Reminders/Things due this week

• Forum 1.1 (Helms et al) due Wednesday Feb 8 (submitting your post by noon is prefered, but midnight is okay)
• Check out the Gradebook on Sakai on Tuesday. There should be grades for your Exercise 1.1 - Exercise 1.4 if you have submitted your responses. If you do not see your grades, check your Dropbox and submit your Exercises. Also pls notify me to update your grades!

Learning objectives of this week

• Continue discussing issues arise in multiple linear regression: assumption violations and outliers:
  • What are the six assumptions of linear regression?
  • What are consequences of violating assumptions?
  • What are potential remedies?
  • (More on case influence and outliers) What are “global influence” and “local influence”?
• Interpretation and implementation:
  • Learn how to use visualization tools (e.g., diagnostic plots, other plots) to evaluate assumption violations.
  • Be able to interpret the “Fit Diagnostics” section from SAS output.
  • Learn general approach to detect potential outliers

Key concepts

• Assumptions
  • Validity
  • Generalizability
  • Linearity and Additivity
  • Independence
    • Simpson’s paradox
  • Homogeneity/homoscedasticity
  • Normality
• Case influence
  • Influence = Leverage X Discrepancy
  • Global influence & local influence

Reading

GHV ch 11 (11.1-11.3)

Tuesday session

Lecture: Assumptions

Why do we care about assumptions?

What are the sixe assumptions in linear regression