Download the full version of the textbook now at textbookfull.

com

Analysis of Binary Data, Second Edition David

Roxbee Cox

https://textbookfull.com/product/analysis-of-
binary-data-second-edition-david-roxbee-cox/

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

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

Analysis of Survival Data First Edition Cox

https://textbookfull.com/product/analysis-of-survival-data-first-
edition-cox/

textbookfull.com

Visual Six Sigma Making Data Analysis Lean Ian Cox

https://textbookfull.com/product/visual-six-sigma-making-data-
analysis-lean-ian-cox/

textbookfull.com

Biota Grow 2C gather 2C cook Loucas

https://textbookfull.com/product/biota-grow-2c-gather-2c-cook-loucas/

textbookfull.com

Newton the Alchemist: Science, Enigma, and the Quest for

Nature’s "Secret Fire" William R. Newman

https://textbookfull.com/product/newton-the-alchemist-science-enigma-
and-the-quest-for-natures-secret-fire-william-r-newman/

textbookfull.com
Writing an Interactive Story 1st Edition Pierre Lacombe

https://textbookfull.com/product/writing-an-interactive-story-1st-
edition-pierre-lacombe/

textbookfull.com

Marieb Human Anatomy & Physiology (12th edition) Katja

Hoehn

https://textbookfull.com/product/marieb-human-anatomy-physiology-12th-
edition-katja-hoehn/

textbookfull.com

Neuroimmune Diseases From Cells to the Living Brain

Hiroshi Mitoma

https://textbookfull.com/product/neuroimmune-diseases-from-cells-to-
the-living-brain-hiroshi-mitoma/

textbookfull.com

Machine Intelligence and Signal Processing Proceedings of

International Conference Misp 2019 1st Edition Sonali
Agarwal (Editor)
https://textbookfull.com/product/machine-intelligence-and-signal-
processing-proceedings-of-international-conference-misp-2019-1st-
edition-sonali-agarwal-editor/
textbookfull.com

Sensing the Nation s Law Stefan Huygebaert

https://textbookfull.com/product/sensing-the-nation-s-law-stefan-
huygebaert/

textbookfull.com
Using think-aloud interviews and cognitive labs in
educational research 1st Edition Leighton

https://textbookfull.com/product/using-think-aloud-interviews-and-
cognitive-labs-in-educational-research-1st-edition-leighton/

textbookfull.com
 M O N O G R A P H S ON
STATISTICS A N D A P P L I E D P R O B A B I L I T Y

General Editors
D.R. Cox, D.V. Hinkley, D. Rubin and B.W. Silverman
1 Stochastic Population Models in Ecology and Epidemiology
M.S. Bartlett (1960)
2 Queues D.R. Cox and W.L. Smith (1961)
3 Monte Carlo Methods J.M. Hammersley and D.C. Handscomb (1964)
4 The Statistical Analysis of Series of Events D.R. Cox and
P.A.W. Lewis (1966)
5 Population Genetics W.J. Ewens (1969)
6 Probability, Statistics and Time M.S. Bartlett (1975)
7 Statistical Inference S.D. Silvey (1975)
8 The Analysis of Contingency Tables B.S. Everitt (1977)
9 Multivariate Analysis in Behavioural Research A.E. Maxwell (1977)
10 Stochastic Abundance Models S. Engen (1978)
11 Some Basic Theory for Statistical Inference E.J.G. Pitman (1978)
12 Point Processes D.R. Cox and V. Isham (1980)
13 Identification of Outliers D.M. Hawkins (1980)
14 Optimal Design S.D. Silvey (1980)
15 Finite Mixture Distributions B.S. Everitt and D.J. Hand (1981)
16 Classification A.D. Gordon (1981)
17 Distribution-free Statistical Methods 7.5. Maritz (1981)
18 Residuals and Influence in Regression R.D. Cook and S. Weisberg (1982)
19 Applications of Queueing Theory G.F. Newell (1982)
20 Risk Theory, 3rd edition R.E. Beard, T. Pentikainen and
E. Pesonen (1984)
21 Analysis of Survival Data D.R. Cox and D. Oakes (1984)
22 An Introduction to Latent Variable Models B.S. Everitt (1984)
23 Bandit Problems D.A. Berry and B. Fristedt (1985)
24 Stochastic Modelling and Control M.H.A. Davis and R. Vinter (1985)
25 The Statistical Analysis of Compositional Data 7. Aitchison (1986)
26 Density Estimation for Statistical and Data Analysis
B.W. Silverman (1986)
27 Regression Analysis with Applications G.B. Wetherill (1986)
28 Sequential Methods in Statistics, 3rd edition G.B. Wetherill (1986)
29 Tensor Methods in Statistics P. McCullagh (1987)
 30 Transformation and Weighting in Regression R J. Carroll and
D. Ruppert (1988)
31 Asymptotic Techniques for use in Statistics O.E. Barndoff-Nielson
and D.R. Cox (1989)
32 Analysis of Binary Data, 2nd edition D.R. Cox and E.J. Snell (1989)
33 Analysis of Infectious Disease Data N.G. Becker (1989)
34 Design and Analysis of Cross-Over Trials B. Jones and
M.G. Kenward (1989)
35 Empirical Bayes Method, 2nd edition J.S. Maritz and T. tw in (1989)
36 Symmetric Multivariate and Related Distributions K.-T. Fang,
S. Kotz and K. Ng (1989)
37 Generalized Linear Models, 2nd edition P. McCullagh and
J.A. Nelder (1989)
38 Cyclic Designs y.v4. John (1987)
39 Analog Estimation Methods in Econometrics C.F. Manski (1988)
40 Subset Selection in Regression A.J. Miller (1990)
41 Analysis of Repeated Measures M. Crowder and D.J. Hand (1990)
42 Statistical Reasoning with Imprecise Probabilities P. Walley (1990)
43 Generalized Additive Models T.J. Hastie and R.J. Tibshirani (1990)
44 Inspection Errors for Attributes in Quality Control
N.L. Johnson, S. Kotz and X Ww (1991)
45 The Analysis of Contingency Tables, 2nd edition B.S. Everitt (1992)
(Full details concerning this series are available from the Publishers.)
 Analysis of
Binary Data
SECOND EDITION

D.R. COX
Nuffield College
Oxford

and

E.J. SNELL
Department of Mathematics
Imperial College London

CHAPMAN & HALL/CRC

Boca Raton London New York Washington, D.C.
Published in 1989 by
Chapman & Hall/CRC
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 1970 by D.R. Cox, 1989 by D.R. Cox and EJ. Snell

