Leadership in Action: Quantitative Criminal Legal

In the previous year, I met some alumni of the Tufts Computer Science Department through a mentorship program. Some of them were involved with data analytics, specifically quantitative social science. This led to a series of introductions and mentorship opportunities as I was exposed to data scientists working in quantitative social science fields.

Over the summer, I worked closely with some researchers from Harvard’s institute for quantitative social science. For my Leadership-In-Action component, I worked with a group of PhD candidates on racial disparities in the justice system.

In 2019, the District Attorney of Suffolk County in Massachusetts released a wide reaching series of policy changes. This was Rachel Rollin’s Memo. It was in line with many other similar programs throughout the United States.

The main theme of this memo was pretrial diversion, that is the redirection of arrested individuals away from criminal prosecution. In general, entry into criminal court proceedings is correlated with negative life outcomes. Pretrial diversion programs attempt to resolve pathological community issues by redirecting affected individuals away from punitive measures and towards community social programs to rehabilitate and fix the underlying causes of their issues. 

For example, pretrial diversion may mean first time non-violent drug offenses are no longer prosecuted in court if a defendant agrees to attend a local rehab program. Another common example, someone who shoplifts food is likely doing so as a result of poverty. This cycle of poverty would be exacerbated if they had to pay legal fees and pause employment to carry out a prison sentence.

We surveyed community data to find evidence of equity or disparity, as well as find better ways to support community members. The role largely revolved around advocacy for minority group members in using quantitative methods to build evidence for lack or presence of disparity. This information helps inform eligible community members of the effects of recent policy changes as well as options for enhancing positive life outcomes. Our work helped to organize equitable efforts in the space to benefit community members.

A public record does not imply a clean, accessible, and organized public record. In fact, this is an ongoing point of advocacy for several groups. I experienced this issue firsthand and joined the community contributing to its resolution. Much to my initial dismay, legal data is somewhat disorganized in Massachusetts. Database formats and standards change at a whim at the behest of new administrations with little thought of backwards compatibility. To obtain a proper dataset, it is sometimes necessary to contact a myriad of different sources.

For this reason, being proactive and focusing on long-term relationship building was key in the space. Oftentimes, technical issues must be worked around by personal communication and reaching out to key figures for advice. A lot of our work was contingent on third-party willingness to aid our efforts when technological infrastructure was lacking.

Overall, I had the opportunity to apply my heavily quantitative engineering background towards a humanitarian cause. In particular, I had an opportunity to learn a lot about the criminal justice space, as well as quantitative techniques exploring the space. This sparked an interest in quantitative social sciences, as well as the criminal legal space. I would like to pursue further applications of machine learning and mathematics related to the criminal justice system. This has led me to shift my education towards a focus on the power of speech and natural language analysis techniques in computer science.