Download the full version of the textbook now at textbookfull.

com

Practical multivariate analysis 6th Edition

Afifi

https://textbookfull.com/product/practical-
multivariate-analysis-6th-edition-afifi-2/

Explore and download more textbook at https://textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Practical multivariate analysis 6th Edition Afifi

https://textbookfull.com/product/practical-multivariate-analysis-6th-
edition-afifi-2/

textbookfull.com

Multivariate Data Analysis Joseph F. Hair

https://textbookfull.com/product/multivariate-data-analysis-joseph-f-
hair/

textbookfull.com

Multivariate Analysis for the Behavioral Sciences Kimmo

Vehkalahti

https://textbookfull.com/product/multivariate-analysis-for-the-
behavioral-sciences-kimmo-vehkalahti/

textbookfull.com

Principles of Microbial Diversity 1st Edition James W.

Brown

https://textbookfull.com/product/principles-of-microbial-
diversity-1st-edition-james-w-brown/

textbookfull.com
Learning ACT for Group Treatment An Acceptance and
Commitment Therapy skills training manual for Therapists
1st Edition Darrah Westrup
https://textbookfull.com/product/learning-act-for-group-treatment-an-
acceptance-and-commitment-therapy-skills-training-manual-for-
therapists-1st-edition-darrah-westrup/
textbookfull.com

Interpretation and Application of UK GAAP: For Accounting

Periods Commencing On or After 1 January 2015 1st Edition
Steven Collings
https://textbookfull.com/product/interpretation-and-application-of-uk-
gaap-for-accounting-periods-commencing-on-or-after-1-january-2015-1st-
edition-steven-collings/
textbookfull.com

Arthropoda 1st Edition Xenia Melzer [Melzer

https://textbookfull.com/product/arthropoda-1st-edition-xenia-melzer-
melzer/

textbookfull.com

Topology and Physics of Circular DNA (1992) First Edition

Vologodskii

https://textbookfull.com/product/topology-and-physics-of-circular-
dna-1992-first-edition-vologodskii/

textbookfull.com

Antisocial Narcissistic and Borderline Personality

Disorders A New Conceptualization of Development
Reinforcement Expression and Treatment 1st Edition Daniel
J. Fox
https://textbookfull.com/product/antisocial-narcissistic-and-
borderline-personality-disorders-a-new-conceptualization-of-
development-reinforcement-expression-and-treatment-1st-edition-daniel-
j-fox/
textbookfull.com
Building codes illustrated : a guide to understanding the
2015 international building code® Fifth Edition, Updated
And Expanded Edition Ching
https://textbookfull.com/product/building-codes-illustrated-a-guide-
to-understanding-the-2015-international-building-code-fifth-edition-
updated-and-expanded-edition-ching/
textbookfull.com
Practical Multivariate
Analysis
Sixth Edition
Practical Multivariate
Analysis
Sixth Edition

Abdelmonem Afifi
Susanne May
Robin A. Donatello
Virginia A. Clark
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

c 2020 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed on acid-free paper

International Standard Book Number-13: 978-1-138-70222-6 (Hardback)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made
to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all
materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all
material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been
obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future
reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized
in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, mi-
crofilming, and recording, or in any information storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identi-
fication and explanation without intent to infringe.

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com
 Contents

Preface xi

Authors xv

I Preparation for Analysis 1

1 What is multivariate analysis? 3
1.1 Defining multivariate analysis 3
1.2 Examples of multivariate analyses 3
1.3 Exploratory versus confirmatory analyses 6
1.4 Multivariate analyses discussed in this book 6
1.5 Organization and content of the book 9

2 Characterizing data for analysis 11

2.1 Variables: Their definition, classification, and use 11
2.2 Defining statistical variables 11
2.3 Stevens’s classification of variables 12
2.4 How variables are used in data analysis 14
2.5 Examples of classifying variables 15
2.6 Other characteristics of data 15
2.7 Summary 15
2.8 Problems 15

3 Preparing for data analysis 17

3.1 Processing data so they can be analyzed 17
3.2 Choice of a statistical package 18
3.3 Techniques for data entry 19
3.4 Organizing the data 23
3.5 Reproducible research and literate programming 29
3.6 Example: depression study 31
3.7 Summary 33
3.8 Problems 33

4 Data visualization 37
4.1 Introduction 37
4.2 Univariate data 38
4.3 Bivariate data 45
4.4 Multivariate data 50
4.5 Discussion of computer programs 52
4.6 What to watch out for 54
4.7 Summary 56
4.8 Problems 56

v
vi CONTENTS
5 Data screening and transformations 59
5.1 Transformations, assessing normality and independence 59
5.2 Common transformations 59
5.3 Selecting appropriate transformations 62
5.4 Assessing independence 69
5.5 Discussion of computer programs 71
5.6 Summary 71
5.7 Problems 72

6 Selecting appropriate analyses 75

6.1 Which analyses to perform? 75
6.2 Why selection is often difficult 75
6.3 Appropriate statistical measures 76
6.4 Selecting appropriate multivariate analyses 79
6.5 Summary 80
6.6 Problems 80

II Regression Analysis 85
7 Simple regression and correlation 87
7.1 Chapter outline 87
7.2 When are regression and correlation used? 87
7.3 Data example 88
7.4 Regression methods: fixed-X case 89
7.5 Regression and correlation: variable-X case 93
7.6 Interpretation: fixed-X case 93
7.7 Interpretation: variable-X case 94
7.8 Other available computer output 98
7.9 Robustness and transformations for regression 103
7.10 Other types of regression 105
7.11 Special applications of regression 107
7.12 Discussion of computer programs 110
7.13 What to watch out for 110
7.14 Summary 112
7.15 Problems 112

8 Multiple regression and correlation 115

8.1 Chapter outline 115
8.2 When are regression and correlation used? 115
8.3 Data example 116
8.4 Regression methods: fixed-X case 117
8.5 Regression and correlation: variable-X case 119
8.6 Interpretation: fixed-X case 124
8.7 Interpretation: variable-X case 126
8.8 Regression diagnostics and transformations 128
8.9 Other options in computer programs 132
8.10 Discussion of computer programs 136
8.11 What to watch out for 140
8.12 Summary 140
8.13 Problems 141
CONTENTS vii
9 Variable selection in regression 145
9.1 Chapter outline 145
9.2 When are variable selection methods used? 145
9.3 Data example 147
9.4 Criteria for variable selection 149
9.5 A general F test 152
9.6 Stepwise regression 153
9.7 Lasso regression 159
9.8 Discussion of computer programs 163
9.9 Discussion of strategies 164
9.10 What to watch out for 166
9.11 Summary 167
9.12 Problems 168

10 Special regression topics 171

10.1 Chapter outline 171
10.2 Missing values in regression analysis 171
10.3 Dummy variables 177
10.4 Constraints on parameters 184
10.5 Regression analysis with multicollinearity 186
10.6 Ridge regression 187
10.7 Summary 190
10.8 Problems 191

11 Discriminant analysis 195

11.1 Chapter outline 195
11.2 When is discriminant analysis used? 195
11.3 Data example 196
11.4 Basic concepts of classification 197
11.5 Theoretical background 202
11.6 Interpretation 204
11.7 Adjusting the dividing point 207
11.8 How good is the discrimination? 209
11.9 Testing variable contributions 210
11.10 Variable selection 211
11.11 Discussion of computer programs 211
11.12 What to watch out for 212
11.13 Summary 214
11.14 Problems 214

12 Logistic regression 217

12.1 Chapter outline 217
12.2 When is logistic regression used? 217
12.3 Data example 218
12.4 Basic concepts of logistic regression 219
12.5 Interpretation: categorical variables 220
12.6 Interpretation: continuous variables 222
12.7 Interpretation: interactions 223
12.8 Refining and evaluating logistic regression 229
12.9 Nominal and ordinal logistic regression 238
12.10 Applications of logistic regression 243
12.11 Poisson regression 246
12.12 Discussion of computer programs 249
viii CONTENTS
12.13 What to watch out for 249
12.14 Summary 251
12.15 Problems 252

