Biostatistics Manual for Health Research: A Practical Guide to Data Analysis Nafis Faizi

Biostatistics Manual for Health Research: A
Practical Guide to Data Analysis Nafis Faizi
install download
https://ebookmeta.com/product/biostatistics-manual-for-health-
research-a-practical-guide-to-data-analysis-nafis-faizi/
Download more ebook from https://ebookmeta.com

We believe these products will be a great fit for you. Click
the link to download now, or visit ebookmeta.com
to discover even more!
Highway Capacity Manual A Guide for Multimodal Mobility
Analysis 7th Edition Transportation Research Board
https://ebookmeta.com/product/highway-capacity-manual-a-guide-
for-multimodal-mobility-analysis-7th-edition-transportation-
research-board/
Applied Longitudinal Data Analysis for Medical Science:
A Practical Guide 3rd Edition Twisk
https://ebookmeta.com/product/applied-longitudinal-data-analysis-
for-medical-science-a-practical-guide-3rd-edition-twisk/
Introduction to Health Research Methods A Practical
Guide 3rd ed 3rd Edition Jacobsen Kathryn H
https://ebookmeta.com/product/introduction-to-health-research-
methods-a-practical-guide-3rd-ed-3rd-edition-jacobsen-kathryn-h/
Can The As 400 Survive Ibm This Classic 2004 Saga About
The As 400 Will Make As 400 Aficionados Tear It Is A
Great Historical Perspective As To Why Ibm Customers
Were Asking For More 2004th Edition Brian W Kelly
https://ebookmeta.com/product/can-the-as-400-survive-ibm-this-
classic-2004-saga-about-the-as-400-will-make-as-400-aficionados-
tear-it-is-a-great-historical-perspective-as-to-why-ibm-

Historicizing Theories Identities and Nations 1st
Edition Regna Darnell
https://ebookmeta.com/product/historicizing-theories-identities-
and-nations-1st-edition-regna-darnell/
The Twenty First Century and Its Discontents How
Changing Discourse Norms are Changing Culture 4th
Edition Jack Simmons
https://ebookmeta.com/product/the-twenty-first-century-and-its-
discontents-how-changing-discourse-norms-are-changing-
culture-4th-edition-jack-simmons/
Racial Trauma in the School System Naming the Pain 1st
Edition Connesia Handford
https://ebookmeta.com/product/racial-trauma-in-the-school-system-
naming-the-pain-1st-edition-connesia-handford/
Stolen Heart Bride of the Billionaire 4 1st Edition
Jenna Rose
https://ebookmeta.com/product/stolen-heart-bride-of-the-
billionaire-4-1st-edition-jenna-rose/
Communicating Social and Environmental Issues
Effectively 1st Edition Betsy Reed
https://ebookmeta.com/product/communicating-social-and-
environmental-issues-effectively-1st-edition-betsy-reed/

OpenShift for Developers: A Guide for Impatient
Beginners 2nd Edition Joshua Wood
https://ebookmeta.com/product/openshift-for-developers-a-guide-
for-impatient-beginners-2nd-edition-joshua-wood/

BIOSTATISTICS
MANUAL FOR
HEALTH RESEARCH

This page intentionally left blank

BIOSTATISTICS
MANUAL FOR
HEALTH RESEARCH
A Practical Guide to Data Analysis
NAFIS FAIZI
Assistant Professor, Jawaharlal Nehru Medical College,
Aligarh Muslim University, Aligarh, India
YASIR ALVI
Assistant Professor, Hamdard Institute of Medical Sciences and
Research, New Delhi, India

Academic Press is an imprint of Elsevier
125 London Wall, London EC2Y 5AS, United Kingdom
525 B Street, Suite 1650, San Diego, CA 92101, United States
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
Copyright © 2023 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic
or mechanical, including photocopying, recording, or any information storage and retrieval system,
without permission in writing from the publisher. Details on how to seek permission, further
information about the Publisher’s permissions policies and our arrangements with organizations such
as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website:
www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this ﬁeld are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical
treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating
and using any information, methods, compounds, or experiments described herein. In using such
information or methods they should be mindful of their own safety and the safety of others, including
parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume
any liability for any injury and/or damage to persons or property as a matter of products liability,
negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas
contained in the material herein.
ISBN: 978-0-443-18550-2
For information on all Academic Press publications visit our
website at https://www.elsevier.com/books-and-journals
Publisher: Stacy Masucci
Acquisitions Editor: Linda Versteeg-Buschman
Editorial Project Manager: Matthew Mapes
Production Project Manager: Fahmida Sultana
Cover Designer: Mark Rogers
Typeset by TNQ Technologies
All the IBMÒ SPSSÒ Statistics software (“SPSS”) screenshots are permitted for use in this book
through “Reprint Courtesy of IBM Corporation ©”.
MedCalc Software screenshots are permitted for use in the book by M Frank Schoonjans, MedCalc
Software Ltd.

Contents
About the authors ix
Preface xi
List of abbreviations xiii
1. Introduction to biostatistics 1
1. Background 1
2. What is biostatistics? 3
3. Statistical inference 4
4. Aim of the book 5
5. Two foundational concepts 6
6. Data and variables 9
7. Measures of central tendency and dispersion 12
References 16
2. Data management and SPSS environment 17
1. Data management 17
2. Data documentation sheet 20
3. Data capture and cleaning 21
4. SPSS environment 25
5. Data entry and importing in SPSS 30
6. Data transformation in SPSS 37
References 43
3. Statistical tests of significance 45
1. Hypothesis testing 45
2. Statistical tests of significance 47
3. Choosing a statistical test 54
4. Levels of significance and P-values 55
5. Errors in health research 56
6. P-values and effect sizes 59
References 62
4. Parametric tests 63
1. Continuous outcomes 63
2. Parametric tests 63
3. t-Tests: independent and paired 65
4. Independent t-test 65
v

5. Paired t-test 68
6. Parametric tests comparison with >2 groups: analysis of variance 71
7. Repeated-measures ANOVA 77
8. ANOVA, ANCOVA, MANOVA, and MANCOVA 82
References 85
5. Nonparametric tests 87
1. Nonparametric methods 87
2. ManneWhitney U test 88
3. Wilcoxon signed-rank test 93
4. Nonparametric tests comparison with >2 groups: KruskaleWallis test 98
5. Nonparametric tests comparison with >2 related or repeated measures: Friedman test 102
References 107
6. Correlation 109
1. Continuous outcome and exposure 109
2. Correlation versus association 111
3. Pearson’s correlation test 111
4. Spearman’s correlation test 115
5. Correlation versus concordance 120
6. Agreement: Kendall’s s, Kendall’s W, and kappa 120
7. Measuring concordance/agreement 122
References 126
7. Categorical variables 127
1. Categorical variables 127
2. Independent exposure variables: chi-square test 127
3. Alternatives to chi-square test 132
4. Two related exposure variables: McNemar’s test 138
5. More than two related exposure variables: Cochran’s Q test 143
6. Analyzing the summary data 147
References 148
8. Validity 149
1. Validity 149
2. Diagnostic test evaluation 151
3. Diagnostic test evaluation: calculations 158
4. Combining screening tests 161
5. Continuous data and ROC curves 163
References 168
vi Contents

9. Reliability and agreement 171
1. Reliability and agreement 171
2. Reliability methods for categorical variables 174
3. Cohen’s kappa test 175
4. Weighted Cohen’s kappa test 178
5. Fleiss kappa test 183
6. Agreement and concordance: which test to use? 187
7. Reliability for continuous variables: intraclass correlation 187
8. Cronbach’s alpha 191
References 194
10. Survival analysis 195
1. Time to event as a variable 195
2. Survival analysis 196
3. KaplaneMeier survival method 199
4. Cox regression survival method 204
References 211
11. Regression and multivariable analysis 213
1. Regression and multivariable analysis 213
2. Regression analysis 219
3. Linear regression 220
4. Simple linear regression analysis 223
5. Multiple linear regression analysis 231
6. Logistic regression analysis 237
7. Multiple logistic regression analysis 239
8. Multivariable analysis 246
References 247
12. Annexures 249
1. Annexure 1: Choice of statistical tests 249
2. Annexure 2: Notes on data used in the book 249
3. Annexure 3: Guidelines for statistical reporting in journals: SAMPL guidelines 249
4. Annexure 4: Standards for reporting of diagnostic accuracy: STARD guidelines 258
5. Annexure 5: Guidelines for reporting reliability and agreement studies: GRRAS guidelines 261
6. Annexure 6: Proposed agenda for biostatistics for a health research workshop 262
References 263
Index 265
Contents vii

About the authors
Nafis Faizi
Dr. Nafis Faizi is an Assistant Professor and Epidemiologist at Jawaharlal Nehru Med-
ical College, Aligarh Muslim University, India. He is currently a fellow of Health Policy
and Systems Research (India HPSR Fellow) and an Academic Editor of Plos Global Public
Health. He is also an active trainer for Epidemiological Research Unit and a member of
Statistics Without Borders and Global Health Training Network. For the past 9 years, he
has been conducting regular workshops and training in biostatistics, data analysis, and
research writing. His primary qualifications are MBBS and MD in Community Medicine
from India followed by master’s in Public Health (MPH) from the United Kingdom be-
sides multiple executive courses and trainings including those from JPAL, SPSS South
Asia, and International Union against TB and Lung Diseases (Operational Research).
He is also a faculty for International People’s Health University and teaches epidemiology
and biostatistics at Victoria University. Previously, he worked as Scientist E at the Indian
Council of Medical Research-National Institute of Malaria Research (ICMR-NIMR).
He is a member of multiple professional associations including Statistics Without Borders,
International Epidemiological Association, Health Action International, IAPSM, and
IPHA.
Yasir Alvi
Dr. Yasir Alvi is an Assistant Professor at Hamdard Institute of Medical Sciences and
Research, New Delhi, India. An alumnus of Aligarh Muslim University, his primary
qualifications are MBBS and MD in Community Medicine. He has more than 8 years
of teaching and research experience and has published numerous academic articles in
reputed journals focusing on public health, mental health, HIV, tuberculosis, and
COVID-19. He has been the lead investigator and statistical consultant in more than a
dozen projects funded by the WHO, UNICEF, ICMR, and various research institutions.
He is regularly involved in training and human resource development of public health
students and healthcare providers on data analysis, research writing, and biostatistics.
ix

Preface
In 1937, Sir Austin Bradford Hill wrote, “Statistics are curious things. They afford one
of the few examples in which the use, or abuse, of mathematical methods tends to induce
a strong emotional reaction in non-mathematical minds. This is because statisticians
apply, to problems in which we are interested, a technique which we do not understand.
It is exasperating, when we have studied a problem by methods that we have spent labo-
rious years mastering, to find our conclusions questioned, and perhaps refuted, by some-
one who could not have made the observations himself. It requires more equanimity than
most of us possess to acknowledge that the fault is in ourselves.”
Over the past decade, we have provided statistical consulting and training to health
researchers and have encountered their difficulties in biostatistical application. We
have also conducted numerous biostatistics workshops, primarily based on SPSS. The
most recurrent feedback postworkshop was the need for a biostatistics book that could
help in day-to-day research. Clinicians and public health researchers typically have
dual roles in addition to researchdboth in the services sector as well as in teaching
and/or administration. There are excellent books on biostatistics, but most are theory-
laden and do not help with practical applications. We have attempted to write a book
that bridges this gap, provides enough theory, and delves into the applications and inter-
pretations of biostatistical tests. We have also provided boxes in each chapter to highlight
the problems that arise from wrong application or choice of tests, which to our surprise is
quite common.
Our aim is to equip health researchers with all the necessary tools they need to confi-
dently apply biostatistics and interpret their meanings correctly. This book is written as an
instruction manual for applying, comprehending, and interpreting biostatistics rather
than delving deeply into the theoretical underpinnings and heavy statistical calculations.
This book originally began as a design for a handbook in our workshops and training ses-
sions on epidemiological research, but has evolved into its current shape thanks to con-
tributions from training participants, peers, and students. The book has been designed for
12 sessions to be conducted over 3 intensive days (an agenda is provided in annexure).
We provide hands-on data with details for practice. Ideally, such a three-day session
would be most beneficial for researchers who are preparing to write their research pro-
tocols, dissertations, or scientific papers.
This book has been written as an aid to the few biostatistics enthusiasts who stand as
troubleshooters for the entire medical college, hospital, or institute. The book would
serve as a very helpful rapid reference guide for epidemiology units, research advisory
committees, and medical education units/departments. We believe that an experienced
xi

epidemiologist or health researcher can also conduct a workshop based on this manual. If
anyone plans to hold such a workshop, we ask that the book be provided as a part of the
workshop kit so that all participants can beneﬁt from it and refer to it in the future. We
continue to conduct workshops on biostatistics based on this manual, and would love to
help in conducting such sessions.
Please feel free to advise, suggest, comment, and criticize the contents as well as in-
tents of this work. We would be especially grateful if you could point out any errorsd
inadvertent or otherwisedin the book. This book is devoted to encouraging students
and scholars to conduct research with a sense of curiosity. It remains the most effective
source of hope in these trying times.
Naﬁs Faizi and Yasir Alvi
xii Preface