Chapman & Hall/CRC is an imprint of Taylor & Francis Group
No claim to original U.S. Government works
Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4
International Standard Book Number-10: 0-412-30620-4 (Hardcover)
International Standard Book Number-13; 978-0-412-30620-4 (Hardcover)
This book contains information obtained from authentic and highly regarded sources. Reprinted material is
quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts
have been made to publish reliable data and information, but the author and the publisher cannot assume
responsibility for the validity of all materials or for the consequences of their use.
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, microfilming, and
recording, or in any information storage or retrieval system, without written permission from the publishers.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Cox, D.R. (David Roxbee)

Analysis of binary data/D.R.Cox and E.J, Snell - 2nd ed.
p. cm. - (Monographs on statistics and applied probability)
Includes bibliographical references and index.
ISBN 0-412-30620-4
1. Analysis of varience. 2. Probabilities. 3. Distribution. (Probability theory) I. Snell, E.J. II.Title.
QA279.C68 1989
519.5’352—dcl9

T&F informa Visit thè Taylor & Francis Web site at

http://www.taylorandfrancis.com
Taylor & Francis Group and thè CRC Press Web site at
is the Academic D ivision o f T&F Informa pic. http://www.crcpress.com
 Contents

Preface to first edition IX

Preface to second edition xi

1 Binary response variables 1

1.1 Introduction 1
1.2 Examples 2
1.3 Dependency relations for probabilities 13
1.4 Some statistical considerations 18
1.5 Some numerical comparisons 20
1.6 Some generalizations of the logistic model 23
Bibliographic notes 24

2 Special logistic analyses 26

2.1 Introduction 26
2.2 Simple regression 33
2.3 2 x 2 contingency table 43
2.4 Matched pairs 52
2.5 Several 2 x 2 contingency tables 56
2.6 Multiple regression 66
2.7 Residuals and diagnostics 69
2.8 Factorial arrangements 84
2.9 Cross-over designs 92
2.10 Rasch model 94
2.11 Binary time series 96
Bibliographic notes 102

3 Some complications 106

3.1 Introduction 106
3.2 Anomalous dispersion 106
vi CONTENTS

3.3 Empirical Bayes methods 115

3.4 Errors of measurement 119
Bibliographic notes 124

4 Some related approaches 126

4.1 Introduction 126
4.2 Prospective studies 127
4.3 Retrospective studies 128
4.4 Relation between discriminant analysis
and logistic regression 132
4.5 Some different theoretical formulations 140
Bibliographic notes 148

5 More complex responses 149

5.1 Introduction 149
5.2 Paired preferences 149
5.3 Nominal data 153
5.4 Ordinal data 158
5.5 Multivariate binary data 161
5.6 Mixed binary and continuous responses 163
Bibliographic notes 166

Appendix 1 Theoretical background 167

A 1.1 Introduction 167
A 1.2 Ordinary least squares theory 167
A 1.3 Extensions of least squares 170
A 1.4 Likelihood functions 171
A 1.5 Techniques for inference 179
A 1.6 Some warnings 183
A 1.7 Some further developments 183
A 1.8 Conditioning 184
A 1.9 Bayesian approach 186

Appendix 2 Choice of explanatory variables in multiple

regression 188
A2.1 Introduction 188
A2.2 Types of explanatory variable 189
A2.3 Formation of explanatory variables 190
A2.4 Small numbers of explanatory variables 192
A2.5 Large numbers of explanatory variables 193
CONTENTS vii

Appendix 3 Review of computational aspects 195

A3.1 Introduction 195
A3.2 Specific programs 195
A3.3 Expert systems 196

Appendix 4 Further results and exercises 198

References 215

Author index 228

Subject index 231

 Preface to first edition

This monograph concerns the analysis of binary (or quantal) data, i.e.
data in which an observation takes one of two possible forms, e.g.
success or failure. The central problem is to study how the probability
of success depends on explanatory variables and groupings of the
material.
Many particular methods, especially significance tests, have been
proposed for such problems and one of the main themes of the
monograph is that these methods are unified by considering models
in which the logistic transform of the probability of success is a linear
combination of unknown parameters. These linear logistic models
play here much the same role as do normal-theory linear models in
the analysis of continuously distributed data.
Some knowledge of the theory of statistics is assumed. I have
written primarily for statisticians, but I hope also that scientists and
technologists interested in applying statistical methods will, by
concentrating on the examples, find something useful here.
I am very grateful to Dr Agnes M. Herzberg and Dr P.A.W. Lewis
for extremely helpful comments. I acknowledge also the help of Mrs
Jane Gentleman who programmed some of the calculations.
D.R. Cox
London
April 1969
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
 Preface to second edition

We have added new material partly to amplify matters dealt with only
very cryptically in the first edition and partly to describe some of the
more recent developments, for example on regression diagnostics. In
addition the contents of the first edition have been rearranged; for
example, the method of least squares with empirically estimated
weights is now of much less importance than it used to be because
computational developments have, for many problems, brought
maximum likelihood fitting within the painless grasp of most users of
statistical analysis. By giving some prominence to examples we have
aimed to make the book accessible to a range of readers. One of the
Appendices summarizes the theoretical background.
When the first edition was written it was feasible to give a relatively
complete annotated bibliography of work on the analysis of binary
data. The number of papers on this topic is now so large and is
increasing so rapidly that no such bibliography has been attempted.
Instead, in the Bibliographic Notes at the end of the each chapter we
have aimed to give just a few key references for further reading and for
details omitted from the main text.
We are grateful to Professor N. Wermuth, Mainz, for thoughtful
comments on a portion of the manuscript and to Professor J.K.
Lindsey, Liège, for advice over ordinal data and to Professor P.
McCullagh.
The first author is grateful to Science and Engineering Research
Council for a Senior Research Fellowship held at Department of
Mathematics, Imperial College, London.
D. R. Cox
E. J. Snell
London
July 1988
 CH APTER 1

Binary response variables

1.1 Introduction
Suppose that on each individual we have an observation that takes
one of two possible forms. The following are examples:
1. an electronic component may be defective, or may be
nondefective;
2. a test animal may die from a specified dose of a poison, or may
survive;
3. a subject may give the correct reply in an experimental situation,
or may give a wrong reply;
4. a test specimen may fracture when struck with a standardized
blow, or may not;
and so on. If for the ith individual we can represent this observation,
or response, by a random variable, Y;, we may without loss of
generality code the two possible values of 7; by 1 and 0 and write
E{Yi) = prob(y; = 1) = 9i, prob(7i = 0) = 1 - 0,, ( 1. 1)
say. It is often convenient to call 7^ = 1 a ‘success’ and 7^ = 0 a ‘failure’.
It is reasonable to call such observations binary; an older term is
quantal.
We assume that such binary observations are available on n
individuals, usually assumed to be independent. The problem is to
develop good methods of analysis for assessing any dependence of 0,
on explanatory variables representing, for example, groupings of the
individuals or quantitative explanatory variables.
We have followed the usual terminology and have distinguished
between (a) response variables and (b) explanatory variables, the
variables of the second type being used to explain or predict variation
in variables of the first type.
Sometimes a binary response variable arises by condensing a more
complex response. Thus a component may be classed as defective
2 BI NARY R E S P O N S E V AR I ABL E S

when a quantitative test observation falls outside specification limits

or, more generally, when a set of test observations falls in an
unacceptable region. When this is done we need to consider whether
there is likely to be serious loss in treating the problem in terms of a
binary response.
In addition to a binary response variable there may be further
response variables for each individual. Thus in a psychological
experiment, as well as the rightness or wrongness of a reply, the time
taken to make the reply may be available. Joint analysis of the
response variables is then likely to be informative.