13 Regression analysis with survival data 255

13.1 Chapter outline 255
13.2 When is survival analysis used? 255
13.3 Data examples 256
13.4 Survival functions 256
13.5 Common survival distributions 262
13.6 Comparing survival among groups 262
13.7 The log-linear regression model 264
13.8 The Cox regression model 266
13.9 Comparing regression models 274
13.10 Discussion of computer programs 276
13.11 What to watch out for 276
13.12 Summary 278
13.13 Problems 278

14 Principal components analysis 281

14.1 Chapter outline 281
14.2 When is principal components analysis used? 281
14.3 Data example 282
14.4 Basic concepts 282
14.5 Interpretation 285
14.6 Other uses 292
14.7 Discussion of computer programs 294
14.8 What to watch out for 294
14.9 Summary 295
14.10 Problems 296

15 Factor analysis 297

15.1 Chapter outline 297
15.2 When is factor analysis used? 297
15.3 Data example 298
15.4 Basic concepts 298
15.5 Initial extraction: principal components 300
15.6 Initial extraction: iterated components 303
15.7 Factor rotations 305
15.8 Assigning factor scores 309
15.9 Application of factor analysis 310
15.10 Discussion of computer programs 310
15.11 What to watch out for 312
15.12 Summary 313
15.13 Problems 314

16 Cluster analysis 317

16.1 Chapter outline 317
16.2 When is cluster analysis used? 317
16.3 Data example 318
16.4 Basic concepts: initial analysis 318
16.5 Analytical clustering techniques 324
CONTENTS ix
16.6 Cluster analysis for financial data set 328
16.7 Discussion of computer programs 333
16.8 What to watch out for 336
16.9 Summary 336
16.10 Problems 336

17 Log-linear analysis 339

17.1 Chapter outline 339
17.2 When is log-linear analysis used? 339
17.3 Data example 340
17.4 Notation and sample considerations 341
17.5 Tests and models for two-way tables 343
17.6 Example of a two-way table 345
17.7 Models for multiway tables 347
17.8 Exploratory model building 350
17.9 Assessing specific models 354
17.10 Sample size issues 355
17.11 The logit model 356
17.12 Discussion of computer programs 358
17.13 What to watch out for 358
17.14 Summary 359
17.15Problems 360

18 Correlated outcomes regression 361

18.1 Chapter outline 361
18.2 When is correlated outcomes regression used? 361
18.3 Data examples 362
18.4 Basic concepts 364
18.5 Regression of clustered data with a continuous outcome 369
18.6 Regression of clustered data with a binary outcome 373
18.7 Regression of longitudinal data 375
18.8 Generalized estimating equations analysis of correlated data 379
18.9 Discussion of computer programs 383
18.10 What to watch out for 385
18.11 Summary 386
18.12 Problems 386

Appendix A 389
A.1 Data sets and how to obtain them 389
A.2 Chemical companies’ financial data 389
A.3 Depression study data 389
A.4 Financial performance cluster–analysis data 389
A.5 Lung cancer survival data 390
A.6 Lung function data 390
A.7 Parental HIV data 390
A.8 Northridge earthquake data 391
A.9 School data 391
A.10 Mice data 391

Bibliography 393

Index 411
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
 Preface

The first edition of this book appeared in 1984 under the title “Computer Aided Multivariate Anal-
ysis.” The title was chosen in order to distinguish it from other books that were more theoretically
oriented. By the time we published the fifth edition in 2012, it was impossible to think of a book on
multivariate analysis for scientists and applied researchers that is not computer oriented. We there-
fore decided at that time to change the title to “Practical Multivariate Analysis” to better characterize
the nature of the book. Today, we are pleased to present the sixth edition.
We wrote this book for investigators, specifically behavioral scientists, biomedical scientists,
and industrial or academic researchers, who wish to perform multivariate statistical analyses and
understand the results. We expect the readers to be able to perform and understand the results,
but also expect them to know when to ask for help from an expert on the subject. The book can
either be used as a self-guided textbook or as a text in an applied course in multivariate analysis.
In addition, we believe that the book can be helpful to many statisticians who have been trained
in conventional mathematical statistics who are now working as statistical consultants and need to
explain multivariate statistical concepts to clients with a limited background in mathematics.
We do not present mathematical derivations of the techniques; rather we rely on geometric and
graphical arguments and on examples to illustrate them. The mathematical level has been deliber-
ately kept low. While the derivations of the techniques are referenced, we concentrate on applica-
tions to real-life problems, which we feel are the ‘fun’ part of multivariate analysis. To this end, we
assume that the reader will use a packaged software program to perform the analysis. We discuss
specifically how each of four popular and comprehensive software packages can be used for this
purpose. These packages are R, SAS, SPSS, and STATA. The book can be used, however, in con-
junction with all other software packages since our presentation explains the output of most standard
statistical programs.
We assume that the reader has taken a basic course in statistics that includes tests of hypotheses
and covers one-way analysis of variance.

Approach of this book

We wrote the book in a modular fashion. Part One, consisting of six chapters, provides examples
of studies requiring multivariate analysis techniques, and discusses characterizing data for analysis,
computer programs, data entry, data management, data clean-up, missing values, and transforma-
tions. It also includes a new chapter on graphics and data visualization and presents a rough guide
to assist in the choice of an appropriate multivariate analysis. We included these topics since many
investigators have more difficulty with these preliminary steps than with running the multivariate
analyses themselves. Also, if these steps are not done with care, the results of the statistical analysis
can be faulty.
In the rest of the chapters, we follow a standard format. The first four sections of each chapter
include a discussion of when the technique is used, a data example, and the basic assumptions and
concepts of the technique. In subsequent sections, we present more detailed aspects of the analysis.
At the end of each chapter, we give a summary table showing which features are available in the
four software packages. We also include a section entitled ‘What to watch out for’ to warn the reader
about common problems related to data analysis. In those sections, we rely on our own experiences
in consulting and those detailed in the literature to supplement the formal treatment of the subject.

xi
xii PREFACE
Part Two covers regression analysis. Chapter 7 deals with simple linear regression and is in-
cluded for review purposes to introduce our notation and to provide a more complete discussion
of outliers and diagnostics than is found in some elementary texts. Chapters 8-10 are concerned
with multiple linear regression. Multiple linear regression is used very heavily in practice and pro-
vides the foundation for understanding many concepts relating to residual analysis, transformations,
choice of variables, missing values, dummy variables, and multicollinearity. Since these concepts
are essential to a good grasp of multivariate analysis, we thought it useful to include these chapters
in the book.
Chapters 11-18 might be considered the heart of multivariate analysis. They include chapters on
discriminant analysis, logistic regression analysis, survival analysis, principal components analysis,
factor analysis, cluster analysis, log-linear analysis, and correlated outcomes regression. The mul-
tivariate analyses have been discussed more as separate techniques than as special cases of some
general framework. The advantage of this approach is that it allows us to concentrate on explaining
how to analyze a certain type of data from readily available computer programs to answer realistic
questions. It also enables the reader to approach each chapter independently. We did include inter-
spersed discussions of how the different analyses relate to each other in an effort to describe the ‘big
picture’ of multivariate analysis.

How to use the book

We have received many helpful suggestions from instructors and reviewers on how to order these
chapters for reading or teaching purposes. For example, one instructor uses the following order in
teaching: principal components, factor analysis, and then cluster analysis. Another prefers present-
ing a detailed treatment of multiple regression followed by logistic regression and survival analysis.
Instructors and self-learning readers have a wide choice of other orderings of the material because
the chapters are largely self contained.

