Socioeconomic Drivers of Adult Disproportionate Minority Contact; Data Bias in the Criminal Justice System 3 views

Preface Justice in the United States is heavily influenced by the intersection of physical geography and digital infrastructure utilized by the local government. The technical report, titled Socioeconomic Drivers of Adult Disproportionate Minority Contact, was conducted by a research team to investigate how neighborhood-level factors like poverty and educational attainment influence jail intake rates in the Charlottesville-Albemarle region beyond purely demographic factors such as race. Simultaneously, the STS research paper titled Data Bias in the Criminal Justice System examines the sociotechnical frameworks that allow historical human prejudices to become embedded within predictive policing and risk assessment algorithms. These two projects are fundamentally linked because the technical research identifies the structural inequalities that the STS paper demonstrates are being automated by modern technological systems. Both research efforts are highly relevant to systems engineering because they highlight the necessity of understanding the social context behind the data used to build and deploy complex public safety models. The capstone project provides a methodology for addressing the overrepresentation of specific populations in the legal system by moving beyond race as a solitary variable. This research contributes to a solution by offering the Jefferson Area Criminal Justice System a data-driven understanding of how localized economic factors drive system contact. The methodology for this group project involved the integration of over fifty-nine thousand booking records from the Albemarle-Charlottesville Regional Jail with U.S. Census tract data through a comprehensive data pipeline. The team utilized approximate string matching and geocoding to link these datasets and then applied K-Means clustering and Principal Component Analysis to categorize different regions based on their unique socioeconomic profiles. The overall conclusions of the capstone project indicate that poverty is the most significant structural driver of variation in booking rates throughout the region. The findings show that while racial disparities exist, the strength of the relationship between race and system contact is largely determined by neighborhood socioeconomic status. Racial disparities are significantly more pronounced in high-poverty census tracts while they tend to diminish in more affluent areas. Ultimately, the project demonstrates that disproportionate minority contact is a symptom of broader systemic disadvantages that shape how different communities interact with law enforcement. The STS paper addresses the research question of how historical data bias influences the outputs of modern predictive policing and recidivism risk assessment software. This research is significant because many of these algorithms are proprietary and lack the transparency required for public accountability in the criminal justice system. The methodology utilizes the Social Construction of Technology framework to argue that these tools are not neutral observers but are instead shaped by existing social values and policing patterns. This approach allows for a deeper critique of how technical systems can unintentionally mirror and reinforce human prejudice. The evidence presented in the STS paper highlights how tools like the COMPAS recidivism risk assessment frequently produce biased results because they rely on proxy variables such as criminal history and zip codes. These results indicate a feedback loop where biased data leads to increased surveillance in certain communities, which in turn generates more data to justify that surveillance. The paper concludes that achieving true equity requires a transition toward transparent algorithmic systems and a dedicated effort to de-bias historical data. It also emphasizes the need for a shift in focus from predicting arrests to understanding the underlying community needs and victimization patterns.

BS (Bachelor of Science)

Bias; Criminal Justice; Predictive Policing

School of Engineering and Applied Science Bachelor of Science in Systems and Information Engineering Technical Advisor: Preston White, Loreto Alonzi, Michael Smith STS Advisor: Pedro Francisco Technical Team Members: Jay Sorkin, Mitchell Palmer, Kevin Villalobos, Saiesha Gohil

English

All rights reserved by the author (no additional license for public reuse)

2026-05-07