Weekly Schedule
week date topic chapters homework nature
1 24 jan placing data in context A:1-3 initial reading get up to speed
26 jan comparing groups B:4-6
2 31 jan factorial anova C:7-8 comparing groups learn computer system
2 feb balanced experiments C:8-9
3 7 feb unbalanced designs D:10 factorial anova learn notation
9 feb missing cells D:11
4 14 feb linear models D:12 unbalanced data cautious interpretation
16 feb frequency tables
5 21 feb goodness-of-fit missing cells balanced subsets
23 feb log linear model
6 28 feb generalized LM categorical data goodness-of-fit
1 mar blocking & subsampling H:22
7 6 mar nested designs H:22-23 midterm first 6 weeks
8 mar split-plot design H:23
* 13 mar SPRING BREAK *
15 mar SPRING BREAK *
8 20 mar repeated measures I:25 split plot diff size EUs
22 mar epsilon adjustment
9 27 mar multivariate approach I:26 repeated measures
29 mar cross-over design I:27
10 3 apr random effects G:19 cross-over
5 apr two-factor random G:20
11 10 apr balanced mixed models G:21 random effects
12 apr variance components G:19.4,20.2
12 17 apr restricted ML (REML) G:19.3-5 variance components
19 apr general random models G:20-21
13 24 apr general nested design H:24 nested design
26 apr general repeated meas I:26.3
14 1 may analysis of covariance F:16-17 general rep meas
3 may multiple responses F:18
15 8 may model selection C:9.1-3 ancova
10 may review/overview A-I FINAL out
Homework is generally assigned on Tuesdays and due the following Tuesday.
Discussions on Thursday can help with key points in the homework.
See Weekly Schedule for more detailed information
on topics, and Readings for material in related
texts. Examples in parentheses are examples from R or Splus. To view
them, attach the PDA library and then
run the example as below:
> library( pda )
> example( Tomato )
# topic nature
1 practical data analysis read 1-6, setup computer account (Tomato)
2 analysis of variance 8.1 (BactRoom)
read 7-9
3 unbalanced data 10.2, 11.3a (Hardy, Growth)
read 10-12
4 missing cells 11.3b-e, 11.1a-c (Growth)
read VR 7.0-3
5 counting data redo 11.3 with counts (Count)
read 19.1, 22.1-3, 23
6 counting data analyze BrandX (BrandX)
7 split plot 22.2, 23.1 (Bacteria)
25.3, 26.3(a-c) (Berry)
27.2, 3.2
20.1, seeds & farms (Rantwo)
26.4: consider at least 2 covariance structures (Season)
24.1 (Diet)
Initial readings of
Yandell (A:1-3) and Milliken & Johnson (MJ 4,6)
cover introduction to practical data analysis and design of experiments.
You should be reasonably comfortable with normal F, t,
chi-square distributions, matrix algebra and
general theory of linear models
(e.g. Hocking 2-4, Seber A,3; Searle 7,8; Scheffe A,1).
Some of this material will be reviewed as needed (see Yandell D:12).
Suggested background computer reading includes Venables & Ripley (VR)
1-3 for Splus and di Iorio & Hardy (DH)
or Littell, Freund & Spector (LFS) for SAS.
However, most of the needed material on these languages can be picked up
from examples supplied by the instructor in homework and Internet pages.
Abbreviations identified above (plus M = Miller, H = Hocking)
are used to suggest supplementary reading.
part:topic readings
A:Data in Context MJ 4,6; VR 1-3
B:Comparing Groups H 1.3,1.5,4.1,12; MJ 1,2,3; VR 6.5; LFS 2
C:Factor Effects H 13.1,13.3,14.2,14.4; MJ 5,7-8; VR 6.5; LFS 2,5.2
D:Imbalance H 13.2,13.4; MJ 9-15; VR 6.6; LFS 4
E:Assumptions H 6; VR 6.3; M 1-4; LFS 1.8
F:Covariates H 6.7,10.3,12.4; VR 6.1-6.2,12.3; M 5; LFS 1,6-7,5.5-5.6
G:Fixed & Random Effects H 15-17; MJ 18-23; VR 6.7; M 7; LFS 3.3
H:Nesting Units H 14-17; MJ 5,24-25,28-30; VR 6.7; LFS 3,5.3,5.7-5.8
I:Repeating Measures H 15.2x6,16.4x6; MJ 5,26-29,31-32; LFS 8,5.4
Required
- Yandell (1997)
Practical Data Analysis for Designed Experiments,
Chapman & Hall/CRC Press.
Developed specifically for this course. First part,
especially ch. 2, pertains to statistical consulting.
Recommended
- Littell, Freund & Spector (1991)
SAS System for Linear Models, 3rd ed.
SAS Institute, Cary, NC.
Practical guide for many designs considered in this course. This
package is used widely in government and industry, as well as academia.
- Miller (1997)
Beyond Anova, 2nd ed.
Chapman & Hall/CRC
Press. ISBN: 0412070111.
Good reference on assumptions and placing anova in context of
statistical practice.
- Milliken & Johnson (1992)
Analysis of Messy Data: Designed Experiments, 2nd ed.
Chapman & Hall/CRC Press.
ISBN: 0412990814.
Accessible, with many practical examples, but lacks math details.
- Venables & Ripley (1999)
Modern Applied Statistics with S-Plus, 3rd ed.
Springer-Verlag, New York.
Flexible, extendable interactive data analysis system with nice
graphics; complements SAS in many situations.
- Littell, Milliken, Stroup and Wolfinger (1996)
SAS System for Mixed Models.
SAS Institute, Cary, NC.
Ideal supplement for the last half to third of the course.
Describes proc mixed and proc glm
usage in great detail with examples.
ISBN: 1-55544-430-X.
Suggested
- DiIorio FC & Hardy KA (1996)
Quick Start to Data Analysis with SAS,
Duxbury.
Fairly comprehensive introduction.
Beyond basics of data handling.
Introduces many common statistical procedures by example,
but somewhat uneven in their detail.
- Miller (1997)
Simultaneous Statistical Inference, 2nd ed.
Springer-Verlag, New York.
Invaluable reference for multiple comparisons. Hsu (1996)
Multiple Comparisons,
Chapman & Hall,
London,
is an accessible update to compliment this reissued classic.
- Littell, Milliken, Stroup & Wolfinger (1996?) SAS System
for Mixed Models. SAS
Institute, Cary, NC.
ISBN: 1-55544-779-1.
Background & Theory
- Hocking (1996)
Methods and Applications of Linear Models,
Wiley, New York.
- Scheffe (1959)
The Analysis of Variance,
Wiley, New York.
- Searle (1987)
Linear Models for Unbalanced Data,
Wiley, New York.
- Searle, Casella & McCulloch (1992)
Variance Components,
Wiley, New York.
- Seber (1977)
Linear Regression Analysis,
Wiley, New York.
Last modified: sun 23 jan 2000 by Brian Yandell
(yandell@stat.wisc.edu)