Swathi is a senior double majoring Computer Science and Public Policy. She is interested in the ways technological tools are utilized in the public sector and the complications that such implementation may cause. She is particularly concerned with the ways technology can impact marginalized groups, such as women, refugees, and racial minorities.
Are Numbers Objective? An Analysis of Computational Risk Assessment’s Impact on Underrepresented Communities
Faculty Advisor: Professor Philip M. Napoli
Abstract: This paper seeks to examine the effect of a specific class of pretrial risk assessment instrument on various identity groups. Risk assessment tools (RAT’s) are known to take particular characteristics about a defendant in order to give a numeric score of the person’s risk of reoffending or their risk of failing to appear in court. Justice system staff members can then use risk scores to inform bail, pretrial, or sentencing decisions. While these tools are controversial, they have been implemented in judicial systems across the country. This paper will examine the effect of the Virginia Pretrial Risk Instrument (V-PRAI) and its effect on individuals’ from different racial groups, gender, and age groups while holding their class of charge constant. This evaluation looks particularly at the Durham County Court system and its implementation of V-PRAI. Using multiple linear regression, the model takes into consideration a defendant’s age, sex, interviewer, race, and charge type in order to understand which factors may confound an unbiased assessment of a defendant’s risk.