List of abbreviations
ANCOVA Analysis of Covariance
ANOVA Analysis of Variance
ASA American Statistical Association
AUC Area Under Curve
BMI Body Mass Index
COPE Committee on Publication Ethics
DDS Data Documentation Sheet
GIGO Garbage In, Garbage Out
GRRAS Guidelines for Reporting Reliability and Agreement Studies
ICC Intra Class Correlation
ICMJE International Committee of Medical Journal Editors
IEC Institutional Ethics Committee
IQR Interquartile Range
KS or KeS test KolmogoroveSmirnov test
LIMO Linearity, Independence, Multicollinearity, and Outliers
LINE-M Linearity, Independence, Normality, Equality of variances, and Multicollinearity
LR Likelihood Ratio
LSD or Fisher’s LSD Least Significant Difference
MANOVA Multivariate Analysis of Covariance
NHST Null Hypothesis Significance Testing
NPV Negative Predictive Value
OR Odds Ratio
PCA Principal Component Analysis
PPV Positive Predictive Value
RMANOVA Repeated Measures Analysis of Variance
ROC curve Receiver Operating Characteristic Curve
SAMPL Statistical Analyses and Methods in the Published Literature
SD Standard Deviation
SE or SEM Standard Error of the Mean
SPSS Statistical Package for Social Sciences
STARD Standards for Reporting of Diagnostic Accuracy
SW or SeW test ShapiroeWilk test
TSA Time Series Analysis
Tukey’s HSD Tukey’s Honestly Significant Difference test
VIF Variance Inflation Factor
WHO World Health Organization
xiii

CHAPTER 1
Introduction to biostatistics
Statistical thinking will one day be as necessary a qualification for efficient citizenship as the abil-
ity to read and write.
H.G. Wells
1. Background
In 2010, while reading a popular biostatistics book, we came across a statement by Pro-
fessor Frederick Mosteller on statistics that struck a chord with us and continues to do so.
He said, “It is easy to lie with statistics, but it is easier to lie without them” (Pagano & Gauvreau,
2000). In human health and medicine, statistics are vital to gauge uncertainties and var-
iations to measure, interpret, and analyze them. This in turn helps to determine whether
they are due to biological, genetic, behavioral, or environmental variability or simply due
to chance (Indrayan & Malhotra, 2017).
However, the contribution of data and statistics in medical education and public
health is sometimes taken for granted. Despite the importance of data, little efforts are
made in many hospitals (both private and public) to analyze the data and calculate vital
information such as average duration of stay of malaria patients in the hospital ward, anal-
ysis of the catchment area of different specialties, time taken by a patient to finally see a
doctor in a government hospital, and other data. These data have an effect on everyday
practice, decision-making, and the working of hospitals. Healthcare is knowledge-based,
and knowledge is created through careful transformation and treatment of data, scientific
temper, and available expertise. While times have changed, we still find Florence Night-
ingale’s note on hospitals relevantd“In attempting to arrive at the truth, I have applied every-
where for information, but in scarcely an instance have I been able to obtain hospital records fit for any
purposes of comparison” (Nightingale, 1863).
With digitalization and smart technologies, data is everywhere but not often con-
verted to meaningful information and even lesser to beneficial knowledge. T.S. Eliot’s
the Rock had a great reminder- “..Where is the life we have lost in living, where is
the wisdom we have lost in knowledge, where is the knowledge we have lost in infor-
mation..” “Poor data is worse than no data”, and observations based on experience and
eminence alone are worse than poor data, as they are poorly recorded and often, biased
(Faizi et al., 2018). Knowledge in the era of evidence-based medicine critically depends
on carefully conducted research and its statistical analysis, rather than the Alice in
Biostatistics Manual for Health Research
ISBN 978-0-443-18550-2, https://doi.org/10.1016/B978-0-443-18550-2.00006-2
© 2023 Elsevier Inc.
All rights reserved. 1

Wonderland saying, ‘I’m older than you, and must know better” (Carroll, 1991). However,
even within these limitations, we have different statistical tools which help in understand-
ing these uncertainties and limitations better. We must strive to “become aware of the nature
and extent of these imperfect informations” instead of getting “paralyzed by this lack of knowledge’
(Cohen et al., 1987).
Let us consider this very important example from the 1950s regarding the then
considered treatment of coronary artery disease (CAD) (Belle et al., 2004). CAD is a
widely prevalent disease in which the coronary arteries get occluded, leading to angina
pectoris (pain in the chest). Further narrowing leads to deprivation of blood supply to
heart muscles, which eventually leads to Myocardial Infarction (MI), commonly known
as a Heart Attack. In the 1950s, it was widely believed that a large blood supply would be
forced to the heart by ligating internal mammary arteries (which supplies blood to the
chest). Not only was this considered promising, but it was also carried out with reason-
ably successful results. It gathered a fair amount of support till adequately designed studies
were conducted. We reproduce the results of one such study to emphasize the impor-
tance of statistics (Dimond et al., 1960). This study took 18 patients randomly selected
for internal mammary arteries (IMA) ligation or sham operation. The sham operation
consisted of a similar incision with exposure of IMA but no ligation. Both the cardiologist
and the patients were blind to the actual procedure. Table 1.1 shows the results of the
study. Please note that the results in Table 1.1 are based on the research paper (Dimond
et al., 1960), but the groups have been clubbed for the sake of illustration.
Even a preliminary look at the data indicates no real difference between the sham
operation and ligation of IMA. Based on this observation alone, some may even be temp-
ted to say that the sham operation was actually better, as every patient felt cured/
benefitted after the operation. However, there is always an element of chance in our ob-
servations, which must be accounted for, before interpreting the results. Numerical ob-
servations alone could be simply due to chance. This is something we will discuss later in
this book. For now, we apply Fisher’s exact test on this table, and the p-value is 0.28.
Such a p-value means that the difference between the effects is not significant and could
be due to chance. This means that the perception of cure/benefit was not significantly
different with the Sham operation. Therefore, such statistical inference is vital. With
Table 1.1 Surgical benefit postligation of internal mammary artery (patients’ opinion).
Perception of surgical effects on
symptoms
Ligation of internal mammary
artery Sham operation
Cured/benefitted 9 5
Disappointed/no long-term
improvement
4 0
Total 13 5
2 Biostatistics Manual for Health Research

the careful application of epidemiology and research, statistics has truly transformed mod-
ern medicine, saving millions of lives. So, the next time you feel that it is too difficult to
attempt climbing this learning curve, ask yourself whether you would consent for such an
operation that ligates your artery in a futile attempt to improve angina pectoris.
2. What is biostatistics?
Biostatistics (biostatistics) is defined as “statistical processes and methods applied to the collection,
analysis, and interpretation of biological data and especially data relating to human biology, health,
and medicine” (Merriam-Webster, n.d.). However, if biostatistics is limited to human
biology, health, and medicine, one is tempted to ask why not learn statistics itself?
2.1 Why biostatistics rather than statistics?
There are three distinct reasons for the focus on biostatistics in health research, rather than
statistics (Belle et al., 2004). The three reasons are methods, language, and application.
The first reason is the statistical methods. Some statistical methods have a distinct and
everyday use in biostatistics unlike in statistics, requiring due attention and concerndfor
example, survival life-table analyses. Second, every subject has its language and the lan-
guage of biostatistics is closer to biomedical and healthcare areas. The steep language
curve to bridge the language of both these disciplines is easier to scale with biostatistics
rather than statistics. The third is the application of biostatistical analysis. Owing mainly
to its directed and tailored language, biostatistical analysis is directly understood and
applied in healthcare and related policies. The application of biostatistics is ubiquitous
in healthcare and is easily understood and applied among healthcare workers.
2.2 Role of biostatistics
Biostatistics is an essential tool for health research and clinical decision-making as well as
healthcare research. While this manual focuses on its role in data analysis and interpretation,
it is essential to note that biostatistics has a significant role in the following steps of research:
1. Research planning and designing
2. Data collection
3. Data entry and management
4. Data analysis
5. Data interpretation and presentation.
In health research, more often than not, the biostatistician is consulted during or after
the data entry for further analysis. In many cases, this is too late. Often, even resuscitation
attempts could be rendered futile at such a stage. The famous statistician Ronald Fisher
aptly commented in the first session of the Indian Statistical Conference, in Calcutta in
1938, “To call in the statistician after the experiment is done may be no more than asking him
to perform a post-mortem examination: he may be able to say what the experiment died of.”
Introduction to biostatistics 3

3. Statistical inference
Our world often revolves around the careful recognition of patterns and associations. Some-
times, they are entirely wrong, such as the many conspiracy theories and disinformation
campaigns circulating daily on social media. Statistical and epidemiological illiteracy aids
in making such mistakes. Such progressions are even pathological, as we see in cases of para-
noia. This extreme error of perception is called apophenia, a tendency to interpret random
patterns as meaningful. If apophenia is one extreme, the other extreme is unofficially called
randomania, which is a failure to recognize or appreciate meaningful patterns when they
exist. In statistical inferential terms, apophenia is a bigger problem as it is a Type 1 error,
whereas randomania is akin to a Type 2 error. We will discuss this more in Chapter 3.
Statistical inference or inferential statistics is the “process through which inferences about a
population are made based on certain statistics calculated from a sample of data drawn from that pop-
ulation” (Johnson et al., 2012). How to draw statistical inferences is the primary purpose
of this book. In simpler terms, it answers a simple question, whether the numerical or
observational difference in the data is significant, or is it due to chance? Essentially there
are three forms of such inferences:
1. Point estimation: In point estimation, we are interested in a single number for the
target population from a sample. For example, What is the mean weight of infant
boys in rural Delhi? Through a designed cross-sectional study, we find that the
mean weight is 10 kg.
2. Interval Estimation is about estimating the unknown parameter of the population
that lies within the two intervals, as calculated from the sample. For example, the 95%
confidence interval of the mean weight of infant boys in Delhi was found from the
sample to be 9.1e11.0. This means that there is a 95% chance that the values of the
actual population will lie between these values.
3. Hypothesis testing: This starts with a claim/assumption about the data-null hypoth-
esis, and we check through data whether the claim is true or false. We will learn more
about this in Chapter 3.
In terms of logic and reasoning, we must be aware that the inferential process in sta-
tistics and most science is inductive and not deductive like mathematics. While we refrain
from a detailed discussion on this, it is essential to clearly understand inductive and deduc-
tive reasoning. Taking full responsibility of oversimplification, we leave it to Sherlock
Holmes and Aristotle to explain this further in Fig. 1.1.
In dire opposition to the norms of police investigations, Sherlock Holmes gathered
evidence without proposing a theory. He clearly explains his process akin to statistical
inferential reasoning, “Let me run over the main steps. We approached the case, you remember,
with an absolutely blank mind, which is always an advantage. We had formed no theories. We were
simply there to observe and draw inferences from our observations" (Doyle, 1893).

4. Aim of the book
This book is designed as a manual for doctors and health researchers, focusing on a prac-
tical understanding of applied biostatistics in health research. The book also deals with
introductory biostatistics essential for any researcher. Over the years, we came across
questions without satisfactory answers, such as the difference between association and
correlation or the exact reason why we cannot apply multiple t-tests instead of
ANOVA. We have tried to cover such questions in boxes. However, we have tried
our best to refrain and restrict ourselves (sometimes with great difﬁculty) to only as
much theory as is required to interpret, choose, and analyze statistical tests correctly.
There is an abundance of theory-laden textbooks on biostatistics which are necessary
to develop advanced skills in biostatistics. The contents of this book are as relevant and
essential for understanding, reading, and writing papers, as it is for conducting research.
The book has been written in the form of a manual, as we believe that applied biosta-
tistics can be best learned in a workshop-based environment. The book has been arranged
accordingly so that each chapter is loosely based on a session (or atleast two in case of
regression), and we also propose a workshop-based agenda should the medical education,
and/or research units need to adopt a methodology that we have used for quite some
time (see Annexures, Chapter 12). We use SPSS as the statistical software to perform
the statistical calculations. In the next chapter, we introduce the SPSS software and pro-
vide screenshots of how to analyze the statistical tests under discussion in the following
Figure 1.1 The inductive Holmes and the deductive Aristotle. Comparing deductive approach of
Aristotle and inductive approach of Sherlock Holmes.

chapters. The second chapter also engages with relevant data management principles,
including data entry, which is deprioritized to the level of deliberate and wilful neglect,
despite its huge importance. The third chapter is essential to learn a few crucial concepts
before analyzing any data and applying statistical tests. We are vehemently opposed to
“garbage in and garbage out (GIGO)” practices where we commit gross injustice by play-
ing with our data and submitting it to software to create sophisticated figures and statistics.
At best, they contribute to creating a pile of rubbish research (along with noise pollution),
and at worse, they find their way to better journals and affect guidelines and policy. The
latter imperils human health and the former insults human mind. After the first three
chapters, the rest of the chapters deal with statistical tests in different situations.
To summarize, the main objectives of this manual are to introduce biostatistics in a
way that readers can:
1. Choose and apply appropriate statistical tests.
2. Interpret and present the findings of the statistical tests correctly.
3. Understand and review the results and statistical interpretations of most research pub-
lished in medical journals.
4. Use SPSS for statistical analysis with ease and without anyone’s help.
5. Two foundational concepts
There are two concepts that lay down the foundation for what we will learn later (they
never fail to amaze us). The first is the law of large numbers, and the other is the central
limit theorem (Box 1.1).
5.1 Law of large numbers
The law of large numbers is a probability concept that states, “The more you sample, the
truer your sample mean is to the average mean of the population.” This means that as we increase
the sample size of our research, the chances of our sample average being closer to the
actual population average increase. This is extremely important for research as it estab-
lishes the consistency of the estimator and predicts the validity of the results in a wider
population. In other words it helps in a strong internal and external validity as well as
generalizability. Remember the story of the hare and the tortoise? The one in which
BOX 1.1 Central limit theorem and law of large numbers
Both these theorems are important theorems about the sample mean. The law of large
numbers states that the sample mean approaches the population mean as n gets large. On
the other hand, the central limit theorem states that multiple sample means approach a normal
distribution as n gets large (n denotes sample size).

the slow and steady tortoise won the race? While the message could be true, the evidence
behind it is not. The story lacks an adequate sample size for such a bold inference. Fiction
and anecdotes can be helpful in highlighting messages but should not be a hindrance to
evidence. A place for everything and everything in its place.
Another way to understand this is through the good old coin with two facesdhead
and tail. The probability of getting head is 50%, but this may not be evident when Mr.
Lucky goes for a toss. Mr. Lucky has the habit of winning tosses. The law of large number
states that Mr. Lucky has not been observed so much. If he participates in a “large” num-
ber of tosses and says head every time, the chances are that he would lose almost 50% of
the time (Table 1.2). Gambler’s fallacy is an interesting fallacy (Box 1.2).
Table 1.2 Law of large numbers in a coin toss.
Coin toss Result
Probability of heads
(out of 100 %)
Number of tosses Heads Tails
1 1 0 1/1 ¼ 100%
10 7 3 7/10 ¼ 70%
100 65 35 65/100 ¼ 65%
1000 601 399 601/1000 ¼ 60.1%
10,000 5402 4598 5402/10,000 ¼ 54.02%
100,000 50,100 49,900 50,100/100,000 ¼ 50.10%
BOX 1.2 Gambler’s fallacy
A common erroneous claim arises due to the belief that any number of the observed sample
could represent a large population. Also known as the Monte Carlo fallacy, this is a famous story
from the Monte Carlo casino in 1913. Gamblers mistakenly believe that if s/he is losing (or not
getting the number) s/he expects, it would even out in the next turn, and they will win that turn.
In Monte Carlo Casino, the ball fell on black 26 times in a row. As this was an extremely uncom-
mon occurrence, the gamblers thought that the next will not be black. They lost millions putt-
ing a bet against black, believing that the streak was extremely unlikely and had to be followed
by a streak of red.
This is due to their mistaken belief that if something is more frequent now, it will become
less likely in the future. This is a fallacy as the next event or turn is “independent” of the previous
event. This is similar to the joke on the poor deduction of surgical optimism. The surgeon
proclaimsd“Nine times out of 10, this operation is unsuccessful. This makes you a lucky patient
as you are patient number 10.”