1.2 Examples
It is convenient to begin with a few simple specific examples
illustrating the problems to be considered.

Example 1.1 The 2 x 2 contingency table

Suppose that there are two groups of individuals, 0 and 1, of sizes «q
and « 1 and that on each individual a binary response is obtained. The
groups may, for example, correspond to two treatments. Suppose
further that we provisionally base the analysis on the assumption that
all individuals respond independently with probability of success
depending only on the group, and equal, say, to (/>o and (/>i in the two
groups. Note that we use 0 for a probability referring to a group,
reserving 9 for the probability of success for an individual.
In this situation we need consider only the random numbers of
successes R q and in the two groups; in fact {Rq^R^) form sufficient
statistics for the unknown parameters ((j)o,(t>i)- It is conventional to
set out the numbers of successes and failures in a 2 x 2 contingency
table as in Table 1.1. Nearly always it is helpful to calculate the
proportions of successes in the two groups.
In Table 1.1 the columns refer to a dichotomy of the response
variable, i.e. to a random variable in the mathematical model; the
rows refer to a factor classifying the individuals into two groups, the
classification being considered as non-random for the purpose of the
analysis. It is possible to have contingency tables in which both rows
and columns correspond to random variables.
As a specific example, the data of Table 1.2 refer to a retrospective
survey of physicians (Cornfield, 1956). Data were obtained on a group
of lung cancer patients and a comparable control group. The numbers
1.2 EXAMPLES

Table 1.1 A 2 x 2 contingency table

Failures Successes Total Propn successes

Group 0 Ro ^o/^o
Group 1 Ri/n,

Total n Q n ^ — R q — Ri R q -{-Ri no + Wj

Table 1.2 Numbers of smokers in two groups of physicians

Smokers Non-smokers Total Propn non-smokers

Controls 32 11 43 0.256
Lung cancer
patients 60 3 63 0.048

Total 92 14 106

of individuals in the two groups are approximately equal and are in

no way representative of population frequencies. Hence it is reason­
able to make an analysis conditionally on the observed total numbers
in the two groups.
An essential initial step is to calculate the sample proportions of
successes in the two groups Ro/riQ and Riln^, and these are shown in
Table 1.2. Further analysis of the table is concerned with the precision
of these proportions.
When two groups are to be compared using binary observations, it
will often be sensible to make an initial analysis from a 2 x 2
contingency table. However, the assumptions required to justify
condensation of the data into such a form are not to be taken lightly.
Thus in Section 2.4 we shall deal with the methods to be followed
when pairs of individuals in the two groups are correlated. The most
frequent inadequacy of an analysis by a single 2 x 2 contingency table
is, however, the presence of further factors influencing the response,
i.e. nonconstancy of the probability of success within groups. To
ignore such further factors can be very misleading.
For many purposes it is not necessary to give the numbers of
failures in addition to the numbers of successes and the numbers of
trials per group. In most of the more complex examples we shall
therefore omit the first column of the 2 x 2 table.
4 BI NARY R E S P O N S E V ARI ABL E S

There are many extensions of the 2 x 2 table; there follow some

examples.

Example 1.2 Several 2 x 2 contingency tables

Suppose that to compare two treatments we have several sets of
observations, each of the form of Example 1.1. The different sets may
correspond to levels of a further factor or, as in the following specific
example, may correspond to different blocks of an experimental
design. In general, in the sth set of data, let «o,s be the sample
sizes in the two groups and let /?o,s and be the total numbers of
successes.
Table 1.3 (Gordon and Foss, 1966) illustrates this. On each of 18
days babies not crying at a specified time in a hospital ward served as
subjects. On each day one baby chosen at random formed the
experimental group and the remainder were controls. The binary
response was whether the baby was crying or not at the end of a
specified period. In Table 1.3, not crying is taken as a ‘success’ and
the observed numbers rQ^ and ^ are therefore the numbers of
babies in the two groups not crying; the common convention is

Table 1.3 The crying of babies

No. of
No of control No. not experimental No. not
Day, s babies, Mq.s crying, Tq., babies, ^ crying,

1 8 3 1 1
2 6 2 1 1
3 5 1 1 1
4 6 1 1 0
5 5 4 1 1
6 9 4 1 1
7 8 5 1 1
8 8 4 1 1
9 5 3 1 1
10 9 8 1 0
11 6 5 1 1
12 9 8 1 1
13 8 5 1 1
14 5 4 1 1
15 6 4 1 1
16 8 7 1 1
17 6 4 1 0
18 8 5 1 1
1.2 EXAMPLES 5

followed of denoting observed values of random variables by

lower-case letters. The special feature of this example is that ^ = 1,
so that 5 takes values 0 and 1; usually there are several individuals

in each group.
The object of the analysis is to assess the effect of the treatment on
the probability of success. The tentative basis for the analysis is that
there is in some sense a constant treatment effect throughout the
experiment, even though there may be some systematic variation
from day to day. The experiment has the form of a randomized block
design, in fact a matched pair design, but the binary nature of the
response and the varying numbers of individuals in the groups
complicate the analysis.
For the reasons indicated after Example 1.1 it would not in general
be a sound method of analysis to pool the data over days, thus
forming a single 2 x 2 contingency table with entries ZR q.s»
One simple, if approximate, method of analysis that is not distorted
by systematic differences between groups is to calculate for the sth
group the difference in the proportions of successes, i.e.

0 ,s
( 1.2)
«0,/
This is an unbiased estimate for the sth set of the difference between
the probabilities of success. When (1.2) is averaged over all the sets, an
unbiased estimate of the mean difference between groups results. A
difficulty of this analysis is that the quantities (1.2) have in general
different precisions for different s.

