Regression Analysis

This intermediate level course aims at providing theoretical and practical training for epidemiologists, clinicians and other professionals of related health disciplines in statistical modeling with particular emphasis on straight line linear and multiple regression. Included topics are: review of straight line regression and correlation, ANOVA for straight line regression, appropriateness of straight line model, polynomial regression, multiple regression analysis, partial F-test, dummy/indicator variables, statistical interaction, comparing straight line regressions, analysis of covariance, estimation and interpretation, goodness-of-fit, model selection, collinearity and outlier diagnostics. 30 hrs.

Objectives:

  • Students will learn the fundamental methods of statistical regression modeling for continuous response variables.
  • Students will learn how to build and interpret a variety of multiple regression models, including models with continuous, nominal/ indicator, and polynomial explanatory regressor variables.
  • Students will become familiar with outlier and collinearity diagnotics to refine models, as well as statistical software packages for computing multiple regression models.

The following texts (book) will be used:  Applied Regression Analysis and Other Multivariable Methods by Kleinbaum, Kupper, Muller and Nizam.

Course code: ESP09

Faculty: Brian Marx

Date: August 18 – 22, 2014

Time: 08:45 – 16:00

Course days: Monday to Friday

Course material: Online, download instructions will be sent in August by e-mail.

Prerequisites: Course Introduction to Data-analysis or equivalent knowledge