Lesson 31: Multiple Regression Analysis

Overview Section

In this lesson, we'll explore using the REG procedure for performing regression analyses with a continuous response variable and more than one predictor variable. We'll also explore using the LOGISTIC procedure for performing logistic regression analyses with a binary response variable and one or more predictor variables.

Objectives

Upon completion of this lesson, you should be able to:

  • use the REG procedure to analyze data arising from a designed experiment
  • use the REG procedure to analyze data arising from nonexperimental research
  • use the variable selection methods available in the REG procedure to find the "best" model (or models!) for a set of data
  • create and use dummy variables in a regression model
  • use the variance inflation factor to look for multicollinearity
  • use the LOGISTIC procedure to analyze regression data having a response variable with just two levels (such as yes/no or dead/alive)
  • capture the necessary output from the LOGISTIC procedure in order to create a receiver operating characteristic (ROC) curve

Textbook Reference Section

 Chapter 9 of the textbook.