Later, further examples will be given whose analysis requires the

combination of data from several 2 x 2 contingency tables.

Example 1.3 A 2 x 2^ system

Suppose that there are p two-level factors thought to affect the
probability of success. Let a binary response be observed on each
individual and suppose that there is at least one individual corre­
sponding to each of the 2^ cells, i.e. possible factor combinations;
usually there are an appreciable number of individuals in each cell.
It would be possible to think of such data as arranged in a (p + 1)-
dimensional table, with two levels in each dimension. Alternatively
and more usefully, we can think of a 2 x 2^ table in which the two
columns correspond to success and failure and the 2^ rows are the 2^
standard treatments of the factorial system. One of the problems of
 BI N AR Y R E S P O N S E V AR I A BL E S

Table 1.4 2 x 2 ^ system. Study of cancer knowledge

No. No. Propn No. No. Propn

successes trials successes successes trials successes

1 84 477 0.176 d 2 12 0.167

a 75 231 0.325 ad 7 13 0.538
b 13 63 0.206 bd 4 7 0.571
ab 35 94 0.372 abd 8 12 0.667
c 67 150 0.447 cd 3 11 0.273
ac 201 378 0.532 acd 27 45 0.600
be 16 32 0.500 bed 1 4 0.250
abc 102 169 0.604 abed 23 31 0.742

analysis is to examine what order of interactions, in some suitable

sense, is needed to represent the data; that is, we think first of 2^
probabilities of success, one for each cell, and then try to represent
these usefully in terms of a smaller number of parameters.
Table 1.4 is a specific example of a 2 x 2"^ system based on an
observational study by Lombard and Doering (1947) (see Dyke and
Patterson (1952) for a detailed discussion of the analysis of these
data). In this study, the response concerned individuals’ knowledge
of cancer, as measured in a test, a ‘good’ score being a success and
a ‘bad’ score a failure. There were four factors expected to account
for variation in the probability of success, the individuals being
classified into 2"^ cells depending on presence or absence of exposure
to A, newspapers; B, radio; C, solid reading; D, lectures. In Table 1.4
the standard notation for factor combinations in factorial experi­
ments is used; thus ac denotes the cell in which A and C are at their
upper levels and B and D at their lower levels.
Some general conclusions can be drawn from inspection of the cell
proportions of successes. This is an example where both the response
variable and the factors are reduced to two levels from a more
complex form.

Example 1.4 Serial order

Suppose that a series of independent binary responses is observed and
that it is suspected that the probability of success changes systemati­
cally with serial order. One application is to some types of data in
experimental psychology, where a subject makes a series of responses,
each either correct or incorrect, and where the probability of correct
response is suspected of changing systematically. In this context, the
1.2 EXAMPLES 7

treating of successive responses as independent may, however, be

seriously misleading. Another application is in human genetics where
each child in a family is classified as having or not having a particular
genetic defect. Here, except for possible complications from multiple
births, each family leads to a sequence of binary responses. In this
application, data from many families are required. We then have a
number of sequences, usually not all of the same length, and it is
required to examine the data for systematic changes, with serial order,
in the probability of a genetic defect. It is not usually reasonable to
suppose that in the absence of such effects the probability of a defect is
the same for all families.

Example 1.5 Stimulus binary response relation

The following situation is of wide occurrence. There is a stimulus
under the experimenter’s control; each individual is assigned a level of
the stimulus and a binary response then observed. One important
field of application is bioassay, where, for example, different levels of
stimulus may represent different doses of a poison, and the binary
response is death or survival. Similar situations arise in many other
fields.
In such applications it is often possible to choose a measure x of
stimulus level such that the probability of success is zero for large
negative x, unity for large positive x and is a strictly increasing
function of x. In fact it has the mathematical properties of a
continuous cumulative distribution function; see Fig. 1.1(a). If the x
scale is suitably chosen, the distribution function will be symmetric;
for example, in the particular application mentioned above it is often
helpful to take x as log dose.
Table 1.5 gives some illustrative data; at each of a number of
dose levels a group of individuals is tested and the number dying
recorded. In Fig. 1.1(b) the proportions dying are plotted against log

Table 1.5 Simple form ofhioassay

Concn Log2 concn No. of deaths No. of indivs Propn deaths

Co 0 2 30 0.067
2co 1 8 30 0.267
4co 2 '15 30 0.500
8t'o 3 23 30 0.767
16co 4 27 30 0.900
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
 BI N AR Y R E S P O N S E V AR I A BL E S

(b) Log^ (concenrrahon/CQ)

Figure 1.1 Stimulus-response curves, (a) Idealized theoretical curve, (h)

empirical curve from Table 1.5.

concentration and, except for random fluctuations, give a curve

similar to the idealized theoretical one of Fig. 1.1(a). The object of
analysing such data is to summarize and assess the properties of
the stimulus response curve. Sometimes, as in many bioassay
problems, the aspect of primary importance is the level of x at which
1.2 EXAMPLES

the probability of success is or sometimes the level Xp at which

some other specified value, p, of the probability of success is achieved
In other applications, for example in experimental psychology, the
steepness of the response curve is the aspect of primary interest. We
shall refer tc x variously as a stimulus, or as an explanatory variable.

Example 1.6 Grouped data with binary response

Suppose that there are /c = p -h 1 groups of individuals, each indivi­
dual having a binary response. If individuals’ responses are inde­
pendent, and the probability of success is constant within each group,
we can, just as in the discussion of the 2 x 2 table, condense the data,
giving simply the number of trials n^ and the number of successes
for the 5th group (5 = 0 ,..., g). In contingency table form we then have
Table 1.6.
Now suppose that the groups are meaningfully ordered and that it
is reasonable to expect that any change in the probability of success is
monotonie with group order. If, further, scores Xq, . . . , x^ can be
allocated to the groups such that a smooth relation between the
probability of success and the value of x is reasonable, the situation is

Table 1.6 A contingency table for grouped data

Propn
Failures Successes Total successes

Group 0 «0 -7 Ro Ro no Rq/^O
Group g ^9 R jn ,

Table 1.7 Nasal carrier rate and tonsil size

Propn of
Non-carriers Carriers Total carriers x

Tonsils present.
not enlarged 497 19 516 0.0368 - 1
Tonsils enlarged
+ 560 29 589 0.0492 0
Tonsils enlarged
269 24 293 0.0819 1
10 BI N AR Y R E S P O N S E V AR I A BL E S

then formally the same as that of Example 1.5. A distinction is that in

the previous example the main interest is in the location and shape of
a response curve, whereas in the present example there is usually
substantial interest in the null hypothesis that the probability of
success is constant.
Table 1.7 gives a specific example quoted by Armitage (1955). The
data refer to children aged 0-15 and the binary response concerns
whether a child is or is not a carrier for Streptococcus pyogenes; the
children are grouped into three sets depending on tonsil size. No
objective x variable was available and the tentative scores —1,0, 1
were therefore assigned to the three groups. A difficulty in the
interpretation of such data is that other possibly relevant explanatory
variables - age, sex, etc. - should be considered.