5.2 Central limit theorem
The central limit theorem states that given certain conditions if multiple samples are taken
from a population, the means of the samples would be normally distributed, even if the population is
not normally distributed. Unlike the Law of Large Numbers, central limit theorem is a sta-
tistical theory, not a probability concept. In other words, if the means/averages of all the
different studies from a population are plotted, we get a normally distributed curve. The
mean/average of such a curve is the average of the population.
The normal distribution curve is also called the Gaussian curve after the name of Carl
Friedrich Gauss who discovered it in 1809. Many scientists are of the opinion that
Gaussian curve is a more appropriate name as the underlying distribution has both the
so-called “normal” and “abnormal” values. In 1810, Marquis de Laplace proved the cen-
tral limit theorem, validating and upholding its importance (Laplace, 1810).
As we would see later, the normal distribution curve is a bell-shaped curve with
certain important properties that have profound implications for statistical tests (Fig. 1.2).
For now, it would sufﬁce to know two of its properties:
1. The graph is symmetrical around its highest point. The highest point is the mean (m,
symbol for mean or average).
2. The distribution follows the 68e95-99.7 rule. This means that in a normally distrib-
uted data, 68% of the population has a value within mean 1 standard deviation
(m 1 SD), 95% have a value within mean 2 standard deviations (m 2 SD),
and 99.7% within a mean 3 standard deviations (m 3 SD).
While discussing the 68-95-99.7 rule, another less important but interesting theorem
is Chebyshev’s inequality theorem. Chebyshev’s inequality theorem states that regardless of
Figure 1.2 The 68e95-99.7 rule. Bell’s curve showing the area covered by 1,2, and 3 standard devi-
ations. (From Wikipedia commons (CC-BY-4.0).)

the probability distribution, at least 1 1/k 2
of the distribution’s values are always
within k standard deviations of the mean. In other words, regardless of the distribution,
at least 75% of the values will always lie within two standard deviations (1 1/k2
, with
k ¼ 2), and at least 88.8% of the values within three standard deviations (1 1/k2
, with
k ¼ 3).
6. Data and variables
6.1 Data
The word data refers to observations and measures often collected through research.
When data are carefully arranged and analyzed, it becomes information. Data are classi-
fied as qualitative and quantitative. While qualitative data are nonnumeric, quantitative
data are numeric. In statistical software and study questionnaires, we prefer entering data
in numbers and assigning values for each number. For example, blood groups A, B, O,
and AB could be coded as 1, 2, 3, and 4. This makes it more efficient as it is less prone to
error while collecting and entering, as we will discover later. An important subclassifi-
cation of data type is NOIR, that is, nominal, ordinal, interval, and ratio data. While
nominal and ordinal data are qualitative, interval and ratio are quantitative (Table 1.3).
Table 1.3 Types of data.
Data type Nominal Ordinal Scale
Nature Qualitative Qualitative Quantitative
Assessment Labels Ordered/ranked Countable/
measurable
Expression Proportions/
percentages
Proportions/
percentages
Average or mean
Measure of central
tendency
Not applicable Mode Mean, median, mode
Examples Normal, anemic Mild, moderate and
severe anemia
Hemoglobin levels
(Hb g%)
Obese, normal Morbidly obese,
obese, overweight
Body mass index
(BMIs in kg/m2
)
Normotensive,
hypertensive
Mild, moderate, and
severe
hypertension
Blood pressure (mm
g)
Vaccine vial monitor
(VVM): useable or
unusable
4 stages of VVM Not applicable
Pediatric, adults,
geriatric
Age groups: lowest,
lower, higher,
highest (0e10, 10
e20, 20e30, 30
e40)
Age

6.2 Qualitative data
Qualitative data are unmeasurable and uncountable data or attributes. It is categoricald
described or labeled data into different categories or states. Quantitative data are either
countable or measurable.
6.2.1 Nominal data
Nominal data are qualitative data with only names/labels/categories without comparable
or intuitive order. Example: Blood Groups A, B, O, AB.
6.2.2 Ordinal data
Ordinal data are qualitative data with categories that could be ordered or ranked. How-
ever, the difference between the categories cannot be measured or is not known. For
example, mild, moderate, and severe fever.
6.3 Quantitative data
6.3.1 Scale data: interval and ratio
Both interval and ratio are numerical data with little difference between them. In fact,
many statistical software (including SPSS) consider no difference between them, as it
considers both as scale data, the data that can be scaled. In interval data, as its meaning
suggests (interval ¼ gap ¼ space in between), not only orders but exact differences are
also known. Height, weight, blood pressure, etc. are all examples of scaled data.
The difference between interval and ratio is best appreciated when any attribute does
not have an absolute or meaningful zero. For example, temperature is measured in de-
grees such as Celsius degrees. A temperature of 0C is not “no” temperature as zero sug-
gests. This is in stark opposition to the Kelvin scale, where 0 K actually means no
temperature and is, therefore, a true zero (Table 1.4). Zero Kelvin is the lowest temper-
ature limit of a thermodynamic scale. It is the temperature at the hypothetical point at
which all molecular activity ceases, and the substance has no thermal energy. So, in ratio
variables the zero has a meaningful role. For example, if your Car has an inside temper-
ature of 12C and the outside temperature is 24C, it would be wrong to say that the
outside temperature is double than inside the car. This is because the reference point
Table 1.4 Temperature measures in Kelvin and Celsius.
Kelvin Celsius Fahrenheit
0 K 273.15 459.67
255.4 K 17.8 0
273.15 K 0 32
373.15 K 100 212
Data typedratio Data typedinterval

to compare them needs 0 as a reference point, and the 0o
on the celsius scale is meaning-
less. However, 24 K is double that of 12 K because it has a meaningful 0 as a reference
point. The temperature in Kelvin is a ratio type data (Kelvin is also the standard unit (SI)
of temperature measurement).
Both interval and ratio data actually measure an attribute, unlike nominal or ordinal
data where the attribute cannot be measured but can only be labeled or ranked,
respectively.
6.3.2 Discrete and continuous
Quantitative data are also classified into discrete and continuous data. Discrete data are
countable, whereas continuous data are measurable. What does this mean? Discrete
data are data that can take a restricted number of fixed specified values, for example,
number of children born to a woman (can be 1, 2, etc. but not 1.5). Continuous data
can take an unrestricted number of measurable values, although they may have an upper
or lower limit. For example, weight cannot be 0 kg. Another essential point to note is
that measures have units. For example, height can be measured in centimeters or inches.
Countable data or discrete data do not have units of measurement. The ordinal (qualita-
tive) and discrete (quantitative) data difference is explained in Box 1.3.
6.4 Cardinal and ordinal data
The term Cardinal measures answers the question of how many, whereas Ordinal measures
only rank individuals or households, as explained before. So, cardinal measures are contin-
uous measures such as age in years or weight in kilograms, etc.
6.5 Variable
Research use variables. Variables are any characteristics, numbers, or quantity that can be
measured/counted, labeled, or ordered. In other words, variables are used as data items,
and the term is used interchangeably with data. However, in health research, the term
variable is strictly used to identify and describe the different characteristics of the sample.
BOX 1.3 Ordinal data versus discrete data
Whereas ordinal data are qualitative, the discrete data are quantitative data. In ordinal data, the
ranking or order is only important (e.g., mild, moderate, or severe anemia), whereas in discrete
data, both order and magnitude are important (e.g., number of children born to a woman). For
discrete data, two points are important to note: (1) numbers in discrete data represent actual
countable quantities rather than labels that can be ordered or arranged, and (2) in discrete
data, a natural order exists among the possible values.

In research, we describe variables as an outcome and exposure variables (Table 1.5).
To test hypotheses, the outcome variables are also called dependent variables, whereas
the exposure variables are called independent variables. Although, in the strict sense, “in-
dependent” means that the said variable independently affects the dependent variable. In
regression and other mathematical models, the outcome variable is placed on the left-
hand side of the equation and the predictors/exposures on the right-hand side. In tables,
the independent variables are usually identiﬁed as rows and dependent as columns.
7. Measures of central tendency and dispersion
7.1 Introduction
Measures of central tendency describe the location of the middle of the data. It is a mea-
sure of the data average, a single value that could represent the whole data. Variability
measures the spread of the data around its middle or central values. In other words, it
is a measure of spread or dispersion.
7.2 Central tendency
The central tendency in inferential statistics is most commonly measured as arithmetic
mean or mean. The mean is the mathematical average of the observations and is calcu-
lated by dividing the sum of the observations by the total number of observations
(Table 1.6). In normally distributed data (or even nearly normal), the mean is the best
representative measure of central tendency. However, since it is calculated
Table 1.5 Common terms, synonyms, and conventions for different variables.
Variable Outcome variable Exposure variable
Synonyms/alternative terms
Dependence effect Dependent variable Independent variable
Questionnaire based Response variable Explanatory variable
Case-control design Case-control groups Risk factors
Interventional design Outcomes Exposure/Treatment group
Regressions Regressand Regressor
Covariates (continuous)
Factors (categorical)
Conventions
Mathematical/Regression
equations
Left-hand side Right-hand side
Tables Columns Rows
Graphs y-variable x-variable
Vertical Horizontal

mathematically, it tends to be affected most by any one very large/very small value (out-
liers) in the dataset. Outliers also affect the normal distribution of the data. Mean is rep-
resented as
̄
(called x bar) and is calculated mathematically as follows:
̄
¼ Sxi/n, where S
represents the summation of all individual values, that is, xi and n represents the total
number of values.
If the data are not normally distributed or have outliers, the median becomes a better
representative measure of central tendency. This can be appreciated in the second and
third datasets in Table 1.6. Usually median is not calculated mathematically but is simply
the middlemost value when the data are arranged in an ascending or descending order.
When the middlemost values (as in even number of data) are two, their average value is
the median.
Sometimes, we need to know the most commonly occurring value to describe a phe-
nomenon in health sciences. Such a measure is called mode. A dataset could be unimodal
or bimodal, or even multimodal. The data in Table 1.6 show a unimodal presentation.
However, diseases like Hodgkin’s lymphoma have a bimodal presentation, with most
common occurrence around age 15 and age 55 years.
Mean, median, and mode tend to overlap in a normally distributed data. However,
the mean tends to be affected the most with the addition of even one outlier and reﬂects
skewing (Fig. 1.3). One outlier toward the right (large value) moves the mean toward the
right. This rightward shift or long right tail is called positive skewing, where mean be-
comes rightmost or higher than median and mode and is no longer the best representative
of central tendency. Similarly, one outlier toward the left (small value) moves the mean
toward the left. This leftward shift or long left tail is called negative skewing, where mean
becomes leftmost or lower than median and mode and is no longer the best representa-
tive of central tendency.
Table 1.6 Measures of central tendency.
Dataset (weight in kg)
Mean (
x [ Sxi/n)
(arithmetic
mean)
Median (middle
most value)
Mode (most
commonly occurring
value)
7,8,9,10,10,10,11,12.13 7 þ 8þ..þ13/
9 ¼ 10
5th value ¼ 10 10
7,8,9,10,10,10,11,12.13,30 7 þ 8..þ30/
10 ¼ 12
5th þ 6th value/
2 ¼ 10
10
1,7,8,9,10,10,10,11,12.13 1 þ 7þ..þ13/
10 ¼ 9.1
5th þ 6th value/
2 ¼ 10
10