What’s new in the Sixth Edition

During the nearly thirty-six years since we wrote the first edition of this book, tremendous advances
have taken place in the field of computing and software development. These advances have made
it possible to quickly perform any of the multivariate analyses that were available only in theory at
that time. They also spurred the invention of new multivariate analyses as well as new options for
many of the standard methods. In this edition, we have taken advantage of these developments and
made many changes as described below.
For each of the techniques discussed, we used the most recent software versions available and
discussed the most modern ways of performing the analysis. In each chapter, we updated the ref-
erences to today’s literature (while still including the fundamental original references). In terms
of statistical software, we discontinued description of S-Plus because of the more wide-spread use
of the similar package R. Also, we no longer include Statistica since it is largely not used by our
intended readers.
In addition to the above-described modifications, we included comments to distinguish between
exploratory and confirmatory analyses in Chapter 1 and throughout the book. We also expanded
the discussion of missing values in Chapter 3 and added a discussion of literate programming and
reproducible research.
As mentioned above, we added a new chapter (Chapter 4) on graphics and data visualization.
In Chapter 9, we updated our discussion of variable selection and added a description of Lasso,
a more recent method than the ones already included. In Chapter 10, we added a description of
MICE, a multiple imputation approach for dealing with missing values. In Chapter 18, we added a
description of the generalized estimating equations (GEE) method for handling correlated data and
compared it to the mixed model approach. Finally, in each chapter we updated and/or expanded the
summary table of the options available in the four statistical packages to make it consistent with the
most recent software versions.
PREFACE xiii
Data sets used for examples and problems are described throughout the book as needed and
summarized in Appendix A. Two web sites are also available. The first one is the CRC web site:
http://www.crcpress.com/product/isbn/9781138702226. From this site, you can down-
load all the data sets used in the book by clicking on the Downloads/Updates tab. The other web
site that is available to all readers is: https://stats.idre.ucla.edu/other/examples/pma6.
This site, developed by the UCLA Institute for Digital Research and Education (IDRE), includes
the data sets in the formats of various statistical software packages available in the links included in
the Appendix A part of the table of contents in that web page. It also includes illustrations of exam-
ples in most chapters, complete with code for three of the four software packages used in the book.
Please note that the current site is done for the 5th edition and it is hoped that it will be updated for
the sixth edition. We encourage readers to obtain data from either web site and frequently refer to
the solutions given in the UCLA web site for practice.

Acknowledgements
We would like to express our appreciation to our colleagues and former students and staff that
helped us over the years, both in the planning and preparation of the various editions. These include
our colleagues Drs. Carol Aneshensel, Roger Detels, Robert Elashoff, Ralph Frerichs, Mary Ann
Hill, and Roberta Madison. Our former students include Drs. Stella Grosser, Luohua Jiang, Jack
Lee, Steven Lewis, Tim Morgan, Leanne Streja, and David Zhang. Our former staff includes Ms.
Dorothy Breininger, Jackie Champion, and Anne Eiseman. In addition, we would like to thank Ms.
Meike Jantzen and Mr. Jack Fogliasso for their help with the references and typesetting.
We also thank Rob Calver and Lara Spieker from CRC Press for their very capable assistance
in the preparation of the sixth edition.
We especially appreciate the efforts of the staff of the UCLA Institute for Digital Research and
Education in putting together the UCLA web site of examples from the book (referenced above).
Our deep gratitude goes to our spouses, Marianne Afifi, Bruce Jacobson, Ian Donatello, and
Welden Clark, for their patience and encouragement throughout the stages of conception, writing,
and production of the book. Special thanks go to Welden Clark for his expert assistance and trou-
bleshooting of earlier electronic versions of the manuscript.

Abdelmonem Afifi
Susanne May
Robin A. Donatello
Virginia A. Clark
 Authors

Abdelmonem Afifi, Ph.D., has been Professor of Biostatistics in the School of Public Health, Uni-
versity of California, Los Angeles (UCLA) since 1965, and served as the Dean of the School from
1985 until 2000. His research includes multivariate and multilevel data analysis, handling miss-
ing observations in regression and discriminant analyses, meta-analysis, and model selection. Over
the years, he taught well-attended courses in biostatistics for public health students and clinical re-
search physicians, and doctoral-level courses in multivariate statistics and multilevel modeling. He
has authored many publications in statistics and health related fields, including two widely used
books (with multiple editions) on multivariate analysis. He received several prestigious awards for
excellence in teaching and research.

Susanne May, Ph.D., is a Professor in the Department of Biostatistics at the University of Wash-
ington in Seattle. Her areas of expertise and interest include clinical trials, survival analysis, and
longitudinal data analysis. She has more than 20 years of experience as a statistical collaborator
and consultant on health related research projects. In addition to a number of methodological and
applied publications, she is a coauthor (with Drs. Hosmer and Lemeshow) of Applied Survival
Analysis: Regression Modeling of Time-to-Event Data. Dr. May has taught courses on introductory
statistics, clinical trials, and survival analysis.

Robin A. Donatello, Dr.P.H., is an Associate Professor in the Department of Mathematics and

Statistics and the developer of the Data Science Initiative at California State University, Chico. Her
areas of interest include applied research in the public health and natural science fields. She has
expertise in data visualization, techniques to address missing and erroneous data, implementing
reproducible research workflows, computational statistics, and data science. Dr. Donatello teaches
undergraduate and graduate level courses in statistical programming, applied statistics, and data
science.

Virginia A. Clark, Ph.D., was professor emerita of Biostatistics and Biomathematics at UCLA.
For 27 years, she taught courses in multivariate analysis and survival analysis, among others. In
addition to this book, she is coauthor of four books on survival analysis, linear models and analysis
of variance, and survey research, as well as an introductory book on biostatistics. She published
extensively in statistical and health science journals.

xv
 Part I

Preparation for Analysis

1
 Chapter 1

What is multivariate analysis?

1.1 Defining multivariate analysis

The expression multivariate analysis is used to describe analyses of data that are multivariate in
the sense that numerous observations or variables are obtained for each individual or unit studied. In
a typical survey 30 to 100 questions are asked of each respondent. In describing the financial status
of a company, an investor may wish to examine five to ten measures of the company’s performance.
Commonly, the answers to some of these measures are interrelated. The challenge of disentangling
complicated interrelationships among various measures on the same individual or unit and of inter-
preting these results is what makes multivariate analysis a rewarding activity for the investigator.
Often results are obtained that could not be attained without multivariate analysis.
In the next section of this chapter several studies are described in which the use of multivariate
analysis is essential to understanding the underlying problem. Section 1.3 provides a rationale for
making a distinction between confirmatory and exploratory analyses. Section 1.4 gives a listing and
a very brief description of the multivariate analysis techniques discussed in this book. Section 1.5
then outlines the organization of the book.

1.2 Examples of multivariate analyses

The studies described in the following subsections illustrate various multivariate analysis tech-
niques. These are used later in the book as examples.

Depression study example

The data for the depression study have been obtained from a complex, random, multiethnic sample
of 1000 adult residents of Los Angeles County. The study was a panel or longitudinal design where
the same respondents were interviewed four times between May 1979 and July 1980. About three-
fourths of the respondents were re-interviewed for all four interviews. The field work for the survey
was conducted by professional interviewers from the Institute for Social Science Research at the
University of California in Los Angeles.
This research is an epidemiological study of depression and help-seeking behavior among free-
living (noninstitutionalized) adults. The major objectives are to provide estimates of the prevalence
and incidence of depression and to identify causal factors and outcomes associated with this con-
dition. The factors examined include demographic variables, life events stressors, physical health
status, health care use, medication use, lifestyle, and social support networks. The major instrument
used for classifying depression is the Depression Index (CESD) of the National Institute of Mental
Health, Center of Epidemiological Studies. A discussion of this index and the resulting prevalence
of depression in this sample is given in Frerichs et al. (1981).
The longitudinal design of the study offers advantages for assessing causal priorities since the
time sequence allows us to rule out certain potential causal links. Nonexperimental data of this type
cannot directly be used to establish causal relationships, but models based on an explicit theoretical

3
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
4 CHAPTER 1. WHAT IS MULTIVARIATE ANALYSIS?
framework can be tested to determine if they are consistent with the data. An example of such model
testing is given in Aneshensel and Frerichs (1982).
Data from the first time period of the depression study are described in Chapter 3. Only a subset
of the factors measured on a subsample of the respondents is included in this book’s web site in
order to keep the data set easily comprehensible. These data are used several times in subsequent
chapters to illustrate some of the multivariate techniques presented in this book.

Parental HIV study

The data from the parental HIV study have been obtained from a clinical trial to evaluate an inter-
vention given to increase coping skills (Rotheram-Borus et al., 2001). The purpose of the interven-
tion was to improve behavioral, social, and health outcomes for parents with HIV/AIDS and their
children. Parents and their adolescent children were recruited from the New York City Division of
Aids Services (DAS). Adolescents were eligible for the study if they were between the ages of 11
and 18 and if the parents and adolescents had given informed consent. Individual interviews were
conducted every three months for the first two years and every six months thereafter. Information
obtained in the interviews included background characteristics, sexual behavior, alcohol and drug
use, medical and reproductive history, and a number of psychological scales.
A subset of the data from the study is available on this book’s web site. To protect the identity of
the participating adolescents we used the following procedures. We randomly chose one adolescent
per family. In addition, we reduced the sample further by choosing a random subset of the original
sample. Adolescent case numbers were assigned randomly without regard to the original order or
any other numbers in the original data set.
Data from the baseline assessment will be used for problems as well as to illustrate various
multivariate analysis techniques.

