File(s) under embargo
Reason: author request
until file(s) become available
Data and Humans in Algorithmic Risk Assessment
Algorithmic risk assessment instruments (RAIs) are increasingly used to aid decision makers in high-stakes domains. Characterizing the risks and opportunities of RAIs requires understanding how the data, the design of the tools, their evaluation, and their use affect the resulting decisions. In this thesis, we focus on two components of this design pipeline: the role of selection biases in the data (Part I) and of the interactions between RAIs and human decision makers (Part II).
Part I of the thesis covers analyses of the selection biases in criminal justice data that stem from the discrepancy between offense and arrest. We propose sensitivity analyses for fairness evaluations of recidivism RAIs when not all offenses result in arrests. Then, we present three studies that estimate racial disparities in the likelihood of arrest for a criminal offenses The first study focuses on disparities in police enforcement for marijuana violations by comparing data from police records with estimates of underlying criminal behavior obtained from survey data. The second study employs police records of violent crimes. The third study extends this analysis by introducing methods to adjust arrest rates computed on police records for unreported offenses, using victimization data.
Part II of the thesis discusses analyses of the influence of RAIs on human judgment. We first investigate how the deployment of an RAI affected decisions in child maltreatment hotline screenings. In particular, we take advantage of a glitch in the implementation of the tool to study whether decision makers adhere to the algorithmic acommendations when the risk is misestimated. Then, we present the results of two vignette studies involving predictive tasks that crowdworkers perform with the aid of RAIs. In the first study, we analyze the ecological validity of various study design choices. In the second study, we test whether homophily between humans and RAI, as well as monetary incentives, affect the behavior of the participants. We conclude with a user study that assesses the impact of human-RAI workflow configurations on the decision-making of expert veterinary radiologists Human food production and consumption is a leading cause that contributes to climate change. Although systematic transformations are imperative to tackle sustainable issues, individual food-related behaviors can be significant drivers for larger systematic changes. However, the complexity of individual food choices proposes challenges for sustainable dietary transitions due to cultural and market forces, high barriers to nutritional knowledge, and a variety of motivations behind a single food choice. The thesis investigated how persuasive design might support a shift toward more sustainable food practices for individuals with different dietary motivations. Through the literature and artifact review, I studied how persuasive design can implicitly change food-related behaviors with the psychological and social theories behind it, and identified the opportunities, challenges, and gaps in current food systems and existing tools. I conducted contextual inquiries, diary studies, and generative activities with participants to better understand their behaviors, mindsets, and attitudes in a variety of scenarios throughout their food experiences. As a result, I explored potential concepts and designed an alternative food experience to demonstrate how people may learn climate footprint in food through social interactions. In conclusion, persuasive messages for food sustainability can be more effective if the design leverages multiple persuasian techniques and encourages active participation from users.
DepartmentStatistics and Data Science
- Doctor of Philosophy (PhD)