Instant Ebook Access, One Click Away – Begin at ebookgate.

com

An Introduction to Biostatistic 3rd Edition Thomas

Glover

https://ebookgate.com/product/an-introduction-to-
biostatistic-3rd-edition-thomas-glover/

OR CLICK BUTTON

DOWLOAD EBOOK

Get Instant Ebook Downloads – Browse at https://ebookgate.com

Click here to visit ebookgate.com and download ebook now
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Signs An Introduction to Semiotics 2nd Edition Thomas A.

Sebeok

https://ebookgate.com/product/signs-an-introduction-to-semiotics-2nd-
edition-thomas-a-sebeok/

ebookgate.com

The Holocaust An Introduction Thomas Dalton

https://ebookgate.com/product/the-holocaust-an-introduction-thomas-
dalton/

ebookgate.com

An Introduction To Harmonic Analysis 3rd Edition Yitzhak

Katznelson

https://ebookgate.com/product/an-introduction-to-harmonic-
analysis-3rd-edition-yitzhak-katznelson/

ebookgate.com

An Introduction to Enterprise Architecture 3rd Edition

Scott A. Bernard

https://ebookgate.com/product/an-introduction-to-enterprise-
architecture-3rd-edition-scott-a-bernard/

ebookgate.com
An Introduction to Geographical Information Systems 3rd
Edition Ian Heywood

https://ebookgate.com/product/an-introduction-to-geographical-
information-systems-3rd-edition-ian-heywood/

ebookgate.com

Practicing Archaeology An Introduction to Cultural

Resources Archaeology Second Edition Thomas William
Neumann
https://ebookgate.com/product/practicing-archaeology-an-introduction-
to-cultural-resources-archaeology-second-edition-thomas-william-
neumann/
ebookgate.com

Media Today An Introduction to Mass Communication 3rd

Edition Joseph Turow

https://ebookgate.com/product/media-today-an-introduction-to-mass-
communication-3rd-edition-joseph-turow/

ebookgate.com

An Introduction to the Analysis of Algorithms 3rd Edition

Michael Soltys

https://ebookgate.com/product/an-introduction-to-the-analysis-of-
algorithms-3rd-edition-michael-soltys/

ebookgate.com

Understanding Jurisprudence An Introduction to Legal

Theory 3rd Edition Raymond Wacks

https://ebookgate.com/product/understanding-jurisprudence-an-
introduction-to-legal-theory-3rd-edition-raymond-wacks/

ebookgate.com
An Introduction to Biostatistics
Third Edition
An Introduction to Biostatistics
Third Edition

Thomas Glover
Hobart and William Smith Colleges

Kevin Mitchell
Hobart and William Smith Colleges
For information about this book, contact:
Waveland Press, Inc.
4180 IL Route 83, Suite 101
Long Grove, IL 60047-9580
(847) 634-0081
[email protected]
www.waveland.com

Copyright © 2016, 2008, 2002 by Thomas Glover and Kevin Mitchell

10-digit ISBN 1-4786-2779-1

13-digit ISBN 978-1-4786-2779-4

All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or
transmitted in any form or by any means without permission in writing from the publisher.

Printed in the United States of America

7 6 5 4 3 2 1
 CONTENTS

Preface ix

1 Introduction to Data Analysis 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Populations and Samples . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Variables or Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Measures of Central Tendency: Mean, Median, and Mode . . . . . . . 5
1.5 Measures of Dispersion and Variability . . . . . . . . . . . . . . . . . . 8
1.6 Descriptive Statistics for Frequency Tables . . . . . . . . . . . . . . . . 14
1.7 The E↵ect of Coding Data . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.8 Tables and Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.9 Quartiles and Box Plots . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.10 Accuracy, Precision, and the 30–300 Rule . . . . . . . . . . . . . . . . 25
1.11 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2 Introduction to Probability 37
2.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.2 Use of Permutations and Combinations . . . . . . . . . . . . . . . . . 40
2.3 Introduction to Set Theory and Venn Diagrams . . . . . . . . . . . . . 44
2.4 Axioms and Rules of Probability . . . . . . . . . . . . . . . . . . . . . 47
2.5 Probability Rules and Mendelian Genetics (Optional) . . . . . . . . . 56
2.6 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3 Probability Distributions 71
3.1 Discrete Random Variables . . . . . . . . . . . . . . . . . . . . . . . . 72
3.2 The Binomial Distribution . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3 The Poisson Distribution . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.4 Continuous Random Variables . . . . . . . . . . . . . . . . . . . . . . . 87
3.5 The Normal Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.6 The Standard Normal Distribution . . . . . . . . . . . . . . . . . . . . 92
3.7 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
vi CONTENTS

4 Sampling Distributions 109

4.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.2 Distribution of the Sample Mean . . . . . . . . . . . . . . . . . . . . . 111
4.3 Confidence Intervals for the Population Mean . . . . . . . . . . . . . . 117
4.4 Confidence Intervals for the Population Variance . . . . . . . . . . . . 124
4.5 Confidence Intervals for a Population Proportion . . . . . . . . . . . . 127
4.6 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

5 Introduction to Hypothesis Testing 137

5.1 An Overview: The Famous Cornflakes Example . . . . . . . . . . . . . 137
5.2 Typical Steps in a Statistical Test of Hypothesis . . . . . . . . . . . . 142
5.3 Type I versus Type II Errors in Hypothesis Testing . . . . . . . . . . . 144
5.4 Binomial Example of Hypothesis Testing (Optional) . . . . . . . . . . 150
5.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

6 One-Sample Tests of Hypothesis 155

6.1 Hypotheses Involving the Mean (µ) . . . . . . . . . . . . . . . . . . . . 155
6.2 Hypotheses Involving the Variance ( 2 ) . . . . . . . . . . . . . . . . . 162
6.3 Nonparametric Statistics and Hypothesis Testing . . . . . . . . . . . . 166
6.4 The One-Sample Sign Test . . . . . . . . . . . . . . . . . . . . . . . . 167
6.5 Confidence Intervals Based on the Sign Test . . . . . . . . . . . . . . . 170
6.6 The One-Sample Wilcoxon Signed-Rank Test . . . . . . . . . . . . . . 172
6.7 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

7 Tests of Hypothesis Involving Two Samples 189

7.1 Comparing Two Variances . . . . . . . . . . . . . . . . . . . . . . . . . 189
7.2 Testing the Di↵erence Between Two Means of Independent Samples . 193
7.3 Confidence Intervals for µ1 µ2 . . . . . . . . . . . . . . . . . . . . . . 198
7.4 The Di↵erence Between Two Means with Paired Data . . . . . . . . . 200
7.5 The Wilcoxon Rank-Sum (Mann-Whitney U ) Test . . . . . . . . . . . 204
7.6 The Sign Test and Paired Data . . . . . . . . . . . . . . . . . . . . . . 207
7.7 The Wilcoxon Signed-Rank Test for Paired Data . . . . . . . . . . . . 209
7.8 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

8 k-Sample Tests of Hypothesis: The Analysis of Variance 231

8.1 Model I ANOVA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
8.2 Mean Separation Techniques for Model I ANOVAs . . . . . . . . . . . 243
8.3 Model II ANOVA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
8.4 The Kruskal-Wallis Test . . . . . . . . . . . . . . . . . . . . . . . . . . 255
8.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262

9 Two-Factor Analysis 276

9.1 Randomized Complete Block Design ANOVA . . . . . . . . . . . . . . 276
9.2 Factorial Design Two-Way ANOVA . . . . . . . . . . . . . . . . . . . . 287
9.3 The Friedman k-Sample Test: Matched Data . . . . . . . . . . . . . . 296
9.4 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
CONTENTS vii

10 Linear Regression and Correlation 315

10.1 Simple Linear Regression . . . . . . . . . . . . . . . . . . . . . . . . . 318
10.2 Simple Linear Correlation Analysis . . . . . . . . . . . . . . . . . . . . 330
10.3 Correlation Analysis Based on Ranks . . . . . . . . . . . . . . . . . . . 336
10.4 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

11 Goodness of Fit Tests 357

11.1 The Binomial Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
11.2 Comparing Two Population Proportions . . . . . . . . . . . . . . . . . 361
11.3 The Chi-Square Test for Goodness of Fit . . . . . . . . . . . . . . . . . 364
11.4 The Chi-Square Test for r ⇥ k Contingency Tables . . . . . . . . . . . 369
11.5 The Kolmogorov-Smirnov Test . . . . . . . . . . . . . . . . . . . . . . 376
11.6 The Lilliefors Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
11.7 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386

A Proofs of Selected Results 403

A.1 Summation Notation and Properties . . . . . . . . . . . . . . . . . . . 403
A.2 Expected Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
A.3 The Formula for SSTreat in a One-Way ANOVA . . . . . . . . . . . . . 413
A.4 ANOVA Expected Values . . . . . . . . . . . . . . . . . . . . . . . . . 413
A.5 Calculating H in the Kruskal-Wallis Test . . . . . . . . . . . . . . . . 416
A.6 The Method of Least Squares for Linear Regression . . . . . . . . . . . 417

B Answers to Even-Numbered Problems 419

Answers for Chapter 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
Answers for Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
Answers for Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
Answers for Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
Answers for Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
Answers for Chapter 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
Answers for Chapter 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
Answers for Chapter 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444
Answers for Chapter 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
Answers for Chapter 10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
Answers for Chapter 11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473

C Tables of Distributions and Critical Values 483

C.1 Cumulative Binomial Distribution . . . . . . . . . . . . . . . . . . . . 484
C.2 Cumulative Poisson Distribution . . . . . . . . . . . . . . . . . . . . . 490
C.3 Cumulative Standard Normal Distribution . . . . . . . . . . . . . . . . 492
C.4 Student’s t Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . 494
C.5 Cumulative Chi-Square Distribution . . . . . . . . . . . . . . . . . . . 497
C.6 Wilcoxon Signed-Rank Test Cumulative Distribution . . . . . . . . . . 499
C.7 Cumulative F Distribution . . . . . . . . . . . . . . . . . . . . . . . . 502
C.8 Critical Values for the Wilcoxon Rank-Sum Test . . . . . . . . . . . . 511
C.9 Critical Values of the q Statistic for the Tukey Test . . . . . . . . . . . 515
C.10 Fisher’s Z-transformation of Correlation Coefficient r . . . . . . . . . 517
C.11 Correlation Coefficient r Corresponding to Fisher’s Z-transformation . 520
C.12 Cumulative Distribution for Kendall’s Test (⌧ ) . . . . . . . . . . . . . 523
viii CONTENTS

C.13 Critical Values for the Spearman Rank Correlation Coefficient, rs . . . 526
C.14 Critical Values for the Kolmogorov-Smirnov Test . . . . . . . . . . . . 527
C.15 Critical Values for the Lilliefors Test . . . . . . . . . . . . . . . . . . . 528

References 529

Index 531

Guide to Hypothesis Testing 537

 PREFACE

Our goal in writing this book was to generate an accessible and relatively complete
introduction for undergraduates to the use of statistics in the biological sciences. The
text is designed for a one quarter or one semester class in introductory statistics for
the life sciences. The target audience is sophomore and junior biology, environmental
studies, biochemistry, and health sciences majors. The assumed background is some
coursework in biology as well as a foundation in algebra but not calculus. Examples
are taken from many areas in the life sciences including genetics, physiology, ecology,
agriculture, and medicine.
This text emphasizes the relationships among probability, probability distributions,
and hypothesis testing. We highlight the expected value of various test statistics under
the null and research hypotheses as a way to understand the methodology of hypoth-
esis testing. In addition, we have incorporated nonparametric alternatives to many
situations along with the standard parametric analysis. These nonparametric tech-
niques are included because undergraduate student projects often have small sample
sizes that preclude parametric analysis and because the development of the nonpara-
metric tests is readily understandable for students with modest math backgrounds.
The nonparametric tests can be skipped or skimmed without loss of continuity.
We have tried to include interesting and easily understandable examples with each
concept. The problems at the end of each chapter have a range of difficulty and come
from a variety of disciplines. Some are real-life examples and most others are realistic
in their design and data values. Throughout the text we have included short “Concept
Checks” that allow readers to immediately gauge their mastery of the topic presented.
Their answers are found at the ends of appropriate chapters. The end-of-chapter
problems are randomized within each chapter to require the student to choose the
appropriate analysis. Many undergraduate texts present a technique and immediately
give all the problems that can be solved with it. This approach prevents students
from having to make the real-life decision about the appropriate analysis. We believe
this decision making is a critical skill in statistical analysis and have provided a large
number of opportunities to practice and it.
The material for this text derives principally from a required undergraduate bio-
statistics course one of us (Glover) taught for more than twenty years and from a
second course in nonparametric statistics and field data analysis that the other of
x PREFACE

us (Mitchell) taught during several term abroad programs to Queensland, Australia.

Recent shifts in undergraduate curricula have de-emphasized calculus for biology stu-
dents and are now highlighting statistical analyses as a fundamental quantitative skill.
We hope that our text will make teaching and learning these processes less arduous.

Supplemental Materials
We have provided several di↵erent resources to supplement the text in various ways.
There is a set of Additional Appendices that are available online at http://waveland.
com/Glover-Mitchell/Appendices.pdf. These are coordinated with this text and
contain further information on several topics, including:

• additional post-hoc mean comparison techniques beyond those covered in the

text for use in the analysis of variance;

• a method for determining confidence intervals for the di↵erence between medians
of independent samples based on the Wilcoxon rank-sum test;

• a discussion of the Durbin test for incomplete block designs; and

• a section covering three di↵erent field methods.

Also available is set of 300 additional problems to supplement those in the text.
This material is available online for both students and instructors at http://waveland.
com/Glover-Mitchell/ExtraProblems.pdf.
An Answer Manual for instructors is available free on CD from the publisher.
Included on this CD are

• a PDF file containing both questions and answers for all the problems in the text
and another file containing only the answers for all the problems in the text;

• a PDF file containing the supplementary problems mentioned above and another
file containing both the questions and the answers to all of the supplementary
problems; and

• a PDF file of the Additional Appendices.

The material in this textbook can be supported by a wide variety of statistical

packages and calculators. The selection of these support materials is dictated by
personal interests and cost considerations. For example, we have successfully used
SPSS software (http://www.spss.com/) in the laboratory sessions of our course. This
software is easy to use, relatively flexible, and can complete nearly all the statistical
techniques presented in our text.
A number of free online statistical tools are also available. Among the best is The R
Project for Statistical Computing which is available at http://www.r-project.org/.
We have created a companion guide, Using R: An Introduction to Biostatistics for
our text that is available free online at http://waveland.com/Glover-Mitchell/
r-guide.pdf, and is also included on the CD for instructors. In the guide we work
through almost every example in the text using R. A useful feature of R is the ability
to access and analyze (large) online data files without ever downloading them to your
own computer. With this in mind, we have placed all of the data files for the examples
and exercises online as text files. You may either download these files to use with R
or any other statistical package, or in R you may access the text files from online
and do the necessary analysis without ever copying them to your own computer. See
http://waveland.com/Glover-Mitchell/data.pdf for more information.
Our students have used Texas Instrument calculators ranging from the TI-30 to
the TI-83, TI-84, and TI-89 models. The price range for these calculators is consid-
erable and might be a factor in choosing a required calculator for a particular course.
Although calculators such as the TI-30 do less automatically, they sometimes give
the student clearer insights into the statistical tests by requiring a few more computa-
tional steps. The ease of computation a↵orded by computer programs or sophisticated
calculators sometimes leads to a “black box” mentality about statistics and their cal-
culation.
For both students and instructors we recommend D. J. Hand et al., editors, 1994,
A Handbook of Small Data Sets, Chapman & Hall, London. This book contains
510 small data sets ranging from the numbers of Prussian military personnel killed
by horse kicks from 1875–1894 (data set #283) to the shape of bead work on leather
goods of Shoshoni Indians (data set #150). The data sets are interesting, manageable,
and amenable to statistical analysis using techniques presented in our text. While a
number of the data sets from the Handbook were utilized as examples or problems
in our text, there are many others that could serve as engaging and useful practice
problems.