Example 1.7 Multiple regression with binary response

The previous three examples have involved a dependence between a
probability of success and a single regressor or explanatory variable,
x; they are thus comparable with normal-theory regression problems
with a single regressor variable. Suppose now that there are for the ith
individual, p, explanatory variables Xji,...,Xjp, regarded as non-
random, and a binary response. It is necessary to assess the relation
between the probability of success and the variables x,i,...,x^p.
This type of problem occurs especially in medical contexts. The
binary response may represent success or failure of a particular
treatment; death or survival over a specified time following treatment;
or death from a particular cause, as contrasted with death from some
other cause. The variables x,i,...,X jp represents quantitative, or
qualitative, properties of an individual thought to influence the
response; possible x variables are age, time since diagnosis, initial
severity of symptoms, sex (scored as a zero-one variable), aspects of
medical history of the individual, etc.
One particular example is a study of the factors affecting the
probability of having coronary heart disease (see Walker and Duncan
(1967) who give a number of earlier references). Another example
concerns perinatal mortality (Feldstein, 1966).

Example 1.8 Regression on two explanatory variables

Table 1.8 summarizes a two-factor 5 x 4 industrial investigation in
which the number, r, of ingots not ready for rolling out of m tested is
1.2 EXAMPLES 11

Table 1.8 Number, r, of ingots not ready rolling out of m tested

(First figure in each cell is r, second m)
Heating time, Xi
Soaking
time, X2 7 14 27 51 Total

1.0 0,10 0,31 1,56 3,13 4,110

1.7 0,17 0,43 4,44 0,1 4,105
2.2 0,7 2,33 0,21 0,1 2,62
2.8 0,12 0,31 1,22 0,0 1,65
4.0 0,9 0,19 1,16 0,1 1,45

Total 0,55 2,157 7,159 3,16 12,387

shown for combinations of heating time, and soaking time, X2 . This

is a rather simpler situation than most of those sketched in
Example 1.7 because there are only two explanatory variables. Note
also that the data are grouped into sets all with the same values of the
explanatory variables.

In some situations, especially with observational data, we need

to analyse changes in the probability of success, there being a
considerable number of potential explanatory variables. Two broad
approaches are possible, illustrated by Examples 1.3 and 1.7.
In the first approach the values of the explanatory variables are
coarsely grouped, in the extreme case each explanatory variable
taking only two values, as in Example 1.3. With p explanatory
variables there will thus be at least 2^ cells, for each of which the
proportion of successes can be found. An advantage of this approach
is that quite complicated ‘interactions’ can be detected. Possible
disadvantages stem from the necessity of coarse grouping and from
the fact that if p is at all large many of the cells will either be empty or
contain very few observations.
In the second approach, a regression-like model is taken, express­
ing a smooth and simple dependence of the probability of success on
the values of the explanatory variables. No grouping is necessary, but
a disadvantage is that relatively complicated interactions may be
difficult to detect. In practice both methods are useful, separately and
in combination. With both, problems concerning alternative choices
12 B I NAR Y R E S P O N S E VAR I A BL E S

of explanatory variables, familiar from normal-theory regression

methods, arise in essentially the same form.

Example 1.9 A binary time series

In the previous examples it is a reasonable provisional assumption to
suppose that the responses of different individuals are independent. In
the analysis of a binary time series we are directly concerned with the
lack of independence of different observations.
A specific example concerns daily rainfall. It is often reasonable first
to classify days as wet (success) or dry (failure); there results a sequence
of Is and Os, a binary time series. The amounts of rainfall on wet days
can be analysed separately.
More generally, if we consider a response 1 as the occurrence of an
event and a response 0 as non-occurrence, a binary time series is a
series of events in discrete time. If, further, the proportion of Is is
low, the series approximates to a series of point events in continuous
time; in particular, a completely random binary series, in which all
responses have independently the same probability, 6, of giving a
success, tends, as 0->O, to a Poisson process. Cox and Lewis (1966)
have summarized statistical techniques for the analysis of point
events in continuous time.

In this section a number of relatively simple problems have been

described which can be generalized in various ways:
1. we may have situations of more complex structure. Thus, instead
of the single response curve of Example 1.5, we might have several
response curves and be interested in comparing their shapes;
2. we may have multivariate binary responses and consider prob­
lems analogous to those of multivariate normal theory;
3. we may have responses taking not just two possible values, but
some small number greater than two. Can the techniques to be
developed for analysing binary responses be extended?
Some of the more complex problems will be considered later. Others
are indicated as further results or exercises.
As in other fields of statistical analysis, problems of two broad types
arise. We require techniques for efficient analysis and assessment of
uncertainty in the context of an assumed probabilistic model. Also we
need techniques for tabular and graphical display and condensation
of data, sometimes with the objective of finding a suitable model for
Random documents with unrelated
content Scribd suggests to you:
that you, a man of science—"
"As a scientist, your Borsu fascinates me, of course. I'd like to
interrogate him for years. I'd like to dissect his mind and body until I
know everything about him and his people. But you're asking me a
different question. You're asking—do I want Borsu as a neighbor?"
Woodward stood up. His face was pale, and the leg that wasn't there
throbbed with pain. He was sweating, a gray sweat that coursed
down his seamed cheek and soiled the collar of his shirt.
"I don't feel well," he said. "If you'd excuse me—"
"Of course," Ridgemont said solicitously. "We can talk more about this
later, when you feel up to it...."
"Yes," Woodward said.

