RESEARCH
It is well established that our various child protective systems tend to disadvantage children and adolescents of color. There is a consensus that much of this stems from, among other things such as racism and implicit biases, differences in both decision-making and the effectiveness of the agencies and people that make up these systems.
Much has been written on the biases that occur within risk assessment measures and methodologies. The reliance on prior criminal history as a predictive factor in risk assessment tools does not adequately consider the fact that some communities are more policed than others, increasing a youth’s chances of arrest and re-arrest. As a result, the static risk factors in any assessment place young people of color at a disadvantage when decisions are made regarding supervision, management, sentencing, etc.
The authors of this study (Kristina Childs, Jennifer Peck and Caitlin Brady) start from a place of recognizing that youth carry multiple identities. Rather than just look at one factor, they set out to examine the intersectionality of race, ethnicity and gender.
Focusing on one risk assessment tool, the Positive Achievement Change Tool (PACT), Childs and her colleagues studied 20,617 adolescents referred to the juvenile justice system between 2007 and 2014 where the primary disposition was probation. Results of the study showed small but statistically significant differences across risk categories. Black boys were least likely to be rated as low-risk; and Hispanic girls were most likely to be rated as low-risk. Black boys had the highest rate of re-referral while Hispanic girls had the lowest rate. Overall, Black and Hispanic boys had the highest base rates for re-arrest and also the highest inaccurate classification (e.g., were scored at low risk and yet were rearrested).
The authors also discuss the difficulty in making conclusions about test bias when the base rates of recidivism are not equal across the groups studied.
IMPLICATIONS FOR PROFESSIONALS
There is an old saying that “intelligence is not found in knowing the answers, but in asking the right questions.” This study helps professionals do exactly that. When conducting assessments and making supervision decisions, professionals can be better placed to question the measures they use and the best use of the findings. After all, this study shows that there is an entire narrative underlying the question of prior history of involvement with the legal system. It may involve the over-policing of neighborhoods, the disenfranchisement of communities, under-assessed strengths and protective factors, a lack of access to diversion programs, or any other number of circumstances. It is encouraging that so much recent research has examined the processes by which biases get embedded in personal and system-wide decision-making.
IMPLICATIONS FOR THE FIELD
As a field, we need to accept the fact that we are not getting risk assessment and classification right yet. The nuances of this and similar studies illustrate that while some forms of racial bias can be quite overt indeed, others are far more implicit. Well-intentioned and effective professionals often work in systems in which biased decision-making create harms for children, adolescents, and their families at an institutional level. The good news in all of this is that we are far better positioned to effect changes at both the personal and policy levels thanks to the increasing consensus in the literature about bias.
CITATION:
Childs, K. K., Peck, J. H., & Brady, C. M. (2022). Predictive bias in juvenile risk assessment: Considering race/ethnicity and sex. Crime & Delinquency, 00111287221143936.
ABSTRACT:
This study examines the joint effects of race/ethnicity and sex on the classification quality of the Positive Achievement Change Tool (PACT). We assess predictive parity, calibration, and error rate balance using a statewide sample of youth who completed probation from 2007 to 2014. Similar rates of predictive parity were observed across groups. Calibration was observed consistently across groups, but only when low risk youth were used as the reference category. Black and Hispanic boys revealed the highest base rates of re-arrest and inaccurate classification (i.e., scored low risk but were rearrested). Our findings highlight the difficulty in making conclusions about test bias, especially when base rates of recidivism are not equal across groups. Implications for juvenile justice policy and practice are discussed.
