Statistical Genetics of Quantitative Traits Linkage, Maps and

QTL

Visit the link below to download the full version of this book:

https://medipdf.com/product/statistical-genetics-of-quantitative-traits-linkage-
maps-and-qtl/

Click Download Now

George Casella
University of Florida
Dept. Statistics
103 Griffin-Floyd Hall
Gainesville FL 32611-8545
USA

Editorial Board
George Casella Stephen Fienberg Ingram Olkin
Department of Statistics Department of Statistics Department of Statistics
University of Florida Carnegie Mellon University Stanford University
Gainesville, FL 32611-8545 Pittsburgh, PA 15213-3890 Stanford, CA 94305
USA USA USA

ISBN 978-0-387-75964-7 e-ISBN 978-0-387-75965-4

DOI: 10.1007/978-0-387-75965-4

Library of Congress Control Number: 2008920950

© 2008 Springer Science+Business Media, LLC

All rights reserved. This work may not be translated or copied in whole or in part without the
written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street,
New York, NY, 10013, USA), except for brief excerpts in connection with reviews or scholarly
analysis. Use in connection with any form of information storage and retrieval, electronic
adaptation, computer software, or by similar or dissimilar methodology now known or hereafter
developed is forbidden.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if
they are not identified as such, is not to be taken as an expression of opinion as to whether or
not they are subject to proprietary rights.

Printed on acid-free paper.

9 8 7 6 5 4 3 2 1

springer.com
To my mentors, who taught me more than I realized
Preface

Statistical design is one of the fundamentals of our subject, being at the core of
the growth of statistics during the previous century. Design played a key role in
agricultural statistics and set down principles of good practice, principles that
still apply today. Statistical design is about understanding where the variance
comes from, and making sure that is where the replication is. Indeed, it is
probably correct to say that these principles are even more important today.
Fisher (1947) compared a dataset to a sample of gold ore. The finest analysis
could only extract the proportion of gold contained in the ore. But a good
design could produce a sample with more gold.
There are plenty of “Design of Experiments” books available, many of
which do a fine job of describing not only how to design experiments, but
also how to analyze them. So why bother with another book? There were two
main reasons.
The first reason is the observation that many of our “standard” analyses
have become driven by the default setting of one’s favorite computer package
and, unfortunately, many times these default settings provide an incorrect
analysis. More frightening is the fact that sometimes such default analyses
have found their way into textbooks.
The second reason is that, although design books have gotten broader
in coverage of designs and often have much to say about analysis, the basic
theoretical underpinnings are not always covered, and if they are, they are
not covered in sufficient detail to understand how to construct the correct
analysis and to understand why the computer package default analysis may
be incorrect. Without fully understanding what the correct analysis should be,
it is impossible to design a good experiment.
So ... the goal is to describe the principles that drive good design, which
are also the principles that drive good statistics. Moreover, this will be done
with detail and attention paid to the theoretical background – only by having
more than a passing familiarity with the fundamental theory can one truly
understand statistical design.
VIII Preface

This book is not an encyclopedia of designs. There is no attempt to cover

all designs, and no attempt to teach data analysis at all. Although we will
analyze many datasets, we will usually use common anova techniques and not
preach about transformations, heterogeneity, missing data, and all the other
good stuff that a good data analyst needs to know. We assume that the design
has been run with enough success so that rescue techniques are not needed
for the analysis. We also assume that the student has been exposed to such
data analysis strategies.
We will cover the most popular designs in depth, both with theory and
examples, and datasets from consulting sessions and research publications. We
emphasize basic principles and careful modeling and, armed with such tools,
the student should be able to apply these principles in any situation
This book grew out of a course on Statistical Design taught at Cornell
during the 1990s, and at Florida in the 2000s. Most of the examples and
datasets are from consulting sessions with graduate students writing theses
or professors writing papers, and the subjects span everything from planting
alfalfa in the field to harvesting brain stem cells. (Most of my career has been
spent at Colleges of Agriculture – Rutgers, Cornell, and split between Arts
and Sciences and Agriculture at Florida. This will show in the examples.) I
have found that although the data and the lab techniques have changed, the
statistical principles remain quite constant.
Now, for the more important details:

◦ The level of the text is for first or second year graduate students. The
students should be familiar with standard statistical methodology (anova,
blocking, multiple regression) that one would get from a typical one-year
methods sequence (from books such as Ott and Longnecker 2000, Rawlings
et al. 1998).
◦ The material in the text is about right for a one-semester course. There
is probably a bit more than can be comfortably covered, so some picking
and choosing will be necessary. However, marching through the text is a
reasonable strategy.
◦ The chapters cover, for the most part, the standard material of a design
book, with mostly real examples, and applications of design in real sit-
uations. Although we cover many microarray designs, we do not have a
special section, instead treating them as the topics apply. The only un-
usual chapter is Chapter 4, where the concept of blocking is explored
further, and the effect of a random factor is examined. To me, blocks are
not about being samples from a larger population (which can be difficult
to justify) but rather about the correlation structure that they induce.
◦ Most chapters have a section Technical Notes, which contains the under-
lying theory in detail. The level in these notes is “anything goes”; we
use a lot of matrix algebra, some calculus, and also some statistical con-
cepts such as likelihood and sufficiency. These are in-depth looks that will
enhance the understanding of the advanced student, but skipping these
 Preface IX

sections will not hamper the beginner. To fully appreciate these sections it
would be good if the student has had a course from a book like Statistical
Inference by Casella and Berger (2001).
◦ The exercises are divided into “Essential” and “Accompaniment”. Every-
one should do the essentials. The accompaniments tend to be more of
a theoretical nature, going into the details of the procedures. I strongly
suggest that, if the students are able, they should do these exercises too.
I have always found that, for me, true understanding only comes from
slogging through the details.
◦ The datasets are on the book web page found at

www.stat.ufl.edu/∼casella
and, for most datasets in the examples there will be an accompanying R
program. These will not be sophisticated analyses, but rather will serve
as a starting point. I make no guarantees about the R programs. Smaller
datasets may only appear in text.

My best advice is that if you really want to understand statistical design,

read (or even better, reread) Fisher. His ideas, especially about blocking, have
greatly influenced my thinking. In fact, what I perceived as mishandling of
the randomized complete block design was one of the driving forces behind
the text. Fisher, of course, got it right.
There are many gray areas in design – when to pool, how to replicate, etc. –
some of which cannot be fully answered with statistical fact. This is where
we enter the realm of opinions, where judgments are made more on anecdotal
evidence and experience rather than formal calculations. After doing this stuff
for over 20 years opinions form about how to do things – I have taken the
liberty of sharing those thoughts.
Finally, thanks to all of my mentors. First there was Leon Gleser, my PhD
advisor, who taught me to work hard and try to learn as much as possible.
My design mentors were many, starting with Virgil Anderson at Purdue, and
Walt Federer at Cornell, who not only made me really understand split-plot
designs, but also taught me to say “a model” and never “the model”. And Carl
Lowe, the Plant Breeding Professor at Cornell who taught the course “Field
Plot Techniques”, and knew more about field plot layouts than anyone on
the planet. When he retired he gave me all of his notes and examples, many
of which appear as examples and exercises (and in the Appendix). I also
thank the students and colleagues who suffered through my learning process,
who listened as I thought through all of this, read through notes, and solved
problems: Mihai Giurcanu, Jamie Jarabek, David Lansky, Michael Meredith,
Deborah Reichert, and Andy Scherrer.
My most special mentor was Myra Samuels. A true sadness in my life is
that she died in 1992. She taught me to always ask questions until you really
understand. Thanks, Myra.
X Preface

And to all of my other mentors – you have all helped me more than you
know – Bob Bechhoffer, Jim Berger, Larry Brown, Shanti Gupta, Jean-Pierre
Habicht, George McCabe, Doug Robson, Bill Strawderman. And all of the
others that I have learned from. Thanks.

George Casella
Gainesville, Florida

February 18, 2008

You may be right
I may be crazy
But it just may be a lunatic you’re looking for.
Billy Joel
You may be right

I finally fixed it because I had, and still have, persistence.

Richard P. Feynman
Surely You’re Joking, Mr. Feynman
Contents

1 Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Variance and Covariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Partitioning Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5 Randomization, Layouts, and Designs . . . . . . . . . . . . . . . . . . . . . . 16
1.6 Replication: True and Technical . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.8 Technical Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
1.9 Miscellanea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

2 Completely Randomized Designs . . . . . . . . . . . . . . . . . . . . . . . . . . 43

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.1.1 A Oneway CRD Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.1.2 CRD and the Two-sample t-test . . . . . . . . . . . . . . . . . . . . . 44
2.1.3 CRD Anova . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.2 Model and Distribution Assumptions . . . . . . . . . . . . . . . . . . . . . . 48
2.3 Expected Squares and F -tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.4 Estimating Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.5 Deeper into Factorials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.5.1 Investigating Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.5.2 Higher-Order Factorials . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.6 Adjusting for Covariates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.8 Technical Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
2.9 Miscellanea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

