So if regression doesn't prove causation, what does?

• Administrivia:
  • Extra office hours by Nuria: 10 - 1 May 26 and 27th.
  • Extra office hours by Michael: 10 - 12 on Thursday May 20th.

Observational studies vs controled experiments

• Observational studies
  • An observation study is one where you watch the world around you
  • Examples: Astrophysics, traditional evolution, economics, sociology

• controled Experiments (using randomization)
  • Manipulate the world and see the result
  • Examples: chemistry, psychology, microbiology, modern genetic

• Controled experiments are the domain of ANOVA
  • Example: how long does it take to find something on a web page?
    • number of items
    • distractions
    • layout
    • colors (yellow on white)
    • etc
  • If you just watch people and measure the time till they click on something
    • you don't know if they were actually searching
    • maybe they knew where they were going
    • reading a book!
  • Controlled experiment: Give subjects a target to find on the page
    • Can measure success
    • can measure time until the find it
    • can measure subjective experience
  • Changing X variables
    • build many different pages and see which works best (large one-way ANOVA)
    • Build pages that change 2 different attributes (two-way ANOVA)
    • possible control for subject effect (three-way! ANOVA)
  • Generalizes a simple controled experiment
  • If you find a ROW effect or a COL effect you know that assignemtn to ROW/COL is random and hence uncorrelated with any possible lurking variable.
  • So many row and many column effects can be tested at the same time. This can save lots of time!

• Example: ANOVA project (sample 2 way ANOVA writeup)

Last modified: Tue Apr 18 13:07:21 2000