Acknowledgments
For the preparation of this third edition, thanks are due to the following people:
Don Rosso and Dakota West at Waveland Press for their support and guidance; Ann
Warner of Hobart and William Smith Colleges, for her meticulous word processing of
the early drafts of this manuscript; and the students of Hobart and William Smith
Colleges for their many comments and suggestions, particularly Aline Gadue for her
careful scrutiny of the first edition.

Thomas J. Glover
Kevin J. Mitchell
Geneva, NY
 1

Introduction to Data Analysis

Concepts in Chapter 1:
• Scientific Method and Statistical Analysis
• Parameters: Descriptive Characteristics of Populations
• Statistics: Descriptive Characteristics of Samples
• Variable Types: Continuous, Discrete, Ranked, and Categorical
• Measures of Central Tendency: Mean, Median, and Mode
• Measures of Dispersion: Range, Variance, Standard Deviation, and Standard
Error
• Descriptive Statistics for Frequency Data
• E↵ects of Coding on Descriptive Statistics
• Tables and Graphs
• Quartiles and Box Plots
• Accuracy, Precision, and the 30–300 Rule

1.1 Introduction
The modern study of the life sciences includes experimentation, data gathering, and
interpretation. This text o↵ers an introduction to the methods used to perform these
fundamental activities.
The design and evaluation of experiments, known as the scientific method, is
utilized in all scientific fields and is often implied rather than explicitly outlined in
many investigations. The components of the scientific method include observation,
formulation of a potential question or problem, construction of a hypothesis, followed
by a prediction, and the design of an experiment to test the prediction. Let’s consider
these components briefly.

Observation of a Particular Event

Generally an observation can be classified as either quantitative or qualitative. Quan-
titative observations are based on some sort of measurement, for example, length,
weight, temperature, and pH. Qualitative observations are based on categories reflect-
ing a quality or characteristic of the observed event, for example, male versus female,
diseased versus healthy, and mutant versus wild type.

Statement of the Problem

A series of observations often leads to the formulation of a particular problem or
unanswered question. This usually takes the form of a “why” question and implies
2 CHAPTER 1: Introduction to Data Analysis

a cause and e↵ect relationship. For example, suppose upon investigating a remote
Fijian island community you realized that the vast majority of the adults su↵er from
hypertension (abnormally elevated blood pressures with the systolic over 165 mmHg
and the diastolic over 95 mmHg). Note that the individual observations here are quan-
titative while the percentage that are hypertensive is based on a qualitative evaluation
of the sample. From these preliminary observations one might formulate the question:
Why are so many adults in this population hypertensive?

Formulation of a Hypothesis
A hypothesis is a tentative explanation for the observations made. A good hypothesis
suggests a cause and e↵ect relationship and is testable.
The Fijian community may demonstrate hypertension because of diet, life style,
genetic makeup, or combinations of these factors. Because we’ve noticed extraordinary
consumption of octopi in their diet and knowing octopods have a very high cholesterol
content, we might hypothesize that the high level of hypertension is caused by diet.

Making a Prediction
If the hypothesis is properly constructed, it can and should be used to make predic-
tions. Predictions are based on deductive reasoning and take the form of an “if-then”
statement. For example, a good prediction based on the hypothesis above would be:
If the hypertension is caused by a high cholesterol diet, then changing the diet to a low
cholesterol one should lower the incidence of hypertension.
The criteria for a valid (properly stated) prediction are:

1. An “if” clause stating the hypothesis.

2. A “then” clause that

(a) suggests altering a causative factor in the hypothesis (change of diet);

(b) predicts the outcome (lower level of hypertension);
(c) provides the basis for an experiment.

Design of the Experiment

The entire purpose and design of an experiment is to accomplish one goal, that is,
to test the hypothesis. An experiment tests the hypothesis by testing the correctness
or incorrectness of the predictions that came from it. Theoretically, an experiment
should alter or test only the factor suggested by the prediction, while all other factors
remain constant.
How would you design an experiment to test the diet hypothesis in the hypertensive
population?
The best way to test the hypothesis above is by setting up a controlled experiment.
This might involve using two randomly chosen groups of adults from the community
and treating both identically with the exception of the one factor being tested. The
control group represents the “normal” situation, has all factors present, and is used
as a standard or basis for comparison. The experimental group represents the “test”
situation and includes all factors except the variable that has been altered, in this case
SECTION 1.2: Populations and Samples 3

the diet. If the group with the low cholesterol diet exhibits significantly lower levels
of hypertension, the hypothesis is supported by the data. On the other hand, if the
change in diet has no e↵ect on hypertension, then a new or revised hypothesis should
be formulated and the experimental procedure redesigned. Finally, the generalizations
that are drawn by relating the data to the hypothesis can be stated as conclusions.
While these steps outlined above may seem straightforward, they often require
considerable insight and sophistication to apply properly. In our example, how the
groups are chosen is not a trivial problem. They must be constructed without bias and
must be large enough to give the researcher an acceptable level of confidence in the
results. Further, how large a change is significant enough to support the hypothesis?
What is statistically significant may not be biologically significant.
A foundation in statistical methods will help you design and interpret experiments
properly. The field of statistics is broadly defined as the methods and procedures for
collecting, classifying, summarizing, and analyzing data, and utilizing the data to test
scientific hypotheses. The term statistics is derived from the Latin for state, and orig-
inally referred to information gathered in various censuses that could be numerically
summarized to describe aspects of the state, for example, bushels of wheat per year,
or number of military-aged men. Over time statistics has come to mean the scientific
study of numerical data based on natural phenomena. Statistics applied to the life
sciences is often called biostatistics or biometry. The foundations of biostatistics
go back several hundred years, but statistical analysis of biological systems began
in earnest in the late nineteenth century as biology became more quantitative and
experimental.

1.2 Populations and Samples

Today we use statistics as a means of informing the decision-making processes in the
face of the uncertainties that most real world problems present. Often we wish to
make generalizations about populations that are too large or too difficult to survey
completely. In these cases we sample the population and use characteristics of the
sample to extrapolate to characteristics of the larger population. See Figure 1.1.
Real-world problems concern large groups or populations about which inferences
must be made. (Is there a size di↵erence between two color morphs of the same species
of sea star? Are the o↵spring of a certain cross of fruit flies in a 3 : 1 ratio of normal to
eyeless?) Certain characteristics of the population are of particular interest (systolic
blood pressure, weight in grams, resting body temperature). The values of these
characteristics will vary from individual to individual within the population. These
characteristics are called random variables because they vary in an unpredictable
way or in a way that appears or is assumed to depend on chance. The di↵erent types
of variables are described in Section 1.3.
A descriptive measure associated with a random variable when it is considered
over the entire population is called a parameter. Examples are the mean weight of
all green turtles, Chelonia mydas, or the variance in clutch size of all tiger snakes,
Notechis scutatus. In general, such parameters are difficult, if not impossible, to
determine because the population is too large or expensive to study in its entirety.
Consequently, one is forced to examine a subset or sample of the population and make
inferences about the entire population based on this sample. A descriptive measure
associated with a random variable of a sample is called a statistic. The mean weight
4 CHAPTER 1: Introduction to Data Analysis

Population(s) have traits called random variables.

Summary characteristics of the population random variables
are called parameters: µ, 2, N.
#
Random samples of size n
of the population(s) generate numerical data: Xi ’s.
#
These data can be organized into
summary statistics: X, s2 , n,
graphs, and figures (Chapter 1).
#
The data can be analyzed using
an understanding of basic probability (Chapters 2–4)
and various tests of hypotheses (Chapters 5–11).
#
The analyses lead to conclusions or inferences
about the population(s) of interest.

FIGURE 1.1. The general approach to statistical analysis.

of 25 female green turtles laying eggs on Heron Island or the variability in clutch size
of 50 clutches of tiger snake eggs collected in southeastern Queensland are examples
of statistics.
While such statistics are not equal to the population parameters, it is hoped that
they are sufficiently close to the population parameters to be useful or that the poten-
tial error involved can be quantified. Sample statistics along with an understanding
of probability form the foundation for inferences about population parameters. See
Figure 1.1 for review.
Chapter 1 provides techniques for organizing sample data. Chapters 2 through 4
present the necessary probability concepts, and the remaining chapters outline various
techniques to test a wide range of predictions from hypotheses.
Concept Checks. At the end of several of the sections in each chapter we include one or
two questions designed as a rapid check of your mastery of a central idea of the section’s
content. These questions will be be most helpful if you do each as you encounter it in the
text. Answers to these questions are given at the end of each chapter just before the exercises.

Concept Check 1.1. Which of the following are populations and which are samples?
(a) The weights of 25 randomly chosen eighth grade boys in the Detroit public school
system.
(b) The number of eggs found in each osprey nest on Mt. Desert Island in Maine.
(c) The heights of 15 redwood trees measured in the Muir Woods National Monument,
an old growth coast redwood forest.
(d ) The lengths of all the blind cave fish, Astyanas mexicanus, in a small cavern system
in central Mexico.
SECTION 1.3: Variables or Data Types 5

1.3 Variables or Data Types

There are several data types that arise in statistics. Each statistical test requires that
the data analyzed be of a specified type. Here are the most common types of variables.

1. Quantitative variables fall into two major categories:

(a) Continuous variables or interval data can assume any value in some
(possibly unbounded) interval of real numbers. Common examples include
length, weight, temperature, volume, and height. They arise from measure-
ment.
(b) Discrete variables assume only isolated values. Examples include clutch
size, trees per hectare, arms per sea star, or items per quadrat. They arise
from counting.

2. Ranked (ordinal) variables are not measured but nonetheless have a natural
ordering. For example, candidates for political office can be ranked by individual
voters. Or students can be arranged by height from shortest to tallest and
correspondingly ranked without ever being measured. The rank values have no
inherent meaning outside the “order” that they provide. That is, a candidate
ranked 2 is not twice as preferable as the person ranked 1. (Compare this with
measurement variables where a plant 2 feet tall is twice as tall as a plant 1 foot
tall. With measurement variables such ratios are meaningful, while with ordinal
variables they are not.)

3. Categorical data are qualitative data. Some examples are species, gender,
genotype, phenotype, healthy/diseased, and marital status. Unlike with ranked
data, there is no “natural” ordering that can be assigned to these categories.

When measurement variables are collected for either a population or a sample, the
numerical values have to be abstracted or summarized in some way. The summary de-
scriptive characteristics of a population of objects are called population parameters
or just parameters. The calculation of a parameter requires knowledge of the mea-
surement variables value for every member of the population. These parameters are
usually denoted by Greek letters and do not vary within a population. The summary
descriptive characteristics of a sample of objects, that is, a subset of the population,
are called statistics. Sample statistics can have di↵erent values, depending on how
the sample of the population was chosen. Statistics are denoted by various symbols,
but (almost) never by Greek letters.

1.4 Measures of Central Tendency: Mean, Median, and Mode

Mean
There are several commonly used measures to describe the location or center of a
population or sample. The most widely utilized measure of central tendency is the
arithmetic mean or average.
The population mean is the sum of the values of the variable under study divided
by the total number of objects in the population. It is denoted by a lowercase µ
(“mu”). Each value is algebraically denoted by an X with a subscript denotation i.
6 CHAPTER 1: Introduction to Data Analysis

For example, a small theoretical population whose objects had values 1, 6, 4, 5, 6, 3,

8, 7 would be denoted

X1 = 1, X2 = 6, X3 = 4, X4 = 5, X5 = 6, X6 = 3, X7 = 8, X8 = 7. (1.1)

We would denote the population size with a capital N . In our theoretical population
N = 8.
The population mean µ would be
1+6+4+5+6+3+8+7
= 5.
8
FORMULA 1.1. The algebraic shorthand formula for a population mean is
PN
i=1 Xi
µ= .
N
The Greek letter ⌃ (“sigma”) indicates summation. The subscript i = 1 indicates
to start with the first observation and the superscript N means to continue until and
including the N th observation. The subscript and superscript may represent other
starting and stopping points for the summation within the population or sample. For
the example above,
X5
Xi
i=2

would indicate the sum of X2 + X3 + X4 + X5 or 6 + 4 + 5 + 6 = 21.

XN
PN
Notice also that Xi is written i=i Xi when the summation symbol is embed-
i=1
ded in a sentence. In fact,P to further reduce clutter, the summation sign may not
indexed at all, for example Xi . It is implied that the operation of addition begins
with the first observation and continues through the last observation in a population,
that is,
X X N
Xi = Xi .
i=1

If sigma notation is new to you or if you wish a quick review of its properties, read
Appendix A.1 before continuing.
FORMULA 1.2. The sample mean is defined by
Pn
i=1 Xi
X= ,
n
where n is the sample size. The sample mean is usually reported to one more decimal place
than the data and always has appropriate units associated with it.

The symbol X (read “X bar”) indicates that the observations of a subset of size n
from a population have been averaged. X is fundamentally di↵erent from µ because
samples from a population can have di↵erent values for their sample mean, that is,
they can vary from sample to sample within the population. The population mean,
however, is constant for a given population.
Again consider the small theoretical population 1, 6, 4, 5, 6, 3, 8, 7. A sample of size
3 may consist of 5, 3, 4 with X = 4 or 6, 8, 4 with X = 6.
SECTION 1.4: Measures of Central Tendency: Mean, Median, and Mode 7

Actually there are 56 possible samples of size 3 that could be drawn from the
population in (1.1). Only four samples have a sample mean the same as the population
mean, that is, X = µ:

Sample Sum X

X3 , X6 , X7 4+3+8 5
X2 , X3 , X4 6+4+5 5
X5 , X3 , X4 6+4+5 5
X8 , X6 , X4 7+3+5 5

Each sample mean X is an unbiased estimate of µ but depends on the values

included in the sample and sample size for its actual value. We would expect the
average of all possible X’s to be equal to the population parameter, µ. This is, in fact,
the definition of an unbiased estimator of the population mean.
If you calculate the sample mean for each of the 56 possible samples with n = 3
and then average these sample means, they will give an average value of 5, that is,
the population mean, µ. Remember that most real populations are too large or too
difficult to census completely, so we must rely on using a single sample to estimate or
approximate the population characteristics.

Median
The second measure of central tendency is the median. The median is the “middle”
value of an ordered list of observations. Though this idea is simple enough, it will
prove useful to define it in terms of an even simpler notion. The depth of a value
is its position relative to the nearest extreme (end) when the data are listed in order
from smallest to largest.
EXAMPLE 1.1. The table below gives the circumferences at chest height (CCH) (in
cm) and their corresponding depths for 15 sugar maples, Acer saccharum, measured
in a forest in southeastern Ohio.