3 Complete Block Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.1.1 An RCB Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.1.2 RCB and the Paired t-test . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.1.3 The RCB Anova . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
XIV Contents

3.2 Model and Distribution Assumptions . . . . . . . . . . . . . . . . . . . . . . 98

3.3 Expected Squares and F -tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.4 Estimating Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.5 Modeling the Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.6 Variations on a Theme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.6.1 Replicating the Experiment . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.6.2 Crossed Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.6.3 Latin Squares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.8 Technical Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
3.9 Miscellanea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

4 Interlude: Assessing the Effects of Blocking . . . . . . . . . . . . . . . . 145

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.2 Model and Distribution Assumptions . . . . . . . . . . . . . . . . . . . . . . 146
4.3 Expected Squares and F -tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.4 Estimating Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.5 Modeling the Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
4.6 Reconciliation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.8 Technical Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
4.9 Miscellanea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

5 Split Plot Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.1.1 A Split Plot Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.1.2 Dissecting the Split Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
5.2 CRD on the Whole Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
5.2.1 Model and Distribution Assumptions . . . . . . . . . . . . . . . . 175
5.2.2 Expected Squares and F -tests . . . . . . . . . . . . . . . . . . . . . . . 177
5.2.3 Estimating Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.3 RCB on the Whole Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
5.3.1 Model and Distribution Assumptions . . . . . . . . . . . . . . . . 185
5.3.2 Expected Squares and F -tests . . . . . . . . . . . . . . . . . . . . . . . 188
5.3.3 Estimating Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
5.4 Estimating Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
5.5 Splitting Twice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.6 Variations on a Theme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
5.6.1 Strip Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
5.6.2 Crossover Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
5.6.3 Repeated Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
5.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
5.8 Technical Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
5.9 Miscellanea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
 Contents XV

6 Confounding in Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
6.2 Balanced Incomplete Block Designs . . . . . . . . . . . . . . . . . . . . . . . . 248
6.2.1 Model and Distribution Assumptions . . . . . . . . . . . . . . . . 250
6.2.2 Estimating Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
6.3 Fractional Factorial Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
6.4 Variations on a Theme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
6.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
6.6 Miscellanea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

A Designs Illustrated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

List of Tables

1.1 Dry weight, in grams, of Geranium “Dilys” . . . . . . . . . . . . . . . . . . . . . 2

1.2 Strawberry RCB experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Randomization and anova . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.4 Field layouts and anova . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.1 Two-sample t-tests and anova contrast t-tests . . . . . . . . . . . . . . . . . . . 46

2.2 Oneway anova table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.3 Yields of red clover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4 Twoway anova table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5 Anova for red clover experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.6 Expected mean squares for twoway CRD anova . . . . . . . . . . . . . . . . . 52
2.7 Breakdown of interaction sum of squares . . . . . . . . . . . . . . . . . . . . . . . 56
2.8 Cell means for QTL patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.9 Error Rate Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

3.1 Yield data from strawberry varieties . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

3.2 RCB data layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.3 RCB anova table with subsampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.4 Alfalfa RCB data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.5 Anticoagulant drug data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.6 EMS for RCB anova with random blocks and no subsampling . . . . 102
3.7 EMS for RCB anova with random blocks and replication . . . . . . . . . 111
3.8 EMS for RCB anova with random blocks and replication,
ρε = ρτ β = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.9 RCB replicated with blocks nested in replications. . . . . . . . . . . . . . . 113
3.10 EMS for a replicated RCB design, blocks nested in replications . . . 114
3.11 EMS for a replicated RCB design, blocks crossed with replications 115
3.12 Anova for crossed blocks design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.13 EMS for a crossed block design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.14 Residual contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
XVIII List of Tables

4.1 EMS for RCB anova with fixed blocks – no subsampling . . . . . . . . . 148
4.2 Comparison of the assumptions of two RCB models . . . . . . . . . . . . . 149
4.3 EMS for RCB anova with fixed blocks and replication . . . . . . . . . . . 154
4.4 Comparison of EMS from RCB anovas with fixed and random
blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

5.1 Anova for split plot designwith CRD on whole plot treatments . . . . 176
5.2 Data layout for split plot designwith CRD on whole plot
treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
5.3 EMS for a split plot design with the whole plots in a CRD . . . . . . . 179
5.4 Field layoutfor split plot experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
5.5 Anova for split plot designwith RCB on whole plot treatments . . . . 187
5.6 Data layoutfor split plot design with RCB on whole plot
treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
5.7 EMS for a split plot design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
5.8 Anova for split split plot designwith CRD on whole plot
treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
5.9 EMS for a split split plot designwith the whole plots in a CRD . . . 200
5.10 EMS for a split split plot designwith the whole plots in a RCB . . . 204
5.11 EMS for a strip plot design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

