Course description: Categorical data are often encountered in many disciplines including the fields of clinical and biological sciences. Analysis methods for categorical data are different from the analysis methods for continuous data. There is a rich collection of methods for categorical data analysis. The elegant “odds ratio” interpretation associated with categorical data is a unique one. This online course will cover cross-sectional categorical data analysis theories and methods. From this course, students will learn standard categorical data analysis methods and its applications to the biomedical and clinical studies. This particular course will focus mostly on statistical methods for categorical data analysis arising from various fields of studies including clinical studies; those who take it will come from a wide variety of disciplines. The course will include video lectures, group discussion and brainstorming, homework, simulations, and collaborative projects on real and realistic problems in human health tied directly to the student’s own professional interests. Focus will be given to logistic regression methods. Topics include (but are not limited to) binary response, multicategory response, count response, model selection and evaluation, exact inference, Bayesian methods for categorical data, and supervised statistical learning methods. This course stresses how the core statistical principles, computing tools, and visualization strategies are used to address complex scientific aims powerfully and efficiently, and to communicate those findings effectively to researchers who may have little or no experience in these methods.
Learning Objectives:
1. Gain proficiency, specifically, in logistic regression, and broadly, in generalized linear models
2. Acquire competency in standard and cutting edge categorical data analysis methods
3. Hone skills by applying categorical data analysis methods in analyzing data