Northridge earthquake study

On the morning of January 17, 1994 a magnitude 6.7 earthquake centered in Northridge, CA awoke
Los Angeles and Ventura County residents. Between August 1994 and May 1996, 1830 residents
were interviewed about what happened to them in the earthquake. The study uses a telephone sur-
vey lasting approximately 48 minutes to assess the residents’ experiences in and responses to the
Northridge earthquake. Data from 506 residents are included in the data set posted on the book web
site, and described in Appendix A.
Subjects were asked where they were, how they reacted, where they obtained information,
whether their property was damaged or whether they experienced injury, and what agencies they
were in contact with. The questionnaire included the Brief Symptom Inventory (BSI), a measure
of psychological functioning used in community studies, and questions on emotional distress. Sub-
jects were also asked about the impact of the damage to the transportation system as a result of the
earthquake. Investigators not only wanted to learn about the experiences of the Southern California
residents in the Northridge earthquake, but also wished to compare their findings to similar studies
of the Los Angeles residents surveyed after the Whittier Narrows earthquake on October 1, 1987,
and Bay Area residents interviewed after the Loma Prieta earthquake on October 17, 1989.
The Northridge earthquake data set is used in problems at the end of several chapters of the book
to illustrate a number of multivariate techniques. Multivariate analyses of these data include, for
example, exploring pre- and post-earthquake preparedness activities as well as taking into account
several factors relating to the subject and the property (Nguyen et al., 2006).

Bank loan study

The managers of a bank need some way to improve their prediction of which borrowers will suc-
cessfully pay back a type of bank loan. They have data from the past on the characteristics of persons
1.2. EXAMPLES OF MULTIVARIATE ANALYSES 5
to whom the bank has lent money and the subsequent record of how well the person has repaid the
loan. Loan payers can be classified into several types: those who met all of the terms of the loan,
those who eventually repaid the loan but often did not meet deadlines, and those who simply de-
faulted. They also have information on age, sex, income, other indebtedness, length of residence,
type of residence, family size, occupation, and the reason for the loan. The question is, can a simple
rating system be devised that will help the bank personnel improve their prediction rate and lessen
the time it takes to approve loans? The methods described in Chapter 12 and Chapter 13 can be used
to answer this question.

Lung function study

The purpose of this lung function study of chronic respiratory disease is to determine the effects
of various types of smog on lung function of children and adults in the Los Angeles area. Because
they could not randomly assign people to live in areas that had different levels of pollutants, the
investigators were very concerned about the interaction that might exist between the locations where
persons chose to live and their values on various lung function tests. The investigators picked four
areas of quite different types of air pollution and measured various demographic and other responses
on all persons over seven years old who live there. These areas were chosen so that they are close to
an air-monitoring station.
The researchers took measurements at two points in time and used the change in lung function
over time as well as the levels at the two periods as outcome measures to assess the effects of air
pollution. The investigators had to do the lung function tests by using a mobile unit in the field, and
much effort went into problems of validating the accuracy of the field observations. A discussion
of the particular lung function measurements used for one of the four areas can be found in Detels
et al. (1975). In the analysis of the data, adjustments must be made for sex, age, height, and smoking
status of each person.
Over 15,000 respondents have been examined and interviewed in this study. The data set is being
used to answer numerous questions concerning effects of air pollution, smoking, occupation, etc. on
different lung function measurements. For example, since the investigators obtained measurements
on all family members seven years old and older, it is possible to assess the effects of having parents
who smoke on the lung function of their children (Tashkin et al., 1984). Studies of this type require
multivariate analyses so that investigators can arrive at plausible scientific conclusions that could
explain the resulting lung function levels.
This data set is described in Appendix A. Lung function and associated data for nonsmoking
families for the father, mother, and up to three children ages 7–17 are available from the book’s web
site.

School data set

The school data set is a publicly available data set that is provided by the National Center
for Educational Statistics. The data come from the National Education Longitudinal Study of
1988 (called NELS:88). The study collected data beginning with 8th graders and conducted ini-
tial interviews and four follow-up interviews which were performed every other year. The data
used here contain only initial interview data. They represent a random subsample of 23 schools
with 519 students out of more than a thousand schools with almost twenty five thousand stu-
dents. Extensive documentation of all aspects of the study is available at the following web site:
http://nces.ed.gov/surveys/NELS88/. The longitudinal component of NELS:88 has been
used to investigate change in students’ lives and school-related and other outcomes. The focus
on the initial interview data provides the opportunity to examine associations between school and
student-related factors and students’ academic performance in a cross-sectional manner. This type
of analysis will be illustrated in Chapter 18.
6 CHAPTER 1. WHAT IS MULTIVARIATE ANALYSIS?
1.3 Exploratory versus confirmatory analyses
A crucial component for most research studies and analyses is the testing of hypotheses. For some
types of studies, hypotheses are specified in detail prior to study start (a priori) and then remain
unchanged. This is typically the case, e.g., for clinical trials and other designed experiments. For
other types of studies, some hypotheses might be specified in advance while others are generated
only after study start and potentially after reviewing some or all of the study data. This is often
the case for observational studies. In this section, we make a distinction between two conceptually
different approaches to analysis and reporting based on whether the primary goal of a study is to
confirm prespecified hypotheses or to explore hypotheses that have not been prespecified.
The following is a motivating example provided by Fleming (2010). He describes an experience
where he walked into a maternity ward (when they still had such) while visiting a friend who had
just given birth. He noticed that there were 22 babies, but only 2 of one gender while the other 20
were of the other gender. As a statistician, he dutifully calculated the p-value for the likelihood of
seeing such (or worse) imbalance if in truth there are 50% of each. The two-sided p-value turns out
to be 0.0001, indicating a very small likelihood (1 in 10,000) of such or more extreme imbalance
being observed if in truth there are 50% of each. This is an example of where the hypothesis was
generated after seeing the data. We will call such hypotheses exploratory.
Following Fleming, researchers might want to go out and test an exploratory hypothesis in an-
other setting or with new data. In the example above, one might want to go to another maternity
ward to collect further evidence of a strong imbalance in gender distribution at birth. Imagine that
in a second (confirmatory) maternity ward there might be exactly equal numbers for each gender
(e.g. 11 boys and 11 girls). Testing the same hypothesis in this setting will not yield any statistically
significant difference from the presumed 50%. Nevertheless, one might be tempted to simply com-
bine the two studies. A corresponding two-sided p-value remains statistically significant (p-value
< 0.01).
The above example might appear silly, because few researchers will believe that the distribution
of gender at birth (without human interference) is very different from 50%. Nevertheless, there
are many published research articles which test and present the results for hypotheses that were
generated by looking at data and noticing ‘unusual’ results. Without a clear distinction between
whether hypotheses were specified a priori or not, it is difficult to interpret the p-values provided.
Results from confirmatory analyses provide much stronger evidence than results from ex-
ploratory analyses. Accordingly, interpretation of results from confirmatory analyses can be stated
using much stronger language than interpretation of results from exploratory analyses. Furthermore,
results from exploratory analyses should not be combined with results from confirmatory analyses
(e.g. in meta analyses), because the random high bias (Fleming, 2010) will remain (albeit atten-
uated). To avoid random high bias when combining data or estimates from multiple studies only
data/estimates from confirmatory analyses should be combined. However, this requires clear iden-
tification of whether confirmatory or exploratory analysis was performed for each individual study
and/or analysis.
Many authors have pointed out that the medical literature is replete with studies that cannot be
reproduced (Breslow, 1999; Munafò et al., 2017). As argued by Breslow (1999), reproducibility of
studies, and in particular epidemiologic studies, can be improved if hypotheses are specified a priori
and the nature of the study (exploratory versus confirmatory) is clearly specified.
Throughout this book, we distinguish between the two approaches to multivariate analyses and
presentations of results and provide examples for each.