6.1 Anova tables, 14 incomplete blocks and 7 complete blocks . . . . . . . . 248

6.2 Field layout for a balanced lattice square . . . . . . . . . . . . . . . . . . . . . . . 263
6.3 Reference design and BIBD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
List of Figures

2.1 Interaction plot for fish tissue experiment . . . . . . . . . . . . . . . . . . . . . . 55

2.2 Interaction plot for red clover experiment . . . . . . . . . . . . . . . . . . . . . . 56
2.3 QTL interaction patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.4 Least squares fit for corn yield data . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.5 Adjusted and unadjusted treatment means . . . . . . . . . . . . . . . . . . . . . 68

3.1 Interaction plot of the anticoagulant data . . . . . . . . . . . . . . . . . . . . . . 103

3.2 Relationship of full factorial crossed blocks design
with a Latinsquare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

4.1 Randomization distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

5.1 The three experimental units of a strip plot design . . . . . . . . . . . . . . 209

5.2 Strip plot bioassay experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

6.1 Fractions of a 23 factorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

6.2 Loop designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
6.3 Graphical display of a loop design and BIBD . . . . . . . . . . . . . . . . . . . 271
List of Examples

1.1 Oneway model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Experimental unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 Fish tank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.8 RCB-strawberries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.10 Contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.13 Rehabilitation time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.15 Rehabilitation time continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.16 Fish microarray . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.17 Problematic inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.18 Randomization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.20 Crossed or nested . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.21 Crossing subjects and treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.22 Data layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.23 Field layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.24 Field layout continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.26 Twoway F -ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.27 Irradiation microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.28 Pseudoreplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1 Oneway CRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.2 Oneway CRD continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.3 Red clover CRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.4 Red clover CRD continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5 Fish microarrays revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.6 Red clover CRD revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.7 QTL epistasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.8 2 × 2 factorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.9 Interaction contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.10 Social science experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.11 A 24 factorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.12 Ancova examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
XXII List of Examples

2.13 Corn yield covariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

2.14 Corn yield covariance continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.15 Contrasts for corn yield experiment . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.1 Strawberry blocks revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.2 Blocks as a fixed factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.3 Strawberry blocks – RCB analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.4 Alfalfa blocks - RCB analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.6 Estimation of random effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.8 RCB interaction plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
3.9 Pairwise comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.11 Replicated RCBs, nested . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.12 Replicated RCBs, nested, continued . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.13 Crossed blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.14 Peanut Latin square . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
3.15 Replicating Latin squares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.1 Fixed or random blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.2 Greenhouse experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
4.3 Fixed blocks with interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
4.4 Randomization tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.5 Randomization tests continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
5.1 Dietary split plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
5.2 Dietary split plot continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
5.3 Variety split plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
5.4 Variety split plot continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
5.5 Variety split plot concluded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
5.6 Diet split plot concluded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.7 Ozone split split plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
5.8 Ozone split split plot continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
5.9 RCB split split plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
5.10 Solution data split split . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
5.11 Randomization patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
5.12 Strip plot experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
5.13 Strip plot bioassay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
5.14 Exercise crossover design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
5.15 Three-period crossover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
5.16 Hypertension repeated measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
5.17 Hypertension repeated measures continued . . . . . . . . . . . . . . . . . . . 217
6.1 Diet and blood pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
6.2 Diet and blood pressure continued . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
6.3 Balancing the blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
6.5 Very simple BIBD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
6.6 Very simple BIBD continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
6.7 Estimates in a BIBD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
6.8 Orthopedic BIBD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
6.9 A simple fractional factorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
 List of Examples XXIII

6.10 A simple fractional factorial continued . . . . . . . . . . . . . . . . . . . . . . . 255

6.12 A simple fractional factorial concluded . . . . . . . . . . . . . . . . . . . . . . . 256
6.13 Complete factorial in blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
6.14 A classic fraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
6.15 A classic fraction continued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
6.16 A classic fraction concluded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
6.17 Fractioning a 25 factorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
6.19 A balanced lattice square . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
6.20 A microarray Latin square . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
6.21 Microarray BIBD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
6.22 Avocado gene expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271