Prior work has shown that understanding program and department expectations play a key role in doctoral degree completion. So, how do computer science Ph.D. students form and fulfill expectations in their programs? To answer this, we interview 14 CS Ph.D. students from underrepresented groups on how they identify and meet expectations in their programs. We offer two recommendations for departments - lab-based mentorship and goal clarity - to better support students.

As news sites aim to increase user engagement, how do designers’ structure the personalization of visualizations and what are the potential impacts? We develop a design space of 47 personalized visualizations - or popular media visualizations that allow news readers to segment data by attributes such as race, gender, and income. Through this design space, we identify (un)common design mechanisms that may impact users’ exploration and understanding of topics such as health, politics, and finance. We find that visualizations often lack explicit exploration suggestions or instructions, data notices, and personalized visual insights. We also outline three trajectories for future research.

How can we support the inclusion of ethics in existing computer science courses? We present a novel reposirory of CS ethics modules that can be embedded within existing CS courses. To show this, we pilot our predicitve policing module that aims to provide an ‘ethical laboratory’ for undergraduate students in an intermediate algorithms course. Our findings show that students were able to contextualize ethical issues, but wanted more class time devoted to the activity.

How does the algorithmic management behind online and ride-share gig work impact women working in developing countries? We synthesize literature from academia, government agencies, NGOs, and public media to 1) compile common platform and algorithmic features of gig work platforms, and 2) highlight how these features can impact worker well-being across developing contexts. We pay special attention to how these features impact women, given existing systemic gender oppression across the gobe. Based on this work, we make recommendations for developing actors and platforms - including advocating for intersectional well-being frameworks, standardizing features, removing reliance on algorithmic management, and minimizing hierarchies between workers and users.

How do we evaluate the impact of methods for equitable recommendations, particularly those that aim to support racial equity? Synthesizing prior literature, we outline different methods for evaluating recommendation interventions aimed at improving equity, as well as their challenges and interaction with privacy regulations like GDPR. We extend this analysis through the case study of editorial interventions within the context of online music recommendations.