1.4 Multivariate analyses discussed in this book

In this section a brief description of the major multivariate techniques covered in this book is pre-
sented. To keep the statistical vocabulary to a minimum, we illustrate the descriptions by examples.
1.4. MULTIVARIATE ANALYSES DISCUSSED IN THIS BOOK 7
Simple linear regression
A nutritionist wishes to study the effects of early calcium intake on the bone density of post-
menopausal women. She can measure the bone density of the arm (radial bone), in grams per square
centimeter, by using a noninvasive device. Women who are at risk of hip fractures because of too
low a bone density will tend to show low arm bone density also. The nutritionist intends to sample a
group of elderly churchgoing women. For women over 65 years of age, she will plot calcium intake
as a teenager (obtained by asking the women about their consumption of high-calcium foods during
their teens) on the horizontal axis and arm bone density (measured) on the vertical axis. She expects
the radial bone density to be lower in women who had a lower calcium intake. The nutritionist plans
to fit a simple linear regression equation and test whether the slope of the regression line is zero. In
this example a single outcome factor is being predicted by a single predictor factor.
Simple linear regression as used in this case would not be considered multivariate by some
statisticians, but it is included in this book to introduce the topic of multiple regression.

Multiple linear regression

A manager is interested in determining which factors predict the dollar value of sales of the firm’s
personal computers. Aggregate data on population size, income, educational level, proportion of
population living in metropolitan areas, etc. have been collected for 30 areas. As a first step, a
multiple linear regression equation is computed, where dollar sales is the outcome variable and
the other factors are considered as candidates for predictor variables. A linear combination of the
predictors is used to predict the outcome or response variable.

Discriminant function analysis

A large sample of initially disease-free men over 50 years of age from a community has been
followed to see who subsequently has a diagnosed heart attack. At the initial visit, blood was drawn
from each man, and numerous other determinations were made, including body mass index, serum
cholesterol, phospholipids, and blood glucose. The investigator would like to determine a linear
function of these and possibly other measurements that would be useful in predicting who would
and who would not get a heart attack within ten years. That is, the investigator wishes to derive a
classification (discriminant) function that would help determine whether or not a middle-aged man
is likely to have a heart attack.

Logistic regression
An online movie streaming service has classified movies into two distinct groups according to
whether they have a high or low proportion of the viewing audience when shown. The company
also records data on features such as the length of the movie, the genre, and the characteristics
of the actors. An analyst would use logistic regression because some of the data do not meet the
assumptions for statistical inference used in discriminant function analysis, but they do meet the
assumptions for logistic regression. From logistic regression we derive an equation to estimate the
probability of capturing a high proportion of the target audience.

Poisson regression
In a health survey, middle school students were asked how many visits they made to the dentist in
the last year. The investigators are concerned that many students in this community are not receiving
adequate dental care. They want to determine what characterizes how frequently students go to the
dentist so that they can design a program to improve utilization of dental care. Visits per year are
count data and Poisson regression analysis provides a good tool for analyzing this type of data.
Poisson regression is covered in the logistic regression chapter.
8 CHAPTER 1. WHAT IS MULTIVARIATE ANALYSIS?
Survival analysis
An administrator of a large health maintenance organization (HMO) has collected data for a number
of years on length of employment in years for their physicians who are either family practitioners or
internists. Some of the physicians are still employed, but many have left. For those still employed,
the administrator can only know that their ultimate length of employment will be greater than their
current length of employment. The administrator wishes to describe the distribution of length of
employment for each type of physician, determine the possible effects of factors such as gender and
location of work, and test whether or not the length of employment is the same for two specialties.
Survival analysis, or event history analysis (as it is often called by behavioral scientists), can be used
to analyze the distribution of time to an event such as quitting work, having a relapse of a disease,
or dying of cancer.

Principal components analysis

An investigator has made a number of measurements of lung function on a sample of adult males
who do not smoke. In these tests each man is told to inhale deeply and then blow out as fast and
as much as possible into a spirometer, which makes a trace of the volume of air expired over time.
The maximum or forced vital capacity (FVC) is measured as the difference between maximum
inspiration and maximum expiration. Also, the amount of air expired in the first second (FEV1), the
forced mid-expiratory flow rate (FEF 25–75), the maximal expiratory flow rate at 50% of forced vital
capacity (V50), and other measures of lung function are calculated from this trace. Since all these
measures are made from the same flow–volume curve for each man, they are highly interrelated.
From past experience it is known that some of these measures are more interrelated than others and
that they measure airway resistance in different sections of the airway.
The investigator performs a principal components analysis to determine whether a new set of
measurements called principal components can be obtained. These principal components will be
linear functions of the original lung function measurements and will be uncorrelated with each other.
It is hoped that the first two or three principal components will explain most of the variation in the
original lung function measurements among the men. Also, it is anticipated that some operational
meaning can be attached to these linear functions that will aid in their interpretation. The investigator
may decide to do future analyses on these uncorrelated principal components rather than on the
original data. One advantage of this method is that often fewer principal components are needed
than original variables. Also, since the principal components are uncorrelated, future computations
and explanations can be simplified.

Factor analysis
An investigator has asked each respondent in a survey whether he or she strongly agrees, agrees, is
undecided, disagrees, or strongly disagrees with 15 statements concerning attitudes toward inflation.
As a first step, the investigator will do a factor analysis on the resulting data to determine which
statements belong together in sets that are uncorrelated with other sets. The particular statements
that form a single set will be examined to obtain a better understanding of attitudes toward inflation.
Scores derived from each set or factor will be used in subsequent analyses to predict consumer
spending.

Cluster analysis
Investigators have made numerous measurements on a sample of patients who have been classified
as being depressed. They wish to determine, on the basis of their measurements, whether these
patients can be classified by type of depression. That is, is it possible to determine distinct types of
depressed patients by performing a cluster analysis on patient scores on various tests?
Random documents with unrelated
content Scribd suggests to you:
 The Project Gutenberg eBook of Golf Architecture:
Economy in Course Construction and Green-Keeping
This ebook is for the use of anyone anywhere in the United States and most
other parts of the world at no cost and with almost no restrictions whatsoever.
You may copy it, give it away or re-use it under the terms of the Project
Gutenberg License included with this ebook or online at www.gutenberg.org.
If you are not located in the United States, you will have to check the laws of
the country where you are located before using this eBook.

Title: Golf Architecture: Economy in Course Construction and Green-Keeping

Author: A. Mackenzie

Author of introduction, etc.: Henry Shapland Colt

Release date: May 11, 2021 [eBook #65319]

Language: English

Credits: MFR and the Online Distributed Proofreading Team at

https://www.pgdp.net (This file was produced from images
generously made available by The Internet Archive)

*** START OF THE PROJECT GUTENBERG EBOOK GOLF ARCHITECTURE:

ECONOMY IN COURSE CONSTRUCTION AND GREEN-KEEPING ***
 GOLF ARCHITECTURE

The 140-yard short hole at Sitwell Park: a fiercely criticised green that has become universally
popular.