CCH 18 21 22 29 29 36 37 38 56 59 66 70 88 93 120
Depth 1 2 3 4 5 6 7 8 7 6 5 4 3 2 1

The population median M is the observation whose depth is d = N2+1 , where

N is the population size.
Note that this parameter is not a Greek letter and is seldom computed in practice.
Rather a sample median X̃ (read “X tilde”) is the statistic used to approximate
or estimate the population median. X̃ is defined as the observation whose depth is
d = n+1
2 , where n is the sample size. In Example 1.1, the sample size is n = 15, so the
depth of the sample median is d = 8. The sample median X̃ = X n+1 = X8 = 38 cm.
2

EXAMPLE 1.2. The table below gives CCH (in cm) for 12 cypress pines, Callitris
preissii, measured near Brown Lake on North Stradbroke Island.

CCH 17 19 31 39 48 56 68 73 73 75 80 122
Depth 1 2 3 4 5 6 6 5 4 3 2 1
8 CHAPTER 1: Introduction to Data Analysis

Since n = 12, the depth of the median is 12+1 2

= 6.5. Obviously no observation
has depth 6.5, so this is interpreted as the average of both observations whose depth
is 6 in the list above. So X̃ = 56+68
2
= 62 cm.

Mode
The mode is defined as the most frequently occurring value in a data set. The mode
of Example 1.2 would be 73 cm, while Example 1.1 would have a mode of 29 cm.
In symmetrical distributions the mean, median, and mode are coincident. Bimodal
distributions may indicate a mixture of samples from two populations, for example,
weights of males and females. While the mode is not often used in biological research,
reporting the number of modes, if more than one, can be informative.
Each measure of central tendency has di↵erent features. The mean is a purposeful
measure only for a quantitative variable, whether it is continuous (for example, height)
or discrete (for example, clutch size). The median can be calculated whenever a
variable can be ranked (including when the variable is quantitative). Finally, the
mode can be calculated for categorical variables, as well as for quantitative and ranked
variables.
The sample median expresses less information than the sample mean because it
utilizes only the ranks and not the actual values of each measurement. The median,
however, is resistant to the e↵ects of outliers. Extreme values or outliers in a sam-
ple can drastically a↵ect the sample mean, while having little e↵ect on the median.
Consider Example 1.2 with X = 58.4 cm and X̃ = 62 cm. Suppose X12 had been mis-
takenly recorded as 1220 cm instead of 122 cm. The mean X would become 149.9 cm
while the median X̃ would remain 62 cm.

1.5 Measures of Dispersion and Variability: Range,

Variance, Standard Deviation, and Standard Error
EXAMPLE 1.3. The table that follows gives the weights of two samples of albacore
tuna, Thunnus alalunga (in kg). How would you characterize the di↵erences in the
samples?

Sample 1 Sample 2

8.9 3.1
9.6 17.0
11.2 9.9
9.4 5.1
9.9 18.0
10.9 3.8
10.4 10.0
11.0 2.9
9.7 21.2

SOLUTION. Upon investigation we see that both samples are the same size and
have the same mean, X 1 = X 2 = 10.11 kg. In fact, both samples have the same
median. To see this, arrange the data sets in rank order as in Table 1.1. We have
n = 9, so X̃ = X n+1 = X5 , which is 9.9 kg for both samples.
2
Neither of the samples has a mode. So by all the descriptors in Section 1.4 these
samples appear to be identical. Clearly they are not. The di↵erence in the samples
SECTION 1.5: Measures of Dispersion and Variability 9

TABLE 1.1. The ordered samples of Thunnus

alalunga

Depth Sample 1 Sample 2

1 8.9 2.9
2 9.4 3.1
3 9.6 3.8
4 9.7 5.1
5 9.9 9.9
4 10.4 10.0
3 10.9 17.0
2 11.0 18.0
1 11.2 21.2

X̃1 = 9.9 kg X̃2 = 9.9 kg

is reflected in the scatter or spread of the observations. Sample 1 is much more

uniform than Sample 2, that is, the observations tend to cluster much nearer the
mean in Sample 1 than in Sample 2. We need descriptive measures of this scatter or
dispersion that will reflect these di↵erences.

Range
The simplest measure of dispersion or “spread” of the data is the range.
FORMULAS 1.3. The di↵erence between the largest and smallest observations in a group
of data is called the range:
Sample range = Xn X1
Population range = XN X1
When the data are ordered from smallest to largest, the values Xn and X1 are called the
sample range limits.

In Example 1.3 we have from Table 1.1

Sample 1: range = X9 X1 = 11.2 8.9 = 2.3 kg
Sample 2: range = X9 X1 = 21.2 2.9 = 18.3 kg
The range for each of these two samples reflects some di↵erences in dispersion,
but the range is a rather crude estimator of dispersion because it uses only two of
the data points and is somewhat dependent on sample size. As sample size increases,
we expect largest and smallest observations to become more extreme and, therefore,
the sample range to increase even though the population range remains unchanged.
It is unlikely that the sample will include the largest and smallest values from the
population, so the sample range usually underestimates the population range and is,
therefore, a biased estimator.

Variance
To develop a measure that uses all the data to form an index of dispersion consider
the following. Suppose we express each observation as a distance from the mean
10 CHAPTER 1: Introduction to Data Analysis

xi = Xi X. These di↵erences are called deviates and will be sometimes positive

(Xi is above the mean) and sometimes negative (Xi is below the mean).
If we try to average the deviates, they always sum to 0. Because the mean is the
central tendency or location, the negative deviates will exactly cancel out the positive
deviates. Consider a simple numerical example
X1 = 2 X2 = 3 X3 = 1 X4 = 8 X5 = 6.
The mean X = 4, and the deviates are
x1 = 2 x2 = 1 x3 = 3x5 = 2. x4 = 4
P
Notice that the negative deviates cancel the positive ones so that (Xi X) = 0.
Algebraically one can demonstrate the same result more generally,
n
X n
X n
X
(Xi X) = Xi X.
i=1 i=1 i=1

Since X is a constant for any sample,

n
X n
X
(Xi X) = Xi nX.
i=1 i=1
P P
Xi
Since X = n , then nX = Xi , so
n
X n
X n
X
(Xi X) = Xi Xi = 0. (1.2)
i=1 i=1 i=1

To circumvent this unfortunate property, the widely used measure of dispersion

called the sample variance utilizes the squares of the deviates. The quantity
n
X
(Xi X)2
i=1

is the sum of these squared deviates and is referred to as the corrected sum of
squares, denoted by CSS. Each observation is corrected or adjusted for its distance
from the mean.
FORMULA 1.4. The corrected sum of squares is utilized in the formula for the sample
variance, Pn
i=1 (Xi X)2
s2 = .
n 1
The sample variance is usually reported to two more decimal places than the data and has
units that are the square of the measurement units.

This calculation is not as intuitive as the mean or median, but it is a very good
indicator of scatter or dispersion. If the above formula had n instead of n 1 in
the denominator, it would be exactly the average squared distance from the mean.
Returning to Example 1.3, the variance of Sample 1 is 0.641 kg2 and the variance of
Sample 2 is 49.851 kg2 , reflecting the larger “spread” in Sample 2.
A sample variance is an unbiased estimator of a parameter called the population
variance.
SECTION 1.5: Measures of Dispersion and Variability 11

2
FORMULA 1.5. A population variance is denoted by (“sigma squared”) and is defined
by PN
2 i=1 (Xi µ)2
= .
N
It really is the average squared deviation from the mean for the population. The
n 1 in Formula 1.4 makes it an unbiased estimate of the population parameter. (See
Appendix A.2 for a proof.) Remember that “unbiased” means that the average of all
possible values of s2 for a certain size sample will be equal to the population value 2 .
Formulas 1.4 and 1.5 are theoretical formulas and are rather tedious to apply
directly. Computational formulas utilizeP the fact Pthat most calculators with statistical
registers simultaneously calculate n, Xi , and Xi2 .
P
FORMULA 1.6. The corrected sum of squares (Xi X)2 may be computed more simply
as P
X ( X i )2
CSS = Xi2 .
n
P (
P
Xi )2
Xi2 is the uncorrected sum of squares and n
is the correction term.
To verify Formula 1.6, using the properties in Appendix A.1 notice that
X X 2 X X X 2
(Xi X)2 = (Xi2 2Xi X + X ) = Xi2 2X Xi + X .
P P
Remember that X = nXi , so nX = Xi ; hence
X X X 2 X 2 2 X 2
(Xi X)2 = Xi2 2X(nX) + X = Xi2 2nX + nX = Xi2 nX .
P
Xi
Substituting for X yields
n

X X ✓ P ◆2 X P 2 X P 2
2 2 Xi n( Xi ) ( Xi )
(Xi X) = Xi n = Xi2 = Xi2 .
n n2 n
FORMULA 1.7. Use of the computational formula for the corrected sum of squares gives
the computational formula for the sample variance
P 2 (P Xi )2
2 Xi n
s = .
n 1
Returning to Example 1.3, Sample 2,
X X
Xi = 91, Xi2 = 1318.92, n = 9,
so 2
1318.92 (91) 1318.92 920.11 398.81
s2 = 9
= = = 49.851 kg2 .
9 1 8 8
Remember, the numerator must always be a positive number because it’s a sum of
squared deviations. Because the variance has units that are the square of the measure-
ment units, such as squared kilograms above, they have no physical interpretation.
With a similar derivation, the population variance computational formula can be
shown to be P 2 (P X i )2
2 Xi N
= .
N
Again, this formula is rarely used since most populations are too large to census
directly.
12 CHAPTER 1: Introduction to Data Analysis

Standard Deviation
FORMULAS 1.8. A more “natural” calculation is the standard deviation, which is the
positive square root of the population or sample variance, respectively.
s s
P 2 (P Xi )2 P 2 (P Xi )2
Xi N
Xi n
= and s= .
N n 1
These descriptions have the same units as the original observations and are, in a sense,
the average deviation of observations from their mean.
Again, consider Example 1.3.

For Sample 1: s21 = 0.641 kg2 , so s1 = 0.80 kg.

For Sample 2: s22 = 49.851 kg2 , so s2 = 7.06 kg.

The standard deviation of a sample is relatively easy to interpret and clearly reflects
the greater variability in Sample 2 compared to Sample 1. Like the mean, the standard
deviation is usually reported to one more decimal place than the data and always has
appropriate units associated with it. Both the variance and standard deviation can be
used to demonstrate di↵erences in scatter between samples or populations.

Thinking about Sums of Squares

It has been our experience teaching elementary descriptive statistics that students have
little problem understanding measures of central tendency such as the mean and median.
The sample variance and standard deviation, on the other hand, are often less intuitive to
beginning students. So let’s step back for a moment to carefully consider what these indices
of variability are really measuring.
Suppose a small sample of lengths (in cm) of small mouth bass is collected.

27 32 30 41 35 Xi ’s

These five fish have an average length of 33.0 cm. Some are smaller and others larger than
this mean. To get a sense of this variability, let’s subtract the average from each data point
(Xi 33) = xi generating what is called the deviate for each value. The data when rescaled
by subtracting the mean become

6 1 3 +8 +2 xi ’s

When we add these deviations, their sum is 0, so their mean is also 0. To quantify these
deviations and, therefore, the sample’s variability, we square these deviates to prevent them
from always summing to 0.

36 1 9 64 4 x2i = (Xi X)2

The sum of these squared deviates is

X 2 X
xi = (Xi X)2 = 36 + 1 + 9 + 64 + 4 = 114.
i i

This calculation is called the corrected or rescaled sum of squares (squared deviates).
If we averaged these calculations by dividing the corrected sum of squares by the sample
size n = 5, we would have a measure of the average squared distance of the observations
from their mean. This measure is called the sample variance. However, with samples this
SECTION 1.5: Measures of Dispersion and Variability 13

calculation usually involves division by n 1 rather than n. This modification addresses

issues of bias that are discussed in Section 1.5 and Appendix A.2.
The positive square root of the sample variance is called the standard deviation. In
this context, standard signifies “usual” or “average.” So the sample variance and standard
deviation are just measuring the average amount that observations vary from their center or
mean. They are simply averages of variability rather than averages of observation measure-
ment values like the mean. The fish sample had a mean of 33.0 cm with a standard deviation
of 5.3 cm.

Standard Error
The most important statistic of central tendency is the sample mean. However, the
mean varies from sample to sample (see page 7). We now develop a method to measure
the variability of the sample mean.
The variance and standard deviation are measures of dispersion or scatter of the
values of the X’s in a sample or population. Because means utilize a number of X’s
in their calculation, they tend to be less variable than the individual X’s. An extreme
value of X (large or small) contributes only one nth of its value to the sample mean
and is, therefore, somewhat dampened out.
A measure of the variability in X’s then depends on two factors: the variability
in the X’s and the number of X’s averaged to generate the mean X. We utilize two
statistics to estimate this variability.
FORMULAS 1.9. The variance of the sample mean is defined to be

s2
,
n
and standard deviation of the sample mean or, more commonly, the standard error
s
SE = p .
n

The standard error is the more important of these two statistics. Its utility will be
become clear in Chapter 4 when the Central Limit Theorem is outlined. The standard
error is usually reported to one more decimal place than the data, or if n is large, to
two more places.
EXAMPLE 1.4. Calculate the variance of the sample mean and the standard error
for the data sets in Example 1.3.

SOLUTION. The sample sizes are both n = 9. For Sample 1, s2 = 0.641 kg2 , so the
variance of the sample mean is

s2 0.641
= = 0.71 kg2
n 9
and the standard deviation is s = 0.80 kg, so the standard error is
s 0.80
SE = p = p = 0.27 kg.
n 9

For Sample 2, s2 = 49.851 kg2 , so the variance of the sample mean is

s2 49.851
= = 16.62 kg2
n 9
14 CHAPTER 1: Introduction to Data Analysis

and the standard deviation is s = 7.06 kg, so the standard error is

s 7.06
SE = p = p = 2.35 kg.
n 9

Concept Check 1.2. The following data are the carapace (shell) lengths in centimeters of
a sample of adult female green turtles, Chelonia mydas, measured while nesting at Heron
Island in Australia’s Great Barrier Reef. Calculate the following descriptive statistics for this
sample: sample mean, sample median, corrected sum of squares, sample variance, standard
deviation, standard error, and range. Remember to use the appropriate number of decimal
places in these descriptive statistics and to include the correct units with all statistics.

110 105 117 113 95 115 98 97 93 120

1.6 Descriptive Statistics for Frequency Tables

When large data sets are organized into frequency tables or presented as grouped data,
there are shortcut methods to calculate the sample statistics: X, s2 , and s.
EXAMPLE 1.5. The following table shows the number of sedge plants, Carex flacca,
found in 800 sample quadrats in an ecological study of grasses. Each quadrat was
1 m2 .

Plants/quadrat (Xi ) Frequency (fi )

0 268
1 316
2 135
3 61
4 15
5 3
6 1
7 1

To calculate the sample descriptive statistics using Formulas 1.2, 1.7, and 1.8 would
be quite arduous, involving sums and sums of squares of 800 numbers. Fortunately,
the following formulas limit the drudgery for these calculations.
It is clear that X1 = 0 occurs f1 = 268 times, X2 = 1 occurs f2 = 316 times, etc.,
and that the sum of observations in the first category is f1 X1 , the sum in the second
category is f2 X2 , etc. The sum of all observations is, therefore,
c
X
f1 X1 + f2 X2 + · · · + fc Xc = fi Xi ,
i=1

where
Pc c denotes the number of categories. The total number of observations is n =
i=1 i , and as a result:
f
FORMULA 1.10. The sample mean for a grouped data set is given by
Pc
i=1 fi Xi
P
X= c .
i=1 fi
SECTION 1.6: Descriptive Statistics for Frequency Tables 15

Similarly, the computational formula for the sample variance for a grouped data set
can be derived directly from
Pc
2 fi (Xi X)2
s = i=1 .
n 1
FORMULA 1.11. The sample variance for a grouped data set is given by
Pc P
f i Xi )2
2 i=1 fi Xi2 ( n
s = ,
n 1
Pc
where n = i=1 fi .