That night, in a hotel room in Washington, Dr. Carl Woodward died of

a coronary thrombosis.
Secretary of Science Ridgemont, however, was curious enough about
the doctor's story to send a deputation to his home in Pennsylvania.
As a simple precaution, one of the men in the party was an armed
policeman named Sergeant Kemmer. At the first sight of the blue-
skinned man, Kemmer became alarmed enough to draw his gun.
Borsu recognized the weapon, and its dangerous potential, from his
reading. Frightened, he tried to flee through the front door, and
Kemmer misinterpreted his movement as an attack. He fired three
times. One of the bullets penetrated Borsu's temple, and killed him
instantly.
Three months passed before the next delegation from the aliens
appeared on Earth. This time, their arrival was detected, and the
visitors were brought safely to the local authorities in the Nebraska
community where their vessel landed. Their names were Cor, Basuc,
and Stytin. Stytin was a female, lovely in her blue-skinned
shapeliness.
A team of scientists were dispatched from Washington, Tokyo, and
London to take charge of the alien trio. It was another two weeks
before their marvelous facility with language permitted them to talk
intelligently to their examiners.
On November 8, 1973, Stytin, the blue female, was found assaulted,
mutilated, and murdered in the woods near the town of Ponchi. The
brutal slaying shocked the scientists, who tried to assure Cor and
Basuc that the episode had not been typical of the behavior of the
people of Earth. But Cor and Basuc, who had no memory-record of
killing, became terrified, and fled. Cor was shot and killed by a farmer,
and Basuc was accidently drowned while forging a stream during his
escape.
The death of the four aliens, however, didn't prevent the migration
from beginning. Hunger—not for food alone, but for the blessed green
promise of the Earth—drove the blue aliens to make the journey
before receiving assurance of their welcome. Their tiny two, three,
and four-man craft began dotting the heavens, filling the world with
fears and panics that were only partially allayed by the repeated
assurances of the world's leaders. Despite explanations and pleas for
order, the blue people were frequently slain the moment their ships
touched Earth. There was never an official estimate of the deaths, but
it was certain that well over three thousand of the aliens lost their
lives before ever tasting a drop of cool Earth water, or knowing the
shade of an Earth tree, the peaceful blue of an Earth sky.
Finally, the killings were over. Less than seven thousand Blues
survived the perilous journey, protected upon their arrival by
contingents of armed soldiers, sped to the scene of the landings in
time to stop the citizens from their slaughter.

It was Mostyn Herbert, Secretary-General of the United Nations, who

made the first speech of welcome, before the general assembly.
"The world," he said, "has seen a new migration in these past
months, an event which has brought new hours of infamy to the
human race. The savagery of man to beast, the bestiality of man to
man, has now been exceeded by our shameful record of cruelty
towards these homeless wanderers from a forgotten world. We have
slain almost a third of their number wantonly and without cause. They
proffered to us the wisdom and knowledge of their own civilization,
and asked for nothing in return but sanctuary. We have answered
them with murder, and rape, and a bigotry that has exceeded all
others in the long and reeking history of human injustice. It is time for
the human race to call a halt; not merely for the sake of our alien
visitors, but for the sake of the almighty soul of Man. We must hold
out our hand, and say 'Welcome. Welcome to Earth!'"
Moved by the plea of the Secretary-General, the assembly voted to
form a 12-nation commission to study the problem. Many
governments made offers of hospitality to the aliens; the United
States, Canada, Australia, the Scandinavian and Low Countries, all
expressed a willingness to set aside land areas for the exclusive use
of the aliens. The U.S.S.R. made no offer of land, but suggested that
the Blues could be completely integrated into Russian society. The
choice was left to the Blues and their leader, an Elder named Trecor.
The wisdom of Trecor's decision became a subject of debate for
generations to come. He declined to isolate his people as a "nation,"
separate and apart from the human race. He declined to have them
as boarders within any one sovereign state. Instead, he asked that
the Blues be divided into small communities and dispersed over the
world, where each could work out their individual destiny. His purpose
was a noble one: he wished to make his people truly the neighbors,
even the partners, of the Earthmen.
And so it was.
In the United States, a Blue community began a collective farming
project on acreage deeded to them by the government in Kansas.
Within three years, the crops of winter wheat and corn produced on
the Blue farmlands were so superior in quality that they provoked the
admiration, and envy, of every farmer in the district. In '77, the year of
the Terrible Twisters, only the Blue farmlands were miraculously
spared the destruction of their fields. The ignorant claimed that some
spiritual agency had helped the Blues; the more enlightened credited
the sturdiness of their crops. But both became united in a sullen
resentment of the Blues, the strangers who had committed the
unpardonable sin of prospering in a season of want. From these
beginnings came the illicit organization of terrorists who called
themselves the Dom-Dom, a name originally meaning Defenders of
Mankind. Between the years 1977 and 1991, the Dom-Dom could
take the blame for the violent deaths of more than a thousand Blues.

In New Zealand, another farming community of Blues fared better

than their fellows in Kansas. But in the year 1982, they fell victim to
the still-unnamed plague which Earthmen merely called the Blue
Disease. It seemed to strike only the aliens, but it resembled typhoid
in its symptoms and deadly progress. The Blues themselves became
unable to cope with the disease; their pleas for outside medical help
brought only a handful of Earth physicians. When one of them, a Dr.
Martin Roebuck, died of a seizure that the Blues swore was unrelated
to the plague, the others fled in fear of contagion. Their statement to
the world press claimed that the biological differences between
Earthmen and Blues were too great for Earth medicine to be of value.
And so the Blues of New Zealand died. The white flash of their
funeral pyres lit the night again and again.
In Russia, a non-farming community of Blues, composed mainly of
artists and scientists, lived in a government-constructed "city" and
were carefully nurtured and pampered like talented, precocious
children. After five years of this treatment, the Blues sickened of it
and yearned for a freer life. With the eyes of the world upon them, the
Russians quickly agreed to the Blue demands. Yes, they could do as
they please, live as they please, work as they please. One by one,
their privileges were withdrawn. The Blues found that they had to
provide their own food, their own clothing, maintain their own
shelters; the Russians had given them independence with a
vengeance. They found themselves unable to care for their own
elementary needs; they were like helpless children; they began
quarreling among themselves. For the first time in the remembered
history of the race, a Blue killed another Blue; it was said that the
shame of this episode was the cause of the Elder Trecor's death.
Eventually, the Blues surrendered; they preferred the easy comforts
of their prison, and begged their jailers to lower the bars again.
In the thirty-six years of the Blues' residence on Earth, only four
thousand births were recorded; while ten thousand of the race
perished.
It was in the year 2009, following the Kansas City Massacre by the
Dom-Dom, in which eight hundred Blues died under the
flamethrowers of the terrorists, that the Decision was reached. It was
relayed to the world by an Elder named Dasru, whose prepared
statement was read to the United Nations.
"We came to your world unbidden and unwelcome," the statement
said. "We came to your world asking no privilege, bearing no arms,
wishing for no more than forbearance for our differences, patience for
our ignorance, and sympathy for our homelessness. We offered love
and received hatred. We came in peace and died in war.
"We love the sweet green fields of your planet, its clear water and
skies, its generous soil. But you have never permitted Earth to
become our home, and so we leave you. We leave you, people of the
cruel planet. Rather than suffer your bigotry, and yes, your tolerance,
we leave you. We go to seek another homeland, and in the minds of
our future generations no memory of this hated visit shall remain. We
shall Forget you, Earth; but may you always remember, what drove
us from your world."
Then the exodus began. One by one, the small spacecraft of the
Blues began to rise towards the heavens. Before the next Spring
came to Earth, the Blues were gone.
Ky-Tann cleared his throat, and looked at his young wife. Devia
stared at Deez.
"How long ago?" Ky-Tann asked. "When did this happen?"
"Perhaps three, four thousand years ago," Deez said. "They left the
Earth to its fate, and eventually that fate was extinction. Some defect
in its sun caused an outburst of nuclear fire, and shriveled the planet
to what it has become. But still she stands, their goddess of welcome,
lifting her torch to the empty skies.
"When we dug up that statue, do you know what we found? There
was an inscription on the base. When we learned the story of that
planet's past, the irony of those words was poignant."
"Do you remember them?"
"I could never forget them," Deez said, and his eyes were dark. "Give
me your tired, your poor, your huddled masses yearning to breathe
free, the wretched refuse of your teeming shore; send these, the
homeless, tempest-tossed, to me; I lift my lamp beside the golden
door."
There was silence. Ky-Tann became aware of his wife's tears.
He went to her, and she wiped her eyes. "I'd better put Su-Tann to
bed," she said, trying to smile at Deez. "Did you see her tooth, Deez?
It's her very first."
Ky-Tann took his wife's blue hand, and kissed her blue cheek. "A
beautiful tooth," he said.
 *** END OF THE PROJECT GUTENBERG EBOOK BESIDE THE
GOLDEN DOOR ***