7.3 Dispersion
Dispersion or variability is the measure of spread around the central tendency. There are
many ways of measuring dispersiondrange, interquartile range, standard deviation, and
others. Standard deviation (s) is a mathematical measure of drift from the mean and is
calculated mathematically as s ¼ OS(xi
̄
)2
/n 1, where the numerator is the sum of
squares of individual data differences from mean S(xi
̄
)2
and n is the total number of
observations. The square root of the values is standard deviation, whereas, without the
square root, it is called variance.
The reason standard deviation is measured by squaring the deviations at ﬁrst and then
taking the square root is to value the magnitude of difference from the mean, regardless of
the direction of difference. Whether the values are shifted toward right or left from the
mean, it is equally a measure of spread, and the magnitude is equal. For example, in
Table 1.7, the difference between ﬁrst and last observation from the mean is 3 (7 10)
Figure 1.3 Measures of central tendency and Gaussian curve. Gaussian curve with positive skewing
and negative skewing. (From Wikipedia commons (CC-By-4.0).)
Table 1.7 Calculating the standard deviation of a dataset.
Observations xi Mean (
x [ Sxi/n)
Square of difference from mean
(xi L
x)2
Standard deviation
OS(xi L
x)2
/n L 1
7 10 9 ¼ O28/8 ¼ 1.87
8 10 4
9 10 1
10 10 0
10 10 0
10 10 0
11 10 1
12 10 4
13 10 9
Total 7þ..þ13/9 ¼ 10 S(xi
x)2
¼ 28

and 3 (13 10). Squaring them removes the impact of these signs. The other important
feature is n 1. In small samples such as the ones we use, n 1 is a better mathematical
measure than n. However, in population standard deviations, n should be used instead of
n 1. An example of standard deviation calculation is given in Table 1.7. In any case,
most of these calculations are easily done through software, calculators, and websites
such as easycalculation.com.
The range is the difference between the maximum and minimum values in the data
set. The use of range as a measure of dispersion is very limited in biostatistics. Its most
important usage is to signify if the range is quite large or relatively small. An interquartile
range or interquartile range (IQR) is a better measure, especially in nonnormal distribu-
tions. The IQR is the range of the middle half of the values and does not get affected by
extreme values or outliers like the other measures. It corresponds to the median as a mea-
sure of central tendency.
A quartile divides the dataset into three points which form four groups. The three
points are Q1 (lower quartile), Q2 (median), and Q3 (upper quartile). Each of the four
groups represents 25% of the values (quarter ¼ 1/4). Q1 is the middle value of the lower
half of the data (below median), and Q3 is the middle value of the upper half of the data
(above median). In the dataset in the ﬁrst row of Table 1.8, the Q2 (median) is 10. Q1 is
the middle value/median of the lower half, 8 þ 9/2 ¼ 8.5 and Q3 is the middle value/
median of the upper half, 11 þ 12/2 ¼ 11.5. The interquartile range or IQR is
Q3 Q1. In this example, the IQR would be 11.5 e 8.5 ¼ 3.
If the data have outliers, IQR represents the dispersion better than the standard de-
viation. This can be appreciated in Table 1.8. Whenever the mean is presented in a data-
set, it should be shown as mean standard deviation ( s.d.) or mean(s.d.). When there
are outliers or the distribution is skewed, the median (Mdn) should be presented along
with the IQR or Mdn(IQR).
Table 1.8 Measures of dispersion.
Dataset (weight in kgs.)
Standard deviation
(s [ OS(xi L x
)2
/n L 1) Range Interquartile range
7,8,9,10,10,10,11,12.13 1.87 7e13 3 (Q1 ¼ 8.5, Q2 ¼ 10,
Q3 ¼ 11.5)
7,8,9,10,10,10,11,12.13,30 6.56 7e30 3 (Q1 ¼ 9, Q2 ¼ 10,
Q3 ¼ 12)
1,7,8,9,10,10,10,11,12.13 3.35 1e13 3 (Q1 ¼ 8, Q2 ¼ 10,
Q3 ¼ 11)

References
Belle, Fisher, L. D., Heagerty, P. J., Lumley, T. (2004). Biostatistics: A methodology for the health sciences. John
Wiley Sons.
Carroll, L. (1991). Alice in Wonderland. https://www.gutenberg.org/ﬁles/2344/2344-h/2344-h.htm.
Cohen, B. B., Pokras, R., Meads, M. S., Krusha, W. M. (1987). How will diagnosis-related groups affect
epidemiologic research? American Journal of Epidemiology, 126(1), 1e9. https://doi.org/10.1093/
oxfordjournals.aje.a114639
Dimond, E. G., Kittle, C. F., Crockett, J. E. (1960). Comparison of internal mammary artery ligation and
sham operation for angina pectoris*. The American Journal of Cardiology, 5(4), 483e486. https://doi.org/
10.1016/0002-9149(60)90105-3
Doyle, A. C. (1893). The adventure of the cardboard Box. Project Gutenberg. https://www.gutenberg.org/ﬁles/
2344/2344-h/2344-h.htm.
Faizi, N., Kumar, A. M., Kazmi, S. (2018). Omission of quality assurance during data entry in public health
research from India: Is there an elephant in the room? Indian Journal of Public Health, 62(2), 150e152.
https://doi.org/10.4103/ijph.IJPH_386_16
Indrayan, A., Malhotra, R. K. (2017). Medical biostatistics (4th ed.). Chapman and Hall/CRC. https://
doi.org/10.4324/9781315100265
Johnson, L. L., Borkowf, C. B., Shaw, P. A. (2012). In I. John, J. I. Gallin, F. P. Ognibene (Eds.), Prin-
ciples and practice of clinical research. Academic Press.
Laplace, P. S. (1810). M
emoire sur les approximations des formules qui sont fonctions de tr’es grands nom-
bres et sur leur application aux probabilit
es. Memoires de l’Academie des Sciences (pp. 353e415).
Merriam-Webster. (n.d.). Biostatistics. Retrieved March 18, 2022, from https://www.merriam-webster.
com/dictionary/biostatistics.
Nightingale, F. (1863). Notes on hospitals.
Pagano, M., Gauvreau, K. (2000). Principles of biostatistics.

CHAPTER 2
Data management and SPSS
environment*
For every minute spent in organizing, an hour is earned
Benjamin Franklin.
1. Data management
1.1 Introduction
Data are central to quantitative research, and data management is integral for research.
Data management is the process of compiling, storing, organizing, and securing the data collected
by the researchers. The goal is to manage the data effectively with reliable strategy and
methods such that it is retrievable when required. Data Management has become far
more efﬁcient, sensitive, and reliable in the smart world with improved access to smart-
phones, GPS, and the internet. However, poor data collection or management protocol
can lead to misleading or even wrong results.
Research data management beneﬁts researchers who conduct research and policy ap-
praisals. Most senior researchers have some terrible memories of the research data they
collected in past projects. Research data are not a one-time enterprise; researchers
need the data themselves for future projects and even for institutional collaborations.
Hence, data should be managed through an intuitive, reliable, reproducible, and reuse-
able method. Let us take an example of a senior’s research on the MERS virus during
their graduation or doctorate. Suppose we now want to test the correlations of MERS
with COVID-19. This would be possible only if that data collected a decade ago had
all the relevant details and could be retrievable with intuitive codes and values. The
essence of open research data management is that data are retrievable, interpretable,
and useable for any researcher. This can be possible if we give importance to data man-
agement and prepare a proper plan during protocol preparation and data collection.
Another reason why we advocate for data management is policy compliance. Usually,
the Institutional Ethics Committees mandate that the project proposals explain their data
management plan. Similarly, academic institutions, as well as regional and national
research organizations, require statements on data management. Some organizations
*For datasets, please refer to companion site: https://www.elsevier.com/books-and-journals/book-
companion/9780443185502
Biostatistics Manual for Health Research
ISBN 978-0-443-18550-2, https://doi.org/10.1016/B978-0-443-18550-2.00008-6
© 2023 Elsevier Inc.
All rights reserved. 17

Figure 2.1 Data management cycle and
characteristics.
such as Wellcome Trust and National Health and Medical Research Council need the
data collected by the projects funded by them to be openly shared with other researchers.
Data management and open data sharing are becoming common among research coun-
cils of a few countries, and others may and should follow suit. Additionally, while sub-
mitting the manuscript for consideration for publication, some publishers mandate
declaring the data management and open data sharing plan.
Thus, we strongly encourage researchers to prepare the data management plan as a
part of the protocol and follow it strictly during the entire study duration. This adds value
to their work and supports data integrity. Fig. 2.1 shows the research data management
lifecycle, which helps in keeping track of actions to take in data management.
1.2 Best practices for data management
• Data should be well organized on the data management platform. This helps in
locating and transforming the data whenever required.
• The collected and organized data should be stored securely along with a backup. This
helps in recovering lost data in the face of an unexpected event. We recommend
keeping the data at three different places, and at least one of them be a cloud-
based server.
• The stored data should be accessible to anyone with access. This accessibility should also
be long-term and retrievable in the future.
• Ideally, people should be able to learn where the data are being used to support or
help that project. A suggested citation for data use should be provided along with
the necessary details of the user’s project or research. Publishing the data is highly rec-
ommended unless it has sensitive information.
• The data should be anonymized wherever required.

1.3 Data management plan
The research data management plan should be prepared before the actual project starts as
a part of the protocol explaining how the data will be managed during and after the proj-
ect. We suggest checking for any existing data management format at your university or
research council. A good data management plan and practices repository is available at the
Global Health Training Network. The essential properties of a data management plan are
shown in Box 2.1:
Every researcher should address in their data management plan the enlisted guiding
points. They should describe the type of data they will collect, such as images, text,
spreadsheets, audio and video files, patient records, blood or other specimens, reports,
surveys, etc. It is also essential to determine who will collect and access the sensitive
research data. Choosing the proper storage is critical for data management. Most re-
searchers store the data locally on their personal or official computers. Although it serves
the purpose, there are better tools with improved storage, access, retention, encryption,
and data loss protection. Epicollect is one such platform that helps not only in data collec-
tion, but also with storage. When the data are stored digitally, it should have a predeter-
mined naming pattern of files and folders. It should have sufficient details to sort and find
them in the future. Being concise and consistent is the key to proper data storage and
cleaning. Do pay due importance to version, date of creation, and creator in file naming.
Table 2.1 is one such sample of a naming pattern.
BOX 2.1 Data management template should cover
· Background and Methodology, including Type of data collected
· Metadata or README file
· Plan on naming the files and folders (including updates and versions)
· Ethics and legal compliance
· People responsible and their role
· Data storage and backup, including long-term retention and encryption
· Sharing and public archiving
· Metadata or README file
Table 2.1 Naming a file.
File name
20220313_Ver1.1_COVIDreport_
HCWWHO_YSR.xlsx
20220313 Date of creation
Ver1.1 Version
COVIDreport Content
HCWWHO Standard acronym of the project
YSR Creator initials
Note that we have used _ (underscore) in between the words rather than space as many
software do not accept spaces.
Data management and SPSS environment 19

We should also have a defined time period for data storage, including short-term and
long-term. In a short-term plan, the data are stored until the research is completed and
published, which is often the only concern of most researchers. A long-term data storage
plan is gaining the interest of researchers, as it is helpful for self-use in the future or for
other researchers. However, this is difficult and often costly as storing data or specimens
from the research for a long period of time requires huge space. Moreover, the availability
could be subject to the intellectual property associated with the data. Nevertheless, it is
advisable to use a platform or storage option that facilitates sharing the data with others in
all cases, unless restricted due to unavoidable reasons.
Disclosing how the data can be used by someone who has access to it from the public
archives must be mentioned with suggested citation. Lastly, metadata describes the data
you have collected in an easily understandable form. It is essential to explain a few data
characteristics in a readme file. There may be one metadata file in a project describing all
the variables and measurements as a data documentation sheet or multiple readme files in
each folder you have saved electronically alongside the dataset.
2. Data documentation sheet
A data documentation sheet (DDS) is a ‘codebook’ containing the details of all the vari-
ables (like questions, variable names, input type, possible answers, and code for possible
answers) to be entered. It is the first step in preparing a plan for data collection and entry.
The DDS should be prepared before collecting data or at least before data entry in any
software. In the proceeding exercise, we will try to prepare and understand the concept
of DDS. While we provide this as an exercise, there is no right or wrong method for
designing a codebook.
2.1 Template for DDS
As discussed above, metadata describes the data in an easily understandable form. This is
done by detailing out characteristics of the data variables. In Table 2.2, we provide a tem-
plate of DDS along with the essential attributes of each variable in the questionnaire.
The different components and attributes of a DDS is shown in Box 2.2.
Table 2.2 Data documentation sheet template.
Question
Name of
variable
Variable
type
Possible answer and value/
code of them if applicable
Comments/
missing

2.2 Coding the dataset
A data documentation sheet or codebook contains the details of all the variables to be
entered. Furthermore, it also has code for possible answers, often as Arabic numerals.
This facilitates subsequent analysis of the categorical data as most of the software requires
the data to be in Arabic numerals to perform statistical analysis. Data coding is done
before the data collection, preferably during questionnaire development. The key is to
prepare the unambiguous code. Apart from this, coding also reduces errors during data
collection, entry, and even analysis.
The best practice for coding qualitative data is to start giving codes from 1 and go
further. It is also recommended to use 0 for “negative” and “No” responses, while for
“unknown,” “not applicable,” and/or missing values, we may use 9, 99, or 999.
Exercise 1: Prepare the DDS in the template (given in Table 2.2) of the TB
Patient Treatment card given in Fig. 2.2. The solution is given in Table 2.3.
3. Data capture and cleaning
3.1 Data capture
Traditionally, researchers start their study with the data collection using a paper-based in-
strument, both for primary and secondary data. In most cases, after all the data have been
collected, data from the completed paper-based forms is entered in a software such as
excel, creating a master chart for further processing for analysis. This is often the weakest
link that critically affects data quality (Faizi et al., 2018). Since data entry is often
BOX 2.2 Attributes of a data documentation sheet
· Question: It describes the individual variable in the questionnaire. Try to keep it short and
speciﬁc.
· Variable name: It is the shortened name for the question. It is highly advisable to keep it
intuitive, unique single word, with small letters without spaces or special characters. Limit it
to 8 words for better compatibility with different statistical software.
· Input or variable type: It describes the type of variable. It can be qualitative (categorical:
nominal or ordinal) or quantitative (discrete or continuous), or any distinctive type (date,
dollar, string, etc.)
· Possible answers: It includes all the responses a question may have. For categorical vari-
ables, all the categories become the possible answers. For a continuous variable, it is advis-
able to restrict them from minimum to maximum values.
· Value/codes: The numeric values assigned to each possible answer in categorical variables.
It is instrumental in data entry and analysis.
· Comments: It can include any instruction or words for future purposes. What to do in case
of missing data can be mentioned here.

monotonous and uninspiring activity, it is often assigned to those with little knowledge of
the importance of the research, which in turn, leads to potential errors.
As the saying “Garbage in, Garbage out” goes, we should exercise caution and
concern during data entry to reduce errors at this integral step. Errors in data entry
may lead to poor and misleading interpretations leading to erroneous decisions and
flawed policies. Therefore, data quality is vital for data integrity and research findings.
With the advancement of technology and better access to electronic devices, data collec-
tion and simultaneous data entry are possible nowadays. Data entry can be done on MS
excel, Epidata, and other software after collecting data in a paper-based questionnaire.
Mobile and electronic forms eliminate data entry as a separate step, saving resources
for data entry and removing the possibility of errors during the process. This is termed
data capture, as shown in Fig. 2.3.
Epicollect is one such mobile application that captures data efficiently and greatly fa-
cilitates subsequent handling of the data. Apart from the data capture, Epicollect is also an
efficient data management tool as it assists in the secure storing of the data. It also im-
proves data accessibility, storage, and archiving.
Advantages of Epicollect or other similar applications:
• It saves time, costs, and human resources by combining the process
• It produces more accurate data by preventing data collection and data entry errors
• Sharing of data is possible
• Helps to establish data validity processes by supervisors.
Figure 2.2 TB patient treatment card.