To apply Formulas 1.10 and 1.11, we need to calculate only three sums:
P
• The sample size n = fi
P
• The sum of observations fi Xi
P
• The uncorrected sum of squared observations fi Xi2
Returning to Example 1.5, it is now straightforward to calculate X, s2 , and s.

Plants/quadrat (Xi ) fi f i Xi fi Xi2

0 268 0 0
1 316 316 316
2 135 270 540
3 61 183 549
4 15 60 240
5 3 15 75
6 1 6 36
7 1 7 49

Sum 800 857 1805

Note that column 4 in the table above is generated by first squaring Xi and then
multiplying by fi , not by squaring the values in column 3. In other words, fi Xi2 6=
(fi Xi )2 .
The sample mean is
Pc
i=1 fi Xi 857
X= P c = = 1.1 plants/quadrat,
i=1 f i 800
the sample variance is
Pc
Pc 2 ( i=1 fi Xi )
2
(857)2
2 i=1 fi (Xi ) n 1805 800 2
s = = = 1.11 (plants/quadrat) ,
n 1 800 1
and the sample standard deviation is
p
s = 1.11 = 1.1 plants/quadrat.
Example 1.5 summarized data for a discrete variable taking on whole number
values from 0 to 7. Continuous variables can also be presented as grouped data in
frequency tables.
16 CHAPTER 1: Introduction to Data Analysis

EXAMPLE 1.6. The following data were collected by randomly sampling a large
population of rainbow trout, Salmo gairdnerii. The variable of interest is weight in
pounds.

Xi (lb) fi f i Xi fi Xi2

1 2 2 2
2 1 2 4
3 4 12 36
4 7 28 112
5 13 65 325
6 15 90 540
7 20 140 980
8 24 192 1536
9 7 63 567
10 9 90 900
11 2 22 242
12 4 48 576
13 2 26 338

Sum 110 780 6158

Rainbow trout have weights that can range from almost 0 to 20 lb or more. More-
over their weights can take on any value in that interval. For example, a particular
trout may weigh 7.3541 lb. When data are grouped as in Example 1.6 intervals are
implied for each class. A fish in the 3-lb class weighs somewhere between 2.50 and
3.49 lb and a fish in the 9-lb class weighs between 8.50 and 9.49 lb. Fish were weighed
to the nearest pound allowing analysis of grouped data for a continuous measurement
variable. In Example 1.6,
Pc
i=1 fi Xi 780
X= P c = = 7.1 lb
i=1 f i 110

and
Pc
Pc 2 ( i=1 fi X i )
2
(780)2
2 i=1 fi (Xi ) n 6158 110 2
s = = = 5.75 (lb) .
n 1 110 1
Therefore,
p
s= 5.75 = 2.4 lb.

Again, consider that calculation time is saved by working with 13 classes instead
of 110 individual observations. Whether measuring the rainbow trout to the nearest
pound was appropriate will be considered in Section 1.10.

1.7 The E↵ect of Coding Data

While grouping data can save considerable time and e↵ort, coding data may also o↵er
similar savings. Coding involves conversion of measurements or statistics into easier
to work with values by simple arithmetic operations. It is sometimes used to change
units or to investigate experimental e↵ects.
SECTION 1.7: The E↵ect of Coding Data 17

Additive Coding
Additive coding involves the addition or subtraction of a constant from each observa-
tion in a data set. Suppose the data gathered in Example 1.6 were collected using a
scale that weighed the fish 2 lb too low. We could go back to the data and add 2 lb to
each observation and recalculate the descriptive statistics. A more efficient tack would
be to realize that if a fixed amount c is added or subtracted from each observation in
a data set, the sample mean will be increased or decreased by that amount, but the
variance will be unchanged.
To see why, if X c is the coded mean, then
P P P P P
(Xi + c) Xi + c Xi + nc Xi
Xc = = = = + c = X + c.
n n n n
If s2c is the coded sample variance, then
P P P
2 [(Xi + c) (X + c)]2 (Xi + c X c)2 (Xi X)2
sc = = = = s2 ,
n 1 n 1 n 1
therefore, sc = s.
If the scale weighed 2 lb light in Example 1.6 the new, corrected statistics would
2
be X c = 7.1 + 2.0 = 9.1 lb, and s2c = 5.75 (lb) , and sc = 2.4 lb.

Multiplicative Coding
Multiplicative coding involves multiplying or dividing each observation in a data set by
a constant. Suppose the data in Example 1.6 were to be presented at an international
conference and, therefore, had to be presented in metric units (kilograms) rather than
English units (pounds). Since 1 kg equals 2.20 lb, we could convert the observations to
kilograms by multiplying each observation by 1/2.20 or 0.45 kg/lb. Again, the more
efficient approach would be to realize the following.
If each of the observations in a data set is multiplied by a fixed quantity c, the new
mean is c times the old mean because
P P
cXi c Xi
Xc = = = cX.
n n
Further the new variance is c2 times the old variance because
P P P 2 P
2 (cXi cX)2 [c(Xi X)]2 c (Xi X)2 2 (Xi X)2
sc = = = =c = c2 s2
n 1 n 1 n 1 n 1
and from this it follows that the new standard deviation is c times the old standard
deviation, sc = cs. (Remember, too, that division is just multiplication by a fraction.)
To convert the summary statistics of Example 1.6 to metric we simply utilize the
formulas above with c = 0.45 kg/lb.
X c = cX = 0.45 kg/lb (7.1 lb) = 3.20 kg.
s2c = c2 s2 = (0.45 kg/lb)2 (5.75 lb2 ) = 1.164 kg2 .
sc = cs = 0.45 kg/lb (2.4 lb) = 1.08 kg.
Our understanding of the e↵ects of coding on descriptive statistics can sometimes
help determine the nature of experimental manipulations of variables.
18 CHAPTER 1: Introduction to Data Analysis

EXAMPLE 1.7. Suppose that a particular variety of strawberry yields an average

50 g of fruit per plant in field conditions without fertilizer. With a high nitrogen
fertilizer this variety yields an average of 100 g of fruit per plant. A new “high yield”
variety of strawberry yields 150 g of fruit per plant without fertilizer. How much
would the yield be expected to increase with the high nitrogen fertilizer?

SOLUTION. We have two choices here: The e↵ect of the fertilizer could be addi-
tive, increasing each value by 50 g (Xi + 50) or the e↵ect of the fertilizer could be
multiplicative, doubling each value (2Xi ). In the first case we expect the yield of the
new variety with fertilizer to be 150 g + 50 g = 200 g. In the second case we expect
the yield of the new variety with fertilizer to be 2 ⇥ 150 g = 300 g. To di↵erentiate
between these possibilities we must look at the variance in yield of the original variety
with and without fertilizer. If the e↵ect of fertilizer is additive, the variances with and
without fertilizer should be similar because additive coding doesn’t e↵ect the variance:
Xi + 50 yields s2 , the original sample variance. If the e↵ect is to double the yield, the
variance of yields with fertilizer should be four times the variance without fertilizer
because multiplicative coding increases the variance by the square of the constant
used in coding. 2Xi yields 4s2 , doubling the yield increases the sample variance four
fold.

1.8 Tables and Graphs

The data collected in a sample are often organized into a table or graph as a summary
representation. The data presented in Example 1.5 were arranged into a frequency
table and could be further organized into a relative frequency table by expressing
each row as a percentage of the total observations or into a cumulative frequency
distribution by accumulating all observations up to and including each row. The cu-
mulative frequency distribution could be manipulated further into a relative cumu-
lative frequency distribution by expressing each row of the cumulative frequency
distribution as a percentage of the total. See columns 3–5 in Table 1.2 for the relative
frequency, cumulative frequency,
P and relative cumulative frequency distributions for
Example 1.5. (Here n = fi and r is the row number.)

TABLE 1.2. The relative frequencies, cumulative frequencies, and relative cumu-
lative frequencies for Example 1.5
Pr
fi Pr i=1 fi
n
(100) i=1 fi n
(100)
Xi fi Relative Cumulative Relative cumulative
Plants/quadrat Frequency frequency frequency frequency

0 268 33.500 268 33.500

1 316 39.500 584 73.000
2 135 16.875 719 89.875
3 61 7.625 780 97.500
4 15 1.875 795 99.375
5 3 0.375 798 99.750
6 1 0.125 799 99.875
7 1 0.125 800 100.000
P
800 100.0

Discrete data, as in Example 1.5, are sometimes expressed as a bar graph of

SECTION 1.8: Tables and Graphs 19

relative frequencies. See Figure 1.2. In a bar graph the bar heights are the relative
frequencies. The bars are of equal width and spaced equidistantly along the horizontal
axis. Because these data are discrete, that is, because they can only take certain values
along the horizontal axis, the bars do not touch each other.
40 ....
...............
...
... ...
... ...
................... ... ...
... ... ...
.... ... ...
... ..... ... ...
30 ... .... ... ...
... ... ... .....
... ... ... ...
... ... ... ...
... ... ... ....
...
Relative ...
... .....
...
...
...
...
20 ... ... ... ...
frequency ... ... ... ...
... ................
... .... ... .... ...
... ... ... ..... ... ...
... ... ... ... ... ...
... ... ... ... ... ...
... ... ... .... ... ...
10 ... ... ... ... ... ...
... ..... ... ...
... ... ...
... ...................
... ... ... ... ... ...
... ... ... ...
... ... ..... ... ...
... .... ... ... ... ... ....
... ... ... ..... ... ... ... ... ..................
.. .. .. .. .. ... .. .. .. .. ................. ............... ...............
0
0 1 2 3 4 5 6 7 8
Plants/quadrat

FIGURE 1.2. A bar graph of relative frequencies for Example 1.5.

The data in Example 1.6 can be summarized in a similar fashion with relative
frequency, cumulative frequency, and relative cumulative frequency columns. See Ta-
ble 1.3.

TABLE 1.3. The relative frequencies, cumu-

lative frequencies, and relative cumulative fre-
quencies for Example 1.6
Pr
fi Pr i=1 fi
Xi fi n
(100) i=1 fi n
(100)

1 2 1.82 2 1.82
2 1 0.91 3 2.73
3 4 3.64 7 6.36
4 7 6.36 14 12.73
5 13 11.82 27 24.55
6 15 13.64 42 38.18
7 20 18.18 62 56.36
8 24 21.82 86 78.18
9 7 6.36 93 84.55
10 9 8.18 102 92.73
11 2 1.82 104 94.55
12 4 3.64 108 98.18
13 2 1.82 110 100.00
P
110 100.00

Because the data in Example 1.6 are continuous measurement data with each class
implying a range of possible values for Xi , for example, Xi = 3 implies each fish
weighed between 2.50 lb and 3.49 lb, the pictorial representation of the data set is
a histogram not a bar graph. Histograms have the observation classes along the
horizontal axis. The area of the strip represents the relative frequency. (If the classes
20 CHAPTER 1: Introduction to Data Analysis

of the histogram are of equal width, as they often are, then the heights of the strips
will represent the relative frequency, as in a bar graph.) See Figure 1.3. The strips
in this case touch each other because each X value corresponds to a range of possible
values.
25

20

15
Relative
frequency
10

0
2 4 6 8 10 12 14
Weight in pounds

FIGURE 1.3. A histogram for the relative frequencies for Example 1.6.

While the categories in a bar graph are predetermined because the data are dis-
crete, the classes representing ranges of continuous data values must be selected by
the investigator. In fact, it is sometimes revealing to create more than one histogram
of the same data by employing classes of di↵erent widths.
EXAMPLE 1.8. The list below gives snowfall measurements for 50 consecutive years
(1951–2000) in Syracuse, NY (in inches per year). The data have been rearranged
in order of increasing annual snowfall. Create a histogram using classes of width
30 inches and then create a histogram using narrower classes of width 15 inches.
(Source: http://neisa.unh.edu/Climate/IndicatorExcelFiles.zip)

71.7 73.4 77.8 81.6 84.1 84.1 84.3 86.7 91.3 93.8
93.9 94.4 97.5 97.6 98.1 99.1 99.9 100.7 101.0 101.9
102.1 102.2 104.8 108.3 108.5 110.2 111.0 113.3 114.2 114.3
116.2 119.2 119.5 122.9 124.0 125.7 126.6 130.1 131.7 133.1
135.3 145.9 148.1 149.2 153.8 160.9 162.6 166.1 172.9 198.7

SOLUTION. Use the same scale for the horizontal axis (inches of annual snowfall) in
both histograms. Remember that the area of a strip represents the relative frequency
of the associated class. Since the snowfall classes of the second histogram (15 in) are
one-half those of the first histogram (30 in), then the vertical scale must be multiplied
by a factor of 2 so that equal areas in each histogram will represent the same relative
frequencies. Thus, a single year in the second histogram will be represented by a strip
half as wide but twice as tall as in the first histogram, as indicated in the key in the
upper left corner of each diagram.
In this case, the narrower classes of the second histogram provide more informa-
tion. For example, nearly one-third of all recent winters in Syracuse have produced
snowfalls in the 90–105 inch range. There was one year with a very large amount of
snowfall of approximately 200 in. While one could garner this same information from
the data itself, normally one would use a (single) histogram to summarize data and
not list the entire data set.
SECTION 1.8: Tables and Graphs 21

50 ..........................................
...................................... = 1 yr
45
40
35
30
Relative
25
frequency
20
15
10
5
0
0 30 60 90 120 150 180 210
Snowfall in inches per year

35 .......................
... ..
....................
= 1 yr

30

25

20
Relative
frequency
15

10

0
0 15 30 45 60 75 90 105 120 135 150 165 180 195 210
Snowfall in inches per year

FIGURE 1.4. Two histograms for the data in Exam-

ple 1.8. The areas of the strips represent the relative fre-
quencies. The same area represents the same relative fre-
quency in both graphs.
22 CHAPTER 1: Introduction to Data Analysis

It is worth emphasizing that to make valid comparisons between two histograms,

equal areas must represent equal relative frequencies. Since the relative frequencies of
all the classes in a histogram sum to 1, this means that the total area under each of the
histograms being compared must be the same. Look at Figure 1.4 for an application of
this idea.
Histograms are often used as graphical tests of the shape of samples usually test-
ing whether the data are approximately “bell-shaped” or not. We will discuss the
importance of this consideration in future chapters.

1.9 Quartiles and Box Plots

In the previous sections we have used sample variance, standard deviation, and range
to obtain measures of the spread or variability. Another quick and useful way to
visualize the spread of a data set is by constructing a box plot that makes use of
quartiles and the sample range.

Quartiles and Five-Number Summaries

As the name suggests, quartiles divide a distribution in quarters. More precisely, the
pth percentile of a distribution is the value such that p percent of the observations
fall at or below it. For example, the median is just the 50th percentile. Similarly, the
lower or first quartile is the 25th percentile and the upper or third quartile is
the 75th percentile. Because the second quartile is the same as the median, quartiles
are appropriate ways to measure the spread of a distribution when the median is used
to measure its center.
Because sample sizes are not always evenly divisible by 4 to form quartiles, we
need to agree on how to break a data set up into approximate quarters. Other texts,
computer programs, and calculators may use slightly di↵erent rules which produce
slightly di↵erent quartiles.
FORMULA 1.12. To calculate the first and third quartiles, first order the list of observations
and locate the median. The first quartile Q1 is the median of the observations falling below
the median of the entire sample and the third quartile Q3 is the median of the observations
falling above the median of the entire sample. The interquartile range is defined as
IQR = Q3 Q1 .

