Students will learn statistical methodology for the analysis of
discrete and categorical data. In addition, they will be
introduced to the class of generalized linear models (GLMs),
which subsumes many model classes they will have seen in 
previous course-work. They will learn the theory of GLMs and 
will consider in detail subclasses appropriate for the analysis 
of discrete data, including binary response data (e.g., 
logistic regression models), count data (Poisson and negative 
binomial log-linear models), and categorical data (multinomial 
response models). Students will learn to use these models to 
analyze real data, including how to use statistical software 
appropriately.

Introduction to discrete response data; description and 
inference in two- and three-way contingency tables; review of 
classical linear models; theory of generalized linear models 
(GLMs); fitting methods and algorithms; likelihood-based 
inference methods for GLMs; model diagnostics; and specific sub-
types of GLMs, including applications to binary data, 
categorical data (including ordinal data), and counts. 
Overdispersion and quasi-likelihood will also be discussed.