Table 2.3 Data documentation sheet: solution to Exercise 1.
Question
Name
(variable)
Type
(variable)
Possible answer and
value labels Comments/missing
Serial number sn Numeric Unique field if
missing/
nonunique enter
999
Registration
number
r_no String If missing/
nonunique enter
99
Name of the
patient
name String If missing/
nonunique enter
99
Address address Numeric Rural: 1, Urban: 2
Patient’s gender gender Numeric Male: 1, Female: 2,
Transgender: 3,
Not recorded: 9
Patient’s age in
years
age Numeric 0e125, 126
Treatment center tret_centr Numeric
Date of registration reg_date Date 01/01/2019 to 31/
12/2019, 01/01/
1800
Range of legal dates
enter “01/01/
1800” if date is
missing
Bacteriological
confirmed case
Bact_confir Numeric Yes: 1, No: 2
Date of test test_date Date 01/01/2019 to 31/
12/2019, 01/01/
1800
Range of legal dates
enter “01/01/
1800” if date is
missing
Smear results smear Numeric Negative: 1, Scanty:
2, 1þ : 3, 2þ : 4,
3þ : 5
If missing/
nonunique enter
Weight in kg weight Numeric 1.0e200.0, 999 Enter “999” if
missing
Height in meter height Numeric 0.50e3.00, 9 Enter “9” if missing
Molecular test mol_test Numeric Positive: 1,
Negative: 2, Not
done: 3
If missing/
nonunique enter
9
Figure 2.3 Data capture.

• It can collect different kinds of data (pictures, audio, video, GPS coordinates)
• Epicollect is a free tool and can work offline
Apart from Epicollect, there are a few other data capture tools, such as EpiData, Goo-
gle forms, SurveyMonkey, KoBo Toolbox, etc. But the mobile app and ability to capture
data offline simultaneously by multiple field investigators make Epicollect an excellent
tool.
3.2 Data checking and cleaning
As we discussed the concept of “garbage in, garbage out,” we should observe every re-
cord form for completeness and errors in data collection and data entry if they are per-
formed separately. Checking after the data collection may be feasible for a small
sample, but periodic checking is a must for large datasets.
Many data entry software can be utilized to simultaneously perform data checking
and data entry. Epicollect, EpiData, and KoBo Toolbox and, to some extent SPSS dataset
can be coded to facilitate such operations. For example, restricting the cell to one digit
will prevent an error of double typing the response. In some data-entry software, we
can limit the data entry to only a few codes, and any other entry would be illegal.
Such commands and checks prohibit wrong entries and typos. Thus, data checking
can be done during data entry (interactive checking) and after data entry (batch check-
ing). Although recommended as good research practices, double-entry, matching, and
search for differences (validation of data) are often reserved for critical variables and often
not performed at all (Faizi et al., 2018).
Data checking is followed by data cleaning, where the inaccurate records are cor-
rected by screening, diagnosing, and subsequently editing by modifying, excluding, or
replacing them (Fig. 2.4, modified from Van Den Broeck et al. (2005)).
Figure 2.4 Data cleaning. (Modified from Van Den Broeck, J., Cunningham, S. A., Eeckels, R., Herbst, K.
(2005). Data cleaning: Detecting, diagnosing, and editing data abnormalities. PLoS Medicine, 2(10),
0966e0970. https://doi.org/10.1371/journal.pmed.0020267)

In this step, we take a closer look at the data for any issue it may have in the analysis.
IBM SPSS software can help in data cleaning, which we delve further into in Section 6.
Data transformation in SPSS. Nevertheless, whenever you perform data cleaning, it is
recommended to describe it as a standard part of reporting statistical methods (Ethical
Guidelines for Statistical Practice, 1999). The important terminologies related to data
management are shown in Box 2.3.
4. SPSS environment
4.1 Introduction to the software
Statistical Package for the Social Sciences, or SPSS, is a popular software used by statis-
ticians and researchers worldwide. It is easy-to-understand and user-friendly interface,
minimal to no coding requirements, and a wide range of applicability makes it popular
among health researchers. Given the time and resource constraints, SPSS provides an
excellent platform for complex statistical analysis within a few clicks. In this section we
will go through the SPSS environment and learn some basic commands.
Statistical Package for the Social Sciences is owned by IBM Corporation (IBM) and is
available under a trial version and licensed on a subscription basis. It is available for almost
all operating systems, including Windows, macOS, Linux, and UNIX. A team of devel-
opers regularly works on its revision and updates, with the latest being 28.0. This book is
based on version 26.0 running on a macOS. There are minor differences in different ver-
sions for health researchers unless they use advanced techniques.
BOX 2.3 Important terminologies related to data management
Derived variables: Variables created from one or more variables in the original data by manip-
ulating them. For example, Computing the new variable BMI from the original variable of
weight and height.
Recoded variable: Transforming one variable into a different variable by altering the vari-
able’s coding. For example, Creating a new categorical variable of “age group” with 5-year in-
tervals from the continuous variable of “age.”
Data snooping: Analysing the relationship only after examining the ﬁndings, which was not
planned before, to extract something meaningful from the study.
Data imputation: Data imputation is the substitution of estimated values for missing or
inconsistent data items/ﬁelds. The substituted values are intended to create a data record
that does not fail edits (Organisation for Economic Co-operation and Development (OECD),
2008).

4.2 The first thing you will see
When you open the SPSS software for the first time, you are greeted with a dialog box
that prompts you to open a data file (Fig. 2.5). You may open an existing SPSS data file or
a file in another format, including Excel and text containing data. The SPSS data file can
also be opened directly from your computer. If you have the data or master chart as an
MS Excel file, you have the option of “Open another type of file” and locate the file on
your computer. We will be proceeding with a beginner in mind who is new to the SPSS
environment and does not have a data file.
To create a new SPSS data file, you can dismiss the dialog box by clicking “Close.”
The best practice is to check “Don’t show this dialog box in the future”, so that you don’t
get disturbed by this dialog box. This takes us to the main window with two view modes:
“Data View” and “Variable View” (Fig. 2.6).
Figure 2.5 Opening dialog and SPSS windows.

4.2.1 Data View
The Data viewer is a spreadsheet-like interface with rows and columns where data can be
viewed or entered. The subject/case/participant in the data are represented by individual
horizontal rows, while each column contains variables of interest. The top of the Data
View window you will find a drop-down menu and toolbar, while at the bottom, the
user can toggle between Data View and Variable View.
4.2.2 Toolbar icons from left to right (Fig. 2.7)
The toolbar icons from left to right are as follows:
1. File Open (Folder) allows particular data files to be opened for analysis.
2. File Save (Floppy disk) saves the file in the active window.
Figure 2.6 SPSS initial screen window.
Figure 2.7 Toolbar and menu.

3. File Print prints the file in the active window.
4. Dialog Recall displays a list of recently opened dialog boxes, including recently
performed analysis, any of which can be selected by clicking on their name.
5. Undo
6. Redo
7. Go to Case allows entering the number of the case you want to go to in the data
file.
8. Go to Variables allows entering the number of the variable you want to go to in
the data file.
9. Variables provide data definition information for all variables in the working data
file.
10. Run Descriptive statistics
11. Find (binoculars) allows data to be found easily within the data editor.
12. Split File splits the data file into separate groups for analysis based on the values of
one or more grouping variables.
13. Select Cases provides several methods for selecting a subgroup of cases based on
criteria, including variables and complex expressions.
14. Value Labels allows toggling between actual values and value labels in the Data
Editor.
15. Use Sets allows the selection of sets of variables to be displayed in the dialog boxes.
16. Customize toolbar
4.2.3 Variable View
The Variable viewer displays the properties of individual variables represented in columns
(Fig. 2.8). The properties of variables in SPSS can be described in the following headings:
Name: This is a shortened, unique single word. It must not contain spaces or special
characters or begin with numbers. The default variable name begins with VAR00001,
which should be renamed by double-clicking. For practice, hover on to the first cell
in Variable View, enter “ABC,” and notice changes in other columns.
The best practice is to limit Name to 7e8 words for better compatibility with
other statistical software and to have an intuitive name to find the variable easily.
Type: This shows the type of data attribute. As we saw in the first chapter, data are
classified into two types: qualitative/categorical or quantitative/continuous. As
shown in Fig. 2.8, a variable can be classified into nine different types. The default
variable type in SPSS for data entry is numeric (can take numbers only), which is
also the most used for analysis. The quantitative/continuous data in SPSS can be
entered as “Numeric”, while qualitative/categorical data can be entered as
“Numeric” and “String” (can take any letters or numbers). We could also select
the “Date,” “Dollar,” or “String” as the type of variable, wherever applicable.

The best practice for qualitative and quantitative data entry is using the “Numeric”
type. For qualitative data with finite/known number of possibilities, the “Numeric”
type with value labels is used. Each numerical value is coded for possible answer. In
the data where responses cannot be presumed, like the Name of the patients or
Unique ID, we use “String” and, if required, data cleaning can be done later.
Width, decimal, and columns: Width provides words/digits count that can be
entered in an individual variable, while Column is the size of a particular variable in
Data View. The Decimals show how many decimals values can variable take (only
for the numeric data type). As a best practice, reduce the width according to
requirements.
Label: The label is an expanded option for the variable name. It does not have letter
restrictions and can be used to describe the short variable name. The analysis outputs
mention the label name. This should be used to define the clear and complete name
of the variable.
Values: It is a beneficial tool for data entry. As described earlier, any type of study
variable with multiple finite responses may be entered in SPSS as a numeric type,
with each numeric value representing different responses of the dataset. These
Figure 2.8 Variable View and properties of variables.

responses are added as values corresponding to different numbers. The Value label
command may be activated by clicking “three dots” in the right corner of the value
box. Although not essential, allotting a number to missing variables may be
considered.
Missing: It is essential to inform SPSS that if some values are missing, what would
they be coded as? A unique number should be provided that does not overlap
with the actual value numbers/records. As a best practice, this should be mentioned
in the DDS (Table 2.3).
Measure and role: Measure describes our variable’s scale of measurement. Although
we know that there are four scales; nominal, ordinal, interval, and ratio, the SPSS has
only three, with interval and ratio merged into one and labeled as a scale. A Role de-
scribes what the variable will play in the analysis. “Input,” the default assignment, is
used for the independent variable (predictor), while “Target” is used for the depen-
dent variable (outcome).
5. Data entry and importing in SPSS
5.1 Data entry in SPSS
Exercise 2dPrepare a New SPSS file of the dataset using data given in Table 2.4
and its DDS along with their coding as shown in Table 2.5.
Solution to Exercise 2:
We need to build the form in the Variable View for data entry before entering data in
Data View.
• The first variable in DDS (Table 2.5) is Serial Number. Doubleclick on the first cell in
the Variable View window, rename it to sn (variable name as in DDS), and press
Enter. All the other properties of the “sn” variable will be assigned by default, as
shown in Table 2.6. You have to change a few things according to DDS. By default,
Label column is empty, and you may enter the actual variable name for your conve-
nience. Under Missing, click on the cell corresponding to “sn” to activate it, and then
click three dots on the rightmost part of the selected cell (Fig. 2.9). A dialog box will
appear where you can enter “999” under discrete missing values as in DDS. Click OK
to confirm. Finally, under Measure, change to “Scale” by selecting it from the drop-
down. Your first variable is now ready to be entered into SPSS.
• Registration Number, Name, and Treatment Centre variables are categorical vari-
ables with no prediction of their possible answer. Such variables are labeled as
“String” type with appropriate width as per requirements. Similar to the Serial Num-
ber variable, create these new variables, rename them as discribed in DDS and change
the Type as String. Similarly, enter “99” under discrete Missing value and increase the
Width according to your requirement.

