Skip to main content

Undergraduate Students: Algorithms for Social Accountability

We are looking for undergraduate research students to contribute to our research on the NSF-funded project “Community Responsive Algorithms for Social Accountability (CRASA).” If you are interested, please send an email to with the subject line “UG CRASA: [YourLastName].” Please include a resume and a cover letter indicating why you are interested. You are only eligible if are you are a US citizen or permanent resident (please email if you have any questions about this). The parent project aims to analyze methods to balance societal needs for accountability, current legal standards, and practical issues of algorithm auditing. The descriptions for the currently available two projects are below.

Project 1 – Simulating the Systemic Effects of the Policy Algorithms in Criminal Justice and Law Enforcement: Algorithms play an expanding role in public policy decisions in many areas, including criminal justice and law enforcement. This project aims to explore the systemic effects that various criminal justice and/or proactive policing decision support algorithms cause in society using social simulation methods (i.e., agent-based modeling and system dynamics).

Project 2 – Empirical Evaluation of the Perceived Tradeoff between Fairness/Explainability and Accuracy: Algorithms play an expanding role in public policy decisions in criminal justice, allocation of public resources, public education, and even national defense strategy. However, standards of accountability reflecting current legal obligations and societal concerns have lagged in their extensive use and influence. Many approaches to ensure fairness and explainability require a tradeoff with overall accuracy. This project aims to characterize the (potential) decrease in accuracy when fairer, more accountable, or more explainable models (i.e., usually simpler models such as linear models or simple decision trees) are employed or when input data or model outputs are refined to eliminate demographic discrepancies.