STAT 440T(S) Categorical Data Analysis (Q)

Could the Challenger explosion have been prevented with proper data analysis? This tutorial focuses on methods for analyzing categorical response data which, in contrast to continuous data, consist of observations classified into categories and comprises binary (e.g., O-ring failure) and count data. Traditional tools of statistical data analysis such as linear regression and ANOVA are not designed to handle these types of data and pose inappropriate assumptions such as constant variance and normality. We will develop Generalized Linear Models, designed to address the discrete nature of the observation. Using computer software, we will consider building and fitting such models for applications in the social sciences, biology, medicine, public health, economics and marketing.
Format: Tutorial. All tutorial participants will meet once a week and alternatively present assignments and real data analyses or a discussion and critique thereof. Evaluation will be based on performances on these.
Prerequisite: Mathematics 211, Statistics 201, Statistics 346 or Economics 255, or permission of instructor. Enrollment limit:14 (expected:14). This tutorial is a quantitative/formal reasoning course.

Tutorial meetings to be arranged KLINGENBERG