Another Random Scribd Document
with Unrelated Content

of D. John of Austria, begging mercy for his sins, and delivering up
flag and arms.
They then set out the same day for the Padules, where D. John
was encamped; the Habaqui and the gentlemen commissioners, with
300 Moorish marksmen whom they brought as escort. The Habaqui
rode an Algerian horse, with Arab trappings; he wore a white turban
and a crimson caftan, his only arms a sword set with many precious
stones; he was a spare man with a good figure, with a thin beard
which was beginning to turn white. At his side an ensign of the
escort bore the banner of Aben Aboo, of turquoise damask, with a
half-moon on the point of the staff, and some words in Arabic which
meant, I could not desire more or be contented with less. The
marksmen followed five in a row. Four companies of Spanish
infantry, who were waiting at the limits of the camp, surrounded
them, and on passing the lines the Habaqui gave up the banner of
Aben Aboo to the secretary Juan de Soto, who was riding at his side.
In this way they passed through the ranks of the infantry and horse
soldiers, who played their bands and fired a fine salute of
arquebuses, which lasted a quarter of an hour.
D. John of Austria waited in his tent, attended by all the captains
and gentlemen of the army; he was in full armour, one page held his
helmet, and another, on his left hand, the standard of the
Generalissimo. The Habaqui alighted in front of the tent and went
straight to throw himself at the feet of D. John, exclaiming, Mercy,
my lord, may your Highness grant us mercy in the King's name, and
pardon for our sins, which we know have been great, and taking off
the sword with which he was girded, he placed it in D. John's hand,
saying, These arms and flag I give up to His Majesty in the name of
Aben Aboo and of all the rebels for whom I am empowered to act.
And at that moment Juan de Soto threw down the kinglet's banner
at D. John's feet.
D. John listened to him and looked at him with such quiet and
peaceful dignity that he well represented the justice and mercy of
which he was the guardian. He ordered the Habaqui to rise, and
giving him back his sword, told him to keep it, and with it to serve

His Majesty. D. John afterwards loaded him with favours, and
ordered his gentlemen to do the same: that day the Habaqui dined
in the tent of D. Francisco de Córdoba, and the following one in that
of the Bishop of Guadix, who was in the camp.
The next day the festival of Corpus Christi was celebrated in the
camp, with all the pomp and solemnity possible in such an out-of-
the-way place, and with the joy natural to those who believed that
the disastrous war was ended. By cartloads and armfuls the soldiers
brought flowers and herbs, so plentiful in May in that fertile country,
to adorn the altar and the road by which the Holy Sacrament was to
go. They hung with fair and fragrant garlands the tent in which Mass
was said, and which stood, raised, in a sort of square in the centre
of the camp, and around it they planted green groves and arches of
foliage, with flags and streamers. The soldiers had made it a point of
honour to adorn their tents, and there was not one which was not
beautified with wreaths, flags, and little altars of different kinds;
many of them were ornamented with rich cloths and other precious
things, the booty of war. The Host was carried by the Bishop of
Guadix, under a brocaded canopy, held up by D. John of Austria, the
Knight Commander of Castille D. Francisco de Córdoba, and the
Licentiate Simon de Salazar, Alcaide of the King's Court and
household; in front, two by two, went all the friars and clergy of the
camp, who were numerous, and the knights, captains, and
gentlemen, with torches and tapers of wax, lighted, in their hands.
From one end of the camp to the other the infantry and horsemen
had formed up with their flags flying, and as the Blessed Sacrament
passed, they knelt down, lowering their arms, standards and
banners, kissing the dust; the bands played martial hymns, and
through the air thundered salvos of arquebuses, which did not cease
for at least a quarter of an hour. A friar of St. Francis preached that
day, says Luis del Mármol, who with many tears praised Our Lord
for His great favour and mercy in having made the place Christian by
bringing the Moors to a knowledge of their sins; and besides this he
said many things which consoled the people.

But, unluckily, these rejoicings and consolations were premature,
as very soon afterwards the traitor Aben Aboo went back on his
word, and fortified himself in the Alpujarras, and began to prevent,
with atrocities and punishments, the pacification of the Moors, who
had thronged to submit, and he asked for fresh help from the Kings
of Algiers and Tunis. Loyal and honourable for his part, Hernando el
Habaqui was furious; he went to the Alpujarras swearing to bring
Aben Aboo to reason, or to bring him into the presence of D. John
tied to his horse's tail. But the crafty Moor knew how to lay a snare
into which the loyal Habaqui incautiously fell, and was treacherously
killed, and his corpse hidden for more than thirty days in a dung-
heap, covered up with a matting of reeds.
Few, however, were the followers who remained to Aben Aboo
after this crime was discovered; and pressed without respite, he fled
from cave to cave, always seeing his following diminish, until it
consisted of few more than 200 men, and these tired and worn out.
Gonsalo el Xeniz, who was Alcaide, agreed with a silversmith of
Granada, called Francisco Barrado, to capture Aben Aboo or to kill
him, as he was the cause of so many lives being lost. So, one night,
el Xeniz arranged to meet Aben Aboo in the caves of Berchul, on the
pretext that it was necessary to talk over matters which concerned
everyone. Aben Aboo came alone, as he confided to nobody where
he slept. El Xeniz said to him, Abdala Aben Aboo: what I wish to
say to you is that you should look at these caves, which are full of
unhappy people, sick folk and widows and orphans, and things have
come to such a pass, that if all do not give themselves up to the
King's mercy, they will be killed and destroyed: and by doing the
contrary they will be relieved of their great misery.
When Aben Aboo heard this, he gave a cry as if his soul were
being torn out, and looking furious, he said, What? Xeniz! You have
brought me here for this? You harbour such treason in your breast!
Do not say any more, or let me see you again.
And saying this he left the cave, but a Moor called Cubeyas seized
his arms behind, and a nephew of el Xeniz gave him a blow on the
head with the butt of a musket and stupefied him and threw him to

the ground; then el Xeniz gave him a blow with a stone and killed
him. They took the body, wrapped in a matting of reeds, lying across
a mule, to Berchul, where Francisco Barrado and his brother Andres
were waiting for them. There they opened the corpse, took out the
intestines and filled the body with salt to preserve it; they then put it
on a sumpter mule, with boards at the back and front under the
clothes, to make it appear living. On the right rode the silversmith
Barrado, el Xeniz on the left, bearing the musket and scimitar of the
dead man, surrounded by el Xeniz's relations with their arquebuses
and muskets, and Luis de Arroyo and Jeronimo de Oviedo formed
the rear-guard with a troop of horse. In this manner they entered
Granada with a great crowd of people, who were anxious to see the
body of the dyer of the Albaicin, who had dared to call himself king
in Spain: the arquebuses fired salvos in the square of Bibarrambla
and again in front of the houses of the Audiencia, which were
answered by the artillery of the Alhambra. The President D. Pedro
Deza came out and el Xeniz gave him the musket and scimitar of
Aben Aboo, saying that he did so like the faithful shepherd, who
being unable to bring to his master the animal alive, brought the
skin. Then they cut off the head of the corpse, and abandoned the
body to the boys, who dragged it about and then burned it; the
head was nailed in an iron cage on the gate del Rastro, facing the
road to the Alpujarras, with an inscription underneath, which said:
This is the head
of the traitor Aben Aboo.
No one shall take it away
on pain of death.
Thus ended this celebrated Moorish war, another step by which D.
John of Austria mounted to the summit of his glory.

F
CHAPTER I
rom its narrowness and bareness it seemed a prison, and no
comparison could be found for the scarcity of its furniture; its
triangular shape and massive walls, on which could be seen the
remains of torn-down tapestry, luxurious gilt cornices, and carved,
vaulted ceiling, suggested, as in reality was the case, the corner of a
sumptuous room which, for convenience or by caprice, had been cut
off by a partition. In the centre of this partition rose an altar of dark
wood, without other images or adornments than a life-sized crucifix;
the pallid limbs of the Christ stood out with imposing realism against
the dark background; the dying head was bowed, and its agonised
gaze fixed itself, with a gentle expression of mercy and sorrow, on
those who knelt beneath it. In the opposite corner was one of those
carved fifteenth-century cupboards, of so much value now, but of so
little then; it was open, and in its depths were to be seen many and
terrible instruments of penitence and a few books of prayer; leaning
against the wall was a shut-up folding seat, the only one, and the
only piece of furniture to be seen in this curious room; a great silver
lamp glowed in front of the altar, and by its light could be vaguely
seen the outline of a strange figure, which was moving on the
ground on the frozen stones, giving vent to deep groans and dis-
jointed words.
Little by little the light began to filter through the narrow, arched
window which pierced one of the walls, and then the solitary
personage could be plainly seen; he was old, with a pronounced
aquiline nose, a white beard fell on his chest, and he was so spare
and decrepit, that it might have been said of him as St. Theresa said
of St. Peter Alcantara, That he seemed made of the roots of trees.
He was wrapped in a big black cloak, underneath which a kind of

white gown showed. He was prostrate before the altar, on the cold
stones, and was writhing like a feeble worm, at times leaning his
bald head on the ground, at others raising his withered arms
towards the crucifix, with a movement of love and anguish, like a
sorrowful child who craves the help of its father; then could be seen
the big gold ring with a great seal which moved up and down on his
finger as if it were threaded on a dried-up vine branch. It was full
daylight before the old man finally abandoned his lowly position and
somewhat arranged the disorder of his dress, which was none other
than the habit of a Dominican monk, whose wide folds seemed only
to heighten his tall figure. With a firm step he went to a little door in
the partition, almost hidden by the altar, and through it went into
the adjoining room. This was a sumptuous octagonal oratory, whose
altar was exactly in front of the one in the miserable room where the
old man prayed, so that the rich silver cibary which enclosed the
Blessed Sacrament on the altar of the front room corresponded with
the feet of the crucifix in the back one. There was only one picture
on this magnificent altar, an artistic marvel: the celebrated Madonna
of Fra Angelico, known as the Salus Infirmorum. On the Gospel
side there was a rich canopy of cloth of gold, with faldstool and
cushions covered with the same; and in a line in front of the altar
there were four other faldstools covered with brocade, where four
prelates were praying; they wore white rochets over their purple
cassocks, and stoles embroidered at the neck. On the brilliantly
lighted altar could be seen everything arranged that was necessary
for celebrating the Holy Sacrifice of the Mass. As the old man
entered the oratory, the four prelates rose at once and bowed low
before him, because this old man, who a few seconds before was
moaning like a feeble child, and writhing on the ground before the
crucifix like a vile worm, was no less a person than Christ's Vicar on
earth; called then in the chronology of Roman Pontiffs Pope Pius V,
now in the calendar of saints, St. Pius V.
The Pope knelt under the canopy and buried his wrinkled forehead
in his thin fingers for a long while; then at a sign from him the four
prelates approached and began to robe themselves in the sacred

vestments to celebrate the Holy Sacrifice of the Mass. The Pope was
celebrant, with solemn slowness and deep devotion, although
nothing revealed to the outside world the depth of his internal
emotions.
But on reaching the Gospel of St. John an extraordinary thing
happened; he began to read it slowly, pausing, and marking all the
words, as one who understands and appreciates its deep meaning,
and suddenly, with his face strange and transfigured, and in a voice
which was not his own, he said these words: Fuit homo missus a
Deo, cui nomen erat Joannes! (There was a man sent from God,
whose name was John.) He paused for a minute, turned his face
towards the Virgin, gazing into space, as if seeing celestial visions,
and repeated in a questioning, humble, submissive, loving tone, like
a child asking his mother, Fuit homo missus a Deo, cui nomen erat
Joannes? and in his natural voice, firm, strong, and decided, he
repeated, for the third time, Fuit homo missus a Deo, cui nomen
erat Joannes.
From that moment the weight which was burdening the Pontiff
seemed lifted. The Holy League against the Turk, between the Holy
See, the Signory of Venice and the King of Spain, had been formed,
thanks to the efforts, energy, heroic patience and fervent prayers of
this feeble old man. The united forces of the three powers amounted
to 200 galleys, 100 ships, 50,000 infantry, 4000 horses, and 500
artillery with ammunition and apparatus. The expense of this army
was calculated at 600,000 crowns a day, of which Spain paid half,
Venice two-sixths, and the Holy See the other sixth part. The Pope
had named Marco Antonio Colonna, Duke of Paliano and Grand
Constable of Naples, to be General of his fleet; Venice placed at the
head of her contingent the veteran Sebastian Veniero; and the King
of Spain appointed as General of all his forces by land and sea his
brother D. John of Austria, who had just ended the war with the
Moors.
The Pope in person promulgated the articles of the Holy League
from the altar of St. Peter's. The Roman citizens filled the immense
Basilica, and Pius, standing in front of the altar, surrounded by the

Cardinals and foreign ambassadors, read the text of the document
himself with profound emotion. Then the Te Deum was intoned and
30,000 voices replied at once, and 30,000 hearts were moved with
faith and hope, because the horrors the Turks committed at the
taking of Nikosia, and the danger which threatened Famagusta and
all the island of Cyprus at the moment, made the whole of Europe
fear that Selim would execute, if he were not checked, the plan
which Mahomet II and Solomon the Magnificent had made, of
overcoming Italy and destroying Christianity there.
There remained, however, to be settled a matter of the utmost
importance, and it was this that overburdened the Holy Pontiff at the
time we saw him praying and groaning in the lonely corner, which he
himself had made, behind his oratory, to conceal from men his
converse with Heaven. It was the appointing of a Generalissimo for
the armada of the Holy League, who was worthy to be the leader of
the great enterprise, and who would be a skilful manipulator of this
complicated and difficult machine, on which all Christendom was
gazing and fixing their hopes. The allies did not agree over this, and,
as so often happens in politics, they put personal and wounded
vanity before the holy and noble end that the Pontiff had in view. He
proposed his own general, Marco Antonio Colonna; the Spaniards
wished for their D. John of Austria, the Venetians, without daring to
propose their general, Sebastian Veniero, rejected Colonna, as
having been a failure in the first League; they also objected to D.
John of Austria, on account of the lack of experience which they
imagined he must possess at twenty-four, and proposed the Duke of
Savoy, Emanuele Filiberto, or the Duke of Anjou, afterwards Henri III
of France, who had not revealed as yet his ineptitude and vices. The
arguments about D. John's youth weighed with the Pontiff, and he
inclined to the Duke of Anjou, thinking that his appointment might
possibly gain the help of his brother the King of France, who hitherto
had refused it. However, the time passed in vacillations and doubts,
proposals and refusals, until at last the allies resolved to leave the
appointment absolutely in the hands of the Pontiff, which did not

prevent anyone from using all the means in his power to influence
the august old man in their favour.
However, his holy diplomacy was too far above human cabals for
intrigues to affect his upright policy. The Pope resorted for three
consecutive days to prayer and penitence, as was his humble custom
in difficult circumstances, and on the fourth, on which we saw him
saying Mass before the Madonna of Fra Angelico, he convoked for
that morning the presence of the Cardinals Granvelle and Pacheco
and D. Juan de Zuñiga, the delegates of the King of Spain, and
Michele Suriano and Juan Surenzo, ambassadors from Venice, and
told them distinctly, without evasion, and in contradiction to his
previous opinion, that he named the Lord D. John of Austria
Generalissimo of the Holy League.
The Venetians looked disgusted; but the astute Granvelle was
before them with the only possible objection to D. John: Holy
Father! In spite of his twenty-four years? To which the Pope
answered with great firmness, In spite of his twenty-four years.
The Venetians then knew that they were vanquished, but made it
a condition that the Generalissimo should consult, in cases of
importance, with his two colleagues, thenceforward subordinates,
Marco Antonio Colonna and Sebastian Veniero.
The Pope agreed, shrugging his shoulders as if he granted a thing
of scant importance, and the next day signed the commission of D.
John which the Cardinal Granvelle presented to him, repeating, with
the profound feeling of security which Heaven gives to holy souls,
Fuit homo missus a Deo, cui nomen erat Joannes.