Frontispiece

GOLF ARCHITECTURE
ECONOMY IN COURSE CONSTRUCTION
AND GREEN-KEEPING
 BY
DR. A. MACKENZIE
WITH AN INTRODUCTION
BY
H. S. COLT

SIMPKIN, MARSHALL, HAMILTON,

KENT & CO. LTD., 4 STATIONERS’
HALL COURT : : LONDON, E.C.4

Copyright
First published 1920
 INTRODUCTION
My partner, who is the author of these short essays on Golf Course
Architecture, has asked me to write an introduction. This is,
however, hardly necessary, as the name of Dr. Mackenzie is so well
known in connection with this subject.
Many years ago now the idea came to him, as to a few others,
that it might not be impossible to create a golf course without doing
damage to the natural attractions of the site. Up to that period the
courses which had been designed by man, and not by nature, had in
great measure failed in this direction, and although no doubt they
had provided necessary opportunities for playing the game, the
surroundings in many cases proved a source of irritation rather than
pleasure.
I vividly remember meeting my present partner for the first time. I
had been asked to go to Leeds to advise about the design of the
Alwoodley Golf Course, and stayed at his house. After dinner he took
me into his consulting room, where, instead of finding myself
surrounded by the weapons of his profession as a Doctor of
Medicine, I sat in the midst of a collection of photographs of sand
bunkers, putting greens, and golf courses, and many plans and
designs of the Alwoodley Course. I found that I was staying with a
real enthusiast, and one who had already given close attention to a
subject in which I have always been interested.
And it is this enthusiasm for the natural beauty of nature which
has helped him in all his work, so that in the case of Alwoodley the
player not only has the opportunity of displaying his skill in the
game, but also of enjoying the relaxation which delightful natural
surroundings always give.
 No doubt many mistakes were made in our early attempts, and I
never visit a course which I have designed without seeing where
improvements could be made in the constructional work, and as long
as this is so, I feel that we shall all continue to learn and to make
progress, our instructor being nature herself.
H. S. Colt.
 CONTENTS
CHAPTER PAGE
I. General Principles of Economy in Course Construction
and Green-keeping 17
II. Some Further Suggestions 68
III. Ideal Holes 88
IV. The Future of Golf Architecture 116
 ILLUSTRATIONS
The 140-yard Short Hole at Sitwell Park Frontispiece
PAGE
The Sixteenth Green at Headingley, Leeds 26
The Home Green at Sitwell Park 28
An Artificial Hummock at Moortown, constructed from the
stones removed from the Fairway 32
The Fifteenth Hole on the City or Newcastle Course 40
Diagram of Hole of 370 yards, illustrating the value of one
bunker, B 45
The Artificial Hummocks guarding the Fifth Green at
Alwoodley 49
The Seventeenth Green at Harrogate 52
Grange-over-Sands: the site of one of the greens on the
rocks near the boundary of the course—work just
beginning 62
Grange-over-Sands: ready for turfing—a green constructed
on rocks 63
The “Scraper” at work on Wheatley Park, Doncaster 69
Grange-over-Sands: the turf cutting machine at work 70
Grange-over-Sands: sandhills constructed by means of the
Scraper on terrain originally perfectly flat 71
An Artificial Bunker on the Fulford Course 86
The Second Hole at Headingley 94
The Eighth Green at Moortown 99
The Eighth Hole, “Gibraltar,” Moortown Golf Course 101
The Sixteenth Hole at St. Andrews 103
The Fourteenth Hole at St. Andrews 107
The Seventeenth Hole at St. Andrews 111
Plan of Ideal Two-shot Hole of 420 yards 115
The Fifth Hole at Fulford 124
GOLF ARCHITECTURE
 CHAPTER I
GENERAL PRINCIPLES OF ECONOMY IN COURSE
CONSTRUCTION AND GREEN-KEEPING
Economy in course construction consists in obtaining the best possible
results at a minimum of cost. The more one sees of golf courses, the more
one realises the importance of doing construction work really well, so that it is
likely to be of a permanent character. It is impossible to lay too much stress
on the importance of finality.
Every golfer knows examples of courses which have been constructed and
rearranged over and over again, and the fact that all over the country
thousands of pounds are frittered away in doing bad work which will
ultimately have to be scrapped is particularly distressful to a true economist.
As an example of unnecessary labour and expense, the writer has in mind a
green which has been entirely relaid on four different occasions. In the first
instance, it was of the ridge and furrow type; the turf was then lifted and it
was made dead flat. A new secretary was appointed, and he made it a more
pronounced ridge and furrow than ever; it was then relaid and made flat
again, and has now been entirely reconstructed with undulations of a more
natural outline and appearance.
In discussing the question of finality, it is well to inquire if there are any
really first-class courses in existence which have been unaltered for a
considerable number of years and still remain, not only a good test of golf,
but a source of pleasure to all classes of players. Is there any existing course
which not even the rubber cored ball has spoilt? And, if so, what is the cause
of its abiding popularity? The only one I know of is one which has been
described as “a much-abused old course at a little place called St. Andrews, in
the Kingdom of Fife.” This (as well as some of the other championship courses
to a lesser extent) still retains its popularity among all classes of amateurs. In
fact, it is characteristic of all the best courses that they are just as pleasurable
(possibly even more so) to the long handicap man as to the player of
championship rank. This fact knocks on the head the argument which is often
used that the modern expert tries to spoil the pleasure of the player by
making courses too difficult.
 The successful negotiation of difficulties is a source of pleasure to all
classes of players.
It may be asked, “Who originally constructed St. Andrews?” Its origin
appears to be shrouded in mystery: like Topsy, in Uncle Tom’s Cabin, it simply
“growed.” But the fact of the matter is that St. Andrews differs from others in
that it has always been deemed a sacrilege to interfere with its natural
beauties, and it has been left almost untouched for centuries. No green-
keeper has ever dared to shave down its natural undulations. Most of the
bunkers have been left where nature placed them, and others have originated
from the winds and the rains enlarging divot marks left by the players, and
some of them possibly by the green-keepers converting those hollows where
most players congregated, into bunkers, owing to the difficulty of keeping
them free from divot marks. The bunkers at St. Andrews are thus placed in
positions where players are most likely to go—in fact, in the precise positions
which the ordinary Green Committee would suggest should be filled up. This
is a significant fact, and tends to show that many of our existing ideas in
regard to hazards have been erroneous. Mr. John L. Low pointed out years
ago that no hazard is unfair wherever it is placed, and this particularly applies
if the hazard is visible, as it should be obvious that if a player sees a hazard in
front of him and promptly planks his ball into it he has chosen the wrong spot.
I once heard a Yorkshire tale of an old farmer finding a man in his coal-
house during a recent coal strike. He put his head through the window and
said, “Now I’ve copped you picking out all the big lumps.” A voice from the
darkness came, “You’re a liar, I’m taking them as they come.”
On the old type of course like St. Andrews, the players have to take the
hazards as they come, and do their best to avoid them.
There is nothing new about the ideas of the so-called Golf Architect: he
simply wishes to reproduce the old ideas as exemplified in the old natural
courses like St. Andrews, those courses which were played on before over-
zealous green committees demolished the natural undulations of the fairways
and greens, and made greens like lawns for croquet, tennis, or anything else
except golf, and erected eyesores in the shape of straight lines of cop
bunkers, instead of emphasising the natural curves of the links.
In the old view of golf, there was no main thoroughfare to the hole: the
player had to use his own judgment without the aid of guide posts, or other
adventitious means of finding his way. St. Andrews still retains the old
traditions of golf. For example, I have frequently seen four individuals playing
the long hole (the fourteenth), and deliberately attacking it in four different
ways, and three out of the four were probably right in playing it in the ways
they selected.
At St. Andrews “it needs a heid to play gowf,” as the caddie said to the
professor.
As the truest economy consists in finality, it is interesting to consider the
essential features of an ideal golf course. Some of them are suggested now:
1. The course, where possible, should be arranged in two loops of nine
holes.
2. There should be a large proportion of good two-shot holes, two or three
drive-and-pitch holes, and at least four one-shot holes.
3. There should be little walking between the greens and tees, and the
course should be arranged so that in the first instance there is always a slight
walk forwards, from the green to the next tee; then the holes are sufficiently
elastic to be lengthened in the future if necessary.
4. The greens and fairways should be sufficiently undulating, but there
should be no hill climbing.
5. Every hole should have a different character.
6. There should be a minimum of blindness for the approach shots.
7. The course should have beautiful surroundings, and all the artificial
features should have so natural an appearance that a stranger is unable to
distinguish them from nature itself.
8. There should be a sufficient number of heroic carries from the tee, but
the course should be arranged so that the weaker player with the loss of a
stroke or portion of a stroke shall always have an alternative route open to
him.
9. There should be infinite variety in the strokes required to play the various
holes—viz., interesting brassy shots, iron shots, pitch and run-up shots.
10. There should be a complete absence of the annoyance and irritation
caused by the necessity of searching for lost balls.
11. The course should be so interesting that even the plus man is
constantly stimulated to improve his game in attempting shots he has hitherto
been unable to play.
12. The course should be so arranged that the long handicap player, or
even the absolute beginner, should be able to enjoy his round in spite of the
fact that he is piling up a big score.
13. The course should be equally good during winter and summer, the
texture of the greens and fairways should be perfect, and the approaches
should have the same consistency as the greens.

