Race, Bias, and Risk Assessment

By David S. Prescott, LICSW, & Kieran McCartan, Ph.D.

Note: We are grateful to Tyffani Dent for her contributions to the discussions that led to this post.

On August 1, 2014, in a speech about risk assessment processes, then–United States Attorney General Eric Holder said of the available measures:

Although these measures were crafted with the best of intentions, I am concerned that they may inadvertently undermine our efforts to ensure individualized and equal justice.  By basing sentencing decisions on static factors and immutable characteristics – like the defendant’s education level, socioeconomic background, or neighborhood – they may exacerbate unwarranted and unjust disparities that are already far too common in our criminal justice system and in our society.

His points were well-taken and yet not without considerable push back. The most common response at the time was that the existing tools certainly outperformed the unstructured judgment that in turn was wildly susceptible to bias. This point, too, was a good one. Earlier this Autumn, Jennifer Skeem, associate dean of research and associate professor of social welfare and public policy at the University of California, Berkeley, stated in a major address to the National Forum on Criminal Justice that extensive review of post-conviction risk assessments of federal convicts found “no evidence of predictive bias by race.”

Perhaps that’s one of the issues we need to address first; many aspects of racism take place beyond the awareness of those who work within the structures where racism is found.

As outside observers who have tried to watch developments in risk assessment closely, there is no question that the right risk assessment methods can be useful, but we question whether there isn’t evidence of inherent bias available right in plain view. For example:

-  It’s well-established that people of color are more likely to be arrested, often as a result of over-policing. They are incarcerated at a rate of more than five times that of white people. 

-  It’s also established that people of color are less likely to be referred to diversion programs, and can be subject to bias even within that referral process. All of these points can result in higher scores on risk assessment instruments compared with whites, especially against a backdrop of true crime rates remaining unknown. 

-  Likewise, racial disparities can be found in the bail system. This fact often goes missing in broader discussions of racial disparities in the legal system.

-  Obviously, not all risk assessment methods are created equally; many rely on items that lend themselves, more and less, to racial bias. Items related to family (for example, past family incarceration) and community stability scored outside of an understanding of their context may not accurately reflect a person’s propensity to commit crime.

In some circumstances, further questions arise as to whether many instruments aren’t more effective at predicting who will be arrested than predicting who will commit crimes.

People of color tend to experience intersectionality more than white counterparts, which means that their different socio-political and individuals labels put them at risk of being a victim of crime and of, potentially, being someone who could commit a crime. In addition, people of colour, sometimes because of factors crystalized through intersectionality, are more likely to experience trauma as well as adverse experiences; which matters in how we work with and respond to them. This means that there is an opportunity for better primary, secondary, tertiary & quaternary prevention (see previous blog). However, because of the socio-political aspects of race, vulnerability, trauma, economics, and access to social care in America (and worryingly so in the UK as well) issues related to intersectionality, race and crime never gets truly understood or dealt with.  A clear example of this is a recent report that indicates that UN peacekeepers from multiple countries, of multiple races committed systematic sexual abuse while in Haiti. Experiences of sexual abuse, whether through victimisation or perpetration, does not have a race determinate; but race does play a significant role, though intersectionality and socio-economic-political factors, in the way that we define, prevent and respond to sexual abuse.

Of course, we are not the first, by far, to address this and related topics. We do, however, believe that professionals can become more effective by studying the myriad issues involved that this blog post is only barely able to touch upon. Despite the excellent advances made by our risk assessment instruments, very serious challenges remain.