P
CHAPTER II
ius V wrote at once a brief to D. John of Austria, informing him of
his appointment, and telling him to come quickly to Italy to take
command of the fleet, saying that henceforward he looked on D.
John as a son; as a father he would care for his interest, and would
at once reserve for him the first kingdom conquered from the Turk;
that D. John was never to forget for a moment the great
undertaking which had fallen to his charge, and that he could count
on victory, as he (the Pope) promised it in God's name.
The Pope sent this brief to D. John by his legate a latere to Philip
II, Cardinal Alexandrino, who also bore, at the same time, important
communications for the Kings of France and Portugal. The Cardinal
Alexandrino Michele Bonelli was a nephew of the Pope, and still only
a boy, but he had so much prudence and sagacity and tact in the
management of affairs, that he enjoyed the full confidence of the
Pontiff, who had named him his Secretary of State. However, the
Pope wished to counterbalance the youth of Alexandrino by the
importance and grey hair of those who accompanied him, and sent
in his suite Hipolito Aldobrandini, afterwards Clement VIII,
Alessandro Rierio, Mateo Contarelli, and Francesco Tarugi, all soon
afterwards Cardinals. This learned and splendid company all
disembarked at Barcelona, where they found awaiting them the
Nuncio Giovanni Battista Castagna, afterwards the Pope Urbain VII,
and the General of the Dominicans, Vincenzo Giustiniani; also,
representing the King, the Legate D. Herando de Borja, brother of
the Duque de Gandia, and representing D. John of Austria, his
Master of the Horse, D. Luis de Córdoba.
But it happened that while the embassy of Pius V was
disembarking at Barcelona, by other channels came the dreadful

news of the surrender of Famagusta, the awful death of Marco
Antonio Bragadino, and the horrible treachery committed by Mustafa
on these conquered heroes. For seventy-five days Famagusta
withstood the assault of 250 galleys which blockaded the island, and
of 120,000 Turks with whom Mustafa besieged the walls of the
unhappy town, which had to defend it only 4000 Italian soldiers, 200
Albanians, 800 horse, and between peasants and fishermen 3000
Cypriotes. Till at last, defeated and wanting food, the brave
Governor of the place, Marco Antonio Bragadino, counted the forces
left to him, and found them to be only 1700 soldiers and 1200
Cypriotes, counting sick and wounded, provision for two days, six
barrels of powder, and 120 cannon balls.
Then he thought of capitulating, and Mustafa favourably received
the first overtures they made, loading the officers who went to
propose the capitulation with presents and praises. The besieged
asked that their officers and men of war might be taken to the isle of
Crete with their arms and baggage: that the Turks should supply
galleys for the transport of the troops: that the inhabitants of
Famagusta should be allowed to keep their property and practise
their religion freely.
Mustafa agreed to everything, and even wished the soldiers to
take five cannon and three picked horses, as a testimony to their
heroic defence.
The capitulation was signed by both parties, and the soldiers
began at once to embark on the Turkish galleys.
The next day Bragadino set out from Famagusta to deliver up the
keys to Mustafa, who waited in his tent. He rode a magnificent
horse, preceded by trumpeters in gala armour, with surtout of purple
and a scarlet umbrella which a squire held over his head. The
principal leaders and gentlemen followed, to the number of twenty.
Mustafa received them in his tent with much courtesy, he made
Bragadino sit down at his side on the same divan, and talked for a
long while of the incidents of the siege. But, suddenly throwing off
the mask and revealing his black perfidy, he began to reproach the
Venetian General with having killed several Turkish prisoners in time

of truce, and with insolent arrogance and vehemence, asked him,
And what guarantees, Christian, are you giving me for the safety of
the boats which are taking you to Crete?
Bragadino was indignant at this question, which was an outrage
on the good faith of Venice, and replied that such an insulting
suspicion should have been shown before the capitulation was
signed. Mustafa then rose in a fury, and at a signal, which must have
been previously arranged, his guards threw themselves on Bragadino
and his comrades and loaded them with chains. In front of Mustafa's
tent there was a wide esplanade, and there they were beheaded,
one by one, with such violence that more than once their gore
bespattered Bragadino's purple surtout; three times they made him
kneel down at the block to be beheaded, and as often they took him
away again, just for the pleasure of causing him anguish, contenting
themselves at last by breaking his teeth, cutting off his nose and
ears, and pulling out his nails. Meanwhile the Turkish seamen threw
themselves on the Christian officers and soldiers already embarked,
took away their arms, and chained them to the benches, to convert
them into galley slaves. By dint of tortures the cruel Turks wore out
the noble Bragadino in twelve days. Every morning they beat him,
tied to a tree, and with two baskets of earth hanging from his neck
they made him work at the same forts which the illustrious General
had so gallantly defended. When he met Mustafa out walking, the
soldiers obliged him to kneel down and kiss the dust with his
mutilated lips.
Mustafa converted the cathedral of Famagusta into a mosque, and
to celebrate the sacrilegious ceremony, he ordered the martyred
Bragadino to be brought to his presence. Mustafa was seated on the
high altar, on the very ara, and from there condemned Bragadino to
be flayed alive, crying out in a diabolical rage, Where is your Christ?
See me seated on His altar! Why does He not punish me? Why does
He not set you free?
Bragadino answered nothing, and with the calm dignity of a
martyr began to say the Miserere. They began flaying him by his
feet, fearing that he would not be able to live through the torture,

and they were right; when his executioners reached his waist, and
while the heroic martyr was repeating the words cor mundum crea
in me Deus, he gave a dreadful shudder and died. They filled the
skin with hay, and put it on the yard of a ship, that all the crews
might see it.
These terrible tidings spread fear and consternation everywhere,
but specially in Italy and Spain; because the Ottoman monster, with
its gory claws fixed in defeated Cyprus, was lifting its head and
surveying Europe, seeking new conquest to satisfy its rage and
cupidity. Italy and Spain were the most exposed to fresh attacks of
the monster, with whom no power could then grapple successfully
single-handed, and this is why they welcomed the Holy League with
such enthusiasm, and the anxiety of those who meet with a means
of dissipating a looming danger; and for this also, that the arrival of
Cardinal Alexandrino was looked upon in Spain as an embassy from
Heaven, who was come to confer, as defender of the kingdom, the
invincible sword of the Archangel on D. John of Austria, its best
loved prince.
The Legate's journey from Barcelona to Madrid was one continued
triumphal march, and his entry into the city one of those events
which mark the history of a people. The pontifical ambassador
lodged provisionally at the convent of Atocha, while his official entry
into Madrid was being prepared.
The next day Prince Ruy Gómez de Silva came to visit the Legate
in the name of the King, accompanied by all the principal
personages of the Court, with much pomp and decked out with
many jewels, and two hours later D. John of Austria arrived on the
same errand, with the four Archdukes Rudolph, Ernest, Albert and
Wenceslas, brothers of the Queen Doña Ana, fourth wife of Philip II.
The Legate was very pleased to make D. John's acquaintance, and
talked to him for half an hour, addressing him as Highness, which
displeased Philip, and was the reason why he secretly advised all the
Chancelleries not thus to address his brother, as Philip had not
granted him this honour.

The solemn entry of the Legate was fixed for the next day, and for
it, adjoining the hospital of Anton Martin, and in front of the gate of
that name, was erected a big platform which occupied all the width
of the street, with five wide steps by which to mount on to it,
covered with costly carpets. In the midst of the platform an altar
was raised, with the finest tapestry and ornaments that the palace
could provide, and at the back a gorgeous room in which the Legate
might rest, as from there he was to see all the clergy and monks of
Madrid and the neighbourhood, who had come to receive him and to
offer their homage, pass before him.
At two o'clock D. John of Austria set out in a coach, and went to
the convent of Atocha to pick up the Legate, and enter by the gate
of St. Martin in his company; he was accompanied by his entire
household, in gala attire, and by several Grandees and gentlemen of
the Court, whom the King sent to add to his importance. D. John
was greatly beloved by the people of Madrid, and the naming him
Generalissimo, and the hopes that all Christendom placed in the
brave Prince, had increased their enthusiasm. His coming was
awaited by a great crowd of people, who at once surrounded his
coach and accompanied him to Atocha, applauding him and shouting
for joy. The Legate got into D. John's coach wearing his Cardinal's
cloak, hood and hat, and the enthusiasm of the people grew to such
a pitch, and so loudly did they acclaim D. John, the Legate and the
Pope, that Alexandrino, not accustomed to such a display of feeling,
was first frightened, and then wept for joy, bestowing blessings right
and left, anxious to show his gratitude.
When Alexandrino arrived at the platform, the procession had
already mounted by the street of Atocha, and he seated himself on
the velvet throne, which was placed on the Gospel side, with many
Monsignori, prelates and gentlemen of his household, and a little
before him on his right hand was a Papal Protonotary with the
pontifical standard, which was of white damask, with the tiara and
keys on one side and Christ on the cross on the other. Right and left
of the throne and on the steps, the soldiers of Spain and Germany
guarded him like a royal personage. Then, before the platform,

began to file the Confraternities with their standards, the monks with
their banners, and the parishes with their crosses, and many of the
neighbouring villages had brought their dancers, minstrels, and
clarions, and others were accompanied by Alcaides, Regidors and
Alguacils, all with their wands. On passing they bowed first to the
altar and then to the Legate, who, in return, gave them his blessing.
The King had so nicely calculated the time and the distance, that,
as the procession left by one side of the square, he entered by the
other in a coach, followed by his Spanish and German guard and by
the hundred noble archers. The King went towards the altar and the
Legate came to meet him, taking off his hat and the hood of his
cloak; to which D. Philip replied by bowing, hat in hand.
Then there passed between the two many polite words of
welcome, and then D. Philip and D. John of Austria mounted their
horses, and the Legate a beautiful mule, with cloth of crimson
velvet, a present from the city, and they went together to St. Mary's
to sing a Te Deum and announce the arrival of the Legate.
Twelve trumpeters headed the march with the attendants; two
spare horses covered with crimson velvet with fringes and trimmings
of gold, with saddles and saddle-cloths and bridles of great value;
the family, attendants and retainers, lackeys and pages with their
bags of crimson velvet embroidered with gold. The household of the
Legate and then that of the Alcaides de Corte, many private
gentlemen and members of the Orders, gentlemen purveyors and of
the bedchamber, and a great concourse of nobles and native and
foreign gentlemen. Then followed the Masters of the Horse and
Stewards of the King, Queen, Princess, and of D. John of Austria,
and mixing among them, in different lines, gentlemen and prelates
who had come with Cardinal Alexandrino.
Then a short space, in the midst of which rode, dressed in
mulberry, a Protonotary with the pontifical standard, preceded by
two lictors, and followed by two others wearing the livery of the
Legate and carrying the fasces of the Roman Consuls of old, which
had been granted to the Popes, as a sign of great respect, by the
Emperor Constantine.

The standard was escorted by two of Alexandrino's mace-bearers
and four of the King's, with their coats of arms and crowned maces,
and then followed the Grandees in such numbers, that seldom have
so many been together at one ceremony.
Then came D. John of Austria, and twenty paces behind, the King,
giving the Legate his right hand; but whether it was accidental or
intentional, it happened that on entering the street of Léon D. John
fell back to the King's left, and the three proceeded in a row,
conversing pleasantly, which was so extraordinary and unlike the
rigid etiquette always observed by D. Philip, that it was interpreted
as a public honour the King was doing to the Generalissimo of the
Holy League, and was greeted and welcomed by the populace with
great applause and renewed rejoicing and enthusiasm.
At the porch of St. Mary's the King took leave of the Legate,
without alighting, doffing his hat with great politeness, and the
Legate replied from his mule, in his turn taking off his hood and hat.
Then in the historic church they sang the Te Deum and the Regina
cœli lætare; Alexandrino gave the blessing from the epistle side, and
a Protonotary announced afterwards to the people, from the centre
of the altar, that the Very Illustrious Lord Cardinal Alexandrino,
nephew of the very holy Father and Lord Pius V, came to these
kingdoms of Spain as Legate a latere of His Holiness, and conceded
200 years of pardon to those present.
This ended the ceremony, and D. John of Austria got into his
coach again with the Legate, and conducted him to the lodging
which was prepared in the house of D. Pedro de Mendoza, where
the Presidents of Castille afterwards lived.