The sample IQR describes the spread of the middle 50% of the sample, that is, the
di↵erence between the first and third quartiles. As such, it is a measure of variability
and is commonly reported with the median.
EXAMPLE 1.9. Find the first and third quartiles and the IQR for the cypress pine
data in Example 1.2.

CCH 17 19 31 39 48 56 68 73 73 75 80 122
Depth 1 2 3 4 5 6 6 5 4 3 2 1

12+1
SOLUTION. The median depth is 2
= 6.5. So there are six observations below
the median. The quartile depth is the median depth of these six observations: 6+12
=
3.5. So the first quartile is Q1 = 31+39 2
= 35 cm. Similarly, the depth for the
third quartile is also 3.5 (from the right), so Q3 = 73+75
2
= 74 cm. Finally, the
IQR = Q3 Q1 = 74 35 = 39 cm.
SECTION 1.9: Quartiles and Box Plots 23

A compact way to report the descriptive information involving the quartiles and
the range is with a five-number summary of the data. It consists of the median,
the two quartiles, and two extremes.
EXAMPLE 1.10. Provide the five-number summary for this sample of 15 weights
(in lb) of lake trout caught in Geneva’s Lake Trout Derby in 1994.

Weight

2.26 3.57 7.86

2.45 1.85 3.88
4.60 4.90 3.60
3.89 2.14 1.52
2.83 1.84 2.12

15+1
SOLUTION. The sample size is n = 15. The median depth is d(X̃) = = 8. 2
The first quartile is determined by the seven observations below the median; hence
the quartile depth is 7+1
2
= 4. The ordered data set and depths are

Weight Depth Weight Depth Weight Depth

1.52 1 2.26 6 3.88 5

1.84 2 2.45 7 3.89 4
1.85 3 2.83 8 4.60 3
2.12 4 3.57 7 4.90 2
2.14 5 3.60 6 7.86 1

So X̃ = 2.83 lb, Q1 = 2.12 lb, and Q3 = 3.89 lb. The extremes are 1.52 lb and 7.86 lb.
The five-number summary is usually presented in the form of a chart:

Median: 2.83
Quartiles: 2.12 3.89
Extremes: 1.52 7.86

Two other measures of variability are readily computed from the five-number
summary. The IQR is the di↵erence in the quartiles, IQR = 3.89 2.12 = 1.77 lb.
The range is the di↵erence in the extremes, 7.86 1.52 = 6.34 lb.

Box Plots
The visual counterpart to a five-number summary is a box plot. Box plots can contain
more or less detail, depending on the patience of the person constructing them. Below
are instructions for a moderately detailed version that contains all the essentials.
1. Draw a horizontal or vertical reference scale based on the range of the data set.
2. Calculate the median, the quartiles, and the IQR.
3. Determine the fences f1 and f3 using the formulas below. Points lying outside
these fences will be considered outliers and may warrant further investigation.
f1 = Q1 1.5(IQR)
f3 = Q3 + 1.5(IQR)
When an outlier is detected, one should consider its source. Is it a misrecorded
data point? If it is legitimate, is it special in some way or other?
24 CHAPTER 1: Introduction to Data Analysis

4. Locate the two “adjacent values.” These are the smallest and largest data values
inside the fences.
5. Lightly mark the median, quartiles, and adjacent values on the scale. Choose a
scale to spread these points out sufficiently.
6. Beside the scale, construct a box with ends at the quartiles and a dashed interior
line drawn at the median. Generally this will not be at the middle of the box!
7. Draw a “whisker” (line segment) from the quartiles to the adjacent values that
are marked with crosses “⇥.” Mark any outliers beyond the fences (equivalently,
beyond the adjacent values) with open circles “ .”
EXAMPLE 1.11. Construct a box plot for the lake trout data in Example 1.10.

SOLUTION. We have already computed the quartiles and median. The fences are

f 1 = Q1 1.5(IQR) = 2.12 1.5(1.77) = 0.535 lb

f3 = Q3 + 1.5(IQR) = 3.89 + 1.5(1.77) = 6.545 lb
Only 7.86 lb is an outlier, and this is visually obvious in the box plot shown in
Figure 1.5. The adjacent values are 1.52 lb and 4.90 lb.
8

5 ⇥
..
Weight ..
....
(lb) ...
..
4 ...........................
... ...
... ...
... ...
... ..
3 ...... .... .... ......
.... ...
... ...
... ...
.........................
2 ..
...
⇥ ..

FIGURE 1.5. A box plot for the lake trout

data in Example 1.10. Notice that the range
can be determined as the di↵erence in the most
extreme values, and the IQR can be calculated
as the di↵erence in the quartiles at the top and
bottom edges of the box.

When a box plot, such as the one in Figure 1.5, is not symmetric about the dashed
median line, this is an indication that the data are not symmetrically distributed. We
will discuss the importance of this later in the text.
Some other graphic representations used in preliminary data analysis include stem-
and-leaf diagrams, polygons, ogives, and pictographs. Virtually all statistical packages
available for computers o↵er some of these techniques to rapidly investigate the shape
of a data set. Most of these manipulations are tedious to do by hand and are less useful
than the bar graphs and histograms previously presented. We leave the configuration
of these techniques and their interpretation for other authors and your instructor.
SECTION 1.10: Accuracy, Precision, and the 30–300 Rule 25

1.10 Accuracy, Precision, and the 30–300 Rule

All biologists are aware of the importance of accuracy and precision in data collection
and recording. While these two terms are used synonymously in everyday speech,
they have di↵erent meanings in statistics. Accuracy is the closeness of a measured
or computed value to its true value, while precision is the closeness of repeated mea-
surements of the same quantity to each other. A biased but sensitive instrument may
yield inaccurate but precise readings. On the other hand, an insensitive instrument
might result in an accurate reading, but the reading would be imprecise, since another
reading of the same object would be unlikely to yield an equally accurate value. Unless
there is bias in a measuring instrument, precision will lead to accuracy.
Some measurements are by their nature precise. When we count eggs in a monitor
lizard’s nest and record the number as 9 or 13, these are exact numbers and, therefore,
precise variates. Most continuous variables, however, are approximate with the exact
value unknown and unknowable. Recordings of continuous variable data imply a level
of precision by the number of digits used. For example, if the length of an adult
female monitor lizard is recorded as 97.2 cm, the implied true value of the length is
between 97.15 and 97.25 cm. In other words, the last digit recorded defines an interval
in which the exact value of the variable resides. A measurement of 97 cm implies a
length between 96.5 and 97.5 cm.
In most studies too much precision can slow down data collection while not con-
tributing significantly to the resolution of scientific questions. While it doesn’t make
sense to measure large eucalyptus trees to the nearest millimeter or to weigh sperm
whales to the nearest gram, what level of precision should be recorded? To how many
significant figures should we record measurements? Many biologists use the thirty–
three-hundred (30–300) rule to determine precision for data sets. This rule is easy
to apply and will save a great deal of time and e↵ort. Array the sample by order of
magnitude from largest to smallest measurement. The number of unit steps between
the largest and smallest value should be between 30 and 300. For example, if you
were collecting small shells in the intertidal zone of a beach and the largest was 9
mm and the smallest was 5 mm, the number of units steps would be 4 (a unit step
is a millimeter in this example). If you recorded the lengths to the nearest tenth of
a millimeter with the largest being 9.2 mm and the smallest 5.1 mm in length, the
unit step is now 0.1 mm and there are 41 unit steps (9.2 5.1 = 4.1 mm or 41 tenths
of mm) in the data array. The data set will now give you enough precision for most
statistical analyses and allow for a reasonable error in recording, that is, a mistake of
1 in the last digit recorded is now less than 2.5% as opposed to 25% when the data
were recorded to the nearest millimeter.
If sedge plant heights were measured to the nearest tenth of a centimeter with the
tallest being 194.3 cm and the shortest being 27.1 cm, the unit step would be tenths
of centimeters and the data array would have 1672 unit steps (194.3 27.1 = 167.2
or 1672 tenths of cm). Clearly there is more precision in this data set than is needed.
Recording these plant heights to the nearest centimeter would yield 167 unit steps
(194 27 = 167 cm) and would give enough precision for analysis while saving time
and e↵ort in data collection.
As stated earlier, statistics such as the sample mean and standard deviation should
have one more decimal place than the data, while the sample variance should carry
two more. For example, the mean for the sedge plant height might be 141.3 cm for
26 CHAPTER 1: Introduction to Data Analysis

data collected to the nearest centimeter. After calculation the sample statistics often
have to be rounded to the appropriate number of significant figures. The rules for
rounding are very simple. A digit to be rounded is not changed if it is followed by a
digit less than 5. If the digit to be rounded is followed by a digit greater than 5 or by 5
followed by other nonzero digits, it is increased by one. When the digit to be rounded
is followed by a 5 standing alone or followed by zeros, it is unchanged if it is even
but increased by one if it is odd. So a mean for the sedge data of 141.35 cm would
be rounded to 141.4 cm, while a mean of 141.25 cm would be rounded to 141.2 cm.
Similar rounding should be done for the standard deviation and variance.

Concept Check Answers

1. (Page 4.) Parts (a) and (c) describe samples; (b) and (d) describe populations.
2. (Page 14.) Sample mean: 106.3 cm, sample median: 107.5 cm, corrected sum of
squares: 898.1, sample variance: 99.79 cm2 , standard deviation: 10.0 cm, standard
error: 3.2 cm, and the range is 27 cm: 93–120 cm.

1.11 Problems

1. If X1 = 9, X2 = 8, X3 = 13, X4 = 6, and X5 = 9, evaluate the following:

X 4
X X
(a) Xi (b) Xi (c) (Xi X)
i=2
X X ⇣X ⌘2
(d ) (Xi X)2 (e) Xi2 (f ) Xi
X P 2 X 3
X
( Xi )
(g) Xi2 (h) Xi 9 (i ) (Xi + 2)
n i=1
n
X X X
(j ) 5 (k ) 2Xi (l ) 2 Xi
i=1

2. The poison dart frog, Dendrobates auratus, is native to Costa Rica and other
areas of Central and South America. They were introduced to Hawaii and have
flourished there. These frogs concentrate toxins from their food and also modify
various ingested compounds into toxins called allopumiliotoxins. A small sample
of these frogs was collected on Kauai Island and their overall lengths measured in
cm. The data appear below.

4.1 5.2 4.3 5.1 4.7 4.5 3.9 4.6 4.3

For this sample find the following:

(a) the sum of the X’s;
(b) the sample mean;
(c) the uncorrected sum of squares;
(d ) the corrected sum of squares;
(e) the standard deviation;
SECTION 1.11: Problems 27

(f ) the standard error; and

(g) the range.
(h) Briefly discuss these measurements in light of the 30–300 rule.

3. The red-tailed tropic bird, Phaethon rubricauda, is an extremely rare sea bird
that nests on several islands of the Queensland coast of Australia. As part of
a conservation e↵ort to manage these endangered birds, every nesting pair was
measured and weighed. Below are the body weights of these birds (in kg).

Female 2.45 2.57 2.81 2.37 2.01 2.50 2.32

Male 2.86 2.65 2.75 2.60 2.30 2.49 2.84

(a) Determine the following descriptive characteristics for the weights of the fe-
males: mean, variance, and standard deviation. Is this a sample or population?
Again, pay attention to number of decimal places and appropriate units.
(b) Determine the mean, variance, and standard deviation for the male weights.
(c) Comment on the di↵erences or similarities between the two data sets.

4. As part of a larger study of the e↵ects of strenuous exercise on human fertility and
fecundity, the ages (in years) of menarche (the beginning of menstruation) for 10
Olympic female endurance athletes (runners and swimmers) who had vigorously
trained for at least 18 months prior to menarche were recorded.

13.6 13.9 14.0 14.2 14.9 15.0 15.0 15.1 15.4 16.4

(a) Calculate the following descriptive statistics: sample mean, variance, standard
deviation, and median.
(b) Do you feel that the sample mean is significantly higher than the overall popu-
lation mean for non-athletes of 12.5 years? Provide a rationale for your answer.

5. Beetles of the species Psephenus herricki have an aquatic larval stage called “water
pennies.” Below are the overall lengths (mm) of a sample of water pennies collected
from a stream flowing into Seneca Lake in Geneva, NY. Calculate the sample mean,
variance, standard deviation, standard error, median, and range.

8.1 7.0 7.0 5.5 5.9

7.9 7.3 7.5 7.3 8.0
8.3 8.5 7.3 7.7 8.2
7.2 6.8 8.4 6.8 8.6
6.8 6.3 6.9 6.8 7.4

6. In a study to determine the distribution of Cepaea nemoralis snails at the Hanley

Wildlife Preserve in upstate New York, 100 sampling quadrats were surveyed, with
the results in the table below. Calculate the mean, median, and variance of the
number of snails per quadrat in this study. (Note: The corrected sum of squares
is 345.)
28 CHAPTER 1: Introduction to Data Analysis

No. of snails (Xi ) fi f i Xi fi Xi2

0 69 0 0
1 18 18 18
2 7 14 28
3 2 6 18
4 1 4 16
5 1 5 25
8 1 8 64
15 1 15 225

100 70 394

7. In a new reality television program called “Solitary Survivor,” 15 volunteers were

placed in isolation chambers without clocks, electronic devices, or any other means
of measuring time. After a week their individual daily sleep cycles (day lengths)
were measured and are recorded below.

Day lengths (hr) 26.0 25.5 26.5 24.3 24.2 26.5 27.4 26.6
25.3 26.1 25.9 25.4 26.2 25.1 27.1

(a) Calculate the mean and variance for this sample.

(b) Do you feel that the average day length for these people is significantly longer
than 24 hr? Explain.
8. Larvae of the insect family, Myrmeleontidae, live in sandy soil and dig cone-shaped
pits that trap ants and other prey that fall into them. These so-called ant lions
stay in the soil with their heads just below the bottom of the pit and wait for
unwary prey. An ecology class determined the number of ant lion pits in a sample
of 100 randomly selected 1-m-square quadrats.

Pits/quadrat (Xi ) fi f i Xi fi Xi2

0 5 0 0
1 15 15 15
2 23 46 92
3 21 63 189
4 17 68 272
5 11 55 275
6 5 30 180
7 2 14 98
8 1 8 64

100 299 1185

(a) Is the variable measured here discrete or continuous? Explain.

(b) Calculate the sample mean and variance for these data.
(c) Calculate the median and range for these data.
(d ) Which pair of calculations (b) or (c) is more informative? Why?
 SECTION 1.11: Problems 29

9. An undergraduate ecology student doing research on niche dimensions decided to

repeat part of R. H. MacArthur’s famous study of foraging behavior of warblers
in northeastern coniferous forests. She marked the heights of various branches in
several conifers with colored tape and observed two similar species of warbler with
binoculars and recorded the average foraging height for each bird. The heights in
feet for the individuals observed were:

Bay-breasted warbler 17 10 13 12 13 11 13 16 17 19
Blackburnian warbler 15 17 17 18 15 16 17 24 20 16 24 15

(a) Calculate the mean and standard deviation of the foraging heights for each
species. Comment on the results.
(b) Determine the median and range for each species. Which of the two statistics,
the standard deviation or the range, is a better reflection of the variability in
the foraging height? Explain.