A DECIDED ADVANTAGE
In regard to the first three principles, there can be little difference of
opinion. It is a considerable advantage that a course should be arranged in
two loops of nine holes, as on a busy day players can commence at either the
first or tenth tee.
In regard to the fourth principle. It used to be a common fallacy that greens
should be made dead flat. Even on some of the best golf courses at the
present day you find them made like croquet lawns. There has been
somewhat of a reaction lately against undulating greens, but this, I believe, is
entirely due to the fact that the undulations have been made of a wrong
character, either composed of finicky little humps or of the ridge and furrow
type. Natural undulations are the exact opposite to the artificial ridge and
furrow. The latter has a narrow hollow, and a broad ridge, whereas the former
has a large, bold, sweeping hollow, and a narrow ridge.
The Headingley, Leeds—approximate cost £50: an entirely artificial hole; the site
sixteenth green at
was originally on a severe downhill slope and had to be cut out of rock.
The most interesting putting the writer has ever seen is on the Ladies’
Putting Course at St. Andrews. Even first-class golfers consider it a privilege to
be invited there, and are to be found putting with the greatest enthusiasm
from early morn till late at night. There the undulations are of the boldest
possible type, large sweeping hollows rising abruptly four or five feet up to
small plateaus. A modern golf architect who dared to produce the boldness of
these St. Andrews’ undulations could hardly hope to escape hostile criticism.
In constructing natural-looking undulations one should attempt to study the
manner in which those among the sand-dunes are formed. These are
fashioned by the wind blowing up the sand in the form of waves, which
become gradually turfed over in the course of time. Natural undulations are,
therefore, of a similar shape to the waves one sees by the seashore, and are
of all kinds of shapes and sizes, but are characterised by the fact that the
hollows between the waves are broader than the waves themselves.
If undulations are made of this kind, then there are always plenty of
comparatively flat places where the green-keeper can put the flag, and there
should never be any necessity to cut the hole on a slope.
A test of a good undulation is that it should be easy to use the mowing
machine over it.
If undulations are made of the kind I describe, it is hardly possible to make
them too large or too bold.
Perhaps the most aggravating type of undulation is the finicky little hump or
side-slope which you don’t see until after you have missed your putt, and then
begin to wonder why it has not gone in the hole.
The home green at Sitwell Park: An undulating green with a wide choice of places for the hole in
the hollows or on the flat.

An almost equally common delusion is that fairways should be flat. I quite

agree that there is nothing worse than a fairway on a severe side-slope, but,
on the other hand, there are few things more monotonous than playing every
shot from a dead flat fairway. The unobservant player never seems to realise
that one of the chief charms of the best seaside links is the undulating
fairways, such as those near the club-house at Deal, part of Sandwich, and
most of the old course at St. Andrews, where the ground is a continual roll
from the first tee to the last green, and where one never has the same shot to
play twice; on these fairways one hardly ever has a level stance or a level lie.
It is this that makes the variety of a seaside course, and variety is everything
in golf.
If one considers St. Andrews hole by hole, it is surprising to find at how
many of them the dominating and important incident is associated with an
insignificant-looking hollow or bank, often running obliquely to the line of your
approach.
In constructing undulations of this kind on inland courses, it is well to make
them with as much variety as possible, and in the direction you wish the
player to go to keep the fairway comparatively flat, so as to encourage players
to place their shots, and thus get in a favourable position for their next.
In this connection plasticine is frequently used for making models of
undulations. Plasticine is useful to teach the green-keeper points in
construction he would not otherwise understand—in fact, I believe, I was the
first designer of golf courses to use it for this purpose. The 14th green at
Alwoodley, which was the first one made there, was constructed from a model
in plasticine. It has its disadvantages, however, as a course constructed
entirely from models in plasticine has always an artificial appearance, and can
never be done as cheaply as one in which the green-keeper is allowed a
comparatively free hand in modelling the undulations in such a manner that
not only do they harmonise with their surroundings, but are constructed
according to the various changes in the subsoil discovered whilst doing the
work.

THE FOLLY OF FASHIONS

In regard to the fifth principle that every hole should have a different
character. A common mistake is to follow prevailing fashions. At first we had
the artificial cop bunkers extending in a dead straight line from the rough on
one side to the rough on the other; in modern course architecture these are
fortunately extinct. Secondly, we had the fashion of pot bunkers running down
each side of the course. This was, if anything, an even more objectionable
type of golf than the last. Thirdly, we have had what has been called the
alpinisation of golf courses.
In this connection I would point out that green-keepers should be careful
not to make hillocks so high in the direct line to the hole that they block out
the view: a little to one side of the bee line they may be made as high as one
pleases, but in the direct line hollows should, as a rule, take the place of
hillocks. This is the exact opposite to what is found on many golf courses,
where the hollows are at the sides and the banks in the middle.
 Artificial hummock at Moortown, constructed from the stones removed from the fairway.

The great thing in constructing golf courses is to ensure variety and make
everything look natural. The greatest compliment that can be paid to a green-
keeper is for players to think his artificial work is natural. On Alwoodley and
Moortown practically every green and every hummock has been artificially
made, and yet it is difficult to convince the stranger that this is so. I
remember a chairman of the Green Committee of one of the best-known clubs
in the North telling me that it would be impossible to make their course
anything like Alwoodley, as there we had such a wealth of natural hillocks,
hollows, and undulations. It was only with great difficulty that I was able to
persuade him that, to use an Irishism, these natural features which he so
much admired had all been artificially created. I have even heard one of the
members of our own Green Committee telling a well-known writer on golf that
the hummocks surrounding one of our greens had always been there: he
himself had forgotten that he had been present when the site for them had
been pegged out.

THE QUESTION OF BLIND HOLES

 It is not nearly as common an error to make blind holes as formerly. A blind
tee shot may be forgiven, or a full shot to the green on a seaside course,
when the greens can usually be located accurately by the position of the
surrounding hummocks, but an approach shot should never be blind, as this
prevents an expert player, except by a fluke, from placing his approach so
near the hole that he gets down in one putt.
Blind holes on an inland course where there are no surrounding sandhills to
locate the green should never be permitted, but an even more annoying form
of blindness is that which is so frequent on inland courses—that is, when the
flag is visible but the surface of the green cannot be seen. On a green of this
description no one can possibly tell whether the flag is at the back, middle, or
front of the green, and it is particularly aggravating to play your shot
expecting to find it dead, and to discover that your ball is at least twenty
yards short.
On a seaside course there may be a certain amount of pleasurable
excitement in running up to the top of a hillock in the hope of seeing your ball
near the flag, but this is a kind of thing one gets rather tired of as one grows
older.

IMPORTANCE OF BEAUTY
Another common erroneous idea is that beauty does not matter on a golf
course. One often hears players say that they don’t care a “tinker’s cuss”
about their surroundings: what they want is good golf.
One of the best-known writers on golf has recently been jeering at golf
architects for attempting to make beautiful bunkers. If he prefers ugly
bunkers, ugly greens, and ugly surroundings generally he is welcome to them,
but I don’t think for an instant that he believes what he is writing about, for at
the same time he talks about the beauties of natural courses. The chief object
of every golf architect or green-keeper worth his salt is to imitate the beauties
of nature so closely as to make his work indistinguishable from nature itself.
I haven’t the smallest hesitation in saying that beauty means a great deal
on a golf course; even the man who emphatically states he does not care a
hang for beauty is subconsciously influenced by his surroundings. A beautiful
hole not only appeals to the short handicap player but also to the long, and
there are few first-rate holes which are not at the same time, either in the
grandeur of their undulations and hazards, or the character of their
surroundings, beautiful holes.
 It is not suggested that we should all play round the links after the manner
of the curate playing with the deaf old Scotsman.
The curate was audibly expressing his admiration of the scenery, the
greens, and things in general, until they finally arrived at a green surrounded
by a rookery. The curate remarked, “Isn’t it delightful to hear the rooks?” The
deaf old Scotsman said, “What’s that?” The curate again remarked, “Isn’t it
delightful to hear the rooks?” The old Scotsman replied, “I can’t hear a word
you’re saying for those damned crows.”
The finest courses in existence are natural ones. Such courses as St.
Andrews, and the championship courses generally, are admitted to provide a
fine test of golf. It is by virtue of their natural formation that they do so. The
beauty of gold courses has suffered in the past from the creations of ugly and
unimaginative design. Square, flat greens and geometrical bunkers have not
only been an eyesore upon the whole landscape, but have detracted from the
infinite variety of play which is the heritage of the game.
My reputation in the past has been based on the fact that I have
endeavoured to conserve existing natural features, and where these are
lacking to create formations in the spirit of nature herself.
In other words, while always keeping uppermost the provision of a splendid
test of golf, I have striven to achieve beauty.
It may at first appear unreasonable that the question of æsthetics should
enter into golf-course design; however, on deeper analysis, it becomes clear
that the great courses, and in detail all the famous holes and greens, are
fascinating to the golfer by reason of their shape, their situation, and the
character of their modelling. When these elements obey the fundamental laws
of balance, of harmony, and fine proportion they give rise to what we call
beauty. This excellence of design is more felt than fully realised by the player,
but nevertheless it is constantly exercising a subconscious influence upon him,
and in course of time he grows to admire such a course as all works of beauty
are eventually felt and admired.