D.
CHAPTER III
John's departure once settled and fixed, his first thought was to
say good-bye to Doña Magdalena de Ulloa. Neither years, nor
the natural dazzling of triumph and glory, nor the dark clouds which,
on the contrary, brought disillusion and disenchantment, were ever
able to deaden in D. John his tender love for Doña Magdalena; away
at the bottom of his heart, joined to the religious faith which had
taken such firm root in his soul at Villagarcia, the loyal chivalry,
strong and manly, learned from Luis Quijada, and the active and
practical charity taught by Doña Magdalena herself, there was, so to
speak, like the foundations of the castle of his great nature, the
tender, respectful, confiding love he bore for Doña Magdalena, his
aunt, true remains of the former Jeromín who had become the D.
John who filled the world with his fame, and there always flourished
in him, as in all loyal breasts, the fragrant flower of gratitude.
D. John made a glory of his love and gratitude towards Doña
Magdalena de Ulloa, and in how many of his papers do these natural
and spontaneous gloryings burst forth, like a spring of crystal water
which seeks the first fissure by which to escape. Soon after the
triumph of Lepanto he wrote to the Marqués de Sarria, That my
aunt really is as delighted as she seems to be, I am very certain, as
we share each other's good fortunes, for no son owes his mother
more than I owe her.
So D. John wrote to Doña Magdalena, telling her of his
appointment as Generalissimo, and at the same time begging her to
name a place where he could go to receive her blessing and take
leave of her. He proposed that she should, as she had done before,
leave Villagarcia, where she was, for the convent of Abrojo or
Espina, where, without entering Valladolid, he would go to meet her.

It is certainly a curious circumstance, the reason for which we do not
know, that in none of the many visits D. John paid Doña Magdalena,
did he ever wish to enter Valladolid or stop in Villagarcia, but they
always met at one or other of these convents.
The courier who took D. John's letter brought back Doña
Magdalena's answer, that she would come to Madrid to give him the
blessing he craved and the embrace he desired, and thousands of
other blessings and embraces that she wanted to give him on her
own account. D. John, delighted, ordered the rooms to be prepared
that were always kept in his house for Doña Magdalena, which were
comfortable and apart, in one of the towers which flanked the
palace, which was, as we have said, that of the Conde de Lemus, in
the square of Santiago; it was spacious and magnificent, with two
stories and two towers, very like the Casa de Lujan, which still exists
in the Plaza de la Villa.
D. John and Doña Magdalena had not seen each other since the
death of Luis Quijada, and D. John was very much shocked at the
great change he saw in her. Doña Magdalena was no longer the
beautiful fine lady of whom good Luis Quijada had been so proud at
the entertainments and solemnities of the Court. His death had freed
her from the obligation of complying, like a good wife, with his
wishes, innocent vanities, and the calls of high rank; and now, free
from all such obligations, she had given herself entirely to the saintly
impulses of her austere virtue.
Two pictures of her still exist, which fully show these two phases
of her life. One is in the church of St. Luis at Villagarcia, and the
other in that of St. Isidoro at Oviedo, both founded by the noble
dame. In the first she is seen in all the glory of her youth and
beauty, which was remarkable, in magnificent attire, with costly
jewels and a commanding, though at the same time modest,
attitude: the great lady who hides beneath her velvet and laces the
austere virtues of the saint. In the second picture she wears the
severe dress of the widows of the sixteenth century, more or less
similar to that of many nuns of our own day, still handsome, but
worn by years, penitence and vigils; her weeds of coarse woollen

material, with wide stays stiffened with wood at the waist; she wears
no jewels, nor is there anything white in her dress, not even the coif
or veil which surrounds her pale face; her pose is humble, but at the
same time it has something noble and commanding, even elegant:
the picture of the saint who cannot altogether hide under her
mourning and sackcloth the dignity of the lady of high degree.
It was this last Doña Magdalena in her humility and mourning that
D. John received in his arms when she alighted from her litter, at the
old palace in the square of Santiago. Without a word she pressed
him for a long while to her heart, and then made the sign of the
Cross on his forehead, as she always did in old times to Jeromín
when he got up and when he went to bed. D. John seized the
generous hand, and kissed it again and again, at which those
present were much affected, not only the faithful servants from
Villagarcia, who had come with Doña Magdalena, but all D. John's
household, who had gone to receive her as if she really were his
mother.
For some time Doña Magdalena had known that envy was making
unworthy murmurings against D. John, and with all a mother's
solicitude and fear she had told him of this. D. John's answer to this
letter from Doña Magdalena is the only one that remains of this
interesting correspondence; it breathes the lad's noble confidence
and his absolute faith in the justice of the King, and the tranquillity
of his conscience. After several arguments which prove this, he
adds, You tell me, making me very great, to be careful what I do,
as all eyes are fixed on me, and that I should not be too gay, but
rather avoid all occasions which might be harmful. Again I kiss your
hands for what you are doing for me, and I beg you not to tire in so
doing. To this, Lady, I reply with the simple truth of which I am such
a friend; I give endless thanks to Our Lord that since the loss of my
uncle and father I have always tried to live though absent from one
who was always so good to me as he would wish me to live, and
thus I think that I have not ruled myself so badly or done so little,
that in this respect anyone can affirm the contrary. However much I
should wish to wear smart clothes, the work of a nine months'

campaign would not afford me much opportunity to do so; moreover,
Lady, all times and conditions are not the same, and I see that
sensible people, who are not fools, change as they get older; if there
are others in the world who, in order to speak ill, fall on anybody, it
does not alarm me, whatever they may murmur or say, and as you
write that this has come to such a pitch that you did not even dare
to ask news about me; however, as far as that goes, saints are not
free from the vexations of the world, but I will try to do my utmost
to behave as you think best, whose good advice I pray that I may
always enjoy, because there is no one I wish or ought to please like
her to whom I owe my up-bringing and my present position; this I
shall remember even in my grave. I pray you to forgive such a long
discourse, as the inventions of the times are enough to make a man
do what he least intended, and let me know if those of the Lady
Abbess[11]
are such as to disturb greatly your peace of mind.
These murmurs wounded Doña Magdalena more than if they had
been directed against herself, and her wish to defend D. John and
warn and advise him, were the principal reasons for her coming to
Madrid; because it seemed to her that all this would be easier in her
leisurely visit than to await a passing one from him, which would of
necessity be hurried and agitated. D. John quieted Doña Magdalena,
opening out his heart to her. These rumours, according to him, came
from the Marqués de los Vélez and the Marqués de Mondejar, whose
vanity was wounded, especially the former's, by D. John's victory
over the Moors, which they had not been able to effect with more
time, money and means of action. But these murmurs had had no
influence on the King, so D. John declared. He showed himself a
most loving brother, giving such positive proofs of his confidence in
D. John by appointing him General of the Fleet, and of his paternal
solicitude by counsels and instructions, so that even two days before
he had given a big sheet, corrected by his own hand, in which was
set forth the addresses and formulas to be used in D. John's
correspondence with every sort of person, from the Pope and Kings
to the humblest Councillor or Prior of the Orders. Then Doña
Magdalena asked whether to the names of Mondejar and los Vélez

should not be added another, not so illustrious, but at the same time
more powerful, Antonio Pérez.
D. John strongly repudiated the suspicion. Antonio Pérez had
always been one of his warmest friends. So Doña Magdalena did not
insist further, as she had spoken more by instinct than having certain
proof. She, however, permitted herself to repeat smilingly an Italian
proverb, which Luis Quijada was always quoting, about the honeyed
snares and deceptions of the Court, Chi non sa fingersi amico non
sa essere inimico. Which impressed D. John, coming from her,
although, unfortunately, not as the instinctive cry of alarm should
have done, no doubt an inspiration from Heaven. Then D. John
talked of another person, who was at that time a thorn in his side,
his mother Barbara Blombergh. Away in Flanders, where she lived,
the frivolity and want of decorum of this lady's life had begun to
displease the great Duque de Alba, the Governor of those States,
and he was contemplating taking some violent measures, as she
seemed not to listen to prudent counsels, and the solution D. John
wished was to move her to Spain, for Doña Magdalena to receive
her and constitute herself Barbara's guardian angel.
It grieved Doña Magdalena to see him so sad, and she promised,
and, as we shall see later, performed all he asked; and to distract his
attention from such bitter thoughts, she showed him with glee the
rich neckties and fine shirts she had brought him as a present,
because one of Doña Magdalena's attentions to D. John was that he
never wore any linen that was not sewn by her own hands. She was
always at work, and then sent him large parcels, carefully packed,
wherever he happened to be.
Doña Magdalena's faithful servants came to pay their respects to
D. John, whom they had known as a little boy at Villagarcia. The old
accountant Luis de Valverde, the two squires Juan Galarza and Diego
Ruiz, and the first duenna of honour Doña Petronilla de Alderete, all
came; the other duenna Doña Elizabeth de Alderete was left behind
at Villagarcia to look after Doña Ana of Austria; the duenna came in
very much overcome, and knelt down before D. John to kiss his
hand; but he, touched and smiling and always full of fun, lifted the

frail old woman in the air like a feather, and clasped her in his arms,
and, seeing Jeromín, she dared just to press the smooth, noble
forehead of the future conqueror of Lepanto with her lips. What joy
for her this embrace of her beloved Jeromín, and what an honour
and glory to have kissed the forehead of this august prince, for
whom she—she and nobody else—had sewn and tried on his first
breeches!
The satisfaction lasted the good woman to the end of her days,
and in her will, made three years later at Villagarcia, she left D. John
her savings, 320 ducats, to redeem captives of Lepanto, who were
to give honour to D. John and to pray for her soul.

D.
CHAPTER IV
John started from Madrid to embark at Barcelona on
Wednesday, the 6th of June, 1571, at three o'clock in the
afternoon. He was accompanied only by his Master of the Horse D.
Luis de Córdoba, his gentleman D. Juan de Gúzman, the secretary
Juan de Soto, the valet Jorge de Lima, a caterer, a cook, two D.
Juanillos or fools, two couriers, a guide and three servants, in all
fifteen horses. The rest of his following and servants had been
divided into two parties, one which went on ahead with his Lord
Steward the Conde de Priego, and the other which followed under
the chamberlain D. Rodrigo de Benavides. D. John had arranged this
in order to set out more quietly, and to avoid the manifestations of
the love and enthusiasm of the people of Madrid, which he well
knew not to be to the taste of certain personages. His precaution,
however, was useless, because the people got wind of his departure,
and from the morning waited in the little square of Santiago,
watching for his coming, and when he got to the gate of
Guadalajara, the crowd was so great, that it overflowed into the
country and extended all along the side of the road.
The magnificent Roman gate called Guadalajara still existed then,
its strong blocks of rock united by an enormous arch with railings
and balustrades of the same golden stone. Above this archway, and
standing out bravely between two towers, was the beautiful chapel
with two altars, one to venerate the figure of Our Lady, called la
Mayor, the other that of a Guardian Angel, with a naked sword in his
right hand and a model of Madrid in his left. All travellers used to
pray there, and following the usual custom, D. John alighted and
mounted to the chapel; and he appeared afterwards at the railing to
bow to the people, who were acclaiming him, and such were the

cries of blessing, good-byes and hurrahs, that, according to a writer
of the time, it resounded more than was necessary in some crooked
ears.
D. John slept that night at Guadalajara, in the country house of
the Duque del Infantado, who was waiting there for D. John, with
his two brothers D. Rodrigo and D. Diego de Mendoza, his brother-
in-law the Duque de Medina de Rioseco, and the Conde de Orgaz, all
most intimate friends of D. John. He spent Thursday there, and on
Friday, after dinner, continued his journey, with more haste and
courage, says Vander Hammen, than pleased those who followed
him. D. John truly journeyed with a light heart, and the way seemed
long which separated him from his dreams of glory. His absolute
confidence in Doña Magdalena and her promises had dispelled the
fears he had for his mother's future, and the affectionate farewell,
and fatherly, prudent warnings of his brother the King, had made
him believe that the murmurs and tittle-tattle of those envious of
him had made no impression on the severe monarch. So D. John
was at peace, and he smiled at life, as fortune smiled on him; he
received everywhere honours and ovations, and, what pleased him
more, sincere marks of love and appreciation. A courier overtook him
at Calatayud with a papal brief and letters from Marco Antonio
Colonna, General of the pontifical fleet, and from the Cardinal
Granvelle, temporary Viceroy of Naples, urging him to come to
Messina, which was the meeting-place of the fleets of the Holy
League.
He stopped two days at Montserrat to visit the celebrated
sanctuary of the Virgin, and on Saturday, the 16th of June, he
entered Barcelona at five in the evening, amidst the salutes of
artillery on land and sea, the pealing of bells and the cheers of an
enormous crowd. The Prior D. Hernando de Toledo, who was Viceroy
of Catalonia, received him, with all the magistrates and nobility and
the Knight Commander D. Luis de Requesens, D. John's naval
lieutenant, who had been awaiting him there for three days. The city
overflowed with the noise and animation natural to a seaport on the
eve of the embarkation of a great enterprise. Flags were plentiful at

Biostatistics Manual for Health Research: A Practical Guide to Data Analysis Nafis Faizi

More Related Content

Similar to Biostatistics Manual for Health Research: A Practical Guide to Data Analysis Nafis Faizi (20)

Recently uploaded (20)

Biostatistics Manual for Health Research: A Practical Guide to Data Analysis Nafis Faizi