Courses Details

EPID742: Introduction to Linear, Logistic and Poisson Regression

  • Graduate level
  • Residential
  • Summer term(s) for residential students;
  • 1 Credit Hour(s) for residential students;
  • Instructor(s): Kim, Myra (Residential);
  • Prerequisites: Intro Epidemiology and Biostatistics and Perm. Instr
  • Description: This course will cover regression methods for continuous, binary, and count data. Majority of epidemiologic data involve either binary or count data, and binary data often arise from an underlying continuous data. Therefore, multiple (for continuous data), logistic (for binary data) and Poisson (for count data) regression analyses are all important analytic approaches that frequently provide valuable insights into data collected for epidemiologic studies. All approaches will be covered under the umbrella of generalized linear models (GLM) and presented in a practical and an applied fashion. The discussed material begins with the simplest case with the goal of understanding the fundamental properties of each model. Once these properties are established, more advanced topics such as collinearity, variable selection, non-linear explanatory variables, and goodness-of-fit will be described and applied to example epidemiologic data sets. These approaches provide simple and effective ways to explore complex relationships and illustrate the general process of using a model to draw appropriate conclusions. Prerequisite: Introductory level courses in epidemiology and biostatistics