential backlash. “Some otherwise well-intentioned approaches,” Elek and Hannaford-Agor warn, “can provoke a backlash effect in which the individuals exposed to the intervention are actually more likely to make judgments or behave in ways that manifest prejudice.” This response may be due to participants feeling threatened or angry due to perceived external pressure to comply.

In a legal context, it is easy to imagine that any attempt to sensitize actual or potential jurors to implicit racial bias could lead to the perception by some that the court is engaged in special pleading, showing favoritism toward racial minorities, and that could elicit an even more punitive evaluation. Biased individuals could respond by expressing prejudice as a form of personal independence. Here again, the type of instruction matters. If it is seen as informative and useful, it can reduce bias. But if it is seen as forced compliance, then a rebound effect is more likely.  

One factor that also helps according to the research is a more diverse jury. In addition, many other jury practices — taking notes, providing written questions to witnesses, being able to engage in predeliberation discussion, receiving copies of jury instructions — should help to keep jurors focused on the appropriate criteria for decision. According to the research, that should reduce the potential breeding grounds for bias.

Clear instructions should be part of the solution as well. Thankfully, the National Center for State Courts is putting its money and its research where its mouth is. Before the end of this year, a new study should be out evaluating the effectiveness of different forms of jury instructions in guarding against implicit racial bias. These results should be interesting from the perspective of not only guarding against bias based on race, but in addressing some of the common factors in all bias. 


Other Posts on Bias: 


Elek, J. & Hannaford-Agor, P. (2013). First, Do No Harm: On Addressing the Problem of Implicit Bias in Juror Decision Making. Court Review 49. URL:

Image credit:  Harald Hoyer, Flickr Creative Commons

Related Posts Plugin for WordPress, Blogger...