Updated editions will replace the previous one—the old editions will
be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright in
these works, so the Foundation (and you!) can copy and distribute it
in the United States without permission and without paying copyright
royalties. Special rules, set forth in the General Terms of Use part of
this license, apply to copying and distributing Project Gutenberg™
electronic works to protect the PROJECT GUTENBERG™ concept
and trademark. Project Gutenberg is a registered trademark, and
may not be used if you charge for an eBook, except by following the
terms of the trademark license, including paying royalties for use of
the Project Gutenberg trademark. If you do not charge anything for
copies of this eBook, complying with the trademark license is very
easy. You may use this eBook for nearly any purpose such as
creation of derivative works, reports, performances and research.
Project Gutenberg eBooks may be modified and printed and given
away—you may do practically ANYTHING in the United States with
eBooks not protected by U.S. copyright law. Redistribution is subject
to the trademark license, especially commercial redistribution.

START: FULL LICENSE

THE FULL PROJECT GUTENBERG LICENSE
 PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK

To protect the Project Gutenberg™ mission of promoting the free

distribution of electronic works, by using or distributing this work (or
any other work associated in any way with the phrase “Project
Gutenberg”), you agree to comply with all the terms of the Full
Project Gutenberg™ License available with this file or online at
www.gutenberg.org/license.

Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand, agree
to and accept all the terms of this license and intellectual property
(trademark/copyright) agreement. If you do not agree to abide by all
the terms of this agreement, you must cease using and return or
destroy all copies of Project Gutenberg™ electronic works in your
possession. If you paid a fee for obtaining a copy of or access to a
Project Gutenberg™ electronic work and you do not agree to be
bound by the terms of this agreement, you may obtain a refund from
the person or entity to whom you paid the fee as set forth in
paragraph 1.E.8.

1.B. “Project Gutenberg” is a registered trademark. It may only be

used on or associated in any way with an electronic work by people
who agree to be bound by the terms of this agreement. There are a
few things that you can do with most Project Gutenberg™ electronic
works even without complying with the full terms of this agreement.
See paragraph 1.C below. There are a lot of things you can do with
Project Gutenberg™ electronic works if you follow the terms of this
agreement and help preserve free future access to Project
Gutenberg™ electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright law in
the United States and you are located in the United States, we do
not claim a right to prevent you from copying, distributing,
performing, displaying or creating derivative works based on the
work as long as all references to Project Gutenberg are removed. Of
course, we hope that you will support the Project Gutenberg™
mission of promoting free access to electronic works by freely
sharing Project Gutenberg™ works in compliance with the terms of
this agreement for keeping the Project Gutenberg™ name
associated with the work. You can easily comply with the terms of
this agreement by keeping this work in the same format with its
attached full Project Gutenberg™ License when you share it without
charge with others.

1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the terms
of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.

1.E. Unless you have removed all references to Project Gutenberg:

1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project Gutenberg™
work (any work on which the phrase “Project Gutenberg” appears, or
with which the phrase “Project Gutenberg” is associated) is
accessed, displayed, performed, viewed, copied or distributed:
 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.

1.E.2. If an individual Project Gutenberg™ electronic work is derived

from texts not protected by U.S. copyright law (does not contain a
notice indicating that it is posted with permission of the copyright
holder), the work can be copied and distributed to anyone in the
United States without paying any fees or charges. If you are
redistributing or providing access to a work with the phrase “Project
Gutenberg” associated with or appearing on the work, you must
comply either with the requirements of paragraphs 1.E.1 through
1.E.7 or obtain permission for the use of the work and the Project
Gutenberg™ trademark as set forth in paragraphs 1.E.8 or 1.E.9.

1.E.3. If an individual Project Gutenberg™ electronic work is posted

with the permission of the copyright holder, your use and distribution
must comply with both paragraphs 1.E.1 through 1.E.7 and any
additional terms imposed by the copyright holder. Additional terms
will be linked to the Project Gutenberg™ License for all works posted
with the permission of the copyright holder found at the beginning of
this work.

1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files containing a
part of this work or any other work associated with Project
Gutenberg™.

1.E.5. Do not copy, display, perform, distribute or redistribute this

electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1 with
active links or immediate access to the full terms of the Project
Gutenberg™ License.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if you
provide access to or distribute copies of a Project Gutenberg™ work
in a format other than “Plain Vanilla ASCII” or other format used in
the official version posted on the official Project Gutenberg™ website
(www.gutenberg.org), you must, at no additional cost, fee or expense
to the user, provide a copy, a means of exporting a copy, or a means
of obtaining a copy upon request, of the work in its original “Plain
Vanilla ASCII” or other form. Any alternate format must include the
full Project Gutenberg™ License as specified in paragraph 1.E.1.

1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™ works
unless you comply with paragraph 1.E.8 or 1.E.9.

1.E.8. You may charge a reasonable fee for copies of or providing

access to or distributing Project Gutenberg™ electronic works
provided that:

• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
about donations to the Project Gutenberg Literary Archive
Foundation.”