10. As part of a larger university study on the transition from high school to college,
sleep habits of college freshmen were investigated. Below are the data gathered on
the number of hours slept per night during weekdays of the first semester for 100
freshmen.

Hours Students

5 14
6 17
7 28
8 25
9 10
10 6

100

Find the mean, median, and standard deviation for this data set.

11. Descriptive statistics for an extensive study of human morphometrics has been
completed in the United States. The measurements on height have to be converted
to centimeters from inches for publication in a British journal. The mean in the
study was 68 inches and the standard deviation was 10 inches. What should the
reported mean and variance be? (1 inch = 2.54 cm.)

12. (a) Invent a sample of size 5 for which the mean is 20 and the median is 15.
(b) Invent a sample of size 2 for which the mean is 20 and the variance is 50.

13. Find the median and quartile depths for samples of size n = 22, 23, 24, and 25.

14. (a) Complete a five-number summary for each of the samples in Example 1.3.
(b) Construct parallel box plots for these same data. Interpret these plots.
 30 CHAPTER 1: Introduction to Data Analysis

15. From a large population of tiger salamanders, Ambystoma tigrinum, a random

sample of 81 were collected and their lengths recorded to the nearest centimeter.
Calculate the mean, median, variance, and standard deviation for this sample.

Xi fi

10 2
11 8
12 17
13 22
14 14
15 10
16 7
17 1

81

16. In Problems 3–5, 9, and 15, which data sets satisfy the 30–300 rule? Explain.

17. Why are Problems 6 and 8 exempt from the 30–300 rule?

18. For Problem 8 make a table including the relative frequency, cumulative frequency,
and relative cumulative frequencies. Make a pictorial representation of this data
set. Should you use a bar graph or histogram here?

19. For Problem 15 make a table including the relative frequency, cumulative fre-
quency, and relative cumulative frequencies. Represent the data with a histogram.

20. The smallmouth black bass, Micropterus dolomieu, is a very popular game fish
throughout the temperate zones of North America. In a Bassmaster tournament
on the St. Lawrence River the following fish were caught and weighed to the nearest
10 grams.

1210 1610 1820 1470 1750

1920 1350 1540 1770 1300
1380 1450 1270 2110 2010

(a) For this sample find the mean and standard deviation. Also determine the
median and the range. Which pair of statistics above are more informative?
Provide a rationale for your answer.
(b) Construct a five-number summary for the data above.
(c) Develop a box plot from the five-number summary. Are there any outliers?

21. Suppose the statistics in (a) of the previous problem were going to be reported
in a newspaper article. The writer believes that the statistics would be more
understandable in pounds rather than grams. What values should she report?
 SECTION 1.11: Problems 31

22. In a study of lead concentration in breast milk of nursing mothers, the following
data were reported for mothers aged 21 to 25 living in Riyadh, Saudi Arabia.
(Based on data in: Younes, B. et al. 1995. Lead concentrations in breast milk of
nursing mothers living in Riyadh. Annals of Saudi Medicine, 15(3): 249–251.)

n X ± s (µg/dl)

19 0.777 ± 0.410

(a) Determine the sample variance.

P
(b) Determine Xi .
P
(c) Determine ( Xi ) 2 .
P 2
(d ) Determine Xi .
23. The Sumatran orangutan, Pongo pygmaeus abelii, is an endangered species. When
a small number of these orangutans was discovered on a remote island of Sumatra,
zoologists were sent to evaluate this newfound group. Using tranquilizer darts they
managed to capture and weigh every adult member of the group. Their data (in
kilograms) are given below. Summarize these data with the appropriate indices.
Are the data samples?

Weight (kg)

Females Males

49.7 90.1
48.0 88.0
55.0 79.0
46.3 85.5
44.9 92.3
49.0 90.6
51.3 88.3
89.5
77.3

24. You have been asked to prepare an information pamphlet on the water quality of
the public water supply for a town in in upstate New York. You need to present
data on various water characteristics in a meaningful and easily understood format.
In addition to the pamphlet each consumer will be given a kit to test their own tap
water for contaminants. Supposing the average lead contaminant concentration for
the water supply is 8.0 µg/L, what additional information should be supplied to the
consumer to allow him to make an assessment of the quality of the water coming
from his tap?
25. The timber rattlesnake, Crotalus horridus, is a top predator in the forest ecosys-
tems of eastern North America. It is found nowhere else in the world and has been
the subject of much controversy and myth. The males sexually mature in 5 years,
but the females require 7 to 11 years to mature. Their fecundity is quite low with
females giving birth to 4 to 14 young every 3 to 5 years. Their long development
time and small brood size put these animals at significant risk of extirpation or
 32 CHAPTER 1: Introduction to Data Analysis

extinction due to habitat destruction and indiscriminate hunting. Healthy popula-

tions of C. horridus have adult individuals that range in length from 91 to 137 cm
with a mean length of 115 cm. At a recent “snake rodeo” adult timber rattlers
were captured from a region in the Southern Tier of New York. These snakes’
lengths were crudely measured to the nearest 5 cm and are recorded below.

Length (cm) Frequency

70 5
75 7
80 8
85 10
90 10
95 13
100 15
105 10
110 14
115 11
120 9
125 3
130 0
135 1
140 4

120

Calculate the mean, standard deviation, and range for the rodeo sample. Do
the rodeo data indicate a population that is significantly shorter than the so-
called healthy population? Discuss briefly. Why do you think the 30–300 rule was
violated here?
26. The list below provides snowfall data for 50 consecutive years in Bu↵alo, NY (in
inches per year). The data have been rearranged in order of increasing annual
snowfall. Create two histograms for these data: The first should use classes of
width 20 in, the second should use classes of width 10 in. (HSDS, #278)
25.0 38.8 39.9 40.1 46.7 49.1 49.6 51.1 51.6
53.5 54.7 55.5 55.9 58.8 60.3 63.6 65.4 66.1
69.3 70.9 71.4 71.5 71.8 72.9 74.4 76.2 77.8
78.1 78.4 79.0 79.3 79.7 80.7 82.4 82.4 83.0
83.6 83.6 84.8 85.5 87.4 88.7 89.6 89.8 89.9
90.9 97.0 98.3 101.4 102.4 103.9 104.5 105.2 110.0
110.5 110.5 113.7 114.5 115.6 120.5 120.7 124.7 126.4

Do the two histograms yield di↵erent impressions of the data set?

27. A new bird-like dinosaur was recently discovered in northern China. A member of
the oviraptor group, this new dinosaur was given the genus name Gigantoraptor.
From the various adult fossils so far described, the estimates of this animal’s height
(in meters) at the shoulder are given below.
3.55 2.97 3.40 3.15 3.67 3.62 3.47 3.51 3.66 3.83

(a) Summarize these data with appropriate descriptive statistics.

(b) Comment on the precision and accuracy of the data set and its relationship to
the 30–300 rule.
Exploring the Variety of Random
Documents with Different Content
friends. He took no end of trouble about this business. After his
young friends had obtained situations he continued to look after
them. He took down their names and addresses in a special red
book kept for the purpose, and repeatedly asked them to dine with
him on Sunday afternoons. He usually requested that they should
go to some church or chapel in the evening. In his diary are
repeatedly such entries as the following; “Dined twenty-two of the
boys that I had got situations for, besides the people that were
staying in the house. I never forget that I had none to invite me to
their homes when I first came to London.” How much good such
kindness did it is impossible to tell; for the want of it many a young
man in the City goes to the bad.
Mr. Moore’s second marriage, in 1881, seems rather to have
increased than diminished his philanthropic zeal. A wedding trip of
two months in Italy and elsewhere was but a brief interval of
holiday, to be followed by still harder work in the cause of his Lord
and Master; and then came an illness which rendered necessary for
him more rest of brain and more healthy exercise for his body. In
his knowledge of London he was unrivalled. He knew it by night as
well as by day. Many a time he went down to St. George’s in the
East and to Wapping to look after the poor. He accompanied the
City missionaries into the lowest dens; and as he felt that the only
way of reformation was to get at the children, we cannot be
surprised to learn that in 1866 he became treasurer of the Field Lane
Ragged School, an institution at that time sorely in need of
pecuniary help. But his happiest days were those he spent at his
Border tower at Cumberland. There the house was always full of
visitors, and there the poor were equally welcome as the rich. There
also, he loved to act the part of a distinguished agriculturist and to
preside at cattle shows. His guests were very varied, and included
bishops, Scripture-readers, warehousemen, farmers, City
missionaries, Sunday-school children, pensioners, and statesmen.
He rejoiced in hunting; but all the while he looked after the homes
of the poor, and battled with the immorality which exists quite as
much in the country as in town.
Mr. Moore was a great lover of the Bible, and distributed it by the
thousand, far and near. He always insisted on its being read in
schools. When the Middle-class schools were established in London,
he offered a thousand pounds on condition that the Bible was read
there; but he refused to give it till he found that actually such was
the case. In the case of Christ’s Hospital, after Dr. Jacob’s sermon
on the institution, he became an ardent reformer. As prime warden
of the Fishmongers’, he distinguished himself by the vigour of his
speeches. When Paris was in want, and its people destitute of
bread, he flew to their relief; and no man was more active in giving
relief for the destitute when the Northfleet was sunk. In 1872, he
was proud to be the high sheriff of his native county. Among his last
public works was to give a supper to the cabmen of London, and to
attend the funeral of Dr. Livingstone. And he died as he lived—
engaged in works of mercy. In November, 1876, he left his grand
mansion in Cumberland to attend a meeting of the Nurses’ Institute
in Carlisle. While he was standing opposite the Grey Coat Inn, two
runaway horses, which had escaped from a livery stable, came
galloping up. One of them knocked Mr. Moore down. He was taken
up insensible. Sir William Gull was sent for; but from the first there
was no chance, and in twenty hours he was dead. Great was the
sorrow felt everywhere, and in London and Carlisle public meetings
were held for a George Moore memorial fund. At that in London the
Archbishop of Canterbury presided, and Mr. Samuel Morley was one
of the speakers.
Friend of the church as he was at all times, and especially attached
to the Evangelical clergy, in one thing he differed from them. “The
parsons,” he once said to a meeting of children at Wigton, “will tell
you a good deal about money. They will tell you that it is the root of
all evil; but my opinion is that it is a good thing to make plenty of
money, provided you make a proper use of it.” Such was George
Moore, and such were his views and works. We owe to Dr. Smiles a
biography of him, which is as interesting and instructive as could
well be imagined. It should be read by all City young men; it should
be in every City library. The character therein portrayed ought to be
studied, and revered, and imitated in every home. Few of us can
expect to realise his wealth, but his example is one to be held up to
every City man.
“People who believe,” says a writer in the Daily News, “that genius is
great natural power accidentally directed, may think that the career
of the late Mr. George Moore justifies the well-known definition. Mr.
Moore’s name was very well known, not in England only, but on the
continent, by every one who was labouring to lighten the misery of
the poor. The philanthropic schemes to which he gave the aid of his
energy, his knowledge of men and of life, and his money, were too
many to be numbered here. The French, in particular, cherish a
grateful memory of his benevolent activity, of the help he extended
to the victims in the war of 1870. To many who only heard of Mr.
Moore in his later life, and in the full tide of his helpfulness and
prosperity, it may have been unknown that he was the maker of the
fortune which he distributed with a generous hand. The biography
of him by Mr. Smiles, which has just been published, is a very
interesting account of a career which began in a humble though
honourable estate, and ended by a singular accident in the northern
town where it may be said to have begun. The history of ‘Self-Help’
is not invariably edifying. The chief end of man, after all, is not to
get on in the world, to make a great deal of money, and to have
paragraphs devoted to his glory. This is so far from being the case
that one has even to overcome a slight natural prejudice against the
strength which displays itself mainly in the acquisition of a fortune.
In almost every rank of life leisure has its charms and good gifts,
which a man who never takes rest must miss. The subject of Mr.
Smiles’s book escapes from the vulgar renown of the self-made by
his unselfishness. His energy, his ceaseless labours in his early life,
were not the manifestations of a desire for wealth and for
advancement, but the natural expression of immense natural
strength of mind and body. When success was secured, the same
vigour spent itself in work for other people—for the poor, the weak,
the helpless, the ignorant. Mr. Moore might have devoted himself to
the joys of the collector, of the sportsman, of the ambitious
parvenu. Instead of doing so, he made amusement and enjoyment
subordinate to work for the benefit of others. He had not the
hardness and narrowness of people whose career has been one of
victory over the natural pleasures and innocent impulses of an
indolent race. ‘I don’t think I ever came across any other self-made
man who had so entirely got the chill of poverty out of his bones,’ Dr.
Percival wrote to Mr. Smiles. His geniality and unselfishness soften
the edges of his iron will and determination. People may think that
so much of the material and force that make greatness, might have
been better employed in work of a nobler tone—in science,
literature, law, or art. Mr. Moore took the only career that was open
to him, the career that was most distinctly in contrast with the
pastoral life to which he was bred. He had no education in his
youth, none lay within his reach in the Cumbrian valley where he
was born. With the chances of Dr. Whewell he might have been a
Whewell. With an opening in the East, he might have been, if not a
Clive, a Meadows Taylor. As it happened, the choice lay between the
existence of a farm labourer and that of a tradesman.”
 CHAPTER XI.
ARTISTS AND WRITERS.

