SOC 385K • Discrete Multivariate Models
1:00 PM-2:30 PM
This course deals with the methodology, modeling, and analysis of discrete and categorical dependent variables. Categorical data analysis is appropriate when the dependent variable is measured as a count, or consists of events and trials, or when responses assume binary, ordered and unordered values. Regression-like models for discrete and categorical outcomes are widely used in applied research. The underlying methods for estimating these models, as well as the interpretation of results from these models are different from those of classical linear regression, however there are many similarities.
Students in this course should have prior exposure to statistical methods. This normally includes a course in basic statistics and one on linear regression, or a course that combines the two. It should be noted that this class serves a wide range of students; it is an elective course for students in the MS in statistics program, as well as for students in the social sciences. Some students will have taken courses in mathematical statistics and will have an understanding of matrix algebra and calculus. Other students may have more familiarity with the advanced statistical models in their substantive research area, but might lack the formal statistical/mathematical training. Given this heterogeneity in background, the course material presented here aims to be useful for students at all levels. Those without formal training can acquire a working understanding of the necessary concepts using the resources listed on the syllabus. Most of the material in the exercises and handouts is based on applied, as opposed to theoretical (i.e., mathematical statistical) problems. In keeping with the applied nature of this course, we will provide examples drawn mainly from sociological and demographic research.