Research seminar in information
SI 710-002: Causal Inference (Alain Cohn)
Does social media make us happy? Should you send your kids to private school? Does drinking soda make us unhealthy? These questions are causal in nature, that is, they ask “whether something causes something else to change.” The arguably best way to learn about causality is to run randomized experiments. However, the experimental ideal, that is, random assignment of the treatment, is often impossible or impractical. Thus, we must look for alternative strategies that allow for causal identification when we do not have perfect control over treatment assignment.
This seminar will provide an overview of different empirical methods to identify causal effects using observational data. In addition to randomized experiments, we will cover the most common methods of causal inference, such as controlled regression, matching, instrumental variables, regression discontinuity, and differences-in-differences. The focus is on promoting understanding and intuition of empirical methods, as they are used by practitioners. This means that we will spend more time on the intuitive framework that guides empirical papers rather than the formal details (beyond what is necessary). You will learn about the pros and cons of different methods, how to apply those methods in practice, and how to communicate the empirical findings to the relevant audience.
The seminar will be based largely on Scott Cunningham’s book Causal Inference: The Mixtape. The assignments will involve student presentations, concept quizzes, and programming exercises in which you will be asked to analyze data sets. Solutions will be handed out written in STATA. At the end of the seminar, you will have been exposed to a wide range of empirical methods, you will gain familiarity with the terminology and technical jargon, and you will be able to critically evaluate and replicate published work that uses empirical methods.
SI 710-111: Knowledge Infrastructures (Andrea Thomer)
Knowledge infrastructures are the “robust networks of people, artifacts, and institutions that generate, share, and maintain specific knowledge about the human and natural worlds” (Edwards, 2010). These have historically taken the forms such as scholarly publications, museums, libraries, and archives, but now additionally include data repositories, preprint servers, APIs, digital collections, R packages, linked data, social media, citizen science platforms, and much more. We must study the creation, maintenance, break-down, longevity, controversy and evolution of these complex systems if we are to ensure their longevity, trustworthiness, resilience, and inclusivity.
In this seminar, we will be reading foundational and recent research on knowledge infrastructures from the information sciences, STS, and other related disciplines. Topics include but are not limited to:
- Sustainability, maintenance, and repair of KI
- Break-downs, shut downs, and KI death
- Gatekeeping, trustworthiness and misinformation in KI
- Invisible labor in KI
- Representation, inclusivity and KI governance
Though this seminar is, of course, aimed at those particularly interested in critiquing and sustaining knowledge infrastructures, it may also be of interest to those interested in science and technology studies more broadly, or in the study of the “science of science.” Special attention will be paid to the myriad of methods that must be brought to the study of knowledge infrastructures. You will be encouraged to think through the strengths and weaknesses of different approaches, and consider, for instance, when you might wish to conduct a digital ethnography vs. a scientometric study vs. some novel combination of the two.
This course will also offer you the opportunity to develop their own paper or project related to KI over the course of the semester. This should be tailored to your broader research interests and community. The hope is that your work can be refined after class and submitted to the publication venue of your choice.
SI 710-131: Algorithms & Societal Implications (Nazanin Andalibi)
This doctoral seminar course examines algorithms' roles in our lives and their social, individual, ethical, political, and other implications. While algorithms are not new, we continue to see broader interest in studying their impact on society as evidenced by congressional hearings, calls/reports from the National Science Foundation and Computing Research Association, growing research centers in industry and academia (Microsoft Research, Data & Society, etc), conferences (e.g., ACM Conference on Fairness, Accountability, and Transparency (FAccT*)).
Key topics related to algorithms and their implications include but are not limited to facial and emotion recognition technologies, as well as algorithms' roles in pressing issues such as misinformation, political polarization, discrimination and other identity-based harm, content moderation, and mental health.
The course will take a multidisciplinary approach to algorithms surveying research from computer science, science and technology studies (STS), library and information science, communication, business, law/policy, philosophy, public health, and psychology.
The course will have a particular focus on interrogating contexts in which algorithms fail in respecting the values of marginalized communities, or worse, when they pose risk or cause harm. As such, issues of human rights, justice, wellbeing, privacy, and autonomy, are examples of discussion topics we will have. The students will be invited to add their own interests to the classroom experience and their projects throughout the course. We will also discuss current issues (e.g., in the news) related to algorithms in class for students to develop informed and intellectual positions, and pose meaningful research questions with societal impact that merit scholarly investigation.
This is a seminar course, structured around discussions, readings, writing, and other critical engagement with related topics. Intellectual participation and engagement will be key. In addition to class activities and discussions, students will choose a topic of interest and develop a research proposal paper in that space. I encourage students to use this as an opportunity to develop a project that would be submitted for publication in their venue of choice as aligned with their broader research interests and trajectories