Biostatistical Analysis of Epidemiologic Data II: Poisson and Conditional Logistic Regression Analysis
June 19-23, 2017
8:30 a.m. – 12:00 p.m.
Course Number: 140.677.11
Multivariable analysis of multivariate data is often an essential element of epidemiologic research. Applications of regression techniques are the topic of this course, starting with a review of simple linear regression, as a foundation. Followed by application to non-linear data using more general regression techniques. Then, a complete and extensive description of log-linear regression analysis (also called Poisson regression) and how it works, particularly for the application to count data and tables. Also included is the concept of quasi-independence and the analysis of incomplete tables. Logistic regression techniques are similarly described in detail with emphasis on application to epidemiologic binary outcome data in several contexts. All regression techniques are illustrated with applied examples.
Learning Objectives: Upon successful completion of this course, students will be able to
- Mastery of regression analysis tools
- Understand and apply regression multivariable statistical tools to analyze multivariate data
- Effective use of log-linear/Poisson regression methods in the context of count data and analysis of tables
- Similarly, effective use of logistic regression techniques applied to binary outcomes generated from multivariate data.
Grading Options: Letter Grade or Pass/Fail
Course Materials: Provided in class