Learning objectives of this week

• Conceptual understanding:
  • What are “unique variance explained” and “partial regression coefficient”
  • What is “collinearity” and why we want to reduce it
  • What is “cross validation” and how to conduct cross validation
• Interpretation and implementation:
  • Be able to interpret SAS outputs of a MLR, including intercept, partial regression coefficients, and R-squared.
  • Be able to identify the outcome/DV and predictors for a given scenario provided and write the linear equation specifying their relationships.
  • Understand the basic syntax for SAS procedure: PROC GLM and PROC REG
    • Understand ESTIMATE statement.
    • [Advanced, optional for this course] Be able to implement your own MLR models in SAS using these procedures.

Key concepts

• Partial regression coefficients
• Unique variance explained
• Collinearity and multicolinearity
• Cross validation

Reading

GHV ch 10

Tuesday session

Lecture: Linear regression with multiple predictors: part I (11-12 PM)

Interpreting partial regression coefficients

• Introduce the example
• Specify the equation of a MLR
• What (and why) are partial regression coefficients?

Measuring unique variance explained