Understanding Statistics and Data: Empirics for Law & Policy
(2 Credits) (Prof. Jacob Russell)
This course provides an introductory primer in quantitative evidence, statistics, data sciences, and causal inference. It is designed for students with no background in the subject, requiring no math, statistics, or probability background. Increasingly, we live in a world of empirical data — where arguments are won or lost on the basis of statistical and quantitative evidence. Yet most people have little familiarity with how to interpret data, or how to read a study and to understand its implications and limitations. This is a particular problem given that not all evidence is of equal quality. Even more confusion surrounds ideas about new techniques of data science, including machine learning, neural networks, and artificial intelligence, that have become increasingly common as part of the “big data” revolution.
The goal of this class is to make you a more sophisticated consumer of empirical data. Although we will examine applications in legal practice (e.g., the types of materials that might be the subject of expert reports during litigation), we pull on examples from a range of domains, and the course is designed to teach “data literacy” regardless of one's interests. We will pay particular attention to debates over causality — what kind of evidence is required to make the claim that “X causes Y” instead of “X is associated with Y”? Such causal claims are the coin of the realm for lawyers and policy analysts, yet they are often misused. The past several decades have been described as a “credibility revolution” in empirical research, as scholars develop a deeper understanding of the kinds of research designs that can lead to credible claims of causality, a topic we take up in depth in this course.
No math skills are presumed. Our focus is on reading and evaluating, rather than producing, an empirical analysis, although we will demonstrate some relevant techniques. Topics will range from probability, to study design and analysis, to data visualization. We will also explore machine learning, comparing it to more traditional techniques of statistical analysis. The course will draw on dozens of examples and case studies designed for this class, many drawn from the headlines and from topics of hot public debate, as well as readable materials designed for a non-specialist audience.
This course is open to undergraduates (by permission of instructor), JD students, other graduate students, practicing professionals, and members of the community.