computer code
homework
labs
links

Stat 502

Design and analysis of experiments


Meeting times
Instructor
  • Peter Hoff
  • C-319 Padelford
  • Office Hours: Tue 9:30-10:30, Thu 10:30-11:30 or by appointment.
  • hoff at stat dot washington dot edu

Texts:
  • Hoff, P., "Lecture notes"
  • R Development Core Team, "An Introduction to R" (html, pdf ).
  • Montgomery, D., "Design and Analysis of experiments" (any edition, recommended but not required).

Computing:
The supported software for 502 is the R statistical environment. R code will be provided for most of the in-class data analysis examples.
Approximate schedule:

Week 1 : Experiments, test statistics, completely randomized designs, significance testing
Week 2 : Review of normal theory tests and confidence intervals, basic decision theory, power and sample size
Week 3 : Treatment effects model, ANOVA
Week 4 : SS decomposition, geometric interpretation
Week 5 : Treatment comparisons, model diagnostics
Week 6 : Factorial treatment designs
Week 7 : ANOVA decomposition for the additive model
Week 8 : ANOVA for the interaction model, model comparison, and normal-theory testing
Week 9 : Complete and incomplete block designs, Latin square designs
Week 10 : Fractional Factorial designs, aliasing, confounding, resolution
Week 11 : Split plot designs, different size experimental units, repeated measures


Coursework:
There will be
  • about 7 or 8 regular homeworks,
  • a midterm with take-home and in-class components,
  • a final with take-home and in-class components.
Each one of these components will be worth about the same amount.

Late policy:
Each turned in item receives an initial grade of x, then the actual grade is y=x exp(-d/10), where d is the number of working days after the due date I receive the work. Everyone receives one grace day to be applied to one homework for the entire quarter.