Men who are not supposed to be mercenary often make a great deal
of money. Most of our artists rose from very humble beginnings.
Turner was the son of a hair-dresser. Wilkie was desperately poor;
so was Barry; and William Etty, that great colourist, was the son of a
baker in York—was bound apprentice, wholly against his will, to a
printer in Hull; but he released himself from the shackles of so
uncongenial a pursuit. He was greatly self-taught, for the help he
derived for a hundred guineas, as a private pupil of Sir Thomas
Lawrence, seems rather to have baffled him with despair; yet he
became the most surprising and effective flesh-painter of his age.
The nude style of his figures has often been a topic of remark with a
certain order of critics. Etty himself was wont to say, “‘To the pure
in heart, all things are pure.’ My aim in all my great pictures has
been to paint some great moral on the heart.” He lived, in 1849, to
find all his great works—130 pictures—in the great room of the
Society of Arts: he died that year. By the universal acclamation of
artists he is regarded as our English Titian, and some claim for him a
still higher place, for his canvases have not only the wonderful colour
of that master, but the splendour of Paul Veronese. He died in his
beloved and native city of York; and the poor baker’s boy, by his
industry and genius, had become the master of a considerable
fortune.
Actors and actresses also have made much money. Amongst the
money-making men may emphatically be placed David Garrick, who
was fond of money, and careful about it to the last. Some of our
earlier circus people seem to have made much money.—Batty was
reputed to have died worth half a million.—Ducrow gave himself
extraordinary airs. When the Master Cutler and Town Council of
Sheffield paid Ducrow a visit, with the principal manufacturers and
their families, Ducrow sent word that he only waited on crowned
heads, and not upon a set of dirty knife-grinders.—Philip Astley was
born in 1742, at Newcastle-under-Lyme, where his father carried on
the business of a cabinet-maker. He received little or no education,
and after working a few years with his father, enlisted in a cavalry
regiment. His imposing appearance, being over six feet in height,
with the proportions of a Hercules, and the voice of a Stentor,
attracted attention to him; and his capture of a standard at the
battle of Emsdorff made him one of the celebrities of his regiment.
While serving in the army, he learned some feats of horsemanship
from an itinerant equestrian named Johnson, perhaps the man
under whose management Price introduced equestrian performances
at Sadler’s Wells, and often exhibited them for the amusement of his
comrades. On his discharge from the army, he was presented by
General Elliot with a horse, and thereupon he bought another in
Smithfield, and commenced those open-air performances in Lambeth
which have already been noticed.
After a time he built a rude circus upon a piece of ground near
Westminster Bridge, which had been used as a timber-yard, being
the site of the theatre which has been known by his name for nearly
a century. Only the seats were roofed over, the ring in which he
performed being open to the air. One of his horses, which he had
taught to perform a variety of tricks, he soon began to exhibit, at an
earlier period of each day, in a large room in Piccadilly, where the
entertainment was eked out with conjuring and ombres Chinoises—a
kind of shadow pantomine.
Having saved some money out of these performances, Astley
erected his amphitheatre. At the same time he had to contend with
a fierce competition from what was then the Royal Circus, which
afterwards was called the Surrey Theatre. Astley’s, however, soon
became the popular place of amusement, and as such was visited
and described by Horace Walpole. The fame of the place received a
further illustration in the remark of Dr. Johnson, who, speaking of
the popularity of certain preachers, and the ease with which they get
a crowd to hear them, said, “Were Astley to preach a sermon
standing on his head, or on a horse’s back, he would collect a
multitude to hear him, but no wise man would say he had made a
better sermon for that.”
Let us now turn to a master of homely English—a man whose name
was, at one time, in every one’s mouth, and an author, whose books,
at one time, every one read. His moral works excel in descriptive
power. In politics his savage personalities encircle sarcasm; his
faculty for inventing national nick-names, and mastery of a Saxon
style of inimitable raciness, have given his writings historical
reputation. He has never been equalled among political writers in
his capacity of explaining what he understood. He was the first
journalist who called attention to the condition of the working
classes, I mean William Cobbett.
William Cobbett was born at Farnham, in Surrey, in 1776. His father
was a very poor farmer, who knew enough to teach his boys to read,
and had enough of intellectual originality to think that the triumph of
Washington in the American War of Independence was just. William
began as a mere child to do something towards earning his own
livelihood, and took great delight in the flowers which, while
weeding in great folks’ gardens, he saw. When eleven years old, he
heard some one speak of the splendid flowers in the Royal Gardens
at Kew. Without a word of announcement, and with sixpence-
halfpenny in his pocket, he set off to seek employment in that
irresistible Paradise. When he reached Richmond his funds were
reduced to threepence, and he was very hungry. In a shop-window,
however, he saw the “Tale of a Tub,” price threepence. Mind
triumphed over body; he bought the tale; and sat under a hay-stack
reading it till he fell asleep. He was delighted beyond measure with
the piece, and continued to read and re-read it for many years. The
circumstance was not of happy omen. Swift’s terrible tale we should
pronounce to be as well-fitted to sap the moral and religious
principles of a lad as any book in the English language; and lack of
moral principle was the fatal defect of Cobbett throughout life.
He found employment at Kew, and no doubt gloated over the floral
splendours which he had come to see; but he returned to Farnham,
and grew up in his father’s house. He made an appointment one
day to meet some young friends and accompany them to Guildford
Fair; but coming upon the high road as the London coach was
passing in full career, he made up his mind on the spur of the
moment to start for London. He arrived at the foot of Ludgate Hill
with half-a-crown in his pocket. An honest hop-seller, who knew his
father, took him by the hand, and he found work as an Attorney’s
clerk. He speaks with unlimited abhorrence of the roguery he
witnessed and the misery he endured in this place. “No part of my
life,” he says, “has been totally unattended with pleasure except the
eight or nine months I passed in Gray’s Inn. The office—for so the
dungeon was called where I wrote—was so dark that on cloudy days
we were obliged to burn candles. I worked like a galley-slave from
five in the morning till eight or nine at night, and sometimes all night
long. * * * When I think of the saids and so forths, and the counts
of tautology that I scribbled over—when I think of those sheets of
seventy-two words, and those lines of two inches apart—my brain
turns. Gracious Heaven! if I am doomed to be wretched, bury me
beneath Iceland snows, and let me feed on blubber; stretch me
under the burning Line, and deny me Thy propitious dews; nay, if it
be Thy will, suffocate me with the infected and pestilential air of a
democratic club-room; but save me, save me from the desk of an
attorney!” Anything seemed better than this. William, acting again
on the spur of the moment, enlisted. For more than a year he did
duty at Chatham. Here he mastered grammar—an acquisition which
he always regarded as the basis of his fortunes. He read also in a
circulating library, swallowing enormous quantities of useful or
useless knowledge, and laying it up in a memory of great tenacity.
His father meanwhile was treated by him with heartless neglect.
The old man had been offended by his running away, and appears to
have made no effort to release him from the bondage of the
attorney’s office. When he enlisted, however, his father relented,
and wrote saying that the last hay-rick or pocket of hops at Farnham
would be sold off to buy his discharge. But William vouchsafed no
reply.
Cobbett’s regiment was ordered to Canada, and he accompanied it
to St. John’s, New Brunswick. Here his conduct as a soldier was
exemplary. His talent and activity made him conspicuous, and he
became sergeant-major, raised, though he was still but about
twenty, over the heads of thirty sergeants. In 1791 the regiment
returned to England, and he procured his discharge “in consideration
of his good behaviour, and the services he had rendered his
regiment.” Then occurred one of the most strange and ambiguous
episodes in his life. He lodged charges of pecuniary defalcation
against four of his late officers. A day was appointed for their trial
by court-martial. The functionaries met, the accused were present,
all was ready for commencement, when it transpired that Cobbett
was missing. As he was the accuser, the trial was adjourned to a
stated day in order that an opportunity might be afforded him to
appear. The court again met; he was again absent; the accused
officers, accordingly, were acquitted. They made some show of a
wish to proceed against Cobbett, and what looks very like a feint of
arresting him in his refuge at Farnham. But the upshot was that he
escaped to France, and passed from France, when the revolutionary
atmosphere became too hot for him, to America. Mr. Watson very
properly devotes a good deal of attention to these circumstances,
and we are bound to say that we agree with him in thinking that
Cobbett was bribed with a good round sum to suppress his charges.
It was, of course, an act of flagrant and base dishonesty; but there
is nothing in Cobbett’s life to prove that he shrank from dishonesty,
or was superior to temptation. He was a most affectionate husband
and father, and many of his advices to young men and to the poor
are excellent. His talent was of a coarse kind, but very great. His
activity and indomitable spirit deserve all admiration. He boasted,
probably with truth, that he had never passed an idle day.
Cobbett first distinguished himself in America by publishing a fierce
pamphlet against Priestley. He was soon a noted political writer,
taking the side of ultra-Toryism, and denouncing with furious
emphasis all that savoured of Radicalism or Republicanism. His
talent was indubitable; and as vehement and able rhetoric on the
Church-and-King side was then in demand, he attracted attention.
On returning to England, he was welcomed by the authorities as an
out-and-out Tory, and became the most violent, uncompromising,
and popular of writers on the ministerial side. It is worthy of
recollection that William Cobbett had his windows broken by the
mob for the vehemence of his anti-popular utterances. According to
his own account he met Pitt at dinner in Mr. Windham’s house; and
the fact is not impossible, so highly did ministers at that time prize
the aid of any one who could fight for them against the patriots.
By what steps it is needless to trace, Cobbett gradually sidled round,
and left the cause of the king for that of the mob. His circumstances
became embarrassed, and he fled to America, leaving behind him
debts to the value of upwards of £33,000. He resided at Long
Island, near New York, and continued to edit his Register. In a few
years the irrepressible giant—he stood six foot two, with shoulders
and chest and girth to match—returned to England. He had once
denounced Tom Paine as a miscreant whom no words could
blacken. He now brought Tom Paine’s bones with him, bent upon
having a grand monument built over them in England. In this
instance he signally misunderstood his countrymen. The dead man’s
bones were laughed at, and declared to be those of an old nigger.
Cobbett proposed to sell 20,000 hair-rings at a sovereign a-piece,
with some of Paine’s hair in each; and he was reminded that when
Paine died he was almost bald. Cobbett had at last to shuffle the
bones underground, no one knows where. His own eloquence and
sarcasm made him popular, and procured him a seat in parliament.
He was now the fiercest of democrats. He assailed Protestantism
and detested ministers of religion. His quackery grew worse and
worse until he died in 1835.
Sir Francis Chantrey was a poor lad. He began his career by being a
carver on wood. Rogers used to say—“One day Chantrey said to
him, ‘Do you recollect that about twenty-five years ago a
journeyman came to your house from the wood-carver employed by
you and Mr. Hope, to talk about these ornaments (pointing to some
on a mahogany sideboard), and that you gave him a drawing to
execute them by.’ Rogers replied that he recollected it well. ‘Well,’
said Chantrey, ‘I was that journeyman.’” Chantrey practised portrait-
painting both at Sheffield and after he came to London. It was in
allusion to him that Lawrence said—“A broken-down painter will
make a very good sculptor.”
In 1823, London society was much exercised on the subject of
literary gains. Miss Wynn writes in her “Diaries of a Lady of
Quality”—“I heard to-day from Mr. Rogers that Constable, the
bookseller, told him last May that he paid the author of ‘Waverley’
the sum of £110,000. To that may now be added the produce of
‘Red Gauntlet,’ and ‘St. Ronan’s Well;’ for I fancy Quentin Durward’
was at least printed, if not published. I asked whether the ‘Tales of
my Landlord,’ which do not bear the same name, were taken into
calculation, and was told they were, but of course the poems were
not. All this has been done in twenty years.” In 1803, an unknown
Mr. Scott’s name was found as the author of three very good ballads
in Lewis’s “Tales of Wonder.” This was his first publication.—Pope,
who until now had been considered as the poet who had made the
most by his works, died worth about £800 a-year.—Johnson, for his
last and best work, his “Lives of the Poets,” published after the
“Rambler” and the “Dictionary” had established his fame, got two
hundred guineas, to which was added one hundred more. Mr.
Hayward, in a note, adds—“‘Waverley’ having been published in
1814, the sum mentioned by Constable was earned in nine years, by
eleven novels in three volumes each, and three series of ‘Tales of my
Landlord,’ making nine volumes more; eight novels twenty-four
volumes, being yet to come. Scott’s first publication, ‘Translations
from the German,’ was in 1796. During the whole of his literary life
he was profitably engaged in miscellaneous writing and editing; and
whatever the expectations raised by has continued popularity and
great profits, they were surpassed by the sale of the collected and
illustrated edition of the novels commenced under his own revision
in 1829. Altogether, the aggregate amount gained by Scott in his
lifetime, very far exceeds any sum hitherto named as accruing to
any other man from authorship. Pope inherited a fortune, saved and
speculated; and we must come at once to modern times to find
plausible subjects of comparison. T. Moore’s profits, spread over his
life, yield but a moderate income. Byron’s did not exceed £20,000.
Talfourd once showed me a calculation, by which he made out that
Dickens, soon after the commencement of ‘Nicholas Nickleby,’ ought
to have been in the receipt of £10,000 a-year. Thackeray never got
enough to live handsomely and lay by. Sir E. B. Lytton is said to
have made altogether from £80,000 to £100,000 by his writings’.
We hear of 500,000 francs (£20,000) having been given in France
for Histories—to MM. Thiers and Lamartine for example; but the
largest single payment ever made to an author for a book, was the
cheque for £20,000, on account, paid by Messrs. Longman to
Macaulay soon after the appearance of the third and fourth volumes
of his History, the terms being that he should receive three-fourths
of the net profits.” This note of Mr. Hayward’s, it should be
remembered, was written in 1864. Macaulay cleared a fine sum by
his History, and so did the publishers. During the nine years, ending
with the 25th of June, 1857, Messrs. Longman disposed of 30,978
copies of the first volume of the History; 50,783 copies during the
nine years ending with June, 1866; and 52,392 copies during the
nine years ending with June, 1875. Within a generation of its first
appearance, upwards of 150,000 copies of the History will have been
printed and sold in the United Kingdom alone.
It is to be questioned, when her life comes to be written, whether
any author has been more successful, in a pecuniary point of new,
than Miss Braddon, whose “Lady Audley’s Secret” at once placed her
on the pinnacle of fame and fortune, and yet she began the world as
a ballet-girl.
Few Irishmen, in a literary and political point of view, did better than
the Right Hon. John Wilson Croker. In his “Memoirs,” Charles Mayne
Young thus speaks of his rise and progress:—

“I suspect few people now alive are aware of the

commencement of Croker’s career in London. Horace Smith,
James’s brother, and one of the joint authors of ‘Rejected
Addresses,’ told me that he, his brother, and Cumberland,
formed the staff of the Morning Post when Colonel Mellish was
its sole proprietor. On a certain quarter-day, when he was in
the habit of meeting them at the office and paying them their
salary, he took occasion to pass them unqualified commendation
for the great ability they had brought to bear upon his journal.
He assured them that the circulation of the paper had
quadrupled since their connection with it; ‘but—but—that he
was, nevertheless, under the necessity of dispensing with their
pens for the future.’ The two Smiths were so utterly unprepared
for such a declaration, that they were tongue-tied. Not so the
testy Cumberland, who took care to make himself as clearly
understood as if he had been the veritable Sir Fretful Plagiary.
“‘What,’ he asked his employer, ‘the d—l do you mean? In the
same breath in which you laud your servants to the skies, and
express your sense of obligation to them, you discharge them
oven without the usual month’s warning!’
“Mellish, quite unmoved, replied—‘You must know, good sirs,
that I care for my paper, not for its principles, but as an
investment; and it stands to reason, that the heavier my
outgoings, the less my profits. I do, as I have said, value your
merits highly; but not as highly as you charge me for them.
Now, in future, I can command the services of one man, who
will do the work of three for the wage of one.’
 “‘The deuce you can,’ said Cumberland. ‘He must be a phœnix.
Where, pray, may this omniscient genius be met with?’
“‘In the next room! I will send him to you.’
“As he left, a young man entered, with a well-developed skull, a
searching eye, and a dauntless address.
“‘So, sir,’ screamed out Cumberland, ‘you must have an
uncommon good opinion of yourself! You consider yourself, I
am told, three times as able as any one of us; for you undertake
to do an amount of work, single-handed, which we have found
enough for us all.’ ‘I am not afraid,’ said the young man, with
imperturbable sang froid, ‘of doing all that is required of me.’
They all three then warned him of the tact, discretion, and
knowledge of books and men required—of the difficulties of
which he must expect to find an enterprise of such magnitude
beset, &c., &c. They began then to sound his depth; but on
politics, belles lettres, political economy, even the drama, they
found him far from shallow. Cumberland, transported out of
himself by his modest assurance, snatched up his hat, smashed
it on his head, rammed snuff incontinently up his nose, and
then rushed by Mellish, who was in the adjoining room,
swearing, and saying as he left, ‘Confound the potato. He’s so
tough, there’s no peeling him!’ The tough potato was John
Wilson Croker.”

