Overview. This course is structured around three unequal
components.
Theoretical. This section focuses on proofs of the key
mathematical properties of causal models. We'll begin with some
papers that preceded Judea Pearl's Causality
(2000/2009), and then turn to the view developed therein. Causality
will be an important conceptual centerpiece throughout the
entire course.
Computational. This (shorter) section introduces actual
causal modeling with both artificial and real data sets, using
the R package pcalg.
Our primary purpose will be to consider some fully concrete
examples of causal modeling and its results, to help illustrate
matters that arise in the other two parts of the course.
Philosophical. This section covers some prominent and
recent philosophical work on causal modeling. Although there is
considerable flexibility in what we discuss, we will probably
pay substantial attention to James Woodward's Making Things
Happen (2003). Depending on participants' interests, we
may also consider, e.g., some of Nancy Cartwright's work on this
topic, David Danks' recent (2014) book, etc.
If there is time left over, I will present some material on latent
variable modeling and their roles in causal models.
Requirements. Those enrolled in the course will be asked to
write a term paper for the course, and to give an in-class
presentation.