• You provide a full refund of any money paid by a user who

notifies you in writing (or by e-mail) within 30 days of receipt that
s/he does not agree to the terms of the full Project Gutenberg™
License. You must require such a user to return or destroy all
 copies of the works possessed in a physical medium and
discontinue all use of and all access to other copies of Project
Gutenberg™ works.

• You provide, in accordance with paragraph 1.F.3, a full refund of

any money paid for a work or a replacement copy, if a defect in
the electronic work is discovered and reported to you within 90
days of receipt of the work.

• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.

1.E.9. If you wish to charge a fee or distribute a Project Gutenberg™

electronic work or group of works on different terms than are set
forth in this agreement, you must obtain permission in writing from
the Project Gutenberg Literary Archive Foundation, the manager of
the Project Gutenberg™ trademark. Contact the Foundation as set
forth in Section 3 below.

1.F.

1.F.1. Project Gutenberg volunteers and employees expend

considerable effort to identify, do copyright research on, transcribe
and proofread works not protected by U.S. copyright law in creating
the Project Gutenberg™ collection. Despite these efforts, Project
Gutenberg™ electronic works, and the medium on which they may
be stored, may contain “Defects,” such as, but not limited to,
incomplete, inaccurate or corrupt data, transcription errors, a
copyright or other intellectual property infringement, a defective or
damaged disk or other medium, a computer virus, or computer
codes that damage or cannot be read by your equipment.

1.F.2. LIMITED WARRANTY, DISCLAIMER OF DAMAGES - Except

for the “Right of Replacement or Refund” described in paragraph
1.F.3, the Project Gutenberg Literary Archive Foundation, the owner
of the Project Gutenberg™ trademark, and any other party
distributing a Project Gutenberg™ electronic work under this
agreement, disclaim all liability to you for damages, costs and
expenses, including legal fees. YOU AGREE THAT YOU HAVE NO
REMEDIES FOR NEGLIGENCE, STRICT LIABILITY, BREACH OF
WARRANTY OR BREACH OF CONTRACT EXCEPT THOSE
PROVIDED IN PARAGRAPH 1.F.3. YOU AGREE THAT THE
FOUNDATION, THE TRADEMARK OWNER, AND ANY
DISTRIBUTOR UNDER THIS AGREEMENT WILL NOT BE LIABLE
TO YOU FOR ACTUAL, DIRECT, INDIRECT, CONSEQUENTIAL,
PUNITIVE OR INCIDENTAL DAMAGES EVEN IF YOU GIVE
NOTICE OF THE POSSIBILITY OF SUCH DAMAGE.

1.F.3. LIMITED RIGHT OF REPLACEMENT OR REFUND - If you

discover a defect in this electronic work within 90 days of receiving it,
you can receive a refund of the money (if any) you paid for it by
sending a written explanation to the person you received the work
from. If you received the work on a physical medium, you must
return the medium with your written explanation. The person or entity
that provided you with the defective work may elect to provide a
replacement copy in lieu of a refund. If you received the work
electronically, the person or entity providing it to you may choose to
give you a second opportunity to receive the work electronically in
lieu of a refund. If the second copy is also defective, you may
demand a refund in writing without further opportunities to fix the
problem.

1.F.4. Except for the limited right of replacement or refund set forth in
paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.

1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of damages.
If any disclaimer or limitation set forth in this agreement violates the
law of the state applicable to this agreement, the agreement shall be
interpreted to make the maximum disclaimer or limitation permitted
by the applicable state law. The invalidity or unenforceability of any
provision of this agreement shall not void the remaining provisions.
1.F.6. INDEMNITY - You agree to indemnify and hold the
Foundation, the trademark owner, any agent or employee of the
Foundation, anyone providing copies of Project Gutenberg™
electronic works in accordance with this agreement, and any
volunteers associated with the production, promotion and distribution
of Project Gutenberg™ electronic works, harmless from all liability,
costs and expenses, including legal fees, that arise directly or
indirectly from any of the following which you do or cause to occur:
(a) distribution of this or any Project Gutenberg™ work, (b)
alteration, modification, or additions or deletions to any Project
Gutenberg™ work, and (c) any Defect you cause.

Section 2. Information about the Mission of

Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new computers.
It exists because of the efforts of hundreds of volunteers and
donations from people in all walks of life.

Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project Gutenberg™’s
goals and ensuring that the Project Gutenberg™ collection will
remain freely available for generations to come. In 2001, the Project
Gutenberg Literary Archive Foundation was created to provide a
secure and permanent future for Project Gutenberg™ and future
generations. To learn more about the Project Gutenberg Literary
Archive Foundation and how your efforts and donations can help,
see Sections 3 and 4 and the Foundation information page at
www.gutenberg.org.

Section 3. Information about the Project

Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-profit
501(c)(3) educational corporation organized under the laws of the
state of Mississippi and granted tax exempt status by the Internal
Revenue Service. The Foundation’s EIN or federal tax identification
number is 64-6221541. Contributions to the Project Gutenberg
Literary Archive Foundation are tax deductible to the full extent
permitted by U.S. federal laws and your state’s laws.

The Foundation’s business office is located at 809 North 1500 West,

Salt Lake City, UT 84116, (801) 596-1887. Email contact links and up
to date contact information can be found at the Foundation’s website
and official page at www.gutenberg.org/contact

Section 4. Information about Donations to

the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission of
increasing the number of public domain and licensed works that can
be freely distributed in machine-readable form accessible by the
widest array of equipment including outdated equipment. Many small
donations ($1 to $5,000) are particularly important to maintaining tax
exempt status with the IRS.

The Foundation is committed to complying with the laws regulating

charities and charitable donations in all 50 states of the United
States. Compliance requirements are not uniform and it takes a
considerable effort, much paperwork and many fees to meet and
keep up with these requirements. We do not solicit donations in
locations where we have not received written confirmation of
compliance. To SEND DONATIONS or determine the status of
compliance for any particular state visit www.gutenberg.org/donate.

While we cannot and do not solicit contributions from states where

we have not met the solicitation requirements, we know of no
prohibition against accepting unsolicited donations from donors in
such states who approach us with offers to donate.

International donations are gratefully accepted, but we cannot make

any statements concerning tax treatment of donations received from
outside the United States. U.S. laws alone swamp our small staff.

Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.

Section 5. General Information About Project

Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could be
freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose network of
volunteer support.

Project Gutenberg™ eBooks are often created from several printed

editions, all of which are confirmed as not protected by copyright in
the U.S. unless a copyright notice is included. Thus, we do not
necessarily keep eBooks in compliance with any particular paper
edition.

Most people start at our website which has the main PG search
facility: www.gutenberg.org.

This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg Literary
Archive Foundation, how to help produce our new eBooks, and how
to subscribe to our email newsletter to hear about new eBooks.