THE REAL OBJECT OF THE HAZARD

Most of the remaining principles depend on the proper disposition of
hazards, and I have a rather wider definition of hazards than is given by the
rules of Golf Committee. As a minor kind of hazard undulating ground,
hummocks, hollows, etc., might be included.
 Most golfers have an entirely erroneous view of the real object of hazards.
The majority of them simply look upon hazards as a means of punishing a bad
shot, when their real object is to make the game interesting. The attitude of
the ordinary golfer towards hazards may be illustrated by the following tale
which I have frequently told before, but which will bear repeating:
A player visiting a Scotch course asked his caddie what the members
thought of a stream which was winding in and out between several of the
holes. The caddie replied, “Weel, we’ve got an old Scotch major here. When
he gets it ower he says, ‘Weel ower the bonnie wee burn, ma laddie’; but
when he gets in he says, ‘Pick ma ball oot o’ that domned sewer.’”
The writer was recently playing with his brother, who was home on leave
from abroad. He was clearly enjoying his game, but at Alwoodley we have one
solitary pond into which he topped three balls. On arriving at the club-house
he was asked how he liked the course; he simply remarked, “There were too
many ruddy ponds about.”
It is much too large a subject to go into the question of the placing of
hazards, but I would like to emphasise a fundamental principle. It is that, as
already pointed out, no hazard is unfair wherever it is placed.

The fifteenth hole on the City of Newcastle course: constructed on flat, featureless clay land.
 A hazard placed in the exact position where a player would naturally go is
frequently the most interesting situation, as then a special effort is needed to
get over or avoid it.

GIVING THE PLAYER THRILLS

One of the objects in placing hazards is to give the players as much
pleasurable excitement as possible. On many inland courses there is not a
thrill on the whole round, and yet on some of the championship courses one
rarely takes a club out of the bag without having an interesting shot to play.
This particularly applies to the old course at St. Andrews, and is one of the
reasons why it always retains its popularity with all classes of players. It is
quite true that even this course is condemned by some, but this may be due
to the fact that they have not brains enough, or have not played on it long
enough, to appreciate its many virtues.
There are some leading players who honestly dislike the dramatic element
in golf. They hate anything that is likely to interfere with a constant succession
of threes and fours. They look upon everything in the “card and pencil” spirit.
The average club member on the other hand is a keen sportsman, he looks
upon golf in the “spirit of adventure,” and that is why St. Andrews and courses
modelled on similar ideals appeal to him.
No one would pretend that the old course at St. Andrews is perfect: it has
its disadvantages, particularly in the absence of long carries from the tee, and
in its blind bunkers, but no links in the world grows upon all classes of players
in the same manner. The longer one plays there the keener one gets, and this
is a much truer test of a good course than one which pleases at first and is
boring later on.
A good golf course is like good music or good anything else; it is not
necessarily a course which appeals the first time one plays over it, but one
which grows on the player the more frequently he visits it.
St. Andrews is a standing example of the possibility of making a course
which is pleasurable to all classes of golfers, not only to the thirty handicap
players, but to the plus fourteen man, if there ever was or will be such a
person.
It is an interesting fact that few hazards are of any interest which are out of
what is known among medical men as the direct field of vision. This does not
extend much farther than ten to twenty yards on either side of the direct line
to the hole. Hazards placed outside this limit are usually of little interest, but
simply act as a source of irritation.
Hazards should be placed with an object, and none should be made which
has not some influence on the line of play to the hole.

TOO MANY BUNKERS

On many courses there are far too many bunkers: the sides of the fairways
are riddled with them, and many of these courses would be equally
interesting if half of the bunkers were turfed over as grassy hollows.
It is often possible to make a hole sufficiently interesting with one or two
bunkers at the most. For example:
It is obvious from the diagram that the green-guarding bunker B has a
considerable influence on the line of play to the hole.
The longer the carry a player achieves over the stream the easier his
second shot becomes.
If it were not for this bunker not only the approach but the tee shot would
be uninteresting, as there would be no object in essaying the long carry over
the stream.
 Diagram of hole of 370 yards, illustrating the value of one
bunker, B. Any additional bunker for the tee shot or across
the approach to the green would materially lessen the
interest of the hole. The moral is “Few bunkers placed in
interesting positions!”

Many poor golf courses are made in a futile attempt to eliminate the
element of luck. You can no more eliminate luck in golf than in cricket, and in
neither case is it possible to punish every bad shot. If you succeeded you
would only make both games uninteresting.
 There are many points of resemblance between cricket and golf: the
fielders in cricket correspond to the hazards at golf. The fielders are placed in
the positions where the majority of shots go, and it should obviously be easier
with a stationary ball to avoid the hazards than to avoid the fielders at cricket.
In both games it is only a proportion of bad shots that get punished, but
notwithstanding this the man who is playing the best game almost invariably
comes out on top.
It is an important thing in golf to make holes look much more difficult than
they really are. People get more pleasure in doing a hole which looks almost
impossible, and yet is not so difficult as it appears.
In this connection it may be pointed out that rough grass is of little interest
as a hazard. It is frequently much more difficult than a fearsome-looking
bunker or belt of whins or rushes, but it causes considerable annoyance in lost
balls, and no one ever gets the same thrills in driving over a stretch of rough
as over a fearsome-looking bunker, which in reality may not be so severe.
Narrow fairways bordered by long grass make bad golfers. They do so by
destroying the harmony and continuity of the game, and in causing a stilted
and cramped style by destroying all freedom of play.
There is no defined line between the fairways in the great schools of golf
like St. Andrews or Hoylake.
It is a common error to cut the rough in straight lines. It should be cut in
irregular, natural-looking curves. The fairways should gradually widen out
where a long drive goes; in this way a long driver is given a little more
latitude in pulling and slicing.
Moreover irregular curves assist a player in locating the exact position of a
ball which has left the fairway and entered the rough.

GLORIFIED MOLE-HILLS
Hummocks and hollows should be made of all sorts of different shapes and
sizes, and should have a natural appearance, with plenty of slope at the
bottom like large waves. Most of the hummocks and hollows should be made
so smooth that the mowing machine can be used over them. The glorified
mole-hills one sees on many courses should be avoided.
 The artificial hummocks guarding the fifth green at Alwoodley: approximate cost, £8. The best
way of combining sand and hummocks, with the sand on the slope of the hazard above the
ground level.

Bunkers on an inland course should, as a rule, be made in the opposite way

to what is customary. At the present time most bunkers have the hollows
sanded and the banks turfed. It is suggested that you get a much more
natural appearance if the hollows are partly turfed over and the hummocks
sanded, as in the photographs in these pages. This has the following
advantages: the appearance is much more like a seaside course; the sand
being above the level of the ground, always remains dry. The contrast
between white or yellow sand and the grass helps one to judge distances
much more accurately, and enables the ball to be found more easily, and the
great disadvantage and expense of scything the long grass on the hummocks
to prevent lost balls is done away with.
Ordinary bunkers are, as a rule, made in quite the wrong way. The face is
usually too upright and the ball gets into an unplayable position under the
face. The bottom of the bank of a bunker should have a considerable slope,
so that a ball always rolls to the middle; the top of a bunker may, as it usually
does in nature, be made to overhang a little so as to prevent a topped ball
running through it.