Chapter 13 Introduction

There is an important distinction between multivariable and multivariate models.

  • multivariable : Multiple predictor variables (\(\mathbf{x}\)).
  • multivariate: Multiple response variables (\(\mathbf{Y}\)).

Often analysts will misuse multivariate when they really mean multivariable.

Multivariate techniques are

  • Primarily used as an exploratory technique
  • Restructure interrelated variables
  • Simplify description
  • Reduce dimensionality
  • Avoid multicollinearity problems in regression

We will discuss two different, but related techniques: Principal Component Analysis and Factor Analysis.