That Charles Dickens made a great deal of money, all the world is
well aware. That in the tale of “David Copperfield,” a little of his
childish life was outlined, was known, or rather suspected; but till his
life appeared, no one had the least idea how low down in the world
he and his family were, and how much more creditable to him was
his rise.
If it is good for a man to bear the yoke in his youth, Dickens
certainly had this advantage. We have seldom read a more touching
picture than that which is given of the life of the neglected,
untaught, half-starved boy at this time. It is tragic and affecting
enough in itself, but it is still more impressive as suggesting the
possible lot of hundreds and thousands in this great London of ours.
The one boy, by means of marvellous genius, forces his way to the
front; but who is to tell the story of the obscure multitude who
perish in the struggle? What imagination has ever pictured scenes
as tragic as the following experiences?—

“It is wonderful to me how I could have been so easily cast

away at such an age. It is wonderful to me, that even after my
descent into the poor little drudge I had been since we came to
London, no one had compassion enough on me—a child of
singular abilities, quick, eager, delicate, and soon hurt, bodily or
mentally—to suggest that something might have been spared,
as certainly it might have been, to place me at any common
school. Our friends, I take it, were tired out. No one made any
sign. My father and mother were quite satisfied. They could
hardly have been more so if I had been twenty years of age,
distinguished at a grammar-school, and going to Cambridge.
“The blacking warehouse was the last house on the left-hand
side of the way, at old Hungerford-stairs. It was a crazy,
tumble-down old house, abutting of course on the river, and
literally overrun with rats. Its wainscotted rooms, and its rotten
floors and staircase, and the old grey rats swarming down in the
cellars, and the sound of their squeaking and scuffling coming
up the stairs at all times, and the dirt and decay of the place,
rose up visibly before me, as if I were there again. The
counting-house was on the first floor, looking over the coal-
barges and the river. There was a recess in it, in which I was to
sit and work. My work was to cover the pots of paste-blacking;
first, with a piece of oilpaper, and then with a piece of blue
paper; to tie them round with a string; and then to clip the
paper close and neat, all round, until it looked as smart as a pot
of ointment from an apothecary’s shop. When a certain number
of grosses of pots had attained this pitch of perfection, I was to
paste on each a printed label; and then go on again with more
pots. Two or three other boys were kept at similar duty down
stairs on similar wages. One of them came up, in a ragged
apron and paper cap, on the first Monday morning, to show me
the trick of using the string and tying the knot. His name was
Bob Fagin; and I took the liberty of using his name, long
afterwards, in ‘Oliver Twist.’
“Our relative had kindly arranged to teach me something in the
dinner-hour—from twelve to one, I think it was—every day. But
an arrangement so incompatible with counting-house business
soon died away, from no fault of his or mine; and for the same
reason, my small work-table, and my grosses of pots, my
papers, string, scissors, paste-pot, and labels, by little and little,
vanished out of the recess in the counting-house, and kept
company with the other small work-tables, grosses of pots,
papers, string, scissors, and paste-pots, down stairs. It was not
long before Bob Fagin and I, and another boy whose name was
Paul Green, but who was currently believed to have been
christened Poll (a belief which I transferred, long afterwards,
again to Mr. Sweedlepipe, in ‘Martin Chuzzlewit’), worked
generally side by side. Bob Fagin was an orphan, and lived with
his brother-in law, a waterman. Poll Green’s father had the
additional distinction of being a fireman, and was employed at
Drury-lane Theatre; where another relation of Poll’s, I think his
little sister, did imps in the pantomimes.
“No words can express the secret agony of my soul as I sank
into this companionship; compared these every-day associates
with those of my happier childhood; and felt my early hopes of
growing up to be a learned and distinguished man, crushed in
my breast. The deep remembrance of the sense I had of being
utterly neglected and hopeless—of the shame I felt in my
position—of the misery it was to my young heart to believe that,
day by day, what I had learned, and thought, and delighted in,
and raised my fancy and my emulation up by, was passing away
 from me, never to be brought back any more—cannot be
written. My whole nature was so penetrated with the grief and
humiliation of such considerations, that even now, famous and
caressed and happy, I often forget in my dreams that I have a
dear wife and children—even that I am a man—and wander
desolately back to that time of my life.
“My mother and my brothers and sisters (excepting Fanny in the
Royal Academy of Music) were still encamped, with a young
servant-girl from Chatham workhouse, in the two parlours in the
emptied house in Gower Street North. It was a long way to go
and return within the dinner-hour; and, usually, I either carried
my dinner with me, or went and bought it at some neighbouring
shop. In the latter case it was commonly a saveloy and a penny
loaf; sometimes, a four-penny plate of beef from a cook’s shop;
sometimes a plate of bread and cheese, and a glass of beer,
from a miserable old public-house over the way—the ‘Swan,’ if I
remember right, or the ‘Swan’ and something else that I have
forgotten. Once I remember tucking my own bread (which I
had brought from home in the morning) under my arm,
wrapped up in a piece of paper like a book, and going into the
best dining-room in Johnson’s alamode-beef-house in Charles
Court, Drury Lane, and magnificently ordering a small plate of
alamode-beef to eat with it. What the waiter thought of such a
strange little apparition, coming in all alone, I don’t know; but I
can see him now, staring at me as I ate my dinner, and bringing
up the other waiter to look. I gave him a halfpenny, and I wish,
now, that he hadn’t taken it.”

It was thus Dickens was trained to fight the battle of life. After this
one feels inclined to say, “How great are the blessings of poverty!”
What an impulse it gives the man to raise himself above it, somehow
or other. Hazlitt used to say that “the want of money often places a
man in a very ridiculous position.” There is no doubt about that. It
is also equally clear, that, without money, there can be little comfort,
little independence of thought or action, little real manliness.
Poverty is a wonderful tonic. Volumes might be written in its praise.
Almost all the wonderful things that have been done in the world
have been accomplished by men who were born and bred in
poverty. She is the nurse of genius, the mother of heroes. She has
garlanded the world with gold. Luxury and wealth have ever been
the ruin alike of individuals and nations. The world’s greatest
benefactors have been the money-getting men. Of course there are
a few exceptions; but they are the exceptions that confirm the rule.
 CHAPTER XII.
REFLECTIONS ON MONEY-MAKING.

We have little faith in reflections. If a man cannot draw an inference

for himself, it is little use anyone attempting to draw it for him. The
reader of the preceding pages must have been taught, by example,
how to get money. The art of money-making is a very simple one.
If your income is twenty pounds, and you spend nineteen pounds,
nineteen shillings, and elevenpence-three-farthings, you will never
be troubled about money matters; and, in the course of years, may
have a fortune commensurate with so modest an expenditure.
Having thus acquired a small amount of capital, you must not part
with it to mining-brokers or stock-brokers, however plausible the tale
they tell, and however friendly you may be with them. They are
bound to do business, and for the sake of that, will help their
nearest friend to an investment of the rottenest character. Stock-
brokers may have a sense of honour—may be gentlemen; but I
question much whether a money-broker has any feeling for his
clients, I have known little money made by outsiders speculating on
the Stock Exchange or in mines. I have known many reduced to
beggary and want by such means.
Commerce, in our day, is the high road to wealth. You must begin
at the bottom, and work your way up to the top. It is not talent that
makes a man succeed in business, but the intense determination
which carries a man through every obstacle till the desired end is
attained. It was thus George Moore became a great man. The first
elements in his character were simplicity and directness. He was
prompt, energetic, precise; doing at once what he had to do. He
never cavilled about trifles. There was no shuffling about him—no
humbug. The only thing he could not tolerate was the drone. He
held strong opinions on most subjects, and he adhered to them
firmly. He never did anything by halves; he went into it body and
soul, with the whole of his nature; he went straight to the point.
When he had settled a thing he left it as something done; when two
sides of a question were presented to him, he was quick to decide,
and he was usually right in his decision.
Dr. Smiles writes—“The successful merchant is not merely the man
who is most fertile in commercial combinations, but the man who
acts upon his judgment with the greatest promptitude.” Mr.
Crampton, George Moore’s partner, says—“I never knew him make a
mistake in judgment.”
Another fact to be observed is, that it is the country lads who, as a
rule, are the most successful. At first they fail in accuracy, and
quickness, and promptitude. They are slow compared with town-
bred boys. “The City boy,” writes Dr. Smiles, “scarcely grows up; he
is rushed up; he lives amid a constant succession of excitements,
one obliterating another. It is very different with the country boy; he
is much slower in arriving at his maturity than the town boy, but he
is greater when he reaches it; he is hard and uncouth at first,
whereas the town boy is worn smooth by perpetual friction, like the
pebbles in a running stream. The country boy learns a great deal,
though he may seem to be unlearned; he knows a good deal about
nature, and a great deal about men. He has had time to grow. His
brainpower is held in reserve; hence the curious fact, that, in course
of time, the country-bred boy passes the City-bred boy, and rises to
the highest positions in London life. Look at all the great firms, and
you will find that the greater number of the leading partners are
those who originally were country-bred boys. The young man bred
in the country never forgets his origin.” “There is,” says
Rochefoucauld, “a country accent, not in his speech only, but in his
thought, conduct, character, and manner of existing, which never
forsakes him.”
George Moore had a brother. He was far apter than George; he had
a better education; he had read extensively, and was well versed in
literature; but he wanted that which his brother George had—
intense perseverance. Hence the failure of the one, and the success
of the other. It is thus the determined, persevering man who
succeeds. It was thus Warren Hastings won back the broad lands of
his ancestors.
“In New York,” says an American writer, “fortunes are suddenly
made, and suddenly lost. I can count over a dozen merchants who,
at the time I began to write this book, a few months ago, were
estimated to be worth not less than 250,000 dollars—some of them
half a million—who are now utterly penniless. At the opening of this
year (1868), a merchant, well-known in this city, had a surplus of
250,000 dollars in cash. He died suddenly in July. He made his will
about three months before his death, and appointed his executors.
By that will he divided 250,000 dollars. His executors contributed
1,000 dollars to save a portion of his furniture for his widow, and
that was all that was left her out of that great estate. He did what
thousands have done before him—what thousands are doing now,
and will do to-morrow. He had money enough; but he wanted a
little more. He was induced to go into a nice little speculation in
Wall Street; he put in 50,000 dollars. To save it he put in 50,000
dollars more. The old story was repeated, with the same result.” I
knew a gentleman who began the world as an advertising agent; he
managed to get a share in a newspaper, which eventually became
an immense commercial success. His share of the profits amounted
to some thousands a-year; but this was not enough—he must have
more. He turned money-lender, borrowing at 5 per cent., to lend
money on bad security at a high rate of interest. He died in the
prime of life, a bankrupt, and of a broken heart.
It is not every one who knows when to leave off money-making; but
there is a time when a wise man will remain satisfied with what he
has won. I knew a gentleman in the Corn Exchange, who was worth
£80,000. That was not enough for him, though to many it would
have been a fair fortune. He was determined to make one grand
coup before finally retiring from business, and enjoying the fruits of
his industry and enterprise. He did so against the entreaties of his
friends. The grand coup was a failure, and he died as poor as Job.
Such men are to be met in London every day.
A man who died very rich, was very poor when he was a boy. When
asked how he got his riches, he replied—“My father taught me never
to play till all my work for the day was finished, and never to spend
money till I had earned it. If I had but half-an-hour’s work to do in
a day, I must do that the first thing, and in half-an-hour. After this
was done I was allowed to play. I early formed the habit of doing
everything in its time, and it soon became perfectly easy to do so. It
is to this habit that I owe my prosperity.”
Sir Titus Salt, the millionaire, who made a fortune by the
introduction of alpaca-wool-cloth into the country, was a very early
riser. At Bradford, where he first commenced business, before he
had built his grand manufactory at Saltaire, it used to be said
—“There is Titus Salt; he has made a thousand pounds before other
men were out of bed.”
It was industry that helped to make Franklin a successful man of
business. This industry was, he tells us, a source of credit.
“Particularly I was told, that mention being made of the new
printing-office at the Merchants’ Every-Night Club, the general
opinion was that it must fail, there being already two printers in the
place—Kermer and Bradford. But Dr. Baird gave a contrary opinion:
‘for the industry of that Franklin,’ he said, ‘is superior to anything I
ever saw of the kind. I see him still at work when I go from the
club, and he is at work again before the neighbours are out of bed.
This struck the rest, and he soon after had offers from one of them
to supply us with stationery; but as yet he did not choose to engage
in shop business.’ I mention this,” adds Franklin, “more particularly,
and the more emphatically, though it seems to be talking in my own
praise, that those of my posterity who shall read it, may know the
use of that virtue (industry) when they see its effects in my favour
throughout this relation.”
Again, let us see how men lose money; for the art of keeping money
is of greater importance to a man than that of making it. The great
house of Overend and Gurney fell, and threw all London into a
panic, because the house did not know how to keep money, but
went into all sorts of ruinous speculations, which ultimately brought
it to the ground. “In a little room in one of the by-streets of New
York, up a narrow, dingy flight of stairs, may be seen a man,” says
an American writer, “doing a little brokerage which his friends put
into his hands. That man at one time inherited the name and
fortune of a house which America delighted to honour. That house
was founded by two lads who left their homes to seek their fortunes
in a great city. They owned nothing but the clothes they wore, and
a small bundle tied to a stick, and thrown over their shoulders.
Their clothes were home-spun, were woven under the parental roof,
and cut and made by motherly skill and sisterly affection. They
carried with them the rich boon of a mother’s blessing and a
mother’s prayers. They were honest, industrious, truthful, and
temperate. They did anything they found to do that was honest.
They began a little trade, which increased in their hands, and
extended till it reached all portions of the civilised world. They
identified themselves with every good work. Education, humanity,
and religion blessed their munificence. The founders of the house
died, leaving a colossal fortune, and a name without a stain. They
left their business and their reputation to the man who occupies the
little chamber that we have referred to. He abandoned the
principles on which the fame and honour of the house had been built
up. He stained the name that for fifty years had been untarnished.
He fled from his home; he wandered about the country under an
assumed name. Widows and orphans who had left trust-money in
their hands, lost their all. In his fall he dragged down the innocent,
and spread consternation on all sides. A few years passed, and after
skulking about in various cities abroad, he ventured back. Men were
too kind to harm him. Those whom he had befriended in the days
of his prosperity, helped him to a little brokerage to earn his bread,
and so he lingered on, and died, poor and forgotten, and obscure;—
a warning to the prosperous, not to forget that honesty is the best
policy after all.”
A fast man in business, sooner or later, comes to grief. A young man
in New York represented a New England house of great wealth and
high standing. He was considered one of the smartest and most
promising young men in the city. The balance in the bank, kept by
the house, was very large, and the young man used to boast that he
could draw his cheque any day for 200,000 dollars, and have it
honoured. The New England house used a great deal of paper, and
it could command the names of the best capitalists to any extent.
He was accustomed to sign notes in blank and leave them with the
concern, so much confidence had he in its soundness and integrity.
Yet, strange to say, these notes, with those of other wealthy men,
with nearly the whole financial business of the house, were in the
hands of the young manager in New York. In the meanwhile he
took a turn at Harry Hill’s to relieve the pressure of business. Low
amusements, and the respectable company he found, suited him.
From a spectator he became a dancer. From dancing he took to
drinking. He then tried his hand at play, and was cleaned out every
night, drinking deeply all the while. He became enamoured of a
certain class of women, clothed them in silk, velvets, and jewels,
drove them in dashing teams in the Central Park, secured them fine
mansions, and paid the expenses of their costly establishments, all
the while keeping the confidence of his business associates. In his
jaded, wan, and dissipated look, men saw his attention to business.
The New England manager of the house was the father of the young
man. His reputation was without a stain, and confidence in his
integrity was unlimited. In the midst of his business he dropped
down dead. This brought things to a crisis, and an exposure
immediately followed. The great house was bankrupt, and
everybody ruined that had anything to do with it. Those who
supposed themselves well off, found themselves quite the reverse.
Widows and orphans lost their